Method and measuring system for the interferometric determination of a spatial distribution of an optical property of a test object

Abstract

Method for the interferometric determination of a spatial distribution of an optical property of a test object (14) by means of an interferometric measuring system (10), comprising the steps: - irradiating a test wave (34) generated by a diffractive optical element (32) onto the test object and generating a plurality of interferograms (46) in chronological sequence by superimposing a reference wave (30) with the test wave (34) generated by the diffractive optical element after its interaction with the test object, wherein during the generation of the interferograms a first system parameter (54) of the measuring system is changed to vary a phase difference between the test wave generated by the diffractive optical element and the reference wave and wherein, furthermore, during the generation of the interferograms, at least one further system parameter (56) of the measuring system is varied, wherein a change in the at least one further system parameter (56) has an influence on a phase difference distribution on a detection surface (43) of a detector (43) serving to detect a generated interferogram (46), and - Determining the local distribution of the optical property by evaluating the generated interferograms, wherein during the evaluation, error influences attributable to the varied system parameters (54, 56) are eliminated by taking into account changes in the interferograms caused by the variation of the system parameters when determining the local distribution of the optical property, wherein the evaluation of the generated interferograms (46) is based on at least one sensitivity (58) of the further system parameter (56), which indicates a relationship between the further system parameter (56) and at least one property of the interferograms.

Description

Background of the invention

[0001] The invention relates to a method and a measuring system for the interferometric determination of a spatial distribution of an optical property of a test object. This can, for example, be a spatial distribution of a refractive index of the test object or a shape deviation of an optical surface of the test object from a desired shape.

[0002] Interferometers with a phase-shifting technique are used to measure the shape or fit of surfaces. This involves sequentially recording a series of interference patterns created by superimposing a test wave reflected from the surface with a reference wave. Between two recordings, the reference wave is phase-shifted relative to the measurement wave by a specific difference. In this way, several interference patterns are recorded over one period of the interferometer signal. Complex mathematical methods can be used to reconstruct the surface topography from the recorded interference patterns. Phase shifting can be used to identify surface structures that are much smaller than the wavelength of the measurement wave.

[0003] Examples of such interferometers are Fizeau interferometers, which use a Fizeau element to split a measurement beam into a measurement wave and a reference wave. A piezo system is used to shift the Fizeau element either equidistantly, step by step, or continuously, parallel to the incident direction of the measurement beam. At each position of movement, the resulting interferogram, and thus the respective phase information of the surface, is captured and stored by a camera.

[0004] Unfortunately, the spatial distribution of the test object's optical properties determined from the various interferograms regularly exhibits errors due to fluctuations during the recording of the interferograms, which cannot be adequately corrected using conventional numerical fitting techniques. Furthermore, the mathematical model used to reconstruct the surface topography is often incomplete, or its necessary input parameters are often too imprecise. The resulting measurement accuracy of the interferometric measuring device is therefore often insufficient, especially considering the increasing demands placed on the shape measurement of optical elements for microlithography.

[0005] Document DE102019204096A1 describes a measurement method for interferometrically determining the shape of a surface of a test object. US2006 / 0274325A1 relates to a method for qualifying a diffraction grating of a diffractive optical element. US6956657B2 discloses a method for measuring the surface of a test object by combining several sub-surface measurements into an overall surface. Underlying task

[0006] It is an object of the invention to provide a method and a measuring system by which the aforementioned problems are solved and, in particular, the measuring accuracy in the interferometric determination of the local distribution of the optical property of the test object is improved. Inventive solution

[0007] The aforementioned object can be achieved according to the invention, for example, with a method for the interferometric determination of a spatial distribution of an optical property of a test object by means of an interferometric measuring system. The method according to the invention comprises irradiating a test wave onto the test object and generating a plurality of interferograms in chronological sequence by superimposing a reference wave with the test wave after interaction of the test wave with the test object. During the generation of the interferograms, a first system parameter of the measuring system is changed to vary a phase difference between the test wave and the reference wave, and furthermore, at least one further system parameter of the measuring system is varied during the generation of the interferograms. Furthermore, the method according to the invention comprises determining the spatial distribution of the optical property by evaluating the generated interferograms.During the evaluation, error influences due to the varied system parameters are eliminated by taking into account changes in the interferograms caused by the variation of the system parameters when determining the local distribution of the optical property.

[0008] In other words, during the evaluation, error influences attributable to the first system parameter and at least one additional system parameter are eliminated. By taking into account the changes in the interferograms caused by the variation of the additional system parameter, measurement errors attributable to a misalignment of the measuring system with respect to the additional system parameter can be eliminated from the spatial distribution of the optical property determined as the measurement result. This can improve the overall measurement accuracy of the interferometric determination of the spatial distribution of the optical property of the test object.

[0009] According to one embodiment, the spatial distribution of the optical property of the test object describes a shape of an optical surface of the test object, i.e., the property is a deviation of the individual points of the optical surface from a desired shape. According to another embodiment, the optical property comprises a refractive index of the optical element, i.e., the spatial distribution comprises a refractive index distribution of the optical element.

[0010] The spatial distribution of the optical property is, in particular, a two-dimensional distribution. According to one embodiment, the first system parameter is varied from interferogram to interferogram. The phase difference can be varied such that each of the interferograms is based on a different phase difference. According to one embodiment, the further system parameter is also varied from interferogram to interferogram.

[0011] According to a further embodiment, at least two, in particular at least three, at least five or at least ten further system parameters are varied during the generation of the interferograms.

[0012] According to a further embodiment, the at least one further system parameter is selected from the following group of system parameters: a position and a rotational position of at least one optical element of the measuring system, a position and a rotational position of the test object, a temperature of the at least one optical element, a temperature, a pressure, a humidity and a composition of at least one medium between optical elements of the measuring system, a size, a position of a measuring radiation source of the measuring system, a wavelength, an intensity, a polarization and a degree of coherence of a measuring radiation generated by the measuring radiation source, as well as a position, a rotational position, a temperature and an exposure time of a camera of the measuring system.

[0013] In other words, the at least one further system parameter is a position and / or a rotational position of at least one optical element of the measuring system, a position and / or a rotational position of the test object, a temperature of the at least one optical element, a temperature, a pressure, a humidity and / or a composition of at least one medium between optical elements of the measuring system, a size, a shape and / or a position of a measuring radiation source of the measuring system, a wavelength, an intensity, a polarization and / or a degree of coherence of a measuring radiation generated by the measuring radiation source, and / or a position, a rotational position, a temperature and / or an exposure time of a camera of the measuring system.

[0014] For example, to vary the phase difference between the test wave and the reference wave, the position of a Fizeau element of the measuring system in the axial direction, i.e., in the direction of incidence of the measurement radiation onto the Fizeau element, can be changed as a first system parameter. As a second system parameter, for example, the rotational position of the test object relative to the rotational axis oriented in the axial direction, i.e., in the direction of incidence of the measurement radiation onto the test object, can be varied as a further system parameter.

[0015] According to the invention, the evaluation of the generated interferograms is based on at least one sensitivity of the additional system parameter, which indicates a relationship between the additional system parameter and at least one property of the interferograms. If multiple system parameters are varied, according to one embodiment, the evaluation is based on the respective sensitivities of the varied system parameters.

[0016] According to one embodiment variant, during the evaluation of the generated interferograms, a modification of the at least one property of the interferograms caused by the variation of the system parameters is determined on the basis of the at least one sensitivity, and by means of the determined modification, a time-invariant component of the at least one property is determined.

[0017] According to a further embodiment, the at least one property of the interferograms refers to a spatial brightness distribution, a spatial distribution of a modulation amplitude and / or a spatial phase distribution in the interferograms. In other words, the at least one sensitivity indicates a respective relationship between the system parameter and a spatial brightness distribution, a spatial distribution of a modulation amplitude and / or a local phase distribution in the interferograms. This respective spatial distribution remains constant over time when recording the interferograms with only the first system parameter being varied. Thus, in this recording, only the phase difference would be varied by means of the first system parameter, ieAt least one other system parameter is not varied but remains constant over time, forming a temporally invariant component in the various interferograms. In this text, the temporally invariant component of the spatial brightness distribution is also referred to as a(x, y), the temporally invariant component of the spatial distribution of the modulation amplitude is referred to as b(x, y), and the temporally invariant component of the spatial phase distribution is referred to as Φ(x, y). The respective relationships between a further system parameter and the respective fundamental component are also referred to in this text as the sensitivities of the respective system parameter.

[0018] According to a further embodiment, the at least one sensitivity is determined by simulation and / or experimentally before the interferograms are generated.

[0019] According to a further embodiment, several sensitivities are determined before the interferograms are generated, and the generated interferograms are evaluated on the basis of a selection of the determined sensitivities.

[0020] According to one embodiment, an orthogonal set of sensitivities is compiled when selecting the sensitivities. An orthogonal set of sensitivities refers to sensitivities that have different effects, i.e., sensitivities with the same effect are eliminated. The term "effect" in this context refers to the change in one of the specified properties of the interferograms in response to a change in the system parameter associated with the sensitivity. Sensitivities with the same effect are thus understood to mean sensitivities that essentially define a change in the same properties or the same signature of properties of the interferograms with a corresponding change in the system parameters associated with them.

[0021] According to a further embodiment, a plurality of sensitivities are determined before the interferograms are generated, and during the evaluation of the generated interferograms, a respective correction of the individual interferograms is carried out on the basis of at least one sensitivity selected from the determined sensitivities.

[0022] According to a further embodiment, several sensitivities are determined prior to generating the interferograms, and during the evaluation of the generated interferograms, an intermediate result of the spatial distribution of the optical property determined from the totality of the interferograms is corrected based on at least one sensitivity selected from the determined sensitivities. An intermediate result can, for example, be the spatial distribution of the phase difference between the test wave and the reference wave, which is determined from the measured interferograms and corrected based on the selected sensitivity.

[0023] According to a further embodiment, a step sequence of the varied system parameters is determined by simulation and / or experimentation prior to generating the interferograms. The step sequence of the varied further system parameter can, for example, result in a phase step sequence with respect to the phase difference between the test wave and the reference wave.

[0024] According to a further embodiment, only a respective sub-area of ​​the generated interferograms assigned to a sub-aperture of the test wave is evaluated, and the spatial distribution determined thereby is combined with a further spatial distribution determined from a respective other sub-area of ​​the interferograms. This procedure is particularly useful for calculating errors from the interferograms that are due to striae or air turbulence. These influences can be more easily calculated from smaller interferogram areas than from larger interferogram areas. This is then possible with less complex algorithms. By dividing the evaluation of the generated interferograms into sub-areas and combining the area-by-area evaluation results, such errors can be efficiently calculated from the measurement results.

[0025] According to a further embodiment, after the error influences attributable to the system parameters have been eliminated, any remaining errors in the correspondingly corrected interferograms are eliminated using an optimization algorithm. These remaining errors are also referred to in this text as "penetrating fringes," since these errors appear in the interferogram as fringes that are not influenced by changes in the system parameters and are therefore permanently present and thus "penetrating."

[0026] When calculating the remaining errors, the corrected interferograms are adapted to predefined error signatures. According to one embodiment, an objective function is defined in which errors in the distribution of the phase difference are linked to travel distances of the associated system parameters across one or more of the aforementioned sensitivities, for example, in the form of a sensitivity matrix. This objective function is then optimized, in particular minimized, using an optimization algorithm. The optimization algorithm can be based, for example, on a least-squares optimization method. Based on the travel distance results determined in this process, the errors associated with the associated system parameters are then calculated from the phase difference.

[0027] According to a further embodiment, the generated interferograms form a first interferogram data set, and any errors remaining after eliminating error influences are corrected using at least one additional interferogram data set. For this purpose, for example, the evaluation results of the two interferogram data sets can be combined; alternatively, a different type of mathematical combination can be used. In particular, errors that remain even after correction using an optimization algorithm are corrected in this way.

[0028] The aforementioned object can further be achieved, for example, with a measuring system for the interferometric determination of a spatial distribution of an optical property of a test object. The measuring system comprises an interferometry module for irradiating a test wave onto the test object and generating multiple interferograms in chronological sequence by superimposing a reference wave on the test wave after its interaction with the test object, a first variation device configured to change a first system parameter of the measuring system during the generation of the interferograms in order to vary a phase difference between the test wave and the reference wave, and at least one further variation device configured to vary at least one further system parameter of the measuring system during the generation of the interferograms.Furthermore, the measuring system comprises an evaluation device which is configured to determine the spatial distribution of the optical property by evaluating the generated interferograms and to calculate out error influences attributable to the varied system parameters by taking into account changes in the interferograms which are caused by the variation of the system parameters.

[0029] The features specified with regard to the above-mentioned embodiments, exemplary embodiments, or embodiment variants, etc. of the method according to the invention can be transferred correspondingly to the measuring system according to the invention, and vice versa. These and other features of the embodiments according to the invention are explained in the description of the figures and the claims. The individual features can be implemented either separately or in combination as embodiments of the invention. Furthermore, they can describe advantageous embodiments that are independently protectable and whose protection may only be claimed during or after the application has been filed. Brief description of the drawings

[0030] The above and other advantageous features of the invention are illustrated in the following detailed description of exemplary embodiments of the invention with reference to the accompanying schematic drawings. It shows: Fig. 1 shows an embodiment of a measuring system according to the invention for interferometrically measuring a shape deviation of an optical surface of a test object from a desired shape, with an evaluation device for evaluating recorded interferograms using sensitivities of system parameters of the measuring system, Fig. 2 an exemplary set of sensitivities of system parameters of the measuring system according to Fig. 1, Fig. 3 a set of orthonormalized sensitivities of system parameters of the measuring system according to Fig. 1, Fig. 4 a flow chart to illustrate an embodiment of the method carried out by means of the measuring system according to Fig. 1 executed measuring method according to the invention, Fig. 5 a by means of the measuring system according to Fig. 1 generated phase image with a schlieren distribution, Fig. 6 an illustration of an evaluation of an interferogram for removing penetrating stripes according to an embodiment of the invention, Fig. 7 an illustration of a procedure for removing penetrating stripes according to another embodiment of the invention, Fig. 8 an exemplary set of recorded interferograms, the evaluation of which leads to reduced penetration fringes, as well as Fig. 9 a comparison between scalar phase shifting versus phase shifting with tilting of the reference wave with respect to breakthrough fringes. Detailed description of embodiments according to the invention

[0031] In the exemplary embodiments or embodiments or variants described below, functionally or structurally similar elements are provided with the same or similar reference numerals wherever possible. Therefore, to understand the features of the individual elements of a specific embodiment, reference should be made to the description of other exemplary embodiments or the general description of the invention.

[0032] To facilitate the description, a Cartesian xyz coordinate system is shown in the drawing, from which the respective positional relationship of the components shown in the figures results. Fig. 1 the x-direction runs perpendicular to the drawing plane, the z-direction to the right and the y-direction upwards.

[0033] In Fig. 1 illustrates an embodiment of a measuring system 10 for interferometrically measuring a spatial distribution of an optical property of a test object 14 in the form of an optical surface 12 of a test object 14. The measuring system 10 can be used, in particular, to determine a deviation of the actual shape of the surface 12 from a desired shape. The test object 14 can be, for example, a mirror of a projection objective or of an illumination system of a projection exposure system for EUV microlithography with a non-spherical surface for reflecting EUV radiation with a wavelength of less than 100 nm, in particular a wavelength of approximately 13.5 nm or approximately 6.8 nm. The non-spherical surface of the mirror can, for example, have a freeform surface with a deviation from any rotationally symmetric asphere of more than 5 µm and a deviation from any sphere of at least 1 mm.

[0034] The measuring system 10 comprises an interferometry module 15, a first variation device 50, several further variation devices 52, and an evaluation device 48. The interferometry module 15 contains a measuring radiation source 16 for providing sufficiently coherent measuring radiation 18 in the form of an expanding wave. In this exemplary embodiment, the measuring radiation source 16 comprises a waveguide 20 with an exit surface 21 at which the expanding wave originates. The waveguide 20 is connected to a radiation generation module 22, e.g., in the form of a laser. For this purpose, a helium-neon laser with a wavelength of approximately 633 nm can be provided, for example. However, the measuring radiation 18 can also have a different wavelength in the visible or non-visible wavelength range of electromagnetic radiation.The radiation source 16 with the waveguide 20 represents only one example of a radiation source 16 that can be used for the measuring system 10. In alternative embodiments, instead of the waveguide 20, an optical arrangement with lens elements, mirror elements or the like can be provided to provide a suitable wave from the measuring radiation 18.

[0035] The interferometry module 15 further comprises a beam splitter 24, a collimator 26, a Fizeau element 28, and a diffractive optical element 32 in the form of a computer-generated hologram (CGH). First, the expanding wave of the measurement radiation 18 passes through the beam splitter 24, whereupon it is converted into a plane wave by the collimator 26. The Fizeau element 28 serves as a reference element and has a Fizeau surface 29, at which a portion of the incoming measurement radiation 18 is reflected as a returning reference wave 30. The interferometry module 15 according to Fig. 1 is thus configured as a Fizeau interferometer. Alternatively, other suitable interferometer types can be used, such as an interferometer with a reference arm in which only the reference wave travels (e.g., a Michelson interferometer or a Twyman-Green interferometer).

[0036] The part of the measuring radiation 18 which has passed through the Fizeau element 28 then strikes the diffractive optical element 32 as input wave 31. The diffractive optical element 32 forms a test optics which serves to generate a test wave 34 for irradiation onto the surface 12 of the test object 14.

[0037] The test wave 34 is directed at the test object 14 and has a wavefront that is at least partially adapted to the desired shape of the optical surface 12. The test wave 34 is reflected by the optical surface 12 of the test object 14 and travels back to the diffractive optical element 32 as a returning test wave 34r. Due to the wavefront that is adapted to the desired shape of the optical surface 12, the test wave 34 impinges on the optical surface 12 essentially perpendicularly at every location on the optical surface 12 and is reflected back into itself.

[0038] The test wave 34r returning from the surface 12 passes through the diffractive optical element 32 again and is diffracted again. This results in a retransformation of the returning test wave 34r into an approximately plane wave, the wavefront of which exhibits corresponding deviations from a plane wavefront due to deviations of the surface 12 of the test object from its desired shape.

[0039] Furthermore, the interferometry module 15 contains a detection device 36 with the beam splitter 24 already mentioned above for leading out the combination of the returning test wave 34r and the returning reference wave 30 from the beam path of the irradiated measuring radiation 18 and a camera 38 for detecting an interferogram generated by superimposing the test wave 34r with the reference wave 30r.

[0040] The returning test wave 34r and the returning reference wave 30 impinge as convergent beams on the beam splitter 24 and are reflected by it in the direction of the camera 38. The camera 38 comprises an aperture 40, an eyepiece 42 and a two-dimensional detector 44. Both convergent beams pass through the aperture 40 and the eyepiece 42 and finally impinge on a detection surface 43 of the two-dimensional detector 44. The detector 44 can be designed, for example, as a CCD sensor and detects an interferogram 46 generated by superimposing the test wave 34r with the reference wave 30. As in Fig. 1, such an interferogram comprises a plurality of interference fringes, wherein a deviation of the actual shape of the optical surface 12 from its desired shape leads to a curvature of the interference fringes.

[0041] The evaluation device 48 mentioned above serves to determine the actual shape of the optical surface 12 of the test object 14 from several acquired interferograms 46. For this purpose, the evaluation device 48 has a suitable data processing unit. Alternatively or additionally, the measuring system 10 can contain a data storage device or an interface to a network to enable an external evaluation unit to determine the surface shape using interferograms 46 stored or transmitted via the network.

[0042] The various interferograms 46 processed during the evaluation are determined by changing a first system parameter p, designated by the reference numeral 54 sby means of the first variation device 50 for varying the phase difference between the test wave 34 and the reference wave 30 and by varying at least one further system parameter p, designated by the reference numeral 56 k generated by means of the further variation device 52. The first system parameter p s is used to shift the phase difference step by step and in the embodiment according to Fig. 1, the position of the Fizeau element 28 in the direction of incidence of the measuring radiation 18, i.e., its z-coordinate position. For this purpose, the first variation device 50 can comprise, for example, a piezo element. Shifting the phase difference here means a uniform change in the phase difference distribution on the detection surface 43 of the detector 44, i.e., the phase difference is shifted equally at every point on the detection surface 43. As an alternative to changing the position of the Fizeau element 28, the phase difference can also be shifted, for example, by changing the wavelength of the measuring radiation 18.

[0043] The at least one additional system parameter p k is selected from a group of system parameters, the change of which has an influence on the phase difference distribution on the detection surface 43 of the detector 44. Of this group of system parameters, Fig. 1 a variety of system parameters p kThis includes coordinate positions p1, p2 and p3 of the test object 14 in the x-, y- and z-directions respectively as well as rotational positions of the test object 15, of which Fig. 1 only illustrates the rotational position designated p4 with the x-coordinate axis as the rotational axis. Further rotational positions relative to the y- or z-coordinate axis are possible. Below, the x-, y-, and z-coordinate positions, as well as the rotational positions or tilt positions relative to the x-, y-, and z-coordinate axes as rotational axes, are referred to as settings in all six rigid-body degrees of freedom.

[0044] Other possible system parameters p k include the rigid body degrees of freedom and the temperature of one or more optical elements of the interferometry module 15, for this purpose Fig. 1 illustrates, by way of example, the y-coordinate position p5, a rotational position p6, and the temperature p7 (T) of the collimator 26. Accordingly, settings of such rigid body degrees of freedom can also be applied to other optical elements of the interferometry module 15, such as the Fizeau element 28, the diffractive optical element 32, or the beam splitter 24 as system parameters p k be defined.

[0045] Other possible system parameters p k include settings of rigid body degrees of freedom of the measuring radiation source 16. In the embodiment according to Fig. 1, these are defined by the position and tilt of the exit surface 21. As an example, the y-coordinate position p8 of the exit surface 21 is shown. Furthermore, the wavelength p9 (λ), the intensity p 10 (I), the polarization p 11 (P) and the degree of coherence p 12(K) of the measuring radiation 18 generated by the measuring radiation source 16 as system parameter p k be elected.

[0046] Other possible system parameters p k include settings or parameters of the camera 38, such as a rigid body degree of freedom of the camera 38, for example the Fig. 1 illustrated coordinate position p 13 or rotational position p 14 , the temperature p 15 (T) of the camera 38 and, if applicable, an exposure time of the camera 38.

[0047] Other possible system parameters p k include a temperature p 16 (T), a pressure p 17 (p), a humidity p 18 (f) and a composition p 19 (Z) a medium within the interferometry module 15 and / or specifically the medium between certain optical elements of the interferometry module, such as the medium between the Fizeau element 28 and the diffractive optical element 32, as in Fig. 1. The medium can be air or another gas mixture.

[0048] As already mentioned above, the evaluation device 48 processes several interferograms 46 in which, on the one hand, the phase difference is determined step by step by changing the first system parameter p s is pushed and on the other hand at least one of the further system parameters p k is varied. For example, at least one, at least two, at least three, at least five, at least ten of the other system parameters p s When generating the interferograms 46, only one of the system parameters p s and p k or several of the system parameters p s and p k be changed at the same time.

[0049] The object of the evaluation of the generated interferograms 46 is to develop the phase difference ϕ(x,y) between the reference wave 30 and the returning test wave 34r from a series of M measured interferograms. In this case, the intensity signal J recorded by the camera 38 with the pixels x,y i for the interferogram No. i described by Ji(x,y,ti)=ai'(x,y,ti)+bi'(x,y,ti)cos ϕi'(x,y,ti) with possibly from interferogram to interferogram and thus dependent on the time t i fluctuating properties of the interferogram 46. The fluctuating properties of the interferogram 46 include a fluctuating local brightness distribution a' i (x,y,t i ), also called brightness for short, a spatial distribution of a modulation amplitude b' i (x,y,t i ,), also called modulation for short, and a local phase distribution, also called phase for short: ai'(x,y,ti)=a(x,y)+δai(x,y,ti) bi'(x,y,ti)=b(x,y)+δbi(x,y,ti) ϕi'(x,y,ti)=ϕ(x,y)+Δϕi(x,y,ti)+δϕi(x,y,ti)

[0050] Here, a(x, y), b(x, y), and Φ(x,y) are the stationary mean values ​​of brightness, modulation, and phase. Δϕ i (x, y, t i ) is the desired phase shift in the interferogram i, which is caused by changing the first system parameter p s Furthermore, δa i (x, y, t i ), δb i (x, y, t i ) and δϕ i (x, y, t i ) temporal disturbances of brightness, modulation and phase in the interferogram i, which are caused by the variation of the other system parameters p k caused.

[0051] The time mean of the disturbances is by definition zero (without loss of generality) to avoid ambiguity in a(x, y), b(x, y) and Φ(x,y): ∑iδai(x,y,ti)=∑iδbi(x,y,ti)=∑iΔϕi(x,y,ti)+δϕi(x,y,ti)=0

[0052] As an alternative to (1), the formulation (3) can be used: Ji(x,y,ti)=ai'(x,y,ti)(1+Vi'(x,y,ti)cos ϕi'(x,y,ti)) with Vi'(x,y,ti)=bi'(x,y,ti)ai'(x,y,ti)=V(x,y)+δVi(x,y,ti)

[0053] As already mentioned above, the evaluation techniques used can be based on the fact that the recording of the interferograms 46 is carried out serially at the times t i The defined phase shift Δϕ i (x, y, t i ) between the two waves. Usually, constant phase steps are performed: Δϕ i (x, y, t i ) = i Δϕ with Δϕ = const.

[0054] From the set of M interferograms and the knowledge of the introduced phase shifts Δϕ i (x, y, t i) is calculated by the evaluation device 48 using the sensitivities s explained in more detail below. ak , s bk and s ϕk , or analogously to the sensitivity s Vk when displaying the intensity signal J i According to (3) and (4), the phase difference of interest ϕ(x, y) is determined. From the determined phase difference ϕ(x, y), the deviation of the shape of the optical surface 12 of the test object 14 from its desired shape is determined.

[0055] The (relative) sensitivities each indicate a relationship between the respective further system parameter p k and a respective property of the interferograms. These properties are the brightness a(x,y), the modulation b(x,y) and the phase Φ(x,y) or alternatively the brightness, the contrast V(x,y) and the phase Φ(x,y). Specifically, the sensitivities s ak (x,y), s bk (x,y), s Φk (x,y) or s Vk(x,y) the differential change of a(x, y), b(x, y) and Φ(x,y), alternatively of V(x, y) instead of b(x, y), with the change of a system parameter p k with k = 1...k: sak(x,y)=1a(x,y)∂a(x,y)∂pk sbk(x,y)=1b(x,y)∂b(x,y)∂pk sϕk(x,y)=∂ϕ(x,y)∂pk sVk(x,y)=∂V(x,y)∂pk

[0056] The sensitivities s ak (x,y), s bk (X,y), s Φk (x,y) or s Vk According to one embodiment, (x,y) are determined in advance, ie before measuring the interferograms 46, by simulation and / or experimentally.

[0057] The temporal fluctuations δa(x,y,t), δb(x,y,t), δΦ(x,y,t) and δV(x,y,t) from interferogram to interferogram are due to temporal changes of each effective parameter p k in measuring system 10: δa(x,y,t)=a(x,y)∑k=1Ksak(x,y)δpk(t) δb(x,y,t)=b(x,y)∑k=1Ksbk(x,y)δpk(t) δϕ(x, y, t)=∑k=1Ksϕk(x, y)δpk(t) δV(x, y, t)=∑k=1KsVk(x, y)δpk(t)

[0058] A linear model is presented here, which applies to small deviations. If this model does not provide sufficient accuracy (6a-d), a nonlinear model can be used, which can be extended, for example, by Taylor series expansion. Deviations can occur periodically or stochastically, at high and medium frequencies, as slow drifts, or statically.

[0059] In Fig. 2 is an example set of possible sensitivities 58 of a measuring system 10 according to Fig. 1. The test object 14 is an aspherical specimen that exhibits rotationally symmetric and toric deviations from a sphere. In the upper row of Fig. 2 shows three sensitivities for Φ(x,y) with respect to position. These represent the deviations that result from the translation of the test object 14 in (lateral) x, y and (axial) z. The values ​​shown in the lower row of Fig. The three sensitivities shown in Figure 2 are deviations during rotation of the test object around the x, y, and z axes. This results in phase deviations during rigid-body motion of the test object. A linear model largely applies to the interferometer's measuring range, so doubling the misalignment corresponds to doubling the phase deviation.

[0060] Certain parameter changes (e.g., rotation of the test object 14) lead to a changed assignment of camera pixels to locations on the optical surface 12. According to one embodiment, transformations are introduced to apply the inventive algorithms in order to fix the assignment. Which transformations are applied can be determined by simulation (ray tracing) or experiment, for example, by applying markers ("fiducials") to the optical surface 12 of the test object 14.

[0061] In general, the required coordinate transformations are neither linear nor affine. Since the image of the test object 14 onto the camera 38 is nonlinearly distorted, position changes also lead to nonlinear transformations. In special arrangements, e.g., when testing rotationally symmetrical test objects, the rotation can be described by the usual rotation of the coordinates around the optical axis.

[0062] According to one embodiment, coordinate transformations in the camera plane x, y → x', y' are performed in two variants. According to the first variant, the optical arrangement is changed for the acquisition of an interferogram data set of a series. In this case, the results of each data set are transformed into the agreed coordinate system. According to the second variant, the optical arrangement is changed during the acquisition of the interferograms 46 of a data set. In this case, each individual interferogram 46 is transformed into the agreed coordinate system. In the latter variant, the sensitivities are also transformed into the current coordinate system, if necessary.

[0063] Unknown parameter variations not captured by the model result in stripe structures (“penetrating stripes”) remaining in the results for a(x, y), b(x, y), and Φ(x, y). These stripe structures are similar to the recorded interferograms with an unknown phase position, but can also have twice or a multiple of the number of stripes. The causes can be varied, e.g., fluctuations that are too fast for the camera 38, so that the recorded interferogram signal is unevenly blurred. Air streaks can also lead to stripe images breaking through, exhibiting a streak-like, unpredictable modulation. Breakthrough stripes or stripe images are referred to here as stripes in the interferograms 46 that are not caused by a change in the system parameters p s or p k are influenced and are therefore permanently present or “pervasive”

[0064] Systematic modulations of penetrating stripes s akl (x,y), s blk (x,y), s ϕkl (x, y) are also referred to here as sensitivities, although predictions for them may not be possible. They represent the possible areal modulation of the penetrating stripes with the frequency / . When using phase steps Δϕ i (x, y, t i ), which do not introduce any tilting or deformation of the stripes, the deviations can be represented in the form of penetrating stripes as δad(x, y)=a(x, y)∑k=1K∑l=1Msakl(x, y)(uaklcos(lϕ(x, y))+vaklsin(lϕ(x, y))) δbd(x, y)=b(x, y)∑k=1K∑l=1Msbkl(x, y)(ubkl cos(lϕ(x, y))+vbkl sin(lϕ(x, y))) δϕd(x, y)=∑k=1K∑l=1Msϕkl(x, y)(uϕkl cos(lϕ(x, y))+vϕkl sin(lϕ(x, y))) with the sensitivities (modulations)

[0065] s akl (x, y) as sensitivity k of the fringe frequency / in the brightness a(x, y), s bkl(x, y) as sensitivity k of the strip frequency / in the modulation b(x, y), s ϕkl (x, y) as sensitivity k of the strip frequency / in the phase ϕ(x, y) and the respective coefficients u to be determined akl , v akl , and bkl , v bkl and u ϕkl , vϕ kl .

[0066] The sensitivities (modulations) normalized to one are used as constant data fields, in the case of asymmetric disturbances as linear or cubic terms, and in the case of relaxation oscillations of the test object, for example, additionally as cylindrical functions. The coefficients to be determined by least-squares fitting are akl , v akl describe the strength and phase position of the penetrating fringes contained in a(x, y) with sensitivity k and fringing frequency / . The coefficients u bkl , v bkl and u ϕkl , v ϕkl correspondingly for b(x, y) and Φ(x,y).

[0067] According to one embodiment, before the measurement method is carried out, the required system settings are defined and specified before recording each interferogram 46. For this purpose, a step sequence of the varied system parameters p s and p k determined by simulation and / or experimentation. This affects the phase step settings, but possibly also other parameters such as the test specimen and light source position.

[0068] The phase step algorithm requires the specification of the sequence of phase steps Δϕ i (x, y). It can be taken from the sensitivities. ∂ϕ(x, y)∂pk This results in the phase advance for the parameter k, which is to be used for the phase shift. Several parameters can also be adjusted simultaneously to achieve a resulting phase shift and, if necessary, tilting in x and y. These can be, for example: an axial and a tilting movement of the reference element in the form of the Fizeau element 28, an axial and tilting movement of the test object 14, the wavelength of the measurement radiation source 16, and a refractive index of the medium in the interferometer cavity, i.e. the medium in the area traversed only by the test wave 34 and not by the reference wave 30.

[0069] The sensitivities can be determined by simulation or experiment. In general, for the set of M interferograms, the desired phase steps are specified for each interferogram i: Δϕi(x, y)=∑k=1Ksϕk(x, y)Δpki with K as the number of parameters to be adjusted and Δp kias the manipulated variable of the parameter k to be executed for step no. i.

[0070] It has been shown that, in addition to the necessary phase shift in constant steps, an additional stepwise tilting of the reference wave 30 or the test wave 34 along one of the two tilt axes leads to particularly artifact-free results. These are achieved by resolving ambiguity problems during the iterative evaluation of the interferogram data set, as explained in more detail below.

[0071] In Fig. Figure 9 shows a comparison of the remaining penetrating fringes when using constant (scalar phase shifting) versus constant and tilting phase steps (phase shifting with tilting). Vibrations of the test object 14 with increasing intensity from 0.1 nm to 10 nm were assumed as disturbances during data acquisition (x-axis).

[0072] Simultaneously adjusting other system parameters before each image acquisition, such as the test object position, the position, size, and shape of the measurement radiation source 16, the position of the camera 38, and the position of other optical components, may be advantageous when distinguishing artifacts from the desired test object shape. According to one embodiment, the additional parameter variations are then used to average over artifacts or to separate them from the desired measurement signal.

[0073] According to one embodiment, the brightness distribution on the camera chip is converted into an analog-electronic signal and then into a digital signal during the measurement process. This involves a variety of nonlinear effects in signal processing. Two methods are distinguished when recording the interferograms. These include the so-called phase-stepping method, where the phase shift occurs in a sequence of discrete steps, e.g., Δϕ. i(x, y) = i Δϕ(x, y). Furthermore, the interferograms can be recorded using the so-called synchronous detection method, where the phase shift occurs continuously according to a predetermined path of the phase shifter Δϕ(x, y, t)=∫τ=tt+Tω(x, y, τ)dτ.

[0074] During interferogram recording, external mechanical, thermal, climatic, acoustic, and electrodynamic influences, as well as stray light, can affect the measurement setup. This often leads to small changes in the measurement setup, the positions and rotational orientations of the optical elements, the camera, and the light source; the temperatures of the optical elements and media, as well as the camera and the light source; and the pressure, humidity, and composition of the optical media.

[0075] Depending on the design of the light source, the intensity, wavelength, size and shape, polarization, and degree of coherence of the measuring radiation source 16 can also change during interferogram recording. Depending on the design of the camera and exposure system, the exposure time can also fluctuate during interferogram recording. Transfer and amplifier electronics in the camera system are subject to noise and drift. Changing interfering reflections or external light disturbances can cause coherent or incoherent interference signals to be superimposed on the interferogram.

[0076] According to one embodiment, the selection of sensitivities to be considered in the analysis is carried out very restrictively. The algorithm is "blind" to each "footprint" subtracted from the result. If the test specimen possessed such a deviation, it would not be detectable, as it would be attributed to a misalignment of the setup.

[0077] Once the significant sensitivities have been selected, according to one embodiment the sensitivity data fields are orthogonalized and normalized in order to identify and, if necessary, eliminate ineffective or equally effective sensitivities. In other words, an orthogonal set of sensitivities is compiled when selecting the sensitivities. An orthogonal set of sensitivities is understood to mean sensitivities that have different effects, i.e. sensitivities that have the same effect are eliminated. The term “effect” in this context is understood to mean the change in one of the aforementioned properties of the interferograms 46 as a reaction to a change in the system parameter assigned to the sensitivity. Equivalent sensitivities are thus understood to mean sensitivities that essentially change the same properties orthe same signature of properties of the interferograms with corresponding changes to the system parameters assigned to them.

[0078] Numerical orthonormalization can be performed using methods known to those skilled in the art, such as the Gram-Schmidt orthogonalization method, in which the components of an original sensitivity data field are subtracted from all others in the best-fit manner until no data field is contained within any other. The individual orthogonalized data fields are then normalized (e.g., to ±1).

[0079] In orthonormalization, it has proven useful to introduce the constant and the tilts in x and y as basis functions in every case, because they are elementary in every interferometer (e.g. adjustment degrees of freedom of the reference, choice of the initial wavelength, etc.). In Fig. 3 shows orthonormalized sensitivities 58o. These include the sensitivities 56 from Fig. 2 orthonormalized (S o4 to S o9 ) with the addition of constant and tilts in x and y (S o1 to S o3 ).

[0080] If temporal perturbations do not follow systematic changes, the concept of sensitivities described below cannot be applied. There is no systematic pattern imprinted on the interferograms. Examples include air turbulence and schlieren, which change the phase and brightness distributions in an unpredictable manner from image to image. Schlieren, due to their fine structure, generally cannot be reasonably described by polynomial approximations.

[0081] In this case, the field to be measured can be divided into subapertures, and the method described below can be applied to each individual subaperture, as described in more detail below. Before the measurement, the size and number of subapertures are selected. Their size depends on the number of streaks and the type of streaks (large-scale, small swirls, etc.) and is chosen based on practicality and computational effort.

[0082] The following is based on the Fig. 4, an embodiment of the procedure for determining the shape of the optical surface 12 of the test object 14 is described. In a step S1, as mentioned above, M interferometry J i with target phase steps generated by the first variation device 50 and variations of the further system parameters p generated by the at least one further variation device 52 k generated.

[0083] In step S2, the phase step sequence is estimated in the form of the desired phase step Δϕ(x, y). Depending on the generation of the phase steps, the information about the desired phase step is already available. If the interferometer is very stable, the desired values ​​can be used directly for the subsequent evaluation. Furthermore, in step S2, the temporal fluctuations δa i (x, y), δb i (x,y) and δϕ i (x, y) is set to zero. In a turbulent environment or with an uncalibrated phase shifter, an initial estimate of the achieved phase steps is performed.

[0084] An iteration loop is then executed in steps S3, S4 and S5, which terminates after reaching a termination criterion (step S6). In step S3, the stationary mean values ​​for ϕ i (x, y), a i (x, y) and b i (x, y), for which the individual interferograms J i , alternatively for ϕ i (x, y), a i(x, y) and V i (x, y). In the first iteration step, this is done using the initial values ​​of δa estimated in step S2 i (x, y), δb i (x, y) and δϕ i (x, y). In each iteration step j, an improved determination of ϕ i (x, y), a i (x, y) and b i (x, y), as explained in more detail below.

[0085] Subsequently, in step S4, the temporal fluctuations or the image-to-image deviations δa i (x, y), δb i (x, y) and δϕ i (x, y) are determined by approximation using the sensitivities mentioned above. In other words, the changes in the system parameters p k caused modifications δa i (x, y), δb i (x, y) and δϕ i(x, y) of the interferogram properties comprising brightness, modulation, and phase are determined based on the sensitivities. In step S4, an improved determination of the aforementioned parameters is also carried out in each iteration loop j. Based on the determined deviations δa i (x, y), δb i (x, y) and δϕ i (x, y), in step S3 of the following iteration step, the time-invariant components a(x,y), b(x,y) and ϕ(x,y) of the brightness, the modulation and the phase are determined from equations (2a) to (2c).

[0086] In step S5, residuals are determined, and in step S6, a check is made on the basis of these residuals to determine whether the iteration loop can be terminated. If the loop is terminated, the aforementioned penetrating fringes can be subtracted in an optional step S7. This means that remaining errors in the interferograms corrected by the previous iteration loop, i.e., the interferograms after removing the errors related to the system parameters p k decreasing error influences are eliminated using an optimization algorithm.

[0087] In the following illustrations, the label (j) for the j-th iteration step is omitted. Only (j - 1) is used to label values ​​from the previous iteration step. In the j-th iteration step of the iteration loop comprising steps S3 to S5, the improved values ​​for a(x, y), b(x, y), and ϕ(x, y) result from the previous estimates a (j-1) (x, y) and b (j-1) (x, y), the estimated phase steps Δϕ i (x, y) and the deviations δai(j−1)(x, y), δbi(j−1)(x, y) and δϕi(j−1)(x, y). Equations (1) as well as (2a), (2b) and (2c) are used for the improvement according to the well-known rules of the sum of squares minimization over all interferograms (least squares method): mina, b, ϕ∑i=1M(Ji(x, y)−gi0(x, y) a(x, y)−gi1(x, y) b(x, y) cosϕ(x, y)−gi2(x, y) b(x, y) sinϕ(x, y))2 with gi0(x, y)=1+δai(j−1)(x, y)a(j−1)(xy)=1+∑k=1Ksak(x, y)δpki(j−1) gi1(x, y)=(1+δbi(j−1)(x, y)b(j−1)(xy)) cos(Δϕi(x, y)+δϕi(j−1)(x, y))=(1+∑k=1Ksb,k(x, y)δpk,i(j−1)) cos(Δϕi(x, y)+∑k=1Ksϕk(x, y)δpki(j−1)) gi2(x, y)=−(1+δbi(j−1)(x, y)b(j−1)(xy)) sin(Δϕi(x, y)+δϕi(j−1)(x, y))=−(1+∑k=1Ksb,k(x, y)δpk,i(j−1)) sin(Δϕi(x, y)+∑k=1Ksϕk(x, y)δpki(j−1))

[0088] G i0 (x, y), g i1 (x, y), and g i2 (x, y) are given by the previous estimates, a(x, y), b(x, y) cosϕ(x, y) and b(x, y) sinϕ(x, y) are determined by minimizing the sum of squares of the errors over all interferograms, from which results for a(x, y), b(x, y) and ϕ(x, y) follow.

[0089] The improved determination of the deviations δa i (x, y), δb i (x, y) and δϕ i (x, y) is based on the basic principle that these deviations are the differences between the measured interferogram J i(x, y) and the model formed from the estimates for a(x, y), b(x, y) and ϕ(x,y) according to (11). The three terms c i0 (x, y), c i1 (x, y), c i2 (x, y), from which the deviations can be determined according to (13a,b,c). δJi(x,y)=Ji(x,y)−ci0(x,y)a(x,y)−ci1(x,y)b(x,y)cosϕi'(x,y)−ci2(x,y)b(x,y)sinϕi'(x,y) with ϕi'(x,y)=ϕ(x,y)+Δϕi(x,y) ci0(x,y)=1+δai(x,y)a(x,y)=1+∑k=1Ksak(x,y)δpki(j) ci1(x,y)=(1+δbi(x,y)b(x,y))cosδϕi(x,y)=(1+∑k=1Ksbk(x,y)δpki(j))cos(∑k=1Ksϕk(x,y)δpki(j)) ci2(x,y)=−(1+δbi(x,y)b(x,y))sinδϕi(x,y)=−(1+∑k=1Ksbk(x,y)δpki(j))sin(∑k=1Ksϕk(x,y)δpki(j))

[0090] The solutions for c i0 (x, y), c i1 (x, y), c i2 (x, y) can be determined using various fitting processes that are generally known to those skilled in the art. The results are then the coefficients determined in iteration step j. δpki(j) of the sensitivities k for each interferogram i. Using Eq. (6a,b,c) the coefficients are converted into the remaining deviations δa i (x, y), δb i (x, y), δϕ i (x, y).

[0091] Then a(x, y), b(x, y) and Φ(x,y) are recalculated as above with reference to (9) to (10c).

[0092] In order to calculate the influence of streaks and air turbulence from the measurement result, according to one embodiment, sub-areas of the generated interferograms 46 are evaluated, which are assigned to different sub-apertures of the test wave 34. A sub-aperture of the test wave 34 is understood to be that part of the test wave 34 that illuminates only a partial area of ​​the entire section of the optical surface 12 of the test object 14 illuminated by the test wave 34.

[0093] The subapertures described below have the dimensions Δx and Δy. Three implementation variants for subaperture analysis are explained below.

[0094] In the first version, bicubic C 2 - Splines usage. The coordinates within the subaperture with the center pixel x m , y n are given by: X=xm−Δx2…xm+Δx2 Y=yn−Δy2…yn+Δy2

[0095] The equations (13a,b,c) are populated with, for example, bi-cubic C 2 -Splines ci0(X,Y)=∑j=03∑k=03aijk0XjYk ci1(X,Y)=∑j=03∑k=03aijk1XjYk ci2(X,Y)=∑j=03∑k=03aijk2XjYk

[0096] By fitting to each subaperture, 3 x 16 coefficients a ijk0 , a ijk1 and a ijk2 As with bicubic C 2-splines, the connection conditions of the subapertures are taken into account, ie the polynomial solutions must be twice continuously differentiable in the x- and y-direction. In Fig. Figure 5 gives an example of the Schlieren distribution in a phase image, which is represented by 23 x 23 subapertures and approximated by bicubic C 2 -splines is described.

[0097] According to the second variant of the subaperture evaluation, the following version (16b,c) of bicubic C can also be used as a modified variant of Eq. (15b,c) 2 - Splines with cos / sin weighting are used. This is used when the number of fringes in the interferogram is very high, thus requiring small subapertures. In this variant, the subapertures can be larger because only the modulations of the fringes are fitted (16a still applies): ci1(X,Y)=(∑j=03∑k=03aijk1XjYk)cos(∑j=03∑k=03aijk2XjYk) ci2(X,Y)=(∑j=03∑k=03aijk1XjYk)sin(∑j=03∑k=03aijk2XjYk)

[0098] According to the third variant of subaperture evaluation, the subaperture is continuously weighted. A subaperture is defined with the coordinates X, Y, with the mean coordinate x m , y n and the width Δx, Δy according to (14a,b). For the subaperture, the constants c i0 (x m ,y n ), c i1 (x m , y n ) and c i2 (x m , y n ) determined by a fit to δJi(X,Y)=Ji(X,Y)−ci0(xm,yn)a(X,Y)−ci1(xm,yn)b(X,Y)cosϕi'(X,Y)−ci2(xm,yn)b(X,Y)sinϕi'(X,Y)

[0099] For the fit, the sum of squares to be minimized is assigned a weight w(X, Y) to favor the middle pixels: minci0,ci1,ci2∑i=1M(w(X,Y)δJi(X,Y))2

[0100] The weighting function can be specified, for example, by a Gaussian function: w(x,y)=12πσe−12σ2((xΔx)2+(yΔy)2)

[0101] The width of the weighting function is chosen by σ. The subaperture is then shifted over all valid pixels, so that complete, spatially resolved data sets for c i0 x, y, c i1 (x, y) and c i2 (x, y). The order of subaperture processing is arbitrary.

[0102] As termination criterion for the test according to step S6 in Fig. 4 The quality or the course of the deviation of the model from the measured values ​​serves as the iteration. For example, the spatially resolved standard deviation of all intensity deviations can be calculated from (11): QJ(x,y)=1M∑i=1MδJi(x,y)2

[0103] If all Q J(x, y) are below a threshold or the iteration no longer yields any improvement, the process can be aborted. Alternatively, the mean or maximum value of the data array can be used as the termination criterion. In some cases, the magnitude of the phase improvement |δϕ i (x, y) - δϕ i-1 (x, y)| can be used for a quality measure or progress check according to: Qϕ(x,y)=1M∑i=1M(δϕi−1(x,y)−δϕi(x,y))2, However, the iterative approach of the inventive algorithm is based on adaptation of δJ i (x, y) so that a convergence of Q ϕ (x, y) may not be given.

[0104] Since the "zero state" of the measuring system is unknown, an unknown residue of the sensitivities, also referred to as streaks, may remain in the result. In the above-mentioned optional step S7, the final result, cleaned of residues, is determined. In the process, the corrected interferograms are adapted to predetermined error signatures. According to one embodiment, an objective function is defined in which errors in the distribution of the phase difference over one or more of the above-mentioned sensitivities, for example in the form of a sensitivity matrix, are represented by travel distances of the associated system parameters p. s and p kThis objective function is then optimized, specifically minimized, using an optimization algorithm. The optimization algorithm can be based, for example, on a least-squares optimization method. Using the resulting travel results, the errors associated with the corresponding system parameters are then subtracted from the phase difference.

[0105] According to one embodiment, the removal of the penetrating strips is carried out by means of the following procedure: The breakthrough fringes (see Eq. 7a,b,c) are subtracted from the determined quantities a(x, y), b(x, y) and ϕ(x, y) by separating all long-wavelength terms until the breakthrough fringes become significant, fitting the breakthrough fringes and subtracting them from the remaining signal, and adding the separated long-wavelength terms back to the cleaned signal.

[0106] In Fig. An example is shown in Figure 6. In this example, twice the number of fringes in the interferogram (a) is evident in the result (b). The interfering fringes are also modulated by a quadratic and constant function. After subtracting the corresponding combination of fringe frequency and modulation, the fringe artifacts are eliminated (c).

[0107] In the case of streaks and air turbulence, the penetrating stripes may be finely wave-modulated, so that the deduction of remaining penetrating stripes is carried out on a sub-aperture grid according to a further embodiment.

[0108] The modulations of the penetrating stripes are then calculated according to (7a,b,c) e.g. in bicubic C 2 -splines, fitted to each subaperture and subtracted from the determined quantities a(x, y), b(x, y) and ϕ(x,y) according to the procedure described above.

[0109] Alternatively, the penetrating stripes modulated by schlieren and air turbulence can be subtracted using a moving, weighted subaperture (cf. Fig. 7). The corrections to be deducted in the subaperture for a(x, y), b(x, y), and ϕ(x, y) are assigned a weighting function for the fit as given in Eq. (19). The subaperture is shifted over all pixels to achieve a pixel-resolved overall correction of the penetrating fringes.

[0110] Subtracting sensitivities and streaks carries the risk of subtracting not only the streak artifacts but also a portion of the test specimen deformation to be determined. To reduce this undesired subtraction, according to one embodiment, the aforementioned interferograms 46 form a first interferogram data set, and any errors remaining after eliminating error influences are corrected using at least one further interferogram data set.

[0111] In other words, it is advantageous to average several intererometric measurements with different parameters. Fig.Figure 8 shows an example of an interferogram data set 60 which involves averaging out streaks in the interferograms 46. For example, the streak positions of interferograms 46 are specified therein in N = 11 different rotational positions. By performing a phase measurement in each streak rotational position and calculating the mean value from all 11 measurements, the potential unwanted deduction is reduced to 1 / 11. This is possible because the streaks occupy only a small space in the Fourier spectrum. By changing the number and position of the streaks, the Fourier spectrum can be shifted into other areas without overlap. The maximum possible measurement error is given by the mean value of the deducted streak amounts. The probable measurement error is likely to be significantly lower, since random deformation of the test object in the form of streaks is unlikely.

[0112] According to one embodiment, the possible residual error is systematically minimized by a fitting algorithm in the following way: Given N measurements ϕ i (x, y) with different fringe positions: i = 1, ..., N. The different fringe positions can be achieved by suitable system changes, e.g. changing the tilt of the reference in the form of the Fizeau element 28 or the test object 14, changing the wavelength, changing the position of other optical elements in the beam path, changing the position of the light source etc... The effect of the changes on the phase is described according to (6c) by δϕ(x,y)=∑k=1Ksϕk(x,y)δpk

[0113] The test object 14 is assumed to be undeformable.

[0114] The penetrating stripes are described according to (7c) by δϕd(x,y)=∑k=1K∑l=1Msϕkl(x,y)(uϕklcos(l ϕ(x,y))+vϕklsin(l ϕ(x,y)))

[0115] It is further assumed that all measurements were transferred to the same coordinate system by a suitable (possibly non-linear) coordinate transformation, so that in each measurement the coordinates x,y correspond to identical points on the test object 14. Then, the best-fitting system changes and breakthrough stripes are determined by solving the minimization problem in (22): minϕ0,Δpϕ,k(i),Δpp,k(i),uϕkl(i),vϕkl(i)∑iN{ϕi(x,y)−[ϕ0(x ,y)+∑k=1KΔpk(i)sϕk(x,y)+∑k=1K∑l=1Msϕkl(x,y)(uϕkl(i)cos(l ϕi(x,y))+vϕkl(i)sin(l ϕi(x,y)))]}2

[0116] ϕ0 (x, y) corresponds to the mean solution for the phase, corrected for sensitivities and streaks.

[0117] After finding the solutions for ϕ0,Δpϕ,k(i),Δpp,k(i),uϕkl(i)andvϕkl(i) By means of a suitable fitting process (e.g. minimization of small sum of squares), the solution for ϕ0 (x, y) is obtained, which largely does not contain any adjustment-related artifacts or streaks.

[0118] The method can be applied to all test specimen deformations that can be described by sensitivities, in which equipment requirements (limits, stops, mechanics, stabilizers, etc.) ensure that the corresponding system changes do not occur for at least some of the measurements.

[0119] The above description of exemplary embodiments, embodiments, and variants is to be understood as exemplary. The disclosure thus made enables those skilled in the art, on the one hand, to understand the present invention and the associated advantages, and, on the other hand, also encompasses obvious variations and modifications of the described structures and methods within the understanding of those skilled in the art. Therefore, all such variations and modifications, insofar as they fall within the scope of the invention as defined in the appended claims, as well as equivalents, are intended to be covered by the claims. List of reference symbols 10 Measuring system 12 optical surface 14 test object 15 Interferometry module 16 Measuring radiation source 18 Measuring radiation 20 waveguides 21 Exit surface 22 Radiation generation module 24 beam splitters 26 Collimator 28 Fizeau element 29 Fizeau area 30 Reference wave 31 Input shaft 32 diffractive optical element 34 test shaft 34r returning test wave 36 Recording device 38 Camera 40 aperture 42 eyepiece 43 detection area 44 Detector 46 Interferogram 48 Evaluation device 50 first variation device 52 additional variation devices 54 first system parameters 56 additional system parameters 58 Sensitivity 58o orthonormal sensitivity 60 interferogram data sets

Claims

[1] Method for the interferometric determination of a spatial distribution of an optical property of a test object (14) by means of an interferometric measuring system (10), comprising the steps: - irradiating a test wave (34) generated by a diffractive optical element (32) onto the test object and generating a plurality of interferograms (46) in chronological sequence by superimposing a reference wave (30) with the test wave (34) generated by the diffractive optical element after its interaction with the test object, wherein during the generation of the interferograms a first system parameter (54) of the measuring system is changed to vary a phase difference between the test wave generated by the diffractive optical element and the reference wave and wherein, furthermore, during the generation of the interferograms, at least one further system parameter (56) of the measuring system is varied, wherein a change in the at least one further system parameter (56) has an influence on a phase difference distribution on a detection surface (43) of a detector (43) serving to detect a generated interferogram (46), and - Determining the local distribution of the optical property by evaluating the generated interferograms, wherein during the evaluation, error influences attributable to the varied system parameters (54, 56) are eliminated by taking into account changes in the interferograms caused by the variation of the system parameters when determining the local distribution of the optical property, wherein the evaluation of the generated interferograms (46) is based on at least one sensitivity (58) of the further system parameter (56), which indicates a relationship between the further system parameter (56) and at least one property of the interferograms. [2] Method according to claim 1, wherein at least two further system parameters (56) are varied during the generation of the interferograms. [3] Method according to claim 1 or 2, wherein the at least one further system parameter (56) is selected from the following group of system parameters: a position and a rotational position of at least one optical element of the measuring system, a position and a rotational position of the test object, a temperature of the at least one optical element, a temperature, a pressure, a humidity and a composition of at least one medium between optical elements of the measuring system, a size, a position of a measuring radiation source of the measuring system, a wavelength, an intensity, a polarization and a degree of coherence of a measuring radiation generated by the measuring radiation source, and a position, a rotational position, a temperature and an exposure time of a camera of the measuring system. [4] Method according to one of the preceding claims, in which, during the evaluation of the generated interferograms (46), a modification of the at least one property of the interferograms caused by the variation of the system parameters is determined on the basis of the at least one sensitivity (58) and a time-invariant component of the at least one property is determined by means of the determined modification. [5] Method according to one of the preceding claims, wherein the at least one property of the interferograms designates a local brightness distribution, a local distribution of a modulation amplitude and / or a local phase distribution in the interferograms (46). [6] Method according to one of the preceding claims, in which the at least one sensitivity (58) is determined by simulation and / or experimentally before the generation of the interferograms (46). [7] Method according to one of the preceding claims, in which a plurality of sensitivities (58) are determined before the generation of the interferograms (46) and the evaluation of the generated interferograms is carried out on the basis of a selection of the determined sensitivities. [8] Method according to claim 7, wherein the selection of the sensitivities comprises assembling an orthogonal set of sensitivities (58o). [9] Method according to one of the preceding claims, in which a plurality of sensitivities (58) are determined before the generation of the interferograms (46) and, during the evaluation of the generated interferograms, a respective correction of the individual interferograms is carried out on the basis of at least one sensitivity selected from the determined sensitivities. [10] Method according to one of claims 4 to 9, in which a plurality of sensitivities (58) are determined before the generation of the interferograms (46) and, during the evaluation of the generated interferograms, a correction of an intermediate result of the spatial distribution of the optical property determined from the totality of the interferograms is carried out on the basis of at least one sensitivity selected from the determined sensitivities. [11] Method according to one of the preceding claims, in which, before the generation of the interferograms (46), a sequence of steps of the varied system parameters is determined by simulation and / or experimentally. [12] Method according to one of the preceding claims, in which only a respective partial area of ​​the generated interferograms (46) assigned to a sub-aperture of the test wave (34) is evaluated and the local distribution determined in this way is combined with a further local distribution which is determined from a respective other partial area of ​​the interferograms. [13] Method according to one of the preceding claims, in which, after the error influences attributable to the system parameters (54, 56) have been eliminated, remaining errors in the correspondingly corrected interferograms are eliminated by means of an optimization algorithm. [14] Method according to one of the preceding claims, in which the generated interferograms (46) form a first interferogram data set (60) and, after the elimination of error influences, remaining errors are corrected using at least one further interferogram data set. [15] Measuring system (10) for the interferometric determination of a spatial distribution of an optical property of a test object (14) with: - an interferometry module (15) for irradiating a test wave (34) generated by a diffractive optical element (32) onto the test object and generating a plurality of interferograms (46) in chronological sequence by superimposing a reference wave (30) with the test wave (34) generated by the diffractive optical element (32) after its interaction with the test object, - a first variation device (50) which is configured to change a first system parameter (54) of the measuring system during the generation of the interferograms in order to vary a phase difference between the test wave generated by the diffractive optical element and the reference wave, - at least one further variation device (52) which is configured to vary at least one further system parameter (56) of the measuring system during the generation of the interferograms, wherein a change in the at least one further system parameter (56) has an influence on a phase difference distribution on a detection surface (43) of a detector (43) used to detect a generated interferogram (46), and - an evaluation device (48) which is configured to determine the spatial distribution of the optical property by evaluating the generated interferograms and, in doing so, to calculate out error influences attributable to the varied system parameters by taking into account changes in the interferograms which are caused by the variation of the system parameters, wherein the evaluation of the generated interferograms (46) is based on at least one sensitivity (58) of the further system parameter (56), which indicates a relationship between the further system parameter (56) and at least one property of the interferograms.

Images