Device and method for measuring a wafer
Through the combination of an optical coherent tomography scanner and a scanning device, combined with appropriate optical systems and correction measures, the error and cost problems of measuring wafer characteristic dimensions in the prior art are solved, and a fast, accurate and economical measurement effect is achieved.
Patent Information
- Application Number
- CN202410656898.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-05-25
- Filing Date
- 2024-05-24
- Publication Date
- 2024-11-26
AI Technical Summary
Existing methods of measuring wafers cannot quickly, accurately and inexpensively measure the characteristic dimensions of wafers such as TTV, curvature and warpage, especially on larger wafers, with problems of measurement errors and equipment expensive.
Using optical coherence tomography scanner and scanning device, interference thickness measurement is performed by selecting the wavelength range of at least partially transparent wafer material, combined with a larger optical system to scan the entire wafer without moving the measuring device and wafer, using a separate circuit that can be surrounded by two. The scanning mirror and collision protection device rotating each axis reduce the field curve and calculates the correction distance value to compensate for the field bending of the optical system.
It realizes rapid, accurate and cost-effective measurement of the characteristic dimensions of the wafer, reduces measurement errors, is suitable for measurement of larger wafers, and reduces equipment costs.
Smart Images

Figure CN119022804A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to an apparatus and a method for contactless measurement of geometric variables of a wafer, such as TTV, bow or warp. Background Art
[0002] Wafers are round or square-shaped thin sheets with a thickness of about one millimeter that serve as substrates for integrated circuits, micromechanical components or optoelectronic coatings. Wafers are made from single-crystalline or multicrystalline blanks (so-called ingots) that are cut transversely to their longitudinal axis into individual wafers. In most cases, wafers made of silicon are used, but other materials can also be used, such as glass wafers for the production of microlens arrays or for augmented reality applications.
[0003] The shape of the wafer must meet strict geometric specifications. These specifications include TTV (Total Thickness Variation), which refers to the maximum difference between the thickest and thinnest parts of the wafer. Bow is defined as the maximum deviation of the mid-plane of the wafer from a reference plane. Warp, for those skilled in the art, refers to the deviation of the mid-plane of the wafer from a reference plane, where the bow over the entire wafer surface has been corrected.
[0004] In addition to the common definitions of TTV, bow, and warp, similar but slightly different definitions are occasionally used when characterizing wafers to characterize the wafer's deviation from the ideal shape.
[0005] To comply with these specifications, at least spot measurements of these geometrical variables must be made, preferably during normal production.
[0006] For this purpose, the wafer surface is usually measured with the aid of capacitive or optical methods such as triangulation, laser interferometry or chromatic confocal measurement with light in the visible range. Since these methods cannot measure through the wafer volume, it is advantageous to either turn the wafer around after the first measurement or to measure simultaneously with one sensor on each side of the wafer. For this purpose, the relative position of the two sensors and their distance must be known and controlled very precisely. Positioning errors of the sensors are directly reflected in the measurement result.
[0007] A method is known from the Semi MF65 standard which can be used to determine the TTV with such a two-sided measurement. A first distance A to a first surface of the wafer is measured with a first sensor, and a second distance B to the opposite surface is measured with a second sensor. The structure is calibrated with a wafer of known thickness (the so-called "golden sample") as a preparation for the measurement, so that the total distance G between the sensors can be determined. 总 In this case, the thickness T of any wafer is determined as followsW :
[0008] T W =G 总 –B–A
[0009] Subsequently, an automated scan is performed on the wafer surface, where TTV is defined as the difference between the maximum and minimum thickness:
[0010] TTV=T wmax –T wmin
[0011] FIG. 7 shows the measurement of a wafer 10 according to this known method: a first distance A to the first surface 40 of the wafer is measured by a first sensor 81, and a second distance B to the second surface 58 is measured by a second sensor 82. 总 If known, the wafer thickness T at a certain point can be obtained from these measurements. W .
[0012] In the standards Semi MF534 and MF65, methods for determining bow and warp in this configuration are described. To do this, the median plane of the wafer is first determined. This median plane is defined as the corresponding center point between the two surfaces at any given location of the wafer. To determine the median plane, the above distances A and B are used and the center point is calculated.
[0013] M=1 / 2(G 总 –B+A)
[0014] The bow is determined based on this median plane: Three points of the median plane, located on an equilateral triangle at the edge of the wafer, define a flat comparison surface. Subsequently, the distance between the points on the median plane in the middle of the wafer and the comparison surface is measured. This distance is a measure of the bow.
[0015] The warpage is calculated in a similar way: Here, the deviation RPD of each point of the median plane from the reference plane is determined: The difference between the maximum and the minimum deviation is defined as a measure of the warpage.
[0016] Warp=RPD max -RPD min
[0017] Figure 8 The median plane M is intuitively shown, as well as the variables used to calculate the warpage and curvature. US2012 / 0257207A1 discloses a device for measuring a wafer, which combines a Michelson interferometer for distance measurement with a reflectometer for thickness measurement. The measurement is performed at each point. In order to obtain the measurement value of the entire surface, the wafer is moved relative to the fixed measuring device.
[0018] However, these known measurement methods are relatively slow, since the wafer cannot be moved as fast as desired due to its inertia. Furthermore, the corresponding measuring devices are relatively expensive, since the actuators for moving the wafer must be of particularly high quality in order to achieve smooth and precise movement. Furthermore, the movement of the wafer can cause measurement errors, especially in the case of large wafers. These measurement errors can be statistical, for example due to vibrations or jolts of the actuators, and are therefore difficult or even impossible to eliminate.
[0019] For example, DE 10 2017 128 158 A1 also discloses a measuring device with an optical coherence tomography scanner and a scanning device, which can scan smaller areas without moving the wafer. However, such systems have been limited in terms of their size so that when measuring larger wafers, the wafer and the measuring device also need to be moved relative to each other and the images stitched together. This so-called "stitching" can lead to additional measurement errors, which can lead to significant distortions of the results for bow and warp in particular. The object of the present invention is to provide a device and a method that can be used to measure characteristic dimensions of a wafer, such as in particular TTV, bow and warp, very quickly, accurately and still at a low cost. Summary of the invention
[0020] In terms of the device, the solution of the present invention to achieve the above-mentioned object is a device for measuring a wafer, which has
[0021] an optical coherence tomography scanner adapted to generate a measuring beam and to align the wafer via an optical system,
[0022] wherein the wavelength range of the measuring beam is selected such that the wafer is at least partially transparent thereto,
[0023] a scanning device adapted to deflect the measuring light beam in two spatial directions,
[0024] A control unit is adapted to control the scanning device in such a way that the measuring beam scans the surface of the wafer at a plurality of measuring points in sequence, wherein the distance d between two measuring points is max 140mm≤2d max ≤600mm,
[0025] An analysis unit is adapted to calculate distance values and / or thickness values from the interference signals provided by the optical coherence tomography scanner and to calculate at least one characteristic variable of the wafer from the distance values and / or thickness values.
[0026] Transparent means that a large part of the light, for example more than 20%, passes through the thickness of the wafer (approximately 150 μm to 2000 μm). It must be ensured that enough measuring light enters the wafer, is reflected on the opposite side and passes back through the wafer in the direction of the measuring device. Advantageously, materials with an absorption coefficient of less than 1000 cm -1 , especially less than 100cm -1 or 10cm -1 For sapphire and silicon wafers, the wavelengths are in the so-called VISNIR range of 700 nm to 2000 nm. In particular, the near infrared range (NIR) is used, where wavelengths greater than 1100 nm penetrate thicker wafer materials very well.
[0027] The characteristic variable is preferably TTV, curvature or warpage. In particular, one, more or all of these variables can be determined. Advantageously, the calculated variables are output to the user and / or to a program for automatic processing.
[0028] The invention is based on the recognition that the known measuring device with an optical coherence tomography scanner and a scanning device can be used for wafer measurement if the interferometric thickness measurement is performed using a wavelength range in which the wafer material is at least partially transparent, so that two wafer surfaces can be measured simultaneously. Furthermore, it is advantageous to use a larger optical system that enables a larger measurement field of view, which allows scanning the entire wafer without relative movement of the measuring device and the wafer, and mainly solves the problem of field curvature. The inventors have developed a number of solutions to this problem.
[0029] The inventors have discovered that known measurement devices cause field curvature when the measurement beam is scanned two-dimensionally over the wafer surface. Intuitively, this means that the focus of the measurement beam does not scan across a flat surface, but across a curved surface. As a result, the originally flat wafer surface appears curved.
[0030] Field curvature can be tolerated in very small wafers with a diameter of up to about 80 mm, but it cannot be tolerated in larger wafers because the field curvature increases quadratically with the diameter of the wafer to be measured. A larger wafer is considered to be one in which two of the measuring points that can be reached by the scanning device have a distance d of max Can be 140mm≤d max ≤600mm, preferably 280mm≤d max ≤450mm. For d max=300 mm, the measurement error is typically already about 1 mm due to the field curvature. Therefore, without additional measures, the aforementioned measuring device itself is not suitable for measuring larger wafers when using larger flat field objectives, because the required accuracy in the distance measurement is below 1 μm and is directly reflected in the determination of the curvature and warpage. Therefore, the measurement error of the known measuring device is three orders of magnitude larger in larger wafers.
[0031] The inventors have also found that the additional optical path length difference that causes field curvature is mainly caused by the scanning mirrors. In order to be able to pivot without collision, the two scanning mirrors must be relatively far apart. It turns out that the spatial distance of the rotation axes of the scanning mirrors is the main cause of field curvature.
[0032] The simplest measure to reduce the curvature of field is therefore to use a single scanning mirror that is rotatable about two axes instead of two scanning mirrors that are each rotatable about exactly one axis. However, currently available two-axis scanning mirrors cannot be pivoted as quickly and precisely as two single-axis scanning mirrors. However, the required measuring time is significantly reduced compared to known methods in which the wafer must be moved.
[0033] Another solution to reduce the curvature of field is to use two scanning mirrors that are each rotatably supported about exactly one axis as has been done so far, but to significantly shorten the distance between the scanning mirrors. In this case, however, there is a risk of the scanning mirrors colliding when the rotation angles are large. To eliminate this situation, at least one of the two scanning mirrors can be assigned a collision protection device that is suitable for limiting the rotation angle of the at least one scanning mirror. The consideration behind this is that existing scanning mirrors usually have a certain rotation angle range themselves, and the application in this article only requires a part of this rotation angle range. If the collision protection device limits the initially possible rotation angle (usually ±20°) to the actually required rotation angle (for example ±10°, preferably ±5°), this can reliably eliminate the possibility of collision of the scanning mirrors even when the distance between the scanning mirrors is extremely small.
[0034] The collision protection device can have at least one mechanical stop to limit the rotation angle. This stop can be, for example, a damper made of rubber, against which the at least one scanning mirror hits when the at least one scanning mirror deflects by more than, for example, ±10° or ±5°. Alternatively, the stop can act directly on the rotation axis of the at least one scanning mirror or on a projection formed on the rotation axis.
[0035] As an alternative or additional solution, the collision protection device can be implemented as an electronic limiting device, which is suitable for electronically preventing the supply of control signals to the at least one scanning mirror, which would otherwise exceed a preset rotation angle range. A very reliable solution is, for example, a choke coil for limiting the control current of the scanning mirror.
[0036] As an alternative or in addition, the control software of the scanning mirror is designed in such a way that it prevents the generation of control signals that would otherwise lead to a large deflection of the scanning mirror. To be on the safe side, it is also useful to additionally adopt one of the above measures to prevent the following situation: the software runs abnormally and generates an erroneous signal, causing one of the two scanning mirrors to deflect too much.
[0037] According to another alternative or supplementary measure for reducing the field curvature, an optical system is used which contains at least one, preferably two, anamorphic optical elements for correcting the field curvature. Preferably, the anamorphic element is a cylindrical lens, the axis of symmetry of which matches the orientation of the rotation axis of the scanning mirror in terms of its orientation.
[0038] A preferred measure of the invention is not to reduce the field curvature, but to eliminate its influence during the calibration process. To this end, the analysis device should be suitable for correcting the measured distance values by calculation to compensate for the curvature of the field used by the optical system to focus the measuring beam. This field usually coincides with the focal plane of the optical system. When calibrating, for example, measurements are made on flat glass produced with high precision. The measured deviations from flatness are the correction values, which are added or subtracted from the measured values in subsequent measurements, depending on the specific sign.
[0039] For the correction by calculation, the correction values can be read from a correction table stored in the evaluation unit. Preferably, the correction values for different operating wavelength ranges are stored in the correction table. The correction values for these different operating wavelength ranges do not need to be determined by a calibration measurement of their own, but can also be derived theoretically from the correction values measured for a certain operating wavelength range. In particular, Zernike polynomials can also be used.
[0040] As an alternative to reading from a correction table, the correction value can also be calculated by a mathematical formula. This formula can be derived, for example, by polynomial fitting. In this case, for example, Zernike polynomials can be used as polynomials.
[0041] Of course, a calibration may also be performed in addition to the above-mentioned measures.
[0042] The optical coherence tomography scanner that can be used in the present invention can in particular have:
[0043] a light source that emits light in a wavelength range that makes the wafer at least partially transparent,
[0044] a beam splitter adapted to split the light generated by the light source into a measuring beam and a reference beam,
[0045] reference arm for guiding the reference beam,
[0046] an object arm for guiding a measuring beam using the optical system and the scanning device, and
[0047] A detector is adapted to generate an interference signal by superposition of a reference beam guided in a reference arm and a reflected component of the measuring beam at the wafer.
[0048] In one embodiment, a switchable shading device is arranged in the reference arm, which is suitable for preventing the reference beam from propagating in the reference arm when implementing thickness measurement. The switchable shading device can switch between a "distance mode" and a "thickness mode" for analyzing the interference between the measuring beam and the reference beam. In the thickness mode, the distance of the interface (that is, the thickness of the wafer) is inferred from the interference between the reflections of the measuring beam at two plane-parallel interfaces of the wafer. Therefore, the detector generates an interference signal by superimposing the reflected components of the measuring beam at two different interfaces of the wafer. In order to prevent the unnecessary reference beam in the reference arm from interferingly superimposing the required interference signal in the thickness mode, a shading device is used to prevent the reference beam from propagating in the reference arm. The switching of the shading device is preferably automatically implemented each time the measurement mode is switched.
[0049] In another embodiment, the reference beam in the reference arm is not guided as a free beam or at least not completely, but is at least partially guided in an optical fiber. In a rolled-up optical fiber, the light can be guided in a smaller space over a larger optical path length. If the object arm and the reference arm have different dispersions, measures for dispersion compensation are advantageously taken.
[0050] In terms of method, the solution of the present invention for achieving the object mentioned at the beginning is a method for measuring a wafer, which has the following steps:
[0051] a) generating a measuring beam with an optical coherence tomography scanner, wherein the wavelength range of the measuring beam is selected such that the wafer is at least partially transparent thereto;
[0052] b) directing a measuring beam toward the wafer with an optical system;
[0053] c) deflecting the measuring light beam in two spatial directions by means of a scanning device, wherein the scanning device is controlled in such a way that the measuring light beam scans the surface of the wafer successively at a plurality of measuring points;
[0054] d) calculating the distance value and / or thickness value based on the interference signal provided by the optical coherence tomography scanner
[0055] e) Calculating at least one characteristic variable of the wafer from the distance values and / or thickness values.
[0056] The scanning device is controlled in such a way that the two measuring points have a distance d max 140mm≤d max≤600mm, preferably 280mm≤d max ≤450mm.
[0057] In one embodiment, the optical coherence tomography scanner is characterized by:
[0058] A light source produces light;
[0059] A beam splitter divides the light generated by the light source into the measuring beam and the reference beam;
[0060] directing the reference beam in a reference arm;
[0061] directing the measuring beam in an object arm;
[0062] The detector generates the interference signal by superposition of the reference beam guided in the reference arm and the reflected component of the measuring beam at the wafer.
[0063] When performing thickness measurement, the reference beam can be temporarily blocked from propagating in the reference arm by a light blocking device.
[0064] If the optical system has a field for focusing the measuring beam, the measured distance value can be corrected by calculation to compensate for the curvature of the field. For the correction by calculation, the correction value can be read from a correction table stored in the analysis unit. The correction table can store correction values for different operating wavelengths. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] The embodiments of the present invention are described in detail below with reference to the drawings.
[0066] Figure 1 is a perspective view, not drawn to scale, of a wafer to be measured;
[0067] Figure 2 is a schematic diagram of a measuring device of the present invention;
[0068] Figure 3 is a simplified perspective view of an important part of a scanning device which is part of a measuring device in another embodiment and contains a mechanical stop for a scanning mirror;
[0069] Figure 4 is a simplified perspective view of an important part of a scanning device which is part of a measuring device in another embodiment and contains only a single scanning mirror;
[0070] Figure 5a and Figure 5b are two orthogonal meridional cross sections of an optical system having a cylindrical lens, the system being part of a measuring device in another embodiment;
[0071] Figure 6 A reference arm of an optical coherence tomography scanner according to one embodiment, in which a reference beam passes through a portion of an optical path in an optical fiber.
[0072] Figure 7a According to the prior art, the two-side measurement structure
[0073] Figure 7b Schematic diagram of the measurement structure of the present invention
[0074] Figure 8 Schematic diagram for measuring the variables bending and warping DETAILED DESCRIPTION
[0075] 1. Measuring wafer
[0076] Figure 1 A perspective view of a wafer 10 is shown, but not to scale. The wafer 10 has the shape of a right circular cylinder in the illustrated embodiment, wherein the thickness is greatly exaggerated. A practical wafer 10 would have a diameter of, for example, 300 mm and a thickness of only about 1 mm. It is also possible to use wafers 10 whose surfaces are not perfectly round but rather square.
[0077] Dashed line 12 shows the ideal cylindrical shape of wafer 10. Deviations from this ideal shape may occur due to manufacturing tolerances, which are Figure 1 In order to determine these deviations, the wafer 10 must be measured. When the surface topography of the two wafer surfaces is measured, all common geometrical specifications of the wafer, such as TTV, bow and warp, can be derived therefrom.
[0078] The distribution of the measuring points used to measure the surface topography is adapted to the respective measuring task. Figure 1 In the embodiment shown, the measuring points are arranged along two lines 11, 13, which are perpendicular to each other and intersect in the center of the wafer 10 and extend to the periphery of the wafer 10. However, other measuring modes, such as spiral lines or grid patterns, can also be used. The measuring points can be closely adjacent, for example with a distance of a few microns. In other measuring modes, the distance is in the range of 1 mm.
[0079] 2. Structure and function of the measuring device
[0080] Figure 2 14 is a schematic diagram of a measuring device according to the invention. The measuring device 14 is used to measure a wafer 10 which is carried by a support 18. The support 18 can be designed, for example, as a simple three-point support mechanism, such as Figure 2 20 in the embodiment shown. In the embodiment shown, the support 18 itself is supported on the base 16. The support 18 and the base 16 are generally not part of the measuring device 14.
[0081] As an alternative, a support can be used in which the wafer is not supported horizontally but vertically. This reduces the influence of gravity on the characteristic variables.
[0082] When measuring is performed during the production process, the wafer 10 can also be transported to the measuring device 14 by the transport device, for example.
[0083] The measuring device 14 comprises an optical coherence tomography scanner 22 which generates a measuring beam 24, the structure of which will be described in detail below.
[0084] The scanning device 26, which is generally designated by 26, deflects the measuring beam 24 variably in two spatial directions. For this purpose, the scanning device 26 has a first scanning mirror 28, which is mounted so as to be rotatable along a first rotation axis 30. A second scanning mirror 32 is mounted so as to be rotatable along a second rotation axis 34 which is perpendicular to the first rotation axis 30. The scanning mirrors 28, 32 are driven by a galvanometer drive (not shown), which is controlled by a control unit 36.
[0085] The measuring device 14 also includes an optical system 38, which Figure 2 In the exemplary embodiment shown, the optical system 38 focuses the measuring beam 24 deflected by the scanning device 26 so that it always strikes the surface 40 of the wafer 10 facing the optical system approximately perpendicularly.
[0086] The optical coherence tomography scanner 22 comprises in a known manner a light source 42, a first beam splitter 44 for splitting the light generated by the light source into a measuring beam 24 and a reference beam 46, a reference arm 48 for guiding the reference beam 46, and an object arm 50 using an optical system 38 and a scanning device 26, in which the measuring beam 24 is guided.
[0087] During measurement, the measuring beam 24 propagating in the object arm 50 is focused on the surface 40 of the wafer 10, partially reflected there and returns in the same optical path via the object arm 50 to the first beam splitter 44. There, the reflected component of the measuring beam 24 is superimposed with the reference beam 46 guided in the reference arm 48 and reflected there on the mirror 52. Both light components are directed by the second beam splitter 54 to a detector 56, which converts the optical reference signal into an electrical signal.
[0088] In the embodiment shown, the optical coherence tomography 22 is designed as an FD-OCT (FD stands for Fourier domain). The detector 56 therefore comprises a spectrometer for detecting the spectral intensity distribution. An analysis unit 57 connected to the detector 56 can thereby calculate the distance between the surface 40 and the measuring device 14 (e.g. lens L3) at the point of incidence of the measuring beam 24 in a known manner. For further details of the optical coherence tomography, see DE 10 2017 128 158 A1 already mentioned at the outset.
[0089] The wavelength range of the light generated by the light source 42 can be selected in such a way that the measuring beam 24 can at least partially pass through the wafer 10. In this case, a reflection is also produced on the lower surface 58 of the wafer 10 facing away from the measuring device 14, which reflection is detected by the optical coherence tomography scanner 22. In this case, two measuring beams 24 that have traveled different optical path lengths are superimposed. The detector 56 then determines the difference in optical path length in a known manner via the period of the interference, from which the distance between the two surfaces 40, 58 of the wafer 10 and the thickness of the wafer can be inferred.
[0090] If the aforementioned thickness mode is selected, it should be prevented that light is also propagated in the reference arm 48, since light can also cause interference and thus generate undesired interference signals. For this purpose, the reference arm 48 comprises a switchable shading device, denoted by 60, which can be constructed as a central shutter or focal plane shutter, for example. When switching from the distance mode to the thickness mode, the switchable shading device 60 is automatically closed, so that the light in the reference arm 48 can no longer interfere with the detector 56. If the switch is switched back to the distance mode, the switchable shading device 60 reopens the path of the reference beam 46.
[0091] Optionally, the reference arm 48 has a control device for the shutter device 60, in particular a sensor, which can determine the state of the shutter device 60. For example, the current state can be checked after a power failure by means of the control device.
[0092] exist Figure 2 In the embodiment shown, the measuring light 24 is completely propagated in free space. In other embodiments, the light is guided partly in the optical fiber, which will be discussed below. Figure 6 If the optical coherence tomography scanner 22 is housed in a suitable housing, the light between the optical coherence tomography scanner 22 and the scanning device 26 can also be guided in optical fibers. In this case, the scanning device 26 and the optical system 38 can be housed in a compact and lightweight measuring head, which can be mounted at different locations with little difficulty.
[0093] 3. Field curvature
[0094] The optical system 38 is designed as a flat field optical system so that the collimated light entering the optical system 38 is focused to a flat focal plane. Nevertheless, it is found that the focus of the measuring beam 24 is not exactly within the plane.
[0095] The reason for this undesirable field curvature is that the two scanning mirrors 28, 32 are arranged one behind the other in space, which leads to a path length change that is not easy to understand intuitively. This path length change increases quadratically with increasing rotation angle of the scanning mirrors 28, 32 and increasing distance of the measuring point from the optical axis OA of the optical system 38.
[0096] The optical path length in the reference arm 48 is usually selected so that it corresponds to the optical path length in the object arm 50 to the focus of the measuring beam 24. If the reflecting surface is outside the focus, the coherence tomograph 22 interprets this as the relevant surface being further away or closer than the optical path length preset for the reference arm 48. However, if the focus of the measuring beam 24 is not exactly in the plane, this causes an originally flat surface to appear curved. At a distance of 150 mm from the optical axis OA, the resulting measurement error is already about 1 mm, which is three orders of magnitude higher than the required measurement accuracy of 1 μm.
[0097] To solve this problem, a calibration is performed before the measurement device 14 is delivered. For this purpose, a calibration standard in the form of a flat glass plate produced with high precision is measured. The deviations from flatness measured due to the field curvature are converted into correction values, which are stored in a table in the evaluation unit 57. For each measuring point that the scanning device 26 can pass through on the surface of the calibration standard, a correction value assigned to this measuring point is determined. The assignment of measuring points to correction values can be stored in a table in the evaluation unit 57.
[0098] Alternatively, the correction values can be calculated from a formula derived from the calibration measurements.A combined solution is also possible, in which only the correction values for the individual nodes are stored in a table and the correction values for the positions between the nodes are obtained by interpolation.
[0099] When the actual wafer 10 is measured after this calibration, the corresponding correction value is added or subtracted from the obtained measurement value depending on the specific sign, thereby obtaining a set of corrected measurement data, which can be output to the user of the measurement device 14 for further processing.
[0100] 4. Measures to reduce field curvature
[0101] As an alternative or in addition to calibration, measures can be taken to minimize field curvature.
[0102] Such measures can include, for example, reducing the distance between the two scanning mirrors 28, 32. This distance is usually selected so that the scanning mirrors 28, 32 cannot collide with each other even in the case of large rotation angles. However, for the aforementioned measurement of the wafer 10, usually only a relatively small rotation angle of the order of magnitude of ±10° or even only ±5° is required. The small rotation angle makes it possible to arrange the two scanning mirrors 28, 32 close to each other. However, since the scanning mirrors 28, 32 have a large rotation angle range in principle, a collision protection device must be provided in order to reliably prevent the scanning mirrors 28, 32 from contacting each other during operation. Figure 3 In the embodiment shown, the collision protection device has mechanical stops 64, which can be formed, for example, by rubber lips. The mechanical stops 64 are positioned so that they limit the structurally possible rotation angle of the scanning mirrors 28, 32 themselves to the required rotation angle.
[0103] Depending on the distance and size of the scanning mirrors, only one of the two scanning mirrors 28 , 32 may be provided with a collision protection device.
[0104] In such Figure 4 As shown in the perspective schematic diagram in , the field curvature is smaller when the scanning device 26 does not have two scanning mirrors 28, 32 that are supported so as to rotate around a single rotation axis 30 or 34, but only has a single scanning mirror 28' that can be supported around two preferably orthogonal rotation axes 30, 34.
[0105] Figure 5a and Figure 5b Two orthogonal meridional sections of the optical system 38 of the measuring device 14 according to an alternative embodiment are shown, namely Figure 5a The XZ plane and Figure 5b The YZ plane in the measuring device 14 is also provided with two scanning mirrors. Here, the field curvature is reduced by a suitably designed system 38. For this purpose, the optical system 38 has a first cylindrical lens 66 and a second cylindrical lens 68, wherein the axes of symmetry of the cylindrical lenses 66, 68 are perpendicular to each other. Figure 5a and Figure 5b The two cylindrical lenses 66, 68 are each assigned to one of the two scanning mirrors 28, 32. Therefore, the orientation of the axis of symmetry of the cylindrical lenses 66, 68 matches the orientation of the axis of rotation 30, 34 of the scanning mirror 28 or 32.
[0106] Figure 6 A detail of an optical coherence tomography scanner 22 is shown. In order to allow the measuring beam 24 to scan even relatively large wafers 10 right up to their edge, the optical system 38 must have a large focal length, which results in a relatively large geometrical path length and optical path length in the object arm 50. The optical path length in the reference arm 48 must also be correspondingly large.
[0107] Since free-beam propagation over long distances requires complex beam folding, advantageously, in the reference arm 48 the measuring beam 24 is not guided, or at least not completely, as a free-beam, but is at least partially guided in an optical fiber. Figure 6 In the embodiment, an input coupling lens 70 arranged behind the shading device 60 in the optical path couples the measuring beam into the end 71 of the rolled-up optical fiber 72. A second coupling lens 76 is arranged between the end 74 of the optical fiber 72 and the mirror 52 so as to couple the reference beam reflected at the mirror 52 back into the optical fiber 72.
[0108] The manner in which the reference arm 48 is divided into the optical fiber guiding portion and the free beam guiding portion is not limited to: Figure 6 In an alternative embodiment, the beam splitter 44 can also be realized as a fiber coupler.
[0109] Figure 7b The measurement structure of the present invention is schematically shown Figure 7a The difference between the structure of the prior art and the measurement of the first surface 40 and the second surface 58 of the wafer 10 by two separate measuring heads is shown in FIG. Figure 7b It shows how to measure both sides 40 and 58 on the entire wafer surface of the wafer 10 with a single measuring device 14. In this, both the distance A between the surface 40 of the wafer facing the optical system and the virtual reference plane is measured, and the wafer thickness T is directly measured. W Advantageously, A is measured in distance mode using the reference arm 48, while the wafer thickness T is measured in thickness mode using interference between reflections on the two wafer sides 40 and 58. W .
[0110] This greatly simplifies the formulas for calculating TTV, bow and warp described at the beginning of this article. It is no longer necessary to determine the total distance G 总 , but directly measure T W This also eliminates relevant sources of measurement error. The definition of TTV remains
[0111] TTV=T wmax –T wmin
[0112] The determination of the median plane is also simplified. This median plane can be calculated directly from the distance measurement A relative to an arbitrary reference plane (e.g. the distance to the measuring device or a virtual optical distance to a reference plane) and the thickness:
[0113] M=1 / 2T W +A
[0114] The bow and warp are calculated from this median plane in the manner described above.
Claims
1. A device (14) for measuring a wafer (10), comprising: An optical coherence tomography scanner (22) adapted to generate a measuring beam (24) and to transmit the measuring beam to the optical system (38) aligning the wafer (10), wherein the wavelength range of the measuring light beam (24) is selected such that the wafer (10) is at least partially transparent thereto, a scanning device (26) adapted to deflect the measuring beam (24) in two spatial directions, A control unit (36) adapted to control the scanning device (26) in such a way that the measuring beam (24) The surface of the wafer (10) is scanned at a plurality of measuring points in sequence, wherein the distance d between two measuring points is max 140 mm≤d max ≤600mm, and An analysis unit (57) is adapted to calculate distance values and / or thickness values from the interference signals provided by the optical coherence tomography scanner (22), and to calculate at least one characteristic variable of the wafer (10) from the distance values and / or thickness values.
2. The device according to claim 1, characterized in that The characteristic variable is TTV, bow or warp.
3. The device according to any one of claims 1 or 2, characterized in that The wavelength range is within the range of 700nm-2000nm, in particular within the range of 950nm-2000nm or 1100nm-2000nm.
4. The device according to any one of the preceding claims, characterized in that The wavelength range of the measuring beam (24) is selected such that at least 20% of the light passes through the wafer.
5. The device according to any one of the preceding claims, characterized in that The optical coherence tomography scanner (22) has: Light source (42), a beam splitter (44) adapted to split the light generated by the light source into the measuring beam (24) and the reference beam (46), a reference arm (48) for guiding the reference beam (46), an object arm (50) of the optical system (38) and the scanning device (26) for guiding the measuring beam (24), and A detector (56) is adapted to generate the interference signal by superposition of a reference beam (46) guided in the reference arm (48) and a reflected component of the measuring beam (24) at the wafer (10).
6. The device according to claim 5, characterized in that A switchable light shielding device (60) is arranged in the reference arm (48) and is suitable for preventing the reference beam (46) from propagating in the reference arm (48) when thickness measurement is performed.
7. The device according to any one of the preceding claims, characterized in that The optical system (38) has a field on which the measuring beam (24) is focused, and the analysis unit (57) is adapted to correct the measured distance value by calculation to compensate for the curvature of the field.
8. The device according to claim 7, characterized in that For the correction by calculation, the correction value can be read from a correction table stored in the analysis unit (57).
9. The device according to claim 8, characterized in that The correction table stores correction values for different working wavelength ranges.
10. The device according to any one of the preceding claims, characterized in that The scanning device (26) has exactly two scanning mirrors (28, 32) which are respectively supported so as to be rotatable about exactly one axis (30, 34), and at least one of the two scanning mirrors (28, 32) is assigned a collision protection device (64) which is suitable for limiting the rotation angle of the at least one scanning mirror (28, 32).
11. The device according to claim 10, characterized in that The collision protection device has at least one mechanical stop (64).
12. The device according to claim 10 or 11, characterized in that The collision protection device has an electronic limiting device, which is adapted to electronically prevent the supply of a control signal to the at least one scanning mirror (28, 32), which would otherwise exceed a preset rotation angle range.
13. The device according to any one of the preceding claims, characterized in that The optical system includes at least one anamorphic optical element (66, 68) for correcting for field curvature.
14. The device according to any one of claims 1 to 6, characterized in that The scanning device (26) has exactly one scanning mirror (28') which is mounted so as to be rotatable about two axes (30, 34).
15. A method for measuring a wafer (10), comprising the following steps: a) generating a measuring beam (24) with an optical coherence tomography scanner (22), wherein the wavelength range of the measuring beam (24) is selected such that the wafer (10) is at least partially transparent thereto; b) aligning the measuring beam (24) with the wafer (10) using the optical system (38); c) deflecting the measuring light beam (24) in two spatial directions by means of the scanning device (26), the scanning device being controlled in such a way that the measuring light beam scans the surface (40) of the wafer (10) at a plurality of measuring points in succession; d) calculating a distance value and / or a thickness value based on the interference signal provided by the optical coherence tomography scanner (22); e) calculating at least one characteristic variable of the wafer (10) from the distance value and / or the thickness value.
16. The method according to claim 15, characterized in that The characteristic variable is TTV, bow or warp.
17. The method according to claim 15 or 16, characterized in that In the optical coherence tomography scanner (22) A light source (42) generates light; A beam splitter (44) splits the light generated by the light source (42) into the measuring beam (24) and the reference beam (46); directing the reference beam (46) in a reference arm (48); guiding the measuring beam (24) in an object arm (50); The detector (56) generates the interference signal by superimposing a reference beam (46) guided in the reference arm (48) and a component of the measuring beam (46) reflected by the wafer (10).
