Method for producing and classifying rods made of semiconductor material

Abstract

The invention relates to a method for producing and classifying rods made of semiconductor material, the method comprising: - introducing a reaction gas, which contains at least one semiconductor component in addition to hydrogen, into a reaction chamber of a gas-phase deposition reactor, which reaction chamber is delimited by a reactor shell and a base plate, wherein the reaction chamber contains at least one heated carrier body, which is fastened to the base plate and on which the semiconductor material is deposited to form a rod; - opening the gas phase deposition reactor by vertically lifting the reactor shell from the base plate, wherein, during the lifting and / or during a lowering of the reactor shell that follows the lifting, two-dimensional and / or three-dimensional images of the rod are generated by at least one measuring device attached to a lower edge of the reactor shell; - optionally generating a total image from the obtained two-dimensional or three-dimensional images; - analysing the two-dimensional or three-dimensional images or the total image and determining at least one classification parameter, wherein the rod is fed to a further processing step depending on the classification parameter.

Description

[0001] Wa12233P / Be

[0002] 1

[0003] Methods for the production and classification of rods made of semiconductor material

[0004] The invention relates to a method for the production and classification of rods made of semiconductor material and a gas phase deposition reactor for the production of the rods made of semiconductor material.

[0005] Polycrystalline silicon (polysilicon) serves as the starting material in the production of single-crystal silicon, for example by crucible pulling (Czochralski process) or by zone melting (float zone process). Single-crystal silicon is used in the semiconductor industry for the manufacture of electronic components (chips).

[0006] Furthermore, polysilicon is required for the production of multicrystalline silicon, for example by block casting. The multicrystalline silicon obtained in block form can be used to manufacture solar cells.

[0007] Silicon carbide (SiC) has recently also been used as a semiconductor material for various electronic devices. SiC is characterized by high radiation hardness, a wide band gap, a high saturated electron drift velocity, a high operating temperature, and excellent electronic properties, including the absorption and emission of high-energy protons in the blue, violet, and ultraviolet regions of the spectrum. SiC semiconductors can be used at temperatures up to 250 °C and exhibit high oxidation resistance. This gives them advantages over pure silicon semiconductors. Wa12233P / Be

[0008] 2

[0009] Polycrystalline SiC (Poly-SiC) serves as the starting material for the production of single-crystal SiC semiconductor elements, for example using the PVT process.

[0010] Polysilicon and Poly-SiC can be obtained through a chemical vapor deposition process, which in the case of polysilicon is called the Siemens process.

[0011] In this process, support structures are heated in a bell-shaped reactor by direct electric current, and a reaction gas containing a silicon-containing component and hydrogen is introduced. A carbon-containing component is also required for the production of poly-SiC. The silicon-containing component is usually monosilane (SiH₄) or a chlorosilane / chlorosilane mixture. A typical example is trichlorosilane (SiHCl, TCS). For polysilicon production, SiH₄ or TCS mixed with hydrogen is predominantly used. For poly-SiC production, for example, a methylsilane or a TCS-methane mixture mixed with hydrogen can be used.

[0012] The design of a typical gas-phase deposition reactor for polysilicon production is described, for example, in US 2012 / 0100302 Al. The design of a typical gas-phase deposition reactor for poly-SiC production can be found in US 2023 / 0141427 Al. Generally, essentially identical reactors can be used.

[0013] The base of such a reactor (base plate) is generally equipped with electrodes that hold the support materials. These support materials are usually filaments (also called filament rods or thin rods), which can be made of materials such as silicon, graphite, or silicon carbide (SiC). Wa12233P / Be

[0014] 3

[0015] Typically, two substrates are connected by a bridge to form a pair, which creates an electrical circuit via the electrodes. The surface temperature of the substrates during deposition is usually above 1000 °C. At these temperatures, the silicon-containing or silicon-containing and carbon-containing components of the reaction gas decompose, and elemental silicon or SiC is deposited from the gas phase onto the substrates. This increases their diameter. Once a predetermined diameter is reached, the deposition is stopped, and the resulting rods of semiconductor material (polysilicon or poly-SiC) are removed. After removing the bridge, cylindrical rods can be obtained.

[0016] The morphology of polysilicon and SiC (the rods or fragments produced from them) can influence performance during further processing. Fundamentally, the morphology of a polysilicon or SiC rod is determined by the parameters of the deposition process (e.g., rod temperature, reaction gas composition, specific flow rate). Depending on the parameters, pronounced interfaces, even holes and grooves, can form. These are generally not distributed homogeneously within the rod. For example, varying the parameters can result in polysilicon rods with different (mostly concentric) morphology regions, as described, for instance, in EP 2 662 335 Al. The dependence of the morphology on the rod temperature is expressed, for example, in US 2012 / 0322175 Al.

[0017] The morphology of polysilicon and SiC can vary from compact and smooth to porous and fractured. Compact Wa12233P / Be

[0018] 4

[0019] Polysilicon or SiC is essentially free of cracks, pores, joints and fissures.

[0020] A porous and fractured morphology can, for example, negatively affect the crystallization behavior of polysilicon. This is particularly evident in the CZ process for producing single-crystal silicon for semiconductor applications. Here, the use of fractured and porous polysilicon leads to economically unacceptable yields. In general, particularly compact polysilicon results in significantly higher yields in the CZ process.

[0021] Accordingly, polysilicon, in particular, is distinguished and classified not only according to its purity and fracture size but also according to its morphology. Since the term morphology can encompass various parameters such as porosity (the sum of closed and open porosity), specific surface area, roughness, luster, and color, a reproducible determination of the morphology presents a challenge.

[0022] US Patent 2023 / 0011307 describes a method in which 2D or 3D images are generated after the removal of silicon rods or fragments produced from them. From these images, at least two surface structure parameters are generated using various image processing techniques and combined into a single morphology parameter. This parameter is then used for classification. A disadvantage is that removal of the rods is mandatory. The mechanical stress caused by removal can create additional cracks, which can lead to incorrect conclusions about the deposition process. Furthermore, usually only a small section of the rod surface or internal structure can be examined.

[0023] 5 are examined. Furthermore, it is usually not possible to draw conclusions about the position of the rods.

[0024] US Patent 2022 / 0234900 describes a process in which thermographic images of the rods are generated through a sight glass during polysilicon deposition, and a morphology index is then generated from these images using image processing. This index allows the quality of the deposited polysilicon to be determined and influenced at any given time. A disadvantage is that cracks or fractures in the rods may only appear during cooling, i.e., after the deposition process is complete. Furthermore, a pyrometer mounted in front of a sight glass typically only allows observation of a small section of the rod. Additionally, deposits on the sight glass or dust formation during deposition can affect image quality.

[0025] The invention was based on the objective of providing a method for assessing the morphology or surface properties of semiconductor rods directly after completion of chemical vapor deposition. The semiconductor rods should not require any additional manipulation.

[0026] This task is solved by a method for the fabrication and classification of rods made of semiconductor material, which includes the following steps:

[0027] - Introducing a reaction gas containing at least one semiconductor component in addition to hydrogen into a reaction chamber of a gas phase deposition reactor bounded by a reactor shell and a base plate, wherein the reaction chamber contains at least one heated support body attached to the base plate, on which the Wa12233P / Be

[0028] 6

[0029] Semiconductor material is deposited in the form of a rod,

[0030] - Opening the gas phase deposition reactor by vertically lifting the reactor shell from the base plate, wherein two- and / or three-dimensional images of the rod are generated by at least one measuring device attached to a lower edge of the reactor shell during the lifting and / or during a subsequent lowering of the reactor shell,

[0031] - If necessary, creating a composite image from the obtained two- or three-dimensional images,

[0032] - Analysis of the two- or three-dimensional images or the overall image and determination of a classification parameter, whereby the rod is fed to a further processing step depending on the classification parameter.

[0033] Generally, the best time to evaluate the rods is directly after lifting the reactor shell and before removing (manipulating) the rods. The method according to the invention offers the particular advantage of evaluating the rods without interrupting the actual removal process. After all, the reactor shell must be completely lifted and, if necessary, even removed laterally for rod removal.

[0034] The number of support bodies arranged in the gas-phase deposition reactor is generally irrelevant for the execution of the process according to the invention. Preferably, the gas-phase deposition reactor is a reactor as described in US 2012 / 0100302 Al or US 2023 / 0141427 Al. Accordingly, the support body preferably comprises two filament rods connected by a bridge to form a pair, thus approximately resembling an inverted "U" shape. Wa12233P / Be

[0035] 7

[0036] (or “n”). The free ends of the filament rods are typically connected to electrodes in the reactor base plate. Usually, far more than one support structure is arranged in the reaction chamber. Typical examples of the number of support structures (especially filament rod pairs) in a reactor are 12 (24 pairs), 18 (36 pairs), 24 (48 pairs), 27 (54 pairs), 36 (72 pairs), or 48 (96 pairs). The semiconductor rods, which grow in diameter during deposition, can be described as cylindrical to a good approximation. This approximation is justified because modern vapor deposition reactors are fundamentally designed to ensure the most homogeneous deposition possible. This is achieved in particular by a homogeneous reaction gas flow within the reactor and by a symmetrical arrangement of the support structures.The support bodies or filament rods can generally be cylindrical, but other geometries are also possible.

[0037] After deposition is complete, the semiconductor rods are removed – usually after a cooling period. If the support structure consists of the described filament rod pairs, the bridge is typically removed after removal. Preferably, the section of the semiconductor rod that was connected to the electrode is also removed.

[0038] Preferably, the measuring device comprises an imaging system selected from the group consisting of 2D cameras, 3D cameras, laser scanners, LiDAR (Light Detection and Ranging) cameras, infrared cameras, and combinations thereof. The cameras may be equipped with a fisheye and / or a wide-angle lens. Wa12233P / Be

[0039] 8

[0040] The camera can be monochrome or color. Preferably, it is a digital camera. Both area scan cameras (sensor as an array of pixels) and line scan sensors (with appropriate feed rate from the object to the camera) can be used.

[0041] Regarding various types of image generation, reference can be made to US 2023 / 0011307 Al.

[0042] For example, photogrammetry can be used to create a 3D model (or even a complete image) based on several generated 2D images.

[0043] In general, the generated 2D or 3D images depict the reaction space from several lateral perspectives (due to the vertical (along a longitudinal axis) lifting and possibly rotation (around the longitudinal axis) of the reactor shell).

[0044] The 2D or 3D images are preferably offset from each other at least along the longitudinal axis. The reaction chamber can be depicted perspectively from different azimuth angles (rotating the reactor shell), i.e., by horizontally offsetting the 2D or 3D images.

[0045] Preferably, the measuring device can determine the position of the two- or three-dimensional images using a distance measurement sensor. Accordingly, it is preferred if the measuring device includes a distance measurement sensor with which the distance to the base plate of the gas-phase deposition reactor and / or the distance to the semiconductor rods can be determined. Depending on the distance, image acquisition can be triggered and, if necessary, the focus adjusted. For example, an image series Wa12233P / Be

[0046] 9 are triggered every 1, 5 or 10 cm increasing distance from the base plate.

[0047] The sensor used for distance measurement could be, for example, a laser rangefinder or a gyroscope. Distance measurement can also be performed by triangulation using known points (or landmarks) at the reactor and / or in the surrounding area.

[0048] Alternatively or additionally, the cameras can also record a video during the lifting of the reactor casing. Time-lapse photography, or the generation of individual images at defined time intervals, is also possible (time-lapse). A LiDAR scanner can be used for this purpose, as well as a 2D camera. If a 3D laser scanner is used, the recording frequency usually needs to be increased to achieve higher scanning accuracy.

[0049] To better illuminate the reaction chamber or the semiconductor rod, the measuring device can be equipped with a lighting device. This lighting device can be, for example, an LED panel light, a rod light, an LED strip, or a point light. Preferably, the lighting device produces an area-wide illumination configuration.

[0050] The measuring device may also include means for measuring temperature, in particular surface temperature (e.g. pyrometer).

[0051] According to a preferred embodiment, at least two, preferably at least three, and particularly preferably at least four measuring devices can be mounted on the lower edge of the reactor shell, the measuring devices preferably being equidistant from each other. Wa12233P / Be

[0052] 10

[0053] In the case of a reactor shell with a circular cross-section (bell-shaped reactor shell), the measuring devices are preferably mounted at the same circular arc spacing.

[0054] A complete 360° image of the reaction chamber can be generated by multiple measuring devices during the lifting of the reactor shell. Moving the measuring devices along the reactor circumference (changing the azimuth angle) is not necessary. Alternatively, one or more measuring devices can be arranged to be movable along the reactor circumference. For example, a rail system can be provided for this purpose.

[0055] The reactor shell can rotate around its longitudinal axis during lifting to obtain a complete image of the reaction chamber and, in particular, to minimize the degree of obscuration by rods. This allows for a reduction in the number of measuring devices required in large gas-phase deposition reactors. The measuring device can also be designed to rotate or pivot, thereby increasing its focal area.

[0056] Preferably, the measuring device is attached to an outside of the reactor shell, preferably to a flange or projection circumferentially on the outside.

[0057] The measuring device is preferably detachably connected to the reactor shell. For example, the measuring device can be attached using an electromagnet or a clamping device. In this way, it can be attached only shortly before the reactor shell is lifted. Wa12233P / Be

[0058] 11

[0059] It is also conceivable that the measuring device is arranged on a separate ring (with a similar or identical circumference to the reactor shell) which is moved along the semiconductor rods after the reactor shell has been removed.

[0060] The measuring device is preferably modularly constructed from several sensor systems. In particular, the measuring device is a multi-sensor system. A typical measuring device can, for example, comprise two 2D cameras, two 3D cameras, a lighting device, and a distance sensor. The cameras can each be aligned at different angles to each other, thus each capturing a different focus area.

[0061] Furthermore, the measuring device may contain a computing or processing unit for storing the measurement data and / or for determining the classification parameter. A battery may be provided for independent power supply.

[0062] Preferably, the overall image of the semiconductor rods or the reaction chamber, generated from individual images, is produced using a computing unit integrated into the measuring device. The computing units of several measuring devices are preferably networked, particularly wirelessly. Furthermore, the measuring devices can be connected to a process control station.

[0063] The overall image does not necessarily represent the entire reaction chamber, i.e., a complete image of all semiconductor rods arranged within it. It could also represent a portion of the semiconductor rods. Wa12233P / Be

[0064] 12

[0065] Preferably, the overall image is created by panorama stitching from the 2D and / or 3D images.

[0066] Additionally, after removing the reactor casing, a camera, for example equipped with a fisheye lens, can be swiveled over the semiconductor rods to generate a top-down image. Preferably, the image is generated centrally (in line with the reactor's longitudinal axis) over the semiconductor rods.

[0067] The classification parameter can be a numerical value indicative of a local property of the bars and may also include localization information, i.e., marking an associated area in the 2D and / or 3D images or the overall image. The local property is preferably selected from the following: cracking on the bar surface, contact between two adjacent bars, bar diameter, angle of inclination of the respective bar, surface texture (formation of holes, pores, grooves) of the respective bar, color, and combinations thereof.

[0068] The classification parameter, which determines whether the semiconductor rod is fed into a processing step, is preferably selected from the group consisting of surface texture (formation of holes, pores, grooves), crack formation, rod diameter, angle of inclination, rod position (position on the base plate), color, and combinations thereof. Particularly preferably, at least the surface texture is used as a classification parameter.

[0069] The classification parameter can also be a morphology index, as described in US 2023 / 0011307 Al. Wa12233P / Be

[0070] 13

[0071] The classification parameter can, in particular, be a quality measure determined by regression or include such a quality measure.

[0072] As described at the beginning, depending on the deposition parameters, semiconductor material (especially polysilicon) with different surface properties (morphology) can form, whereby even on the same semiconductor rod, areas of different morphology can occur depending on location. Here, surface properties are understood to refer specifically to the fragmentation of the semiconductor material, which results from the frequency and arrangement of holes, pores, grooves, and cracks. Surface properties can also be understood as the porosity of the semiconductor material. Furthermore, dust deposits and the growth of dendrites are also subsumed under the term surface properties.

[0073] In general, the formation of holes, pores, grooves, and cracks during deposition is evident in a surface structure reminiscent of popcorn. Viewed in profile, a so-called popcorn surface is an accumulation of elevations (peaks, which may exhibit constrictions) and grooves (valleys).

[0074] Early detection of cracks in semiconductor rods is crucial for safety, as these cracks can lead to rod breakage during removal. Therefore, a preferred design includes at least one classification parameter that allows conclusions to be drawn about the number and severity of cracks (crack formation) in the semiconductor rods. In particular, the length, Wa12233P / Be, is considered.

[0075] 14

[0076] The width and depth of a crack are taken into account. The classification parameter can be the crack length, crack width, and / or crack depth. The latter parameters can be determined, for example, by high contrast differences (light to dark and back to light).

[0077] Varying rod diameters (within the same rod or between multiple rods) can indicate an inhomogeneous deposition process. Furthermore, different shredding, crushing, or breaking tools may be required depending on the rod diameter.

[0078] The angle of inclination of a rod can also be a sign of an inhomogeneous deposition process. Furthermore, it can indicate contamination, for example through contact with the inner wall of the reactor casing.

[0079] The rod position refers to the position of the rod in the reaction chamber, i.e., its anchoring point on the base plate. For example, rods near the wall of the reactor vessel may have different properties than rods located in the center of the reactor.

[0080] The color of a rod can provide an indication of dust deposits and its surface condition.

[0081] Determining these classification parameters enables differentiated and optimally tailored further processing based on the corresponding bar quality. For example, individual bars from a batch can be completely rejected, while the others meet the specified quality requirements and are processed further. The classification parameters can also be used to determine whether Wa12233P / Be

[0082] Polysilicon rods from a batch are more likely to be used for single crystal production via the FZ process (higher quality requirements) or the CZ process (lower quality requirements).

[0083] Preferably, the classification parameter is determined using a machine-learned model (machine-

[0084] (Learning system). The generated 2D or 3D images and / or the overall image of the rod are compared with a reference database containing historical, already classified images of semiconductor rods by the machine-learned model. Based on this comparison, the machine-learned model outputs the classification parameter, or the machine-learned model outputs image information from which the classification parameter can be calculated.

[0085] Typically, the generated 2D and / or 3D images and / or the overall image are added to the reference database for this comparison.

[0086] The reference database can, in particular, be a training dataset for retraining the machine-learned model.

[0087] Preferably, after adding the generated 2D and / or 3D images and / or the overall image remain in the database for training purposes of the machine-learned model.

[0088] It can therefore be a computer-implemented method in which at least one camera is arranged in the area of ​​a lower edge (bottom rim) of a removable and optionally rotatable reactor shell of the gas phase deposition reactor Wa12233P / Be

[0089] 16 is controlled to capture the 2D or 3D images synchronized with the lifting of the reactor shell (e.g. by a motorized lifting device), so that multiple perspectives are formed by moving at least one camera together with the reactor hood.

[0090] The output image information can consist of at least one bounding box, which is added to the generated 2D or 3D image and / or the overall image of the rod. This bounding box contains or graphically highlights the event underlying or corresponding to the classification parameter (e.g., crack, hole, pore, color change, dust deposition, inclined rod). Furthermore, the output image information or classification parameter can include a confidence interval. This is usually a percentage indicating the probability with which the machine-learned model believes it has detected the event in question.

[0091] In a particularly preferred embodiment, it is a machine-learned model for detecting cracks, pores, holes, trenches, dendrites, discolorations, i.e., features that can generally be attributed to the surface texture.

[0092] The classification parameter can also be determined using classic software-based image analysis methods. Examples include: - Processing the generated image with image filters, e.g.

[0093] Softening or formation of directional derivatives. Wa12233P / Be

[0094] 17

[0095] - Combination of different generated images or image sections to extract information (e.g. ShapeFrom Shading, i.e. separation of structure and texture).

[0096] - Photometric stereo image for generating depth information from multiple cameras.

[0097] - Segmentation of sub-areas of the generated image, e.g. separation of a sub-area of ​​the rod from the background, binarization with a fixed or dynamic threshold.

[0098] - Calculation of typical key figures (e.g. gray level co-occurrence matrix values ​​or histogram values) for the generated image or the viewed image section.

[0099] Preferably, the classification parameter is determined using software with a computing unit integrated into the measuring device. The completed information can be transmitted to a process control station.

[0100] For the machine-learned model, software such as Yolo v5, v7, v8 and subsequent versions can be used.

[0101] (YouOnlyLookOnce - open-source AI model), OpenCV, LabVIEW, and Halcon can be used. Generally, any CNN (convolutional neural network) can be used (free or commercial versions).

[0102] The further processing step to which the semiconductor rod(s) are fed depending on the classification parameter is preferably selected from the group including fraction size-dependent comminution, storage, packaging, recycling and combinations thereof.

[0103] A further processing step involves feeding the semiconductor rods to various connection methods. Wa12233P / Be

[0104] 18

[0105] High-quality polysilicon rods can be fed into the FZ or CZ process whole (apart from the removal of the electrode area and, if applicable, the bridge). These rods are usually pre-packaged and / or temporarily stored.

[0106] Recycling (essentially sorting) may be particularly appropriate when quality requirements (for example, due to dust accumulation) do not meet specifications. Recycling can involve grinding polysilicon and subsequently using it as a raw material for the production of chlorosilanes.

[0107] Fraction size-dependent comminution means that semiconductor rods, for example, are subjected to different comminution processes depending on their surface properties. These comminution processes can be carried out using different crushing tools.

[0108] Typically, there are five fracture size classes 0 to 4 (BGO to BG4), defined by the grain size of the fragments, where grain size is defined as the longest distance between two points on the surface of a silicon fragment. The fracture size classes group fractions with grain size ranges as follows.

[0109] BGO: 0, 1 to 9 mm

[0110] BG1: 1 to 18 mm

[0111] BG2: 5 to 50 mm

[0112] BG3: 20 to 65 mm BG4: 35 to 150 mm Wa12233P / Be

[0113] 19

[0114] The classification of silicon fragments can be carried out using mesh sieves, where the edge length of the square meshes corresponds to the upper limit of a group size. Optopneumatic classification methods are also known, as described, for example, in US 2009 / 120848 A.

[0115] The semiconductor component contained in the reaction gas preferably comprises at least one silane, in particular a halosilane. The halosilane is preferably selected from the group consisting of the chlorosilanes of the general formulas H n SiC14-n, H m C16-mSi2 , (CHa ) n SiC14-n with n = 1 to 3 and m = 0 to 4. This embodiment is particularly suitable for the production of polysilicon (see US 2012 / 0100302 Al mentioned in the introduction). Preferably, the halosilane is TCS or a mixture of dichlorosilane and TCS. Such a reaction gas composition is particularly suitable for the production of polysilicon according to the Siemens process.

[0116] In another embodiment, the semiconductor component contained in the reaction gas can comprise at least one silane and at least one carbon component. Alternatively or additionally, the semiconductor component can also comprise an organosilane. The semiconductor component can generally consist solely of an organosilane, since this contains both a carbon and a silicon source in a single molecule. This embodiment is particularly suitable for the production of poly-SiC (see US 2023 / 0141427 Al mentioned in the introduction). The organosilane is preferably selected from the group of organochlorosilanes with the general formulas (CHa)nH. m SiC14-nm with n = 0 to 3 and m = 0 or 1 and MenSiaClen-n with n = 1-5. The carbon component can be, for example, methane, ethane, propane, butane, and combinations thereof. Wa12233P / Be

[0117] 20

[0118] Another aspect of the invention relates to a gas phase deposition reactor for the production of rods made of semiconductor material, in particular for carrying out the described process.

[0119] The gas phase deposition reactor comprises a base plate and a reactor shell that define a reaction chamber, at least one inlet for supplying the reaction gas into the reaction chamber, at least one outlet for gas removal from the reaction chamber, and electrode holders arranged on the base plate for at least one support body, wherein at least one measuring device for generating 2D or 3D images is attached to an outer lower edge of the reactor shell.

[0120] Preferably, the gas phase deposition reactor is a (Siemens) reactor for the production of polysilicon and / or for the production of poly-SiC.

[0121] A particularly preferred embodiment of the gas-phase deposition reactor according to the invention comprises 16 cameras, 16 lights, and a laser distance meter for the base plate. Preferably, two cameras and two lights are provided per measuring device, resulting in a total of eight measuring devices. An embodiment in which four measuring devices each comprise four lights and four cameras is also conceivable. The location of the laser distance meter on each measuring device is irrelevant for feasibility. The cameras can be, for example, commercially available action game cameras. The lights can be commercially available LED panel lights. The cameras and / or lights are typically equipped with a battery. Wa12233P / Be

[0122] 21

[0123] The gas phase separation reactor may include lighting devices (preferably four) attached to the base plate, which illuminate the reaction chamber during the lifting of the reactor shell.

[0124] The last described method allows for a non-contact reconstruction of the rod positions before and after deposition. This significantly improves product quality. For example, gas flow optimization can be performed based on the rod position in cases of locally detected morphological deviations (e.g., rod shape, popcorn formation, cracks).

[0125] Regarding the design of the measuring device and the gas phase deposition reactor, reference can be made to the above explanations.

[0126] It is also possible, in principle, to use the measuring device in a reactor for gas-phase reactions, especially for the hydrogenation of chlorosilanes.

[0127] Such a reactor (often also called a converter) typically has a first, usually graphite, pressure-bearing reactor shell that encloses a reaction chamber in which the gas-phase reaction takes place. Heating elements are usually arranged around this reactor shell, and a second reactor shell is arranged around these to prevent heat loss. This second reactor shell thus encloses the first reactor shell and the heating elements. The measuring device can now be positioned at a lower end of this second reactor shell so that, when lifted vertically, both the heating elements and the first Wa12233P / Be

[0128] 22

[0129] Monitor the reactor shell using the image generation described above.

[0130] Accordingly, a method for monitoring a reactor for gas phase reactions using the measuring device is described.

[0131] Furthermore, a reactor for gas-phase reactions, in particular for the production of chlorosilanes, is described, which includes at least one measuring device.

[0132] Fig. 1 shows a section of a gas phase deposition reactor longitudinal section with measuring device.

[0133] Fig. 2 shows a section of a gas phase deposition reactor longitudinal section with measuring device.

[0134] Fig. 3 shows a sequence (AD) of a gas phase deposition reactor with measuring device during the lifting of the reactor shell.

[0135] Fig. 4 shows the cross-section of a gas phase deposition reactor with measuring devices and 21 pairs of rods.

[0136] Fig. 5 shows the cross-section of a gas phase deposition reactor with measuring devices and 12 pairs of rods.

[0137] Fig. 6 shows a measuring device.

[0138] Reference symbol list

[0139] 10 Gas phase deposition reactor

[0140] 11 lower section of the gas phase deposition reactor

[0141] 12 Reactor shell

[0142] 13 Reaction chamber

[0143] 14 Base plate

[0144] 15 Cooling circuit

[0145] 16 Semiconductor rod Wa12233P / Be

[0146] 23

[0147] 17 Bridge

[0148] 18 electrode

[0149] 19 bore

[0150] 20 flange

[0151] 30 Measuring device

[0152] 32 cases

[0153] 33 pins with distance sensor

[0154] 34 Camera

[0155] 35 Light source

[0156] 36 Magnet

[0157] 37 Focus area of ​​the measuring device

[0158] 38 Suspension

[0159] Figure 1 shows a longitudinal section of a gas-phase deposition reactor 10, showing only a section of the wall area of ​​the reactor shell 12. The reactor shell 12 no longer rests on the base plate 14 with cooling circuit 15, but has been moved vertically upwards. Two semiconductor rods 16 (either made of polysilicon or poly-SiC) are shown as examples in the reaction chamber 13, each attached to the base plate 14 via an electrode 18. For clarity, lines for reaction gases and for cooling the reactor shell 12 are not shown. A [missing information - likely a specific component or component] is attached to a lower area 11 of the reactor shell 12.

[0160] A measuring device 30 is attached to the outside of flange 20. The measuring device 30 has a housing 32 which is attached to the flange 20 by a pin 33 resting in a bore 19. A laser-based distance sensor can be located at the tip of the pin 33, which measures the distance to the base plate 14 through the bore. The measuring device 30 can be connected to the flange 20 either detachably or permanently. A camera 34 and an LED light source 35 are arranged in the housing 32. Wa12233P / Be

[0161] 24

[0162] Figure 2 also shows a longitudinal section of a gas-phase deposition reactor 10, showing only a section of the wall area of ​​the reactor shell 12. In contrast to Figure 1, two semiconductor rods 16 are shown as examples, connected to form a pair via a bridge 17. The measuring device 30 is connected to the reactor shell 12 via an electromagnet 36.

[0163] Figure 3 shows four longitudinal sections A, B, C, D of a gas-phase deposition reactor 10 with a measuring device 30 as shown in Figure 2. The reactor shell 12 is raised vertically in the sequence from A to D. The lifting process typically takes about 60 s. At its highest point, the reactor shell 12 can be rotated so that the cameras can view the bridges 17 from multiple perspectives. The measuring device 30 is thus moved vertically along the semiconductor rods 16 and, in this way, has captured the entire rod length, including the bridge 17, in position D.

[0164] Figure 4 shows the cross-section of a gas-phase deposition reactor 10, which, in its reaction chamber 13 (or attached to its base plate 15), comprises forty-two semiconductor rods 16 connected via a bridge 17. The six black markings on the reactor shell 12 represent the focus area 37 of six measuring devices 30, each arranged in the center of a marking. The focus area 37 of each measuring device 30 is selected such that the entire reaction chamber 13 can be captured (shown exemplarily by the four dashed lines).

[0165] Figure 5 shows the cross-section of a smaller gas phase deposition reactor 10 with twenty-four via bridges 17 to Wa12233P / Be

[0166] 25

[0167] pairs of connected semiconductor rods 16. This smaller gas-phase deposition reactor 10 manages with four measuring devices 30. Figure 6 shows a measuring device 30 in side view (left in the image) and in front view (right in the image).

[0168] The measuring device 30 comprises two swiveling cameras 34 and two LED light sources 35, which are attached to a suspension 38. The measuring device 30 is attached to the reactor shell 12 (see Fig. 2) by means of two magnets 36.

[0169] 30 has a modular design and can, for example, allow for the quick exchange of cameras and / or light sources. It is also conceivable to extend the horizontal part of the suspension 38 to accommodate additional cameras, light sources, and sensors. A typical configuration would, for example, consist of four cameras 34 and four light sources 35 arranged in alternating order.

Claims

Wa12233P / Be 26 Patent claims 1. Method for the production and classification of rods made of semiconductor material, comprising - Introducing a reaction gas containing at least one semiconductor component in addition to hydrogen into a reaction chamber of a gas phase deposition reactor bounded by a reactor shell and a base plate, wherein the reaction chamber contains at least one heated support body attached to the base plate, on which the semiconductor material is deposited to form a rod, - Opening the gas phase deposition reactor by vertically lifting the reactor shell from the base plate, wherein two- and / or three-dimensional images of the rod are generated by at least one measuring device attached to a lower edge of the reactor shell during the lifting and / or during a subsequent lowering of the reactor shell, - If necessary, creating a composite image from the obtained two- or three-dimensional images, - Analysis of the two- or three-dimensional images or the overall image and determination of at least one classification parameter, whereby the rod is fed to a further processing step depending on the classification parameter.

2. Method according to claim 1, characterized in that the measuring device comprises an imaging system selected from the group consisting of 2D camera, 3D camera, laser scanner, LiDAR camera, infrared camera and combinations thereof. Wa12233P / Be 27 3. Method according to claim 1 or 2, characterized in that the measuring device determines the position of the two- or three-dimensional images by means of a distance measurement sensor.

4. Method according to one of the preceding claims, characterized in that the measuring device illuminates the rod with a lighting means.

5. Method according to one of the preceding claims, characterized in that at least two, preferably at least three, particularly preferably at least four measuring devices are attached.

6. Method according to one of the preceding claims, characterized in that the reactor shell or the measuring device rotates during lifting.

7. Method according to one of the preceding claims, characterized in that the classification parameter is selected from the group consisting of surface texture, crack formation, bar diameter, angle of inclination, bar position, color and combinations thereof.

8. Method according to one of the preceding claims, characterized in that the determination of the classification parameter is carried out with a machine-learned model, wherein the generated two- or three-dimensional images and / or the overall image of the rod are compared with a reference database of the machine-learned model, wherein the reference database Wa12233P / Be The system contains 28 historical, already classified images of semiconductor rods, and the machine-learned model outputs the classification parameter based on the comparison or outputs image information from which the classification parameter can be determined.

9. Method according to claim 8, characterized in that the generated two- or three-dimensional images and / or the overall image are added to the reference database for training purposes of the machine-learned model.

10. Method according to one of the preceding claims, characterized in that the classification parameter is crack formation.

11. Method according to one of the preceding claims, characterized in that the further processing step is selected from the group consisting of fraction size-dependent comminution, storage, packaging and combinations thereof.

12. Method according to one of the preceding claims, characterized in that the semiconductor component comprises at least one silane, in particular trichlorosilane.

13. Method according to one of the preceding claims, characterized in that the semiconductor component comprises at least one silane and one carbon component and / or at least one organosilane.

14. Gas phase deposition reactor for the production of rods from semiconductor material, in particular for carrying out a method according to at least one of claims 1 to 10. Wa12233P / Be 29 13 , comprising a base plate and a reactor shell defining a reaction chamber, at least one inlet for supplying the semiconductor component into the reaction chamber, at least one outlet for gas removal from the reaction chamber and electrode holders arranged on the base plate for at least one heatable support body, wherein at least one measuring device for generating two- or three-dimensional images is attached to an outer lower edge of the reactor shell .

15. Gas phase deposition reactor according to claim 14, characterized in that the measuring device comprises at least one imaging system selected from the group consisting of 2D camera, 3D camera, laser scanner, LiDAR camera, infrared camera and combinations thereof.

Images