Methods for fault analysis and inspection system

Abstract

The invention relates to a method for defect analysis of a product, in particular a printed circuit board product, semiconductor wafer, galvanic cell or the like, with an inspection system (26) and an inspection system, wherein the inspection system comprises a projection device (32), an optical detection device (28) and a processing device, wherein white light is split into its spectral components by means of a spectrometer device of the projection device and a multichromatic light beam (37) formed from monochromatic light beams is projected onto a product at an angle of incidence β, wherein the optical detection device comprises a detection unit (29) with a camera (27) and a lens (28), wherein a multichromatic light beam (37) is reflected at the product in a detection plane (46) of the detection unit which extends transversely, preferably orthogonally to a product surface (38) of the product.which is captured by means of the camera, wherein the reflected multichromatic light beam is projected onto an image plane (49) of an image sensor of the camera, wherein height information of the product surface is derived by means of the processing device from a spatial distribution of the reflected multichromatic light beam in the image plane and a position of the optical detection unit relative to the product, wherein, within an interval of the spatial distribution of intensity values ​​of the reflected multichromatic light beam in the image plane extending over at least two pixels of the image sensor, the intensity value for each pixel is captured by means of the processing device, wherein, from the respective intensity values, a position of a median and / or mode of the spatial distribution in the image plane is determined by means of the processing device.

Description

[0001] The invention relates to a method for defect analysis of a product, in particular a printed circuit board product, semiconductor wafer, galvanic cell or the like, with an inspection system, and an inspection system, wherein the inspection system comprises a projection device, an optical detection device and a processing device, wherein white light is split into its spectral components by means of a spectrometer device of the projection device and a multichromatic light beam formed from monochromatic light beams is projected onto a product at an angle of incidence β, wherein the optical detection device comprises a detection unit with a camera and a lens, wherein a multichromatic light beam is reflected at the product in a detection plane of the detection unit extending transversely, preferably orthogonally to a product surface of the product, and is detected by means of the camera.wherein the reflected multichromatic light beam is projected onto an image plane of an image sensor of the camera, wherein height information of the product surface is derived by means of the processing device from a spatial distribution of the reflected multichromatic light beam in the image plane and a position of the optical detection unit relative to the product.

[0002] In known methods and inspection systems for product defect analysis, a height profile, for example of a semiconductor wafer, a printed circuit board (PCB), or components mounted on a PCB, is determined using a spectroscopic method. The product to be inspected can be a PCB, a semiconductor wafer, a photovoltaic cell, a galvanic cell, a fuel cell, a battery, or their components, or similar items. Since defects can occur during the manufacturing of PCBs, semiconductor wafers, or similar products, it is regularly necessary to analyze the product for potential defects. For example, conductive traces on the product or electronic components attached to them may be incomplete or incorrectly positioned.Such a defect analysis is regularly performed using an inspection system that includes a camera for capturing images of the product's surface. The camera can be a line scan camera with optical sensors arranged in a single line or row, or with three rows or rows of different sensors (RGB sensors), or an area scan camera with more than three rows or rows.

[0003] From EP 2 307 852 B1, an inspection system or method for defect analysis of PCBs is known, with which a PCB surface can be optically measured. A monochromatic light beam emanating from a white light source is passed through a prism to generate a multichromatic light beam, which is projected onto the printed circuit board product at an angle of incidence. Depending on the height difference of a surface of the printed circuit board product, the surface appears in a spectral color of the multichromatic light beam directed towards that height. A monochromatic light beam thus reflected from the surface in question is detected by means of an optical detection device, in particular a camera. A processing device or a device for data processing with software is then used to analyze the data.By taking into account the angle of incidence of the multichromatic light beam on the detection plane and the position of the printed circuit board product relative to the camera, a computer can calculate height information for the relevant surface of the printed circuit board product.

[0004] A disadvantage of known inspection systems and methods is that the surface of the product, or height differences within the surface, cannot always be reliably and accurately detected. Creating a height profile and obtaining a high-quality analytical image of the product's surface requires a sufficiently high resolution camera or camera chip. However, camera chips or image sensors with very high resolution are comparatively expensive, and there are physical limitations to pixel size.

[0005] Furthermore, the product must be scanned completely at the appropriate resolution, which significantly increases the amount of data to be processed by the processing device. Improving the accuracy of the product analysis therefore requires considerably greater computing power from the processing device and / or the data processing equipment used. At the same time, the product should be analyzed as quickly as possible, although a detailed analysis can slow down the analysis process due to the increased data processing effort.

[0006] Furthermore, a disadvantage is that the height of the product surface cannot be determined with the desired accuracy if the reflected multichromatic light beam projected onto the camera's image sensor lies within the green wavelength range. In particular, precise differentiation of the relevant wavelengths is not possible, as the available image sensors can only detect very small proportions of blue and red light in this case. This significantly complicates the precise determination of a specific wavelength of light.

[0007] The present invention is therefore based on the objective of proposing a method for the failure analysis of a product and an inspection system which enables failure analysis with improved accuracy and speed.

[0008] This problem is solved by a method having the features of claim 1 and an inspection system having the features of claim 15.

[0009] In the inventive method for defect analysis of a product, in particular a printed circuit board product, semiconductor wafer or the like, with an inspection system, the inspection system comprises a projection device, an optical detection device and a processing device, wherein white light is split into its spectral components by means of a spectrometer device of the projection device and a multichromatic light beam formed from monochromatic light beams is projected onto a product at an angle of incidence β, wherein the optical detection device comprises a detection unit with a camera and a lens, wherein a multichromatic light beam is reflected at the product in a detection plane of the detection unit which extends transversely, preferably orthogonally to a product surface of the product, and which is detected by means of the camera.wherein the reflected multichromatic light beam is projected onto an image plane of an image sensor of the camera, wherein height information of the product surface is derived by means of the processing device from a spatial distribution of the reflected multichromatic light beam in the image plane and a position of the optical detection unit relative to the product, wherein the intensity value for each of the pixels is recorded by means of the processing device within an interval of the spatial distribution of intensity values ​​of the reflected multichromatic light beam in the image plane extending over at least two pixels of the image sensor, and wherein a position of a median and / or mode of the spatial distribution of the image plane is determined by means of the processing device from the respective intensity values.

[0010] Depending on the height difference of a surface of the reference object or the product, the wavelength of the multichromatic light beam, which is directed at the detection plane at an angle of incidence β, shifts due to the change in height within the detection plane. Here, a multichromatic light beam is understood to be a beam formed from a bundle of light rays of different wavelengths. The angle of incidence β of this multichromatic light beam is relative to a central axis, or optical axis, of the multichromatic light beam. The optical axis of the multichromatic light beam extends from the spectrometer to a point on the surface within the detection plane. The wavelength of the multichromatic light beam, which varies with the height of the surface, causes the corresponding reflection image in the camera's image plane to also shift relative to the optical axis.The image plane is mapped depending on the wavelength or height. By mapping the image onto the camera's image plane, the processing unit can then calculate height information for the surface. This is achieved by capturing the wavelength or color information of the reflected image. The processing unit then calculates height information for the surface in question based on the position of the optical scanning unit relative to the reference object or product and the image on the image plane or the color information of the reflected image.

[0011] According to the invention, the multichromatic light beam reflected from the product is projected onto the image plane or image sensor over at least two pixels or a greater number of pixels. Since pixels are regularly arranged in sensor rows, the projection can extend over at least two sensor rows. The extent of the projection onto the image sensor corresponds to the interval of the spatial distribution of intensity values ​​of the projection. An intensity value here refers to a light intensity or saturation that results in a more or less strong signal at the respective pixels. The processing device then detects the respective signal or intensity value for each pixel. From the typically differing intensity values ​​of the pixels, the processing device calculates or determines the position of the median and / or mode.The mode of the spatial distribution of light intensity or saturation in the image plane of the image sensor. The median is understood here as the mean or arithmetic mean of the intensity values. The mode is understood here as a maximum of the spatial distribution of intensity values. The median and mode can be congruent or different in their position. As it turns out, determining the position of the median and / or the mode makes it possible to further improve the accuracy of height information determination. This is particularly true because the interval of the spatial distribution of the intensity values, or light intensity or saturation, on the image sensor is taken into account. Preferably, the interval extends over more than two pixels; however, height information can also be determined if the interval lies within two pixels.

[0012] The processing device allows the location of the median and / or mode to be determined independently of the location of individual pixels. The location or position can be referenced to a reference point on the image plane or image sensor that is independent of the pixels. Thus, the location or position of the median and / or mode can be specified without direct reference to a pixel. This also makes it possible to obtain a location or position specification that is not based on a scale defined by the size of the pixels. The processing device can determine a pixel with a maximum value as the highest intensity value of the pixels, and can assign neighboring pixels in the sensor row a percentage of this maximum value corresponding to their intensity value.Since the intensity values ​​within the interval differ and regularly rise and fall in a pattern up to a maximum, the processing device can identify at least one pixel with the highest intensity value of all pixels in the interval. Pixels adjacent to the pixel with the maximum value have a comparatively low intensity value. The processing device can then assign a percentage of the maximum intensity value to each pixel within the interval, based on its respective intensity value and the resulting deviation from the maximum value. This significantly simplifies the calculation of the median and / or mode.

[0013] The processing device allows the position of the median and / or mode within a pixel's height to be determined. For example, the pixel height or length can be 5 µm or less. The median and / or mode can then be located within this length. This enables the inspection system to achieve a resolution exceeding that determined by the pixel's length alone. It also makes it possible to use less expensive image sensors with comparatively large pixel heights and still obtain measurement results with improved accuracy compared to the sensor's physical resolution.

[0014] The intensity of the reflected multichromatic light beam can be spatially distributed in the image plane according to a normal distribution and / or Gaussian distribution curve. This normal distribution and / or Gaussian distribution curve can be symmetrical about the median or mode. The processing device can determine the normal distribution and / or Gaussian distribution curve for each wavelength of the multichromatic light beam detectable by the image sensor. Based on the normal distribution and / or Gaussian distribution curve for a wavelength of the multichromatic light beam, the processing device can precisely determine the exact position of the corresponding median or mode in the image plane. If no normal distribution and / or Gaussian distribution curve is observed, the product under test may be defective. Furthermore, the spectrometer may have a comparatively large aperture, allowing for a greater number of color information points to be captured.Wavelength information is represented in the image plane. This results in a broader normal distribution and / or Gaussian distribution curve for each wavelength, enabling improved differentiation between wavelengths.

[0015] This distribution can be defined via the optics of the spectrometer. For example, a Gaussian distribution can be specified in the optical design, and the reflection from the image sensor or camera can be transferred to the processing device. The spectrometer separates the wavelengths of white light into their component wavelengths. To obtain further information, the individual wavelengths can be represented in terms of their intensity using a Gaussian bell curve. This can involve small wavelength packets for each wavelength, which are kept as narrow as technically possible and practical.

[0016] The reflected multichromatic light beam can be projected onto the camera's image plane by means of a dispersive or diffractive element in the detection unit. The processing device then derives the color information from the spatial or two-dimensional distribution of saturation values ​​of the reflected multichromatic light beam in the image plane. When the multichromatic light beam, which is composed of monochromatic light beams, is projected onto the image plane of the detection device or camera via the dispersive or diffractive element, a point or pixel of the camera in the image plane can detect the light beam, depending on the wavelength of the monochromatic light beam, which represents the height information of the object's surface. The respective color or wavelength information can then be determined based on the position of the point in the image plane.Furthermore, a longitudinal offset of the monochromatic light ray in question can occur, leading to a blurred image in the image plane for that ray. The dispersive or diffractive element can be designed to optically correct this longitudinal chromatic aberration of the multichromatic light ray or the relevant monochromatic components for the image plane. This makes it possible to obtain a substantially sharp image in the image plane, and thus more precise height information, regardless of height information or the wavelength of the monochromatic light ray projected onto the image plane. Height resolution can therefore be significantly improved.

[0017] The processing device allows intensity values ​​to be simultaneously acquired from at least five, preferably eight, pixels of a sensor column of the image sensor with the highest saturation values. A shift of the reflected image in the image plane is then determined by evaluating the spatial distribution or image width of intensity values ​​or saturation values ​​using the processing device. The location of the shift is derived from the spatial distribution or image width of the intensity values, since the pixels of the image sensor with the above-average saturation values ​​represent this location. Therefore, it is sufficient to identify just two pixels with above-average intensity values ​​using the processing device and to acquire the images projected onto these pixels in the image plane. This not only determines the height or image width of the shift but also its position.Not only is a topography possible, but images of the surface of the reference object or product can be captured simultaneously. By using a dispersive or diffractive element, which can be positioned directly in front of the camera or camera chip, line images in different color gradations can be obtained.

[0018] Simultaneous intensity values ​​in the wavelength ranges of red, green, blue (RGB), and preferably cyan, can be captured. This is made possible in particular by the use of an area scan camera. The processing device can then analyze the image information, specifically RGB color information, separately for hue, saturation, and L-value within a color space (HSV, HSL, HSB). With a line scan camera, for example, six individual lines with different color filters (e.g., RGB + magenta, cyan, yellow) can be used to filter out mixed colors, significantly improving the analysis accuracy of the inspection system. It has been found that it is particularly advantageous to be able to filter out and capture cyan (480 nm to 510 nm) in addition to RGB color information. This allows for more precise measurements when predominantly green color information is present.This allows for improved differentiation of green color information due to the additional cyan color information. Capturing the cyan color information can accelerate the determination of altitude.

[0019] Using the processing device, the position of the median and / or mode of the respective wavelength can be determined simultaneously for different wavelengths. A wavelength can also be understood as a limited wavelength range.

[0020] The processing device can derive the surface height information of the product from the color information captured by the product using a lookup table. A lookup table is particularly easy to generate and process using the device. Furthermore, the processing speed of the device can be further optimized by using the lookup table.

[0021] Changing the intensity values ​​in the image plane of the image sensor can trigger asynchronous signals from the individual pixels, which can then be processed by the processing device. Specifically, the pixels can detect a change in light intensity or an intensity value and initiate a signal independently of each other. The image sensor then generates signals that can be output asynchronously. In contrast, signals from, for example, a CCD image sensor are only available within a single sampling frequency. In this case, the state of all light-sensitive pixels is determined synchronously at the respective sampling frequency. If the processing device handles asynchronous signals from the image sensor, changes in height information can also be continuously detected. This enables faster determination of height information.

[0022] Furthermore, only signals from pixels that detect a change in intensity values ​​can be transmitted from the image sensor to the processing device. Consequently, signals from pixels that cannot detect a change in light intensity or intensity values ​​are not transmitted. No signal is generated for these pixels. This allows image information in the image plane to be limited solely to the changed image areas. The amount of image data processed by the processing device can thus be significantly reduced. This also makes it possible to process the image information from the image sensor more quickly through image processing, as the data volume can be significantly reduced. In particular, the image sensor's analog-to-digital converter then has to process a smaller amount of data.

[0023] By means of an offsetting device in the projection apparatus, at least two multichromatic light beams with wavelength ranges differing by 20 nm to 60 nm, preferably by 30 nm to 40 nm, can be projected onto the product and projected onto the image plane of the image sensor with a relative offset. Accordingly, the offsetting device can be configured such that the two multichromatic light beams are projected onto the product and detected by the image sensor essentially simultaneously. Alternatively, it is also possible for the two multichromatic light beams to be projected onto the product and detected by the image sensor successively. The offset of the slightly differing wavelength ranges of the multichromatic light beams by 20 nm to 60 nm, preferably by 30 nm to 40 nm, allows height information of the product surface to be obtained using both multichromatic light beams.This allows for even more precise determination of height information. Problems in determining a wavelength range in the image plane, as can occur with green color information, can be compensated for by the simultaneous wavelength offset of 20 nm to 60 nm, preferably 30 nm to 40 nm, in the image plane. This allows, for example, the acquisition of color information closer to red color information, enabling better differentiation of the relevant green wavelength.

[0024] Alternatively, the projection device can include a lighting unit with a light source, which can be a matrix LED. A matrix LED, in which all LEDs can be individually controlled, can project at least two multichromatic light beams with wavelength ranges differing by 20 nm to 60 nm, preferably by 30 nm to 40 nm, onto the product. By switching on two rows, the matrix LED behaves like a system with two illuminations and two apertures. This significantly simplifies the setup, eliminates the need for an aperture for one illumination, and simplifies adjustment, as the system can switch row by row until the desired image settings are achieved. Furthermore, during scanning, rows can be switched alternately, resulting in a shift of the spectral light in the z-direction or at a certain height, thus amplifying or reducing areas of lower sensitivity.can be improved.

[0025] The product can have at least a partially transparent surface, and at least two multichromatic light beams with differing wavelength ranges can be projected onto the image plane of the image sensor, offset from each other. If the product has at least one transparent, translucent surface, such as a lens array, a substrate, a layer of a coating, or the like, then the thickness or distance of the optically effective interfaces can be measured. A reflection from the product is generated for each interface, with each multichromatic light beam having a different wavelength range. Thus, height information of the interfaces can be obtained using the multichromatic light beams.

[0026] The inspection system according to the invention for the defect analysis of a product, in particular a printed circuit board product, semiconductor wafer, galvanic cell or the like, comprises a projection device, an optical detection device and a processing device, wherein the projection device has at least one spectrometer device by means of which white light can be split into its components and a multichromatic light beam formed from monochromatic light beams can be projected onto a product at an angle of incidence β, wherein the optical detection device has a detection unit with a camera and a lens, wherein the multichromatic light beam reflected from the product can be detected by means of the camera in a detection plane of the detection unit which extends transversely, preferably orthogonally to a product surface of the product.wherein the reflected multichromatic light beam can be projected onto an image plane of an image sensor of the camera, wherein height information of the product surface can be derived by means of the processing device from a spatial distribution of the reflected multichromatic light beam in the image plane and a position of the optical detection unit relative to the product, wherein, by means of the processing device, the intensity value for each of the pixels can be detected within an interval of the spatial distribution of intensity values ​​of the reflected multichromatic light beam in the image plane extending over at least two pixels of the image sensor,wherein, by means of the processing device, the position of a median and / or mode of the spatial distribution in the image plane can be determined from the respective intensity values. For the advantages of the inspection system according to the invention, reference is made to the description of advantages of the method according to the invention.

[0027] The image sensor can be an RGB chip or a grayscale chip, which can have 1024 sensor lines, preferably 32 to 128 sensor lines, and particularly preferably 32 to 64 sensor lines, oriented relatively transversely, preferably orthogonally, to the direction of movement of a product or reference object. For example, an area scan camera can have 1024 pixels x 36 sensor lines, 2048 pixels x 72 sensor lines, 4096 pixels x 128 sensor lines, 16384 pixels x 256 sensor lines, up to 1024 sensor lines or more. A grayscale chip can, in principle, achieve a higher resolution, since all sensor lines of the grayscale chip can be used independently of the wavelength of the light. However, it is also possible to use an RGB chip or a chip with several, e.g., 6, different color filters.

[0028] The projection device can emit light in the wavelength ranges of red, green, blue (RGB), infrared (IR), and / or ultraviolet (UV), preferably in a wavelength range of 400 to 700 nm, and the camera can detect this light. Furthermore, the projection device can include a lighting device, wherein diffuse light can be projected onto the product or reference object by means of the lighting device, and the light of the diffuse light reflected from the product or reference object at the detection plane can be detected by the camera. The lighting device can include a diffuser by means of which a homogeneous distribution of the light on the surface of the product or reference object can be achieved while simultaneously avoiding strong contrasts. The lighting device can emit light in the wavelength ranges of red, green, and blue (RGB), infrared (IR), and / or ultraviolet (UV).It is also possible to select specific color components of the light to mix particular wavelength ranges. The lighting device can consist of a number of light-emitting diodes (LEDs) arranged in a row or a matrix. Furthermore, the lighting device may include a polarizing filter.

[0029] The projection device can include an offset device, wherein the offset device is formed by a double slit aperture, a piezoelectrically actuated optical element or a splitter mirror, wherein at least two multichromatic light beams with wavelength ranges differing by 30 nm to 40 nm, preferably by 30 nm to 40 nm, can be projected onto the product by means of the offset device.

[0030] The image sensor can be designed such that a change in intensity values ​​in the image plane of the image sensor causes asynchronous signals from the respective pixels, which can then be processed by the processing device. The image sensor can be an event-based signal transmission sensor. This ensures that only image information or signals triggered solely by a change in light intensity or intensity values ​​are transmitted. The amount of data to be processed by the processing device can thus be significantly reduced.

[0031] The offset device and the method carried out therewith can in themselves be subject matter of the invention, independent of the method and inspection system according to the invention, or need not have the characterizing features of claim 1.

[0032] The image sensor with event-based signal transmission and the method carried out therewith can also be inventions in themselves, independent of the inventive method and the inventive inspection system, or need not have the characterizing features of claim 1.

[0033] Further advantageous embodiments of an inspection system result from the feature descriptions of the dependent claims relating to method claim 1.

[0034] A preferred embodiment of the invention is explained in more detail below with reference to the accompanying figures.

[0035] They show: Fig. 1: a simplified schematic representation of an embodiment of the inspection system in a side view; Fig. 2: a simplified schematic representation of an image sensor in a longitudinal section view; Fig. 3: a schematic representation of a projection onto the image sensor; Fig. 4: A schematic representation of another projection onto the image sensor.

[0036] The Fig. Figure 1 shows a simplified schematic diagram of an embodiment of an inspection system 26 in a side view. The inspection system 26 comprises a camera 27 or a camera chip. A detection device 28 with a detection unit 29, comprising the camera 27, a lens 30, and a dispersive element 31, is also shown, together with a projection device 32. The projection device 32 has a light source 33 that emits white light, an aperture 34, and another dispersive element 35. The additional dispersive element 35 is designed as a further prism 36, by means of which a multichromatic light beam 37 is projected onto a surface 38 transversely to a direction of movement of a product (not shown in detail here) or a reference object, indicated by an arrow 39.

[0037] The lens 30 comprises a lens assembly 40, shown schematically here, a front lens 41, and an image-side aperture 42. The aperture 42 is specifically designed as a slit aperture 43. The front lens 41 is segment-shaped and has two parallel boundary surfaces 45 arranged coaxially with an optical axis 44. The optical axis 44 passes through a detection plane 46 of the detection unit 29, the detection plane 46 being orthogonal to the surface 38, which corresponds to an object plane 47. The multichromatic light beam 37, or a bundle of corresponding light beams, thus strikes the surface 38, or the object plane 47, at an angle β relative to the detection plane 46 and is reflected from there into the lens 30. The dispersive element 31, designed as a prism 48, is arranged between the lens 30 and the camera 27.The prism 48 disperses the light exiting the lens 30 and projects it onto the camera 27 or its image plane 49.

[0038] Based on the spatial distribution of the reflected multichromatic light beam 37 on the camera 27 or an image sensor 50, height information of the surface 38 relative to the camera 27 is derived by means of a processing device (not shown here). For this purpose, sensor lines of the camera 27, which run parallel to the detection plane 46 (not shown here), are evaluated, whereby five sensor lines with the highest or maximum saturation values ​​can be detected. The height information can then be calculated from the position of the sensor lines relative to the detection plane 46 and the angle of incidence β. Furthermore, the processing device superimposes the sensor lines or their line images. These line images are then combined to form an analysis image of the product or the reference object.

[0039] The Fig. Figure 2 shows a cross-sectional view of the image sensor 50 of the camera 27 in a schematic diagram, as well as a distribution of intensity values ​​or light intensity on the image sensor 50. The image sensor 50 is formed, in particular, from a substrate 51 with a plurality of pixels 52 formed therefrom. The pixels 52 extend in sensor rows and sensor columns over a surface 53 of the image sensor 50. The surface 53 corresponds to the image plane 49. In the Fig. Figure 2 shows a distribution of light intensity, corresponding to a normal distribution or a Gaussian distribution curve, across a series of pixels 52. The distribution of light intensity lies within an interval I of a spatial distribution in the image plane 49 and extends over a number of pixels 52, each with a length L. A median or mode 54 of the normal distribution or Gaussian distribution curve 54 lies on one of the pixels 52. Since different intensity values ​​are recorded by each pixel 52, the processing device can calculate the position of the median 55 with relative accuracy. This allows for a resolution that is significantly more precise than a resolution defined solely by the length L of the pixel 52.

[0040] According to the Fig. Figure 3 shows a first distribution curve 56 and a second distribution curve 57 in the image plane 49. Both distribution curves 56 and 57 are generated by means of an offset device of the projection apparatus (not shown here). Fig. 1. The processing device can determine a median and / or mode 58 or 59 for each of the two distribution curves 56 and 57. Mode 58 is offset from mode 59 by a distance D. This distance can be, for example, 30 nm to 40 nm. Depending on the wavelength of distribution curve 56 and the second distribution curve 57, the processing device can be used to determine the position of mode 58 or 59 in the image plane 49 with particular accuracy using either distribution curve 56 or 57.

[0041] In the Fig. Figure 4 shows a first distribution curve 60 and a second distribution curve 61 in the image plane 49. Both distribution curves 60 and 61 are projected using the projection device. Fig.A multichromatic light beam is generated by projecting it onto a transparent layer of a product, with two multichromatic light beams being reflected from optical interfaces of the layer. The processing device can determine a median and / or mode 62 or 63 for each of the two distribution curves 60 and 61. Mode 62 is offset from mode 63 by an offset H. Depending on the wavelength of distribution curve 60 and the second distribution curve 61, the position of mode 62 or 63 in the image plane 49 can be precisely determined using the processing device. From this position, height information, and thus a layer thickness, can be derived for each distribution curve 60 and distribution curve 61, regardless of their respective intensity. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] EP 2 307 852 B1

[0003]

Claims

[1] Method for defect analysis of a product, in particular a printed circuit board product, semiconductor wafer, electrochemical cell or the like, with an inspection system (26), wherein the inspection system comprises a projection device (32), an optical detection device (28) and a processing device, wherein white light is split into its spectral components by means of a spectrometer device of the projection device and a multichromatic light beam (37) formed from monochromatic light beams is projected onto a product at an angle of incidence β, wherein the optical detection device comprises a detection unit (29) with a camera (27) and a lens (28), wherein a multichromatic light beam (37) is reflected at the product in a detection plane (46) of the detection unit which extends transversely, preferably orthogonally to a product surface (38) of the product, and is detected by means of the camera.wherein the reflected multichromatic light beam is projected onto an image plane (49) of an image sensor (50) of the camera, wherein height information of the product surface is derived by means of the processing device from a spatial distribution of the reflected multichromatic light beam in the image plane and a position of the optical detection unit relative to the product, characterized by , that by means of the processing device within an interval (I) of the spatial distribution of intensity values ​​of the reflected multichromatic light beam in the image plane extending over at least two pixels (52) of the image sensor, the intensity value for each pixel is detected, wherein by means of the processing device a position of a median (55) and / or mode of the spatial distribution in the image plane is determined from the respective intensity values. [2] Method according to claim 1, characterized by, that by means of the processing device the position of the median (55) and / or mode is determined independently of the position of the individual pixels (52). [3] Method according to claim 1 or 2, characterized by , that the processing device determines the pixel (52) with a maximum value as the highest intensity value of the pixels (52), wherein the processing device assigns to neighboring pixels of a sensor row a percentage of the maximum value corresponding to their intensity value. [4] Method according to any of the preceding claims, characterized by , that the processing device determines the position of the median (55) and / or mode within a height (L) of a pixel (52). [5] Method according to any of the preceding claims, characterized by, that a variance of the intensity of the reflected multichromatic light ray (37) in the image plane (49) is spatially distributed according to a normal distribution and / or Gaussian distribution curve (54, 58, 59). [6] Method according to any of the preceding claims, characterized by , that the reflected multichromatic light beam (37) is projected onto the image plane (49) of the image sensor (50) of the camera (27) by means of a dispersive or diffractive element (31) of the detection unit (29). [7] Method of one of the preceding claims, characterized by , that intensity values ​​are simultaneously detected by means of the processing device of at least five, preferably eight pixels (52) of a sensor column of the image sensor (50) with the highest intensity values. [8] Method according to any of the preceding claims, characterized by, that intensity values ​​in the wavelength ranges red, green, blue (RGB) and preferably cyan are simultaneously recorded. [9] Method according to any of the preceding claims, characterized by , that by means of the processing device for different wavelengths the position of the median (55) and / or mode of the respective wavelength is determined simultaneously. [10] Method according to any of the preceding claims, characterized by , that the processing device derives the height information of the surface (38) of the product from colour information captured from the product by means of a conversion table. [11] Method according to any of the preceding claims, characterized by , that by changing the intensity values ​​in the image plane (49) of the image sensor (50) asynchronous signals of the respective pixels (52) are caused, wherein the asynchronous signals are processed by means of the processing device. [12] Method according to any of the preceding claims, characterized by , that only signals from the pixels (52) that detect a change in intensity values ​​are transmitted from the image sensor (50) to the processing device. [13] Method according to any of the preceding claims, characterized by , that by means of an offset device of the projection device (32) at least two multichromatic light beams (37) with wavelength ranges differing from each other by 20 nm to 60 nm, preferably by 30 nm to 40 nm, are projected onto the product and projected onto the image plane (49) of the image sensor (50) offset relative to each other. [14] Method according to any of the preceding claims, characterized by, that the product has at least a partially transparent surface and that at least two multichromatic light rays (37) with different wavelength ranges are projected onto the image plane (49) of the image sensor (50) offset from each other. [15] Inspection system (26) for defect analysis of a product, in particular a printed circuit board product, semiconductor wafer, electrochemical cell or the like, wherein the inspection system comprises a projection device (32), an optical detection device (28) and a processing device, wherein the projection device has at least one spectrometer device by means of which white light can be split into its spectral components and a multichromatic light beam (37) formed from monochromatic light beams can be projected onto a product at an angle of incidence β, wherein the optical detection device comprises a detection unit (29) with a camera (27) and a lens (28), wherein the multichromatic light beam (37) reflected from the product can be detected by the camera in a detection plane (46) of the detection unit which extends transversely, preferably orthogonally to a product surface (38) of the product,wherein the reflected multichromatic light beam can be projected onto an image plane (49) of an image sensor (50) of the camera, wherein height information of the product surface can be derived by means of the processing device from a spatial distribution of the reflected multichromatic light beam in the image plane and a position of the optical detection unit relative to the product, characterized by , that by means of the processing device within an interval (V) of the spatial distribution of intensity values ​​of the reflected multichromatic light beam in the image plane extending over at least two pixels (52) of the image sensor, the intensity value for each of the pixels can be detected, wherein by means of the processing device a position of a median (55) and / or mode of the spatial distribution in the image plane can be determined from the respective intensity values. [16] Inspection system according to claim 15, characterized by, that the image sensor (50) is formed by an RGB chip or a grayscale chip which has up to 1024 sensor lines, preferably 32 to 128 sensor lines, particularly preferably 32 to 64 sensor lines, arranged relatively transversely, preferably orthogonally to a direction of movement (39) of a product. [17] Inspection system according to claim 15 or 16, characterized by , that the projection device (32) comprises an offset device, wherein the offset device is formed by a double slit aperture, a piezoelectrically actuated optical element or a splitter mirror, wherein at least two multichromatic light beams (37) with wavelength ranges differing from each other by 20 nm to 60 nm, preferably by 30 nm to 40 nm, can be projected onto the product by means of the offset device. [18] Inspection system according to any one of claims 15 to 17, characterized by, that the image sensor (50) is designed such that a change in the intensity values ​​in the image plane (49) of the image sensor causes asynchronous signals of the respective pixels (52), wherein the asynchronous signals can be processed by means of the processing device. [19] Inspection system according to any one of claims 15 to 18, characterized by , that the image sensor (50) is an image sensor with event-based signal transmission.

Images