Method and apparatus for measuring transparent multi-layer objects using a coherence tomograph

Abstract

The invention relates to a method for measuring multi-layer objects (30) using a coherence tomograph (11), in which method measurement light in an object arm (18) is directed by an optical system (25) to a surface of the object (30). The interference between measurement light (OL), which is reflected at an optical interface (40, 42, 44) of the object (30), and measurement light (RL), which is guided in the reference arm (16), is detected by a detector (36) and converted into electrical signals. A spectrum is generated therefrom by means of a Fourier transform, wherein each interference is represented in the spectrum by a distance peak (D1, D2, D3). In a process automatically controlled by a control device (39), this measurement is repeated multiple times, the distance between the object (30) and a reference plane (52) being varied in each repetition, the axial position of the reference plane corresponding to an optical path length of the reference arm (18). By analysing the distance peaks (D1, D2, D3) and preferably thickness peaks representing a distance between interfaces, an unambiguous assignment of the distance peaks to the interfaces (40, 42, 44) is determined such that distance values can be assigned to those interfaces.

Description

[0001] Method and device for measuring transparent multilayer objects with a coherence tomograph

[0002] BACKGROUND OF THE INVENTION

[0003] 1. Field of the invention

[0004] The invention relates to a method and a device for measuring transparent multilayer objects using SD OCT, where SD is the acronym for spectral domain and OCT is the acronym for optical coherence tomography.

[0005] 2. Description of the State of the Art For measuring surface profiles of workpieces and other objects, devices based on the principle of SD OCT have been increasingly used for several years. These devices measure the distance to a scattering or at least partially reflective surface of the object without contact at individual measuring points. For this purpose, the interference of light reflected from the surface with light guided in a reference arm of the OCT is detected and spectrally evaluated. Repeating this measurement at a large number of measuring points yields a surface profile of the object.

[0006] If the object is sufficiently transparent to the measuring light, the distances to internal optical interfaces can also be measured with high accuracy. Such optical interfaces typically exist in multilayer objects, e.g., bonded wafers or coated workpieces. If components of the measuring light reflected from two different interfaces interfere, the coherence tomograph measures the distance between the two interfaces, i.e., the thickness of one of several wafers or a coating. If the light propagation is temporarily blocked in the reference arm, only thicknesses, and not distances, are measured. By combining distance and thickness measurements, it is possible, for example, to verify whether the surface of a coated workpiece has the desired shape and whether the coating thickness is constant throughout.

[0007] The first devices of this type could only determine the distance and / or thickness for a single measuring point. To scan an area, the workpiece had to be moved relative to the measuring device (or vice versa). DE 10 2022 104 416 A1 discloses a device in which a scanning unit guides the measuring beam across the workpiece. The applicant markets such devices under the name "Flying Spot Scanner".

[0008] SD OCT measuring devices generate raw measurement data in the form of interference spectra, from which the desired distance or thickness values ​​are derived by Fourier transformation. Each distance or thickness value corresponds to a peak in the transformed spectrum, which can be displayed on a screen.

[0009] For example, when using an SD OCT measuring device to measure distances and thicknesses of a glass plate, three peaks are displayed when using the reference arm: two peaks for the distances to the two interfaces of the plate and one peak for its thickness. If the glass plate is moved parallel to the direction of the measuring beam, the two distance peaks move, while the thickness peak remains stationary, since the thickness of the glass plate does not change with movement in this direction. In this way, the displayed peaks can usually be easily assigned to the quantities being measured. If the light propagation in the OCT's reference arm is interrupted, only the thickness peak is displayed, while the two distance peaks disappear. This allows the thickness peaks to be clearly identified.

[0010] However, even with only two interfaces, the spectrum obtained through the Fourier transform can sometimes become unclear, namely when multiple reflections occur. Multiple reflections occur when light that has already been reflected once at a first interface is reflected again at another interface on its way back to the coherence tomograph. For example, with a glass plate that has an upper and a lower interface, light reflected at the lower interface can be reflected from the inside at the upper interface, directed back towards the lower interface, and then reflected again towards the coherence tomograph. Although the light, reflected a total of three times, usually has only a low intensity, it can still, after interference with the strong signal from the reference arm, lead to a distance peak visible in the spectrum, which appears to represent an interface at a greater distance.At interfaces that strongly reflect the measuring light, multiple reflections with five or seven reflections can even lead to recognizable peaks in the spectrum.

[0011] When measuring an object consisting of two layers, even without multiple reflections, the reference arm produces six peaks that must be assigned to the quantities being measured: three distance peaks, one thickness peak for each layer, and one thickness peak representing the total thickness of the layer system. A three-layer system results in four distance peaks and six thickness peaks, and so on. If additional peaks are then added due to the aforementioned multiple reflections, it becomes difficult or even impossible for the operator to maintain an overview and assign the peaks to the correct interfaces.

[0012] Additional problems can arise during evaluation if thickness and distance peaks overlap completely or partially, so that they are poorly resolved or not detected at all.

[0013] A further difficulty in evaluating the peaks arises from the fact that the position of the OCT's reference plane relative to the object is unknown. This reference plane is located in the object arm of the coherence tomograph at an axial position corresponding to the optical path length of the reference arm. By evaluating the interference patterns, the OCT effectively measures the distances of the interfaces to this reference plane. However, the distance peaks contain no information about whether the interface is located below or above the reference plane. This leads to the distance peaks shifting in different directions during the aforementioned axial movement of the object.

[0014] This can become particularly critical when an interface is located near the reference plane. If the interface is located at two different points, once by an amount x above the reference plane and once by an amount x below the reference plane, this leads to identical distance peaks in the spectrum, even though the distance of the interface between the two points differs by an amount 2x. The complexity of evaluating the spectra can therefore, in the worst case, lead to undetected measurement errors.

[0015] The preceding explanations have made it clear that when measuring multi-layered transparent objects, it can be very tedious, error-prone, and sometimes impossible with reasonable effort to assign the displayed distance peaks to the individual interfaces of the object using the measures described above, namely moving the object and / or dimming the reference arm.

[0016] SUMMARY OF THE INVENTION

[0017] The object of the invention is to provide a method and a device for measuring transparent multilayer objects using a coherence tomograph having an object arm and a reference arm, wherein the detected distance peaks in the spectrum can be more easily and reliably assigned to the distances in the method or device.

[0018] A method according to the invention that solves this problem comprises the following steps: a) Measurement light generated by the coherence tomograph is directed by an optical system in the object arm onto a point on the surface of the object; b) Measurement light reflected at an optical interface of the object passes back into the coherence tomograph via the optical system and interferes there with measurement light guided in the reference arm; c) The interference obtained in step b) is detected by a detector of the coherence tomograph and converted into electrical signals; d) A spectrum is generated from the electrical signals obtained in step c) by Fourier transformation, wherein each interference obtained in step b) is represented by a distance peak in the spectrum;e) In a process automatically controlled by a control unit, steps a) to d) are repeated multiple times, with each repetition changing the distance between the object and a reference plane located in an object arm of the coherence tomograph, the axial position of which corresponds to an optical path length of the reference arm; f) An evaluation unit determines, by analyzing the distance peaks obtained in step e) and preferably also thickness peaks corresponding to a distance between two optical interfaces (40, 42, 44), a unique mapping between the distance peaks and the optical interfaces and assigns distance values ​​to the optical interfaces on the basis of this mapping.

[0019] The invention is based, firstly, on the realization that the steps previously performed by an operator in finding a correct correspondence between distance peaks in the spectrum and the optical interfaces of the object can be automated and thus performed much faster and more reliably. Above all, however, automated correspondence enables the execution of certain test steps that cannot be performed by an operator, or only with great difficulty.

[0020] For example, in more complex spectra, such as those occurring in objects with three or more layers, it is only possible to reliably determine, using computer-aided methods, which distance peaks are due to multiple reflections and thus only seemingly represent a single optical interface.

[0021] If thicknesses are also to be measured, i.e., distances between optical interfaces, the procedure can easily be extended by having the evaluation device, in step f), also determine a unique mapping between the thickness peaks corresponding to a distance between two optical interfaces and the distances between the optical interfaces through analysis, and assign thickness values ​​to the distances between the interfaces on the basis of this mapping.

[0022] Especially in the context of multiple reflections, it is relevant that in step e), for thickness peaks corresponding to a distance between two optical interfaces, it is determined whether there are pairs of distance peaks, each corresponding to a distance from one of the two optical interfaces. For this purpose, the algorithm can, for example, calculate all conceivable differences for all distance peaks, i.e., the peaks that move within the spectrum during the change in step e). Those distance peaks where the difference equals the thickness peak can be assigned to the two optical interfaces that bound the layer. If, however, no two such distance peaks can be identified for a thickness peak, it is assumed that the thickness peak does not represent a real layer but was caused by multiple reflections.

[0023] Conversely, the same applies. If no corresponding thickness peak can be found for two distance peaks, it can be assumed that at least one of the two distance peaks was caused by a multiple reflection.

[0024] However, thickness peaks caused by multiple reflections are often accompanied by corresponding distance peaks. In these cases, a further test criterion can provide clarity, which can only be reliably applied using computer-aided methods. According to this test criterion, step e) determines whether, for a thickness peak representing the distance between two interfaces, there is another thickness peak that corresponds to a multiple of this distance. The underlying assumption is that thickness peaks (recognizable by the fact that they do not change their position in the spectrum when modified in step e) are accompanied by further thickness peaks corresponding to multiples of the distance in the case of multiple reflections. If this condition is met, these further thickness peaks are usually caused by multiple reflections and can be ignored during the assignment. The same applies to the distance peaks that correspond to these further thickness peaks.

[0025] Assigning peaks to interfaces and thicknesses is particularly difficult when the object contains several layers of approximately the same thickness. The thickness peaks then overlap and can no longer be resolved and assigned to the individual layers. In this case, however, the method according to the invention allows at least the distance peaks to be unambiguously assigned to the interfaces, even with complex spectra, so that the desired thicknesses can be determined by calculating the difference between the measured distances.

[0026] The number of layers, and sometimes their approximate (target) thickness, is usually known in advance. This information can be entered into the measuring device, which can then use it to enable a clear assignment in cases of ambiguity. For example, if two thickness peaks appear in the spectrum, representing thicknesses T and 2T, the second thickness peak could be caused by multiple reflections and therefore be ignored. Previously entered information indicating that the object actually has two layers, one of which is approximately twice as thick as the other, can then be used for a clear assignment.

[0027] The method can also be used to determine the position of the reference plane relative to the object. The object can then be positioned so that the reference plane lies outside the plane on one side of the object. This is a significant advantage of the method, as it makes evaluating the peaks considerably simpler and more reliable. If all interfaces lie on one side of the reference plane, the situation described above, where it is impossible to determine from the peak whether an interface is above or below the reference plane, and thus, for example, whether it is concave or convex, cannot occur. Positioning a surface of the object directly on the reference plane should be avoided, as the identical optical path lengths would then result in no interference signal and therefore no distance peak for that interface.

[0028] One way to change the distance between the object and the reference plane in step e) is to move the object using a lifting table controlled by the control unit. This is generally simpler than changing the optical path length in the reference arm using path length modulators, such as those known from DE 10 2013 008 269 A1.

[0029] Often, the task involves measuring objects not just at one point, but at several points, in order to obtain a surface or thickness profile. For this purpose, the measuring light can be guided across the object using a scanning device.

[0030] For objects with uniform thickness layers, it may be sufficient to perform the procedure only once at one point and omit steps e) and f) when measuring the remaining points. Instead, modified distance values ​​are assigned to the optical interfaces based on a previously determined mapping, taking into account shifts in the distance peaks within the spectrum. The underlying rationale is that the distance peaks shift only slightly within the spectrum in this case, ensuring that the initial mapping is not lost. When measuring the remaining points, a specific distance peak, which shifts only slightly from point to point, always represents the same optical interface.

[0031] However, for objects whose layers do not have a uniform thickness, it will usually be necessary to perform the procedure several times, since a previously made assignment between distance peaks and optical interfaces loses its validity, e.g. as a result of a shift in the reference plane or because additional peaks are added by multiple reflections or peaks disappear from the spectrum.

[0032] To ensure that the distance and thickness peaks stand out as clearly as possible against the unavoidable background noise, the measuring light should be focused by the optical system at a focal point located near the object. The closer the focus is to the optical interface being measured, the higher the intensity of the measuring light reflected from the interface back to the coherence tomograph.

[0033] In one embodiment, therefore, prior to step a), an automated pre-adjustment controlled by the control unit determines the focus position relative to the object at which the distance peaks stand out most clearly from the noise background. In the simplest case, this determination can be based on the distance peaks having the highest average intensity. However, it is often more important that all distance peaks can be detected at all than the maximum average intensity. Therefore, exceeding a minimum intensity or a minimum distance from the noise background is often a more suitable criterion for pre-adjustment. Additionally or alternatively, the thickness peaks can also be considered in an analogous manner.

[0034] One way to pre-adjust the focus is to change its position relative to the object and measure the intensity of at least one distance or thickness peak. This makes it very easy to determine the optimal focus position relative to the object without any additional tools.

[0035] The position of the focus relative to the object can be changed, for example, by moving the object using a lifting table. Alternatively, the focus can be changed by an axial movement of the entire optical system (if the measuring light from the light source is collimated) or by an optical system with a variable focal length. For this purpose, the optical system can include movable lenses or lens groups, a liquid lens, and / or another optical element with a variable focal length.

[0036] Instead of the intensity measurement described above, a camera can measure the position of the focus relative to the object. If the measuring light has wavelengths in the infrared spectral range, the camera should be sensitive to this wavelength range so that the focus is recognizable in the images captured by the camera. Using a camera to record the relative position between the focus and the object is particularly useful when the object has steeply inclined surfaces. In such cases, little or no light may be reflected towards the coherence tomograph, preventing a meaningful intensity measurement.

[0037] The camera could, for example, be a (stereo) camera with a camera lens whose optical axis is arranged at an angle between 5° and 90° to an optical axis of the optical system.

[0038] Both approaches – the intensity-based approach and the camera-based approach to determining the focus position relative to the object – can also be combined. This can be achieved, for example, by performing a coarse adjustment based on the focus position measured by the camera relative to the object, in which the focus position is changed. This is followed by a fine adjustment in which the focus position relative to the object is changed again, and the height of at least one peak is measured.

[0039] Additionally, during pre-adjustment, the tilt angle of the stage on which the object is placed can be changed. This allows the object's interfaces to be aligned so that they are essentially perpendicular to the measuring light beam.

[0040] The invention further relates to a method for measuring multilayer objects with a coherence tomograph having an object arm and a reference arm, the method comprising the following steps: a) measuring light generated by the coherence tomograph is directed by an optical system in the object arm onto a point on the surface of the object and thereby focused; b) measuring light reflected at an optical interface of the object passes back into the coherence tomograph via the optical system and interferes there with measuring light guided in the reference arm; c) the interference obtained in step b) is detected by a detector of the coherence tomograph and converted into electrical signals; d) a spectrum is generated from the electrical signals obtained in step c) by Fourier transformation, wherein each interference obtained in step b) is represented by a distance peak in the spectrum;e) Before step a), an automated pre-adjustment is performed by the control unit to determine the position of the focus relative to the object at which the peaks stand out most clearly from a noise background.

[0041] With regard to the device, the aforementioned problem is solved by a device for measuring multilayer objects comprising: a) an optical system; b) a coherence tomograph comprising a detector, an object arm, and a reference arm, configured to generate measurement light and direct it via the optical system in the object arm onto a point on the surface of the object, wherein the detector is configured to detect interference between measurement light reflected at an optical interface of the object and returned to the coherence tomograph via the optical system, and measurement light reflected in the reference arm, and to convert the detected interference into electrical signals; c) an evaluation unit configured to generate a spectrum from the electrical signals generated by the detector by means of a Fourier transform.in which each detected interference is represented by a distance peak in the spectrum, d) a control device configured to perform multiple measurements in which a distance between the object and a reference plane located in an object arm of the coherence tomograph is changed, the axial position of which corresponds to an optical path length of the reference arm, wherein the evaluation unit is configured to make a unique mapping between the distance peaks and the optical interfaces by analyzing the obtained distance peaks and to assign distance values ​​to the optical interfaces on the basis of this mapping.

[0042] The device can have a lifting table that allows the distance between the object and the reference plane to be changed.

[0043] BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Exemplary embodiments of the invention are explained in more detail below with reference to the drawings. These show:

[0045] Figure 1 shows a measuring device according to an embodiment of the invention in a schematic representation;

[0046] Figure 2 shows several graphs illustrating essential steps in the evaluation of the electrical signals generated by the detector of the coherence tomograph;

[0047] Figure 3 shows several distance and thickness peaks in the spectrum obtained from the electrical signals of the coherence tomograph by Fourier transformation;

[0048] Figures 4a to 4e each show a simplified section of the measuring device according to Figure 1 with different distances between the object to be measured and a reference plane;

[0049] Figures 5a to 5e show the spectra obtained from the electrical signals of the coherence tomograph for the distances shown in Figures 4a to 4e; Figures 6a to 6c each show a simplified section of the measuring device according to the

[0050] Figure 1 shows different distances between the focus of the object light and the object;

[0051] Figures 7a to 7c show the spectra obtained from the electrical signals of the coherence tomograph for the distances shown in Figures 6a to 6c;

[0052] Figure 8 shows a section corresponding to Figure 6a with an additional stereo camera;

[0053] Figure 9 shows a further embodiment of a measuring device according to the invention in a representation based on Figure 1 with an additional scanning device; and

[0054] Figure 10 shows a measuring system with a cabin and a measuring device included therein in a simplified perspective view.

[0055] DESCRIPTION OF PREFERRED EXAMPLES

[0056] 1. First embodiment - Construction of the measuring device

[0057] Figure 1 shows a schematic representation of an SD OCT measuring device, designated 10, according to a first embodiment of the invention. The measuring device 10 comprises a coherence tomograph 11, which contains a broadband light source 12, which can, for example, be a superluminescent diode. In the illustrated embodiment, the measuring light used for the measurement is predominantly guided in optical fibers; however, it is also possible to guide the measuring light exclusively as a free beam, as is known in the prior art.

[0058] A first beam splitter 14 divides the measurement light generated by the light source 12 into reference light RL and object light OL. In general, in known SD OCT measuring devices, the measurement light is not divided equally, but asymmetrically. A smaller portion of the measurement light is directed by the first beam splitter 14 as reference light RL into a reference arm 16, and the larger remaining portion of the measurement light is directed as object light OL into an object arm 18. In fiber-based measuring devices, 2x2 fiber couplers are usually used as the beam splitter 14. The reference light RL is reflected as completely as possible in the reference arm 16 and returns to the first beam splitter 14 as the reference light component. Reflection in the reference arm 16 can be achieved using a fiber reflector (e.g., a Bragg reflector). A higher reflectance, independent of wavelength, is achieved with a conventional fixed mirror 20.The reference light RL emerging from the optical fiber can be collimated using a first lens 22 and focused onto the mirror 20 using a second lens 23.

[0059] The reference arm 16 includes a switchable dimming device 24, which allows the propagation of the reference light RL in the reference arm 16 to be interrupted as needed. In the illustrated embodiment, the dimming device 24 is designed as an iris diaphragm.

[0060] The measuring light, directed into the object arm 18 as object light OL, is focused onto the object 30 to be measured by means of an optical system 25, which in the illustrated embodiment is simplified to two lenses 26, 28. The object 30 consists of two adjacent glass plates 32, 34, both of which are transparent to the object light OL. In this context, transparent means that a detectable portion of the object light OL reflected from the interfaces of the lower glass plate 34 can exit the object 30 again. The transmittance of the object 30 should therefore be at least 5% and preferably more than 25%.

[0061] The object light OL is partially reflected or scattered at the interfaces of the two glass plates 32, 34. The reflected object light component passes through the lenses 26, 28 back into the optical fiber and interferes in the first beam splitter 14 with the reference light component from the reference arm 16. The interference signal is coupled out via a second beam splitter 31 and fed to a spectrometer 36, which detects the spectral intensity distribution of the interference signal. In the illustrated embodiment, the spectrometer 36 comprises a grating or other dispersive optical element and a row of light-sensitive cells that detect the intensity of the interference signal deflected by the dispersive optical element in a wavelength-dependent manner.

[0062] From this, an evaluation unit 38 connected to the spectrometer 36 can determine the distances between the interfaces of the glass plates 32, 34 and the measuring device 10, the thicknesses of the individual glass plates 32, 34, and the thickness of the entire object 30. The measuring device also includes a lifting table 33, which has a plate-shaped specimen carrier 35 and an actuator 37 designed as a telescopic cylinder. By actuating the actuator 37, the specimen carrier 35 can be moved variably along the direction indicated by a double arrow, which runs parallel to an optical axis of the optical system 25. The actuator 37 is controlled by a control unit 39.

[0063] 2. First Implementation Example - Evaluation

[0064] Figure 2 schematically illustrates how distance information is acquired in SD OCT. On the left, object 30 is shown with the two glass plates 32 and 34. At each of the three optical interfaces 40, 42, and 44, there is a change in refractive index, resulting in partial reflection of the object light OL, indicated by an arrow.

[0065] To the right of object 30, three interference spectra 1, 2, and 3 are shown. Interference spectrum 1 would be recorded by spectrometer 36 if only the first interface 40 were in the beam path of the object light OL. The reflection of the object light OL at the first interface 40 leads to an intensity modulation of the spectrum in k-space, which is proportional to the difference in optical path lengths traveled by the reference light RL in reference arm 16 and the object light OL in object arm 18. The desired distance information is thus encoded in the frequency at which the intensity oscillates in k-space.

[0066] Corresponding considerations apply to interference spectra 2 and 3 and the optical interfaces 42 and 44 associated with these interference spectra. The optical path length in the reference arm 16 is set such that the modulation of the intensity in k-space is higher in frequency the further the respective optical interface is from the coherence tomograph 40. The increasing modulation frequencies are clearly visible in interference spectra 1, 2, and 3.

[0067] Since the object light OL strikes not only interfaces 40 but also interfaces 42 and 44, the interference spectra 1, 2, and 3 overlap. The spectrometer 36 therefore only records the total spectrum 50 shown to the right, which represents an additive superposition of interference spectra 1, 2, and 3. The spectral components, i.e., the modulation frequencies of spectra 1, 2, and 3, can be obtained from the total spectrum 50 by means of a Fourier transform. This is shown on the right in Figure 2. In the example shown, a peak is obtained for each optical interface 40, 42, and 44, corresponding to a specific distance value z and assigned to one of the interference spectra 1, 2, and 3.

[0068] The spectrometer 36 also detects interferences that do not result from the interference of the reference light component with the object light component described above, but rather from interferences between parts of the object light component itself, from interferences with unwanted reflections at other optical interfaces (e.g., the lenses 22, 23 in the reference arm 16), or from interferences resulting from multiple reflections. A multiple reflection occurs, for example, when object light OL, which has been reflected at the optical interface 42, is partially reflected at the optical interface 40 and then reflected again at the optical interface 42. Compared to an object light component that has only been reflected once at the optical interface 42, such an object light component has traveled an additional optical path length corresponding to twice the thickness of the upper glass plate 32.These partially unwanted interferences also produce peaks after performing the Fourier transformation, which are added to the three peaks shown on the right in Figure 2 and make it difficult to assign the peaks to the optical interfaces 40, 42, 44.

[0069] Figure 3 shows an example of a spectrum I(z) in which the intensity I measured by the light-sensitive cells of the spectrometer 36 is plotted as a function of a distance z in the axial direction, with unavoidable background noise already subtracted for clarity. Five peaks P1 to P5 are visible in the spectrum I(z), as they would be displayed to an operator during the measurement of object 30 with its three interfaces 40, 42, 44. Undesired peaks, such as those that can be caused by the aforementioned multiple reflections, are also omitted for clarity.

[0070] Peak P3 in the middle is taller than the other peaks. This is due – although a user cannot definitively verify this – to the fact that two peaks, indicated by dots, largely overlap here. It is difficult for the operator to determine how the distance values ​​z represented by the peaks should be assigned to the interfaces 40, 42, 44 or the thicknesses of the glass plates 30, 32.

[0071] 3. First Implementation Example - Automated Assignment

[0072] In the following, with reference to Figures 4a to 4e and 5a to 5e, it is explained how, according to the invention, peaks in the spectrum, as shown in Figure 3, can be automatically assigned to the interfaces.

[0073] In Figure 4a, the optical system, indicated by reference numeral 25 at the top, directs the object light OL onto the object 30 and focuses it onto a focal point 54. The object 30 is represented here only by three interfaces 40, 42, and 44, corresponding to distances D1, D2, and D3, respectively. It is assumed that a first layer with a thickness T1 is located between the optical interfaces 40 and 42, and a second layer with a thickness T2 is located between the interfaces 42 and 44.

[0074] A dashed line indicates a reference plane 52 located in the object arm 18 of the coherence tomograph 10, the axial position of which corresponds to the optical path length of the reference arm 16. For example, if the optical path length in the reference arm 16 were lengthened by moving the mirror 20, this would shift the reference plane 52 a corresponding amount towards the object 30. The focal point 54 is only coincidentally located approximately in the reference plane 52; the latter should not be confused with the focal plane of the optical system 25. Further below, with reference to Figures 6a, 6c, and 7a to 7c, it is explained how the position of the focus 54 relative to the object 30 can be optimized.

[0075] The arrow to the right of object 30 indicates that object 30 is moved in several small steps towards the optical system 25 during the execution of the procedure. For this purpose, the control unit 39 controls the lifting table 33 accordingly.

[0076] Figure 5a shows the spectrum I(z) for the distance shown in Figure 4a between object 30 and the reference plane 52. The peaks are simplified here as lines in the spectrum and, for the sake of simplicity, labeled with the quantities they represent. Since the pixel row in the spectrograph 36 has a limited length, not all occurring frequencies can be recorded by the spectrograph 36. For this reason, the spectrum l(z) shown in Figure 5a is limited to a maximum distance value z. max cut off.

[0077] In the axial position of the lifting table 33 assumed in Figure 4a, the object 30 is located so far from the reference plane 52 that only the distance peaks D1, D2, but both thickness peaks T1, T2, appear in the spectrum. The distance between the lower interface 44 and the reference plane 52 is so large that the frequency corresponding to this distance is greater than the maximum distance value z. max and therefore cannot be detected by spectrograph 36

[0078] When object 30 is lifted using the lifting table 33, the distance peaks D1, D2 move to the left in the direction indicated by arrows to smaller distance values ​​z. The thickness peak T1, on the other hand, remains stationary during such a movement of object 30, since the thickness of the layer bounded by the interfaces 40, 42 does not change.

[0079] Figures 4b and 5b correspond to figures 4a and 5a at a later time, when the distance between object 30 and the reference plane 52 is small. Comparing figures 5a and 5b, it can be seen that the two thickness peaks D1 and D2 have shifted further to the left in the spectrum to smaller distance values ​​z, while the thickness peak T1 remains in the same position.

[0080] In this position of the lifting table 33, the distance peak D3, caused by the lower interface 44, is also visible. In addition, a thickness peak T2 has appeared, representing the thickness of the lower layer of the object 30 bounded by the interfaces 42 and 44.

[0081] Figure 5b shows that the following conditions apply to the two thickness peaks T1 and T2:

[0082] T1 = D2 - D1 and Eq. (1)

[0083] T2 = D3 - D2

[0084] Furthermore, a third thickness peak, designated T1 + T2, has appeared. This results from interference of object light components reflected from interfaces 40 and 44 (corresponding to distances D1 and D3). The following relationship applies accordingly to this additional thickness peak T1 + T2:

[0085] T1 + T2 = D3 - D1 Eq. (2)

[0086] The heights of the distance peaks D1, D2, D3 and the thickness peaks T1, T2, and T1+T2 are only schematically indicated in Figures 5a to 5e. In general, distance peaks have a higher intensity than thickness peaks because interference with intense light from the reference arm 16 contributes to the distance peaks. However, the intensities of the distance and thickness peaks generally vary considerably. The intensity is influenced by the reflectivity of the interfaces 40, 42, and 44, and by the position of the focus relative to these interfaces. Often, the interfaces closer to the optical system 25 produce the strongest peaks, but this is not always the case.

[0087] In Figure 5b, a peak labeled 2xT1 is indicated by a dashed line, which results from multiple reflections. This peak arises from the interference of object light reflected at the upper interface 40 with object light reflected at the middle interface 42, reflected again at the upper interface 40, and then reflected once more from the middle interface 42. Thus, the thickness peak 2xT1 appears to correspond to a layer with twice the thickness T1.

[0088] In Figure 5b, arrows also indicate that when the object stage 33 is moved, only the distance peaks D1, D2, D3 move, while all thickness peaks remain stationary.

[0089] When the propagation of the reference light RL in the reference arm 16 is interrupted by actuating the glare control 24, the distance peaks D1, D2, D3 are no longer visible in the spectrum. Therefore, both the movement of the lifting table 33 and the glare control of the reference arm 16 can be used to identify the thickness peaks.

[0090] Figures 4c and 5c illustrate the situation after object 30 is raised further using the lifting table 33. The reference plane 52 is now located inside object 30 for the first time. The distance peak D1 therefore moves to the right in the spectrum according to Figure 5c, which appears to represent an increase in distance. While the distance between the two distance peaks D1 and D2 is equal in the spectra of Figures 5a and 5b, Figure 5c suggests a decrease in the distance between the interfaces 40 and 42.

[0091] It becomes clear that the spectrum shown in Figure 5b or Figure 5c alone is insufficient to correctly assign the distance peaks D1 and D2 to the interfaces 40, 42, and 44. Only by comparing several spectra at different heights of the object 30 can an algorithm recognize that, in the situation illustrated in Figures 4c and 5c, the reference plane 52 is now located between the interfaces 40 and 42. Based on this information, it can then be determined that the distance peak D2 represents the distance of the interface 40 to the underlying reference plane 52. The actual axial position D1' of the upper interface 40 is thus given by

[0092] D1' = R - D1, Eq. (3) where R denotes the z-coordinate of the reference plane 52. For the second interface 42, the actual z-coordinate is further given by

[0093] D2' = R + D2 Eq. (4)

[0094] The same applies to all other interfaces, as these are also located below reference plane 54.

[0095] Figures 4d and 5d show the situation at a later time, where object 30 has been raised to such an extent that the reference plane 52 is located just above the lower boundary surface 44 of object 30. Distance peaks D1 and D2 have shifted to the right, with D1 essentially "overtaking" D2. Distance peak D3 continues to shift to the left, as the reference plane 52 is still located above the lower boundary surface 44.

[0096] Figures 4e and 5e illustrate the situation at a time when object 30 is located entirely above the reference plane 52. The two distance peaks D2 and D3 now shift to the right in the spectrum. The upper interface 40 is so far from the reference plane 52 that the corresponding distance peak D1 lies outside the wavelength range detectable by the spectrograph 36. It is clear from the preceding explanations that the spectra shown in Figures 5a and 5e, in which the reference plane 52 is located relatively far from object 30, do not provide complete information about the distances and thicknesses of object 30. These spectra contain fewer peaks than would be expected with three interfaces.

[0097] The spectra shown in Figures 5b, 5c, and 5d do contain all the peaks. However, as mentioned above, a correct assignment of the peaks to interfaces 40, 42, and 44 cannot be made based on a single one of these spectra, since the spectra differ significantly from each other and it is not immediately apparent which peaks should be assigned to the interfaces.

[0098] Only by considering several spectra together, in which the distance between object 30 and the reference plane 52 was changed, can regularities be identified that allow the peaks to be assigned to the interfaces. A first distinction is made based on which peaks move during this change and which do not. In this way, the distance peaks D1, D2, D3 can be clearly distinguished from the thickness peaks T1, T2, and T1 + T2. For example, considering Figure 5b, the thickness peak T1 can be clearly assigned to the distance between interfaces 40 and 42, since the condition according to equation (1) applies only to this peak. However, this presupposes that the position of the reference plane 52 relative to object 30 is known, as this condition does not appear to be fulfilled in Figure 5c.

[0099] The position of the reference plane 52 can be determined by observing the movements of the distance peaks D1, D2, and D3. As long as all three distance peaks D1, D2, and D3 move to the left when the distance between object 30 and the reference plane 52 decreases, the reference plane 52 is located above object 30. The moment the shortest distance peak D1 disappears at the zero frequency and then moves to the right, the reference plane 52 enters object 30, resulting in the situation shown in Figures 4c and 5c. Since only the thickness peak D1 moves to the right in the spectrum, it is clear that the reference plane 52 is located between interfaces 40 and 42. The correct distances can then be assigned to interfaces 40, 42, and 44 using the conversions described above. The same applies to the spectrum from Figure 5d.Determining the position of the reference plane 52 relative to the object 30 in the manner described above is necessary if this position is not already known for other reasons. One such reason could be that another, but similar, object was previously measured in the same position of the lifting table 33, in which case the position of the reference plane 52 relative to the interfaces 40, 42, 44 can be expected to remain unchanged.

[0100] Another reason could be that the z-coordinate of the surface of object 30 is measured in a different way. For objects with simple geometry, this z-coordinate can be captured, for example, using a camera that takes an image of the object, as will be explained in more detail below in connection with another embodiment.

[0101] Alternatively, the z-coordinate can be derived from the distance peak in the spectrum that has the highest intensity. This distance peak usually originates from the object's surface, since reflections from subsurface optical interfaces are generally weaker than reflections from the surface.

[0102] If the z-coordinate of the surface of object 30 is known, the position of the reference plane 52 relative to object 30 can be determined, since the z-coordinate of the reference plane 52 is also known due to the known optical path length in the reference arm 16.

[0103] The algorithm is also able to determine, for a peak such as the one labeled 2xT 1 in Figure 5b, whether this peak represents an actual structure or is merely the result of multiple reflections. Since thickness peaks generated by multiple reflections represent thicknesses that are multiples of an actual thickness, comparing the thickness peaks generally reveals whether they are caused by multiple reflections or not.

[0104] As can be seen in Figure 5b, the thickness peak 2xT1 represents a thickness twice that represented by the thickness peak D1. If this condition is met, the algorithm assumes that the thickness peak 2xT1 can be ignored, as it is due to a multiple reflection. Since object 30 could contain a layer whose thickness happens to be twice the thickness T1, the algorithm can perform additional checks to rule this out. If the thickness peak 2xT1 represented a real structure, there would have to be two distance peaks whose difference corresponds to the thickness 2xT1. The algorithm can quickly check the existing distance peaks D1, D2, D3 for this and would find that, in the situation shown in Figure 5b, this is not the case.

[0105] An additional indication of multiple reflections can be the intensity of the thickness peak 2xT1, which is usually significantly lower than that of the thickness peak T1 due to the multiple reflections. However, if the intensity of the thickness peak 2xT1 is greater than or equal to the intensity of the thickness peak T1, this suggests that the thickness peak represents a real structure.

[0106] 4. Second embodiment - pre-adjustment

[0107] To best assign the distance and thickness peaks to the interfaces 40, 42, 44, the peaks must stand out clearly from the noise background. Since the reflectivity of the interfaces 40, 42, 44 is generally unchangeable and the intensity of the incident object light OL cannot be increased arbitrarily, it is important that the object light OL is focused as well as possible and strikes the interfaces perpendicularly. a) Perpendicular alignment

[0108] The perpendicular orientation of the interfaces to the object light OL is important because the optical system 25 has only a small numerical aperture. The resulting large depth of field ensures that the focus is maintained over a longer axial measurement range. However, even small deviations from perpendicular light incidence result in noticeably less light returning to the coherence tomograph 11.

[0109] To align the interfaces as perpendicular as possible to the optical axis of the optical system 25, the lifting table 33 can additionally be pivoted about two orthogonal tilting axes and thereby moved into different angular positions. Details can be found in the German patent application filed on the same day by the applicant under file number 10 2024 132 825.8. b) Positioning the focus

[0110] The best possible positioning of the focus 54 relative to the object 30 can be achieved by a pre-adjustment, which is described below with reference to figures 6a to 6c and 7a to 7c.

[0111] Figure 6a shows the lifting table 33 with the object 30 placed on it and the optical system 25, indicated only by a housing. In the position shown in Figure 6a, the lifting table 33 is retracted, so that the object 30 is far from the fixed focus 54 of the object light OL. In Figures 6a to 6c, the distances between the object 30 and the focus 54 are greatly exaggerated for the sake of clarity; in reality, they are only on the order of a few millimeters.

[0112] Figure 7a shows the spectrum I(z) for this position of the lifting table 33. It can be seen that peaks P1 to P5 stand out only slightly from the noise background 56. Therefore, it is not possible to reliably determine whether local intensity maxima are attributable to the noise background or represent peaks associated with interfaces. It is also possible that the distance peaks do not occur at all when the distance between object 30 and focus 54 is large, because then the distance between the reference plane 52 and the interfaces of the object is greater than the coherence length.

[0113] In the position of the lifting table 33 shown in Figure 6b, the focus 54 is closer to the object 30. Therefore, the intensity of the object light OL is higher when it hits the interfaces 40, 42, 44, which leads to more pronounced peaks P1 to P5 in the spectrum, see Figure 7b.

[0114] If the lifting table 33 is extended further, as illustrated in Figure 6c, the focus 54 is already located within the upper glass plate 32. Due to the small numerical aperture, the diameter of the object light beam changes only slightly over the axial extent of the object 30, so that at this position of the focus 54 all peaks P1 to P5 stand out very clearly from the noise background 56, cf. Figure 7c.

[0115] Should a single peak be too weak even in this position of the lifting table 33, minor adjustments can be made to increase the intensity of the peak in question. Even if this significantly reduces the intensity of the stronger peaks, it generally leads to more reliable measurement results overall. The control unit 39, which controls the lifting table 33, receives the necessary information about the peak intensities and the respective signal-to-noise ratio from the evaluation unit 38. The optimization of the z-position of the object 30 can, for example, be carried out according to the condition that the signal-to-noise ratio of no peak falls below a predefined limit.

[0116] The pre-adjustment can be performed independently of the automatic evaluation described above. Preferably, both methods are combined, since both involve a movement of the object 30 using the lifting table 33. For this purpose, only the peak intensities need to be recorded and evaluated during the evaluation. For the z-position of the object 30 relative to the focus 54, where the "best" (usually highest) intensities occur, the automatic evaluation then determines where the reference plane 52 is located relative to the object 30, so that in a further step the peaks can be assigned to the interfaces.

[0117] If object 30 is to be measured at several points (see next section), this z-position of object 30 can generally be maintained, at least for measurements of points located near the point for which the pre-adjustment was performed. If object 30 has a uniform structure and no optical distortions such as field curvature are present, the measurement can also be maintained for the entire object 30. In many cases, however, it will be necessary to repeat the "calibration" described above, including the pre-adjustment, several times to obtain optimal measurement results.

[0118] Figure 8 shows a variant, based on Figures 6a to 6c, in which a stereo camera 58 additionally observes the area above the object 30 from the side. The stereo camera 58 has an image sensor sensitive to infrared light and, in the illustrated embodiment, is oriented such that the optical axis of the camera optics runs at an angle of approximately 85° to the optical axis of the optical system 25. This allows the stereo camera 58 to detect the cross-section of the object light beam on the surface of the object 30. From the size of this cross-section, the location of the focus 54 relative to the object 30 can be calculated.This is particularly helpful in cases where the surface normal of the object is tilted so much relative to the optical axis of the optical system 25 that no or too little light returns to the coherence tomograph 1 to perform the optimization described above as part of the pre-adjustment.

[0119] The position of the focus 54 in the working space between the optical system 25 and the object 30 is usually known. In this case, it suffices to capture the surface of the object 30 with the stereo camera 58 and relate the surface's position, determined by image processing, to the focus 54. Determining the surface's position is simplified if the (preferably visible) light of a pilot laser is directed onto the surface of the object 30. Then, the coordinates of the pilot laser's point of impact directly indicate the height of the object's surface at that point of impact.

[0120] Instead of the side-viewing stereo camera 58, a camera can be used in whose beam path the optical system 25 is located. Details are described in the German patent application with file number 10 2024 120 424.9, which was filed by the applicant on July 18, 2024. With such a partially coaxial beam path, the surface of the object 30 can be observed from above by the camera. Distance measurement can then be easily performed by means of an autofocus measurement, in which the lifting stage 33 is actuated until the image of the surface of the object 30 is maximally sharp on the camera's image sensor.

[0121] 5. Third example - scanning device

[0122] Figure 8 shows a variant of the measuring device 10, based on Figure 1, which additionally includes a scanning device indicated by 62. The scanning device 62 deflects the object light OL variably in two spatial directions. For this purpose, the scanning device 62 has a first scanning mirror 64, which is rotatably mounted about a first axis of rotation 66. A second scanning mirror 68 is rotatably mounted about a second axis of rotation 70, which is oriented perpendicular to the first axis of rotation 66. The scanning mirrors 64 and 68 are driven by galvanometer drives (not shown), which are controlled by the control unit 39. The optical system 25 of the measuring device 10 includes an f-theta lens 72, which is indicated in Figure 8 by three lenses L1, L2, and L3.In the illustrated embodiment, the f-theta lens 72 focuses the object light OL deflected by the scanning device 62 in such a way that the measuring light rays always propagate parallel to the optical axis of the f-theta lens 72.

[0123] 6. Fourth embodiment - Complete measuring system

[0124] Figure 10 shows a complete measuring system with a measuring device 10 according to the invention, which is vibration-isolated and housed in a preferably airtight sealed cabin 100. The cabin 100 is equipped with a door 102 through which an operator or a robot can place objects 30 on the object table 33 and remove them again after measurement. It is also possible to connect the cabin 100 to an existing transport system of a production facility to place objects 30 on the object table 16.

[0125] Also visible are an externally mounted control panel 104, which also serves to display recorded measurement data, and inside the cabin 100 several parts of the measuring device 10, identified by reference numerals and described in more detail above, including the optical system 25, which directs the object light OL towards the microscope slide 35. Arrows indicate the height adjustability and the tiltability of the microscope slide 35, which is mentioned further in section 4a).

[0126] The cabin 100 and damper 106 for vibration-isolated mounting of the measuring device 14 in the cabin 100 protect the measuring device 14 from external influences and vibrations. If necessary, cleanroom conditions can be created inside the cabin 100 using known measures to prevent airborne particles from distorting the measurements.

[0127] In a preferred embodiment, the following procedure is used when measuring an object 30:

[0128] Object 30 is placed on the object stage 33 and door 102 is closed. Subsequently, an automatic two-stage adjustment of the object stage 33 is performed to transfer the wafer into an optimized starting position for its measurement. In the first stage, the aforementioned height adjustment of the object stage 33 is carried out. Object 30 is positioned relative to the focus 54 of the object light OL in such a way as to obtain the strongest possible reflections and correspondingly strong interference signals.

[0129] Secondly, object 30 is positioned relative to the reference plane 52 such that it lies entirely on one side of the reference plane 52 (either above or below it). This facilitates a clear assignment of the peaks to the interfaces and improves the reliability of the topography measurement. If both criteria cannot be met simultaneously, consideration can be given to either changing the position of the focus 54, e.g., using a liquid lens in the optical system 25, or shifting the reference plane 52, e.g., using an optical path length modulator arranged in the reference arm 16.

[0130] In a second step, an angular adjustment is performed as described above in section 4a) with reference to another patent application of the applicant. The aim of the angular adjustment is to align the optical interfaces to be measured as perpendicular as possible to the beam direction, i.e., the optical axis of the optical system 25. For this purpose, the angular position of the specimen carrier 35 is changed until the surface of the object 30, or at least a region of the surface, is oriented more perpendicular to the optical axis of the optical system 25 on average after the change in angular position than before the change in angular position.

[0131] Preferably, the two steps are carried out in the order described above, i.e., starting with the height adjustment. However, the reverse order is also possible.

[0132] The progress of the automated height and / or angle adjustment is preferably displayed on the control panel 104. Optionally, the setting can be confirmed or corrected by an operator.

[0133] After the automated height and / or angle adjustment, the actual measurement of object 30 takes place. This can involve surveying the topography, measuring the thickness of layers, or both. If object 30 is a wafer, parameters such as warping, curvature, or TTV (thickness variation) are preferably determined and displayed on the control panel. Defects in object 30 can also be detected and displayed. Furthermore, the entire topography and / or thickness distribution can be shown. Defects can be highlighted on the control panel display.

[0134] In a further step, it is possible to measure specific regions of interest (ROIs) even more precisely with higher resolution and / or different height and angle settings. The ROIs are selected either manually or by an algorithm, particularly artificial intelligence or machine learning.

[0135] It is also possible to read identification information applied to or burned into the object 30 and display it on the control panel 104. For this purpose, either a separate camera can be used or the coherence tomograph can be used, as disclosed in the German patent application filed by the applicant on October 24, 2024, under file number 10 2024 131 114.2.

Claims

PATENT CLAIMS 1. A method for measuring multilayer objects (30) with a coherence tomograph (11) having an object arm (18) and a reference arm (16), the method comprising the following steps: a) measuring light generated by the coherence tomograph (11) is directed in the object arm (18) by an optical system (25) onto a point on the surface of the object (30); b) measuring light (OL) reflected at an optical interface (40, 42, 44) of the object (30) passes through the optical system (25) back into the coherence tomograph (11) and interferes there with measuring light (RL) guided in the reference arm (16); c) the interference obtained in step b) is detected by a detector (36) of the coherence tomograph (11) and converted into electrical signals;d) A spectrum is generated from the electrical signals obtained in step c) by Fourier transformation, wherein each interference obtained in step b) is represented by a distance peak (D1, D2, D3) in the spectrum; characterized by the following further steps: e) in a process automatically controlled by a control device (39), steps a) to d) are repeated multiple times, wherein in each repetition a distance between the object (30) and a reference plane (52) is changed, which is located in an object arm (18) of the coherence tomograph (11) and whose axial position corresponds to an optical path length of the reference arm (18); f) an evaluation unit (38) determines by analyzing the distance peaks (D1, D2, D3) obtained in step e) and preferably thickness peaks, which represent a distance between interfaces, a unique assignment between the ab-; standpeaks and the optical interfaces (40, 42, 44) and assigns distance values ​​to the optical interfaces based on this assignment.

2. Method according to claim 1, characterized in that in step f) the evaluation device also determines a unique assignment between the thickness peaks corresponding to a distance between two optical interfaces (40, 42, 44) and the distances between the optical interfaces by means of the analysis and assigns thickness values ​​to the distances between the interfaces on the basis of this assignment.

3. Method according to claim 1 or 2, characterized in that in step e) for thickness peaks (T1, T2, T1 +T2) in the spectrum, which correspond to a distance between two optical interfaces (40, 42, 44), it is determined whether there is a pair of distance peaks (D1, D2, D3) which each correspond to a distance to one of the two optical interfaces.

4. Method according to one of the preceding claims, characterized in that in step e) it is determined whether, in addition to a thickness peak (T1, T2, T1 +T2) corresponding to a distance between two optical interfaces, there is another thickness peak (2xT1) corresponding to a multiple of the distance.

5. Method according to one of the preceding claims, characterized in that the evaluation unit (38) determines the position of the reference plane (52) relative to the object (30) in step f).

6. Method according to claim 5, characterized in that in further measurements the object (30) is arranged relative to the reference plane (52) such that the reference plane (52) is always located outside the object (30).

7. Method according to one of the preceding claims, characterized in that in step e) the distance between the object (30) and the reference plane (52) is changed by moving the object (30) with the aid of a lifting table (33) controlled by the control device (39).

8. Method according to one of the preceding claims, characterized in that the measuring light (OL) is directed onto a plurality of different points, and that at the majority of these points steps e) and f) are not carried out, but the optical interfaces (40, 42, 44) are assigned modified distance values ​​based on a previously determined assignment, taking into account shifts of the distance peaks (D1, D2, D3).

9. Method according to one of the preceding claims, characterized in that the optical system (25) focuses the measuring light (OL) in a focus (54) and, prior to step a), determines in an automatically controlled pre-adjustment by the control device (39) at which position of the focus (54) relative to the object (30) the distance peaks (D1, D2, D3) and / or thickness peaks, corresponding to a distance between two optical interfaces (40, 42, 44), stand out most clearly from a noise background (56).

10. Method according to claim 9, characterized in that during pre-adjustment the position of the focus (54) relative to the object (30) is changed and the intensity of at least one distance or thickness peak is measured.

11. Method according to claim 10, characterized in that the position of the focus (54) relative to the object (30) is changed by moving the object (30) with the aid of a lifting table (33).

12. Method according to one of claims 9 to 11, characterized in that a camera (58) measures the position of the focus (54) relative to the object (30).

13. Method according to claim 12, characterized in that, based on the position of the focus (54) relative to the object (30) measured by the camera (58), a coarse adjustment is carried out in which the position of the focus relative to the object is changed, and that a fine adjustment is then carried out in which the position of the focus relative to the object is changed again and the height of at least one distance or thickness peak is measured.

14. Method according to one of claims 12 or 13, characterized in that the camera (58) is a stereo camera with a camera optic whose optical axis is arranged at an angle between 5° and 90° to an optical axis of the optical system (25).

15. Method according to one of claims 9 to 14, characterized in that during pre-adjustment an inclination angle of a substrate (35) on which the object (30) is placed is changed.

16. Device for measuring multilayer objects, comprising a) an optical system (25), b) a coherence tomograph (11) having a detector (36), an object arm (28) and a reference arm (16) and configured to generate measuring light (OL) and to direct it in the object arm (18) via the optical system (25) to a point on the surface of the object (30), wherein the detector (36) is configured to detect interference between measuring light (OL) reflected at an optical interface (40, 42, 44) of the object (30) and returned to the coherence tomograph (11) via the optical system (25) and measuring light (RL) reflected in the reference arm (16), and to convert the detected interference into electrical signals, c) an evaluation unit (38) configured to derive electrical signals from the electrical signals generated by the detector (36) by Fourier transformation to generate a spectrum,in which each detected interference is represented by a distance peak (D1, D2, D3) in the spectrum, d) a control device (39) configured to perform multiple measurements in which a distance between the object (30) and a reference plane (52) located in an object arm (18) of the coherence tomograph (11) is changed and whose axial position corresponds to an optical path length of the reference arm (16), wherein the evaluation unit (38) is configured to make a unique assignment between the distance peaks and the optical interfaces by analyzing the obtained distance peaks (D1, D2, D3) and to assign distance values ​​to the optical interfaces (40, 42, 44) on the basis of this assignment.

Images