Transport system for dry etching of solar cell wafers

Abstract

The present invention relates to a solar cell manufacturing system, and more specifically, to an integrated transport system for dry etching square solar cell wafers of various sizes, such as M6, M10, and M12. In particular, the present invention relates to an automation system in which a loader elevator, a load lock chamber, a transfer module, an alignment module, and an etching chamber are connected in a vacuum continuous processing structure, and to a dry etching equipment comprising a wafer size automatic recognition function, a common chuck, and a replaceable edge ring structure.

Description

Technology Field

[0001] The present invention relates to a solar cell manufacturing system, and more specifically, to an integrated transport system for dry etching square solar cell wafers of various sizes, such as M6, M10, and M12. In particular, the present invention relates to an automation system in which a loader elevator, a load lock chamber, a transfer module, an alignment module, and an etching chamber are connected in a vacuum continuous processing structure, and to a dry etching equipment comprising a wafer size automatic recognition function, a common chuck, and a replaceable edge ring structure. Background Technology

[0002] The solar cell industry is growing rapidly along with the increasing global demand for sustainable energy. In particular, to improve solar cell efficiency and reduce production costs, substrates are continuously being made larger and thinner, and currently, large-area substrates of 156mm x 156mm or larger, ranging from 210mm x 210mm, are becoming the mainstream. At the same time, substrate thickness is being reduced to less than 200μm to reduce material costs, which is significantly increasing the mechanical fragility of the substrates.

[0003] In solar cell manufacturing, dry etching is utilized as a core process for forming anti-reflective films, surface texturing, and pattern formation. Compared to wet processes, plasma-based dry etching allows for precise control and is environmentally friendly, establishing itself as an essential technology in modern solar cell manufacturing. However, due to the characteristics of large-area thin substrates, there are several limitations to applying existing semiconductor process equipment as is.

[0004] Conventional dry etching and semiconductor manufacturing equipment is primarily designed based on circular semiconductor wafers and is therefore unsuitable for square-shaped solar cells. While circular semiconductor wafers generally have a thickness of about 825 μm, solar cell wafers are extremely thin, measuring around 200 μm—only about one-quarter the thickness. These thin substrates are prone to deformation or damage during transport and processing, requiring the application of special transport systems.

[0005] In conventional wafer alignment methods, it is common practice to use a rotary chuck to rotate the substrate while detecting edges or alignment marks with a fixed sensor. While this rotary alignment is effective for small, thick, circular semiconductor wafers, in large-area square solar cell substrates, deformation or damage to the substrate may occur due to centrifugal force and vibration caused by rotation. In particular, the inertial force generated when rotating a large-area substrate measuring 210mm x 210mm can apply significant stress to the edges of the substrate, potentially causing microcracks.

[0006] Processing various cell specifications is another challenge that needs to be addressed. In the solar cell market, substrates of various specifications, such as M6 (166mm), M10 (182mm), and M12 (210mm), are used in combination, and there is a problem requiring the replacement of separate chucks and edge rings or the conversion of equipment for each specification. This hinders the flexibility of the production line and is a factor that increases equipment investment costs.

[0007] Wafer breakage and alignment errors are also frequent problems that occur when transitioning from atmospheric pressure to a vacuum. Thin wafers are sensitive to pressure changes, so deformation may occur during the rapid vacuuming process, and the resulting alignment errors can directly affect the quality of subsequent processes.

[0008] Conventional etching equipment utilizes alignment results solely for robot position correction. When a position or rotation error of the substrate is detected, the trajectory of the transfer robot is modified to place the substrate in the correct position. However, this method makes additional correction impossible after the substrate is loaded into the etching chamber and fails to fundamentally solve the problem of reduced process uniformity at the edges of the substrate.

[0009] In large-area substrates, differences in process conditions between the center and the edges become more pronounced. Plasma density, gas flow distribution, and temperature distribution can vary depending on the substrate location, leading to non-uniformity in etching rates or etching profiles. This non-uniformity can be exacerbated, particularly when the substrate is positioned outside the design location.

[0010] Another problem with existing alignment systems is the trade-off between measurement speed and precision. In rotary alignment, precise measurements are possible only when the substrate is rotated slowly, but this increases overall process time and reduces productivity. Conversely, increasing measurement speed leads to a decrease in precision, which can affect the quality of subsequent processes.

[0011] The automation of wafer size recognition is also a significant challenge. Conventional technology relies on manual setting of wafer specifications or recognition using contact sensors. Manual setting methods are susceptible to errors caused by operator mistakes, while contact sensors pose a risk of damaging thin wafers.

[0012] Recently, the diversification of substrate surface conditions has also posed new challenges to alignment systems. On substrates treated with anti-reflective coatings or textures, optical properties change, making accurate measurements difficult with conventional optical sensors. Furthermore, contamination or particle adhesion on the substrate surface can also affect measurement precision.

[0013] Thermal expansion of the substrate due to temperature changes is also an important factor to consider. If the substrate temperature rises due to plasma heating during the etching process, dimensional changes caused by thermal expansion may occur. This effect can be more pronounced in large-area substrates, and position corrections set at the beginning of the process may not be sufficient.

[0014] Establishing a continuous processing system to improve productivity is also an important task. In batch processing, the total throughput is limited because each wafer must undergo the processes of loading, processing, and unloading individually. If a system capable of continuous processing can be built and operated in a pipeline manner, productivity can be significantly improved.

[0015] Therefore, there is an urgent need to develop new alignment and etching systems suitable for large-area, thin, square solar cell substrates. Such systems must be able to ensure high precision while preventing substrate damage, reduce process time, and improve etching uniformity. Furthermore, an integrated system is required that can flexibly accommodate various wafer sizes and maximize productivity through fully automated continuous processing. The problem to be solved

[0016] The present invention aims to solve the problems of the prior art described above by providing a system for transporting square solar cell wafers under vacuum conditions, including integrated dry etching, which can safely and precisely process large-area thin square solar cell wafers.

[0017] Another objective of the present invention is to prevent mechanical damage to large-area wafers by performing precise alignment using a linear moving sensor without rotating the wafer.

[0018] Another objective of the present invention is to automatically identify wafers of various specifications, such as M6, M10, and M12, and synchronize the entire system through a non-contact automatic wafer specification recognition function using a laser sensor.

[0019] Another objective of the present invention is to reduce equipment investment costs and secure production flexibility by enabling the processing of various wafer sizes with a single piece of equipment through a common chuck and a replaceable edge ring structure.

[0020] Another objective of the present invention is to improve process uniformity by reflecting alignment results in etching process conditions.

[0021] Another objective of the present invention is to ensure process stability and quality through automatic retry and rework branching functions when the allowable range is exceeded.

[0022] Another objective of the present invention is to enable fully automated mass production through a vacuum continuous processing structure from the loader elevator to the etching chamber. means of solving the problem

[0023] To achieve the above objectives, according to one aspect of the present invention, a transport system for vacuum manufacturing equipment including dry etching of a square solar cell wafer may be provided, comprising: a loader elevator for placing a cassette in which a plurality of wafers are stacked in multiple layers; a load lock chamber provided in the loader elevator for switching between atmospheric pressure and vacuum; a transfer module including a rotary robot arm for transporting a wafer from the load lock chamber; an alignment module for receiving and aligning a wafer from the transfer module; and an etching chamber for receiving the aligned wafer from the alignment module and performing dry etching.

[0024] In one embodiment, the loader elevator may include a laser sensor that automatically recognizes the wafer size, a lower lifting structure that controls the wafer loading position according to the output of the laser sensor, and a plurality of cassettes that stack and arrange a plurality of wafers.

[0025] In one embodiment, the alignment module may include a plurality of roller guides arranged diagonally on a square wafer and contacting the corners, a driving unit that moves the roller guides on a linear guide, and a control unit that selectively operates the roller guides according to the size of the wafer.

[0026] In one embodiment, the etching chamber may include a square common chuck and a plurality of replaceable edge rings that are detachably coupled to the common chuck and correspond to different wafer specifications.

[0027] In one embodiment, the etching chamber may additionally include a driving hole for a hole lifter formed in the common chuck to secure the edge ring.

[0028] According to one aspect of the present invention, a system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells may be provided, comprising: a loader elevator for placing a cassette in which a plurality of wafers are stacked in multiple layers; a load lock chamber provided in the loader elevator for switching between atmospheric pressure and vacuum; a transfer module including a rotary robot arm for transporting wafers from the load lock chamber; an alignment module for receiving wafers from the transfer module and aligning them; and an etching chamber for receiving aligned wafers from the alignment module and performing dry etching.

[0029] For example, a system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells may be provided, wherein the loader elevator comprises: a laser sensor that automatically recognizes the wafer size; a lower lifting structure that controls the wafer loading position according to the output of the laser sensor; and a plurality of cassettes that each accommodate wafers of different specifications.

[0030] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for solar cells may be provided, wherein the alignment module comprises a plurality of roller guides that align based on the diagonal corners of a square wafer; a driving unit that moves the roller guides on a linear guide; and a control unit that selectively operates the roller guides according to the size of the wafer.

[0031] For example, a system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells may be provided, wherein the etching chamber comprises a square common chuck; and a plurality of interchangeable edge rings detachably coupled to the common chuck and corresponding to different wafer specifications.

[0032] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the etching chamber further includes a driving hole for a hole lifter formed in the common chuck to fix an edge ring.

[0033] For example, a system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells may be provided, wherein the edge ring has a curved surface formed at the corner; a cross-sectional structure in which the upper part forms an inclined surface and the lower part forms a vertical wall; and a shape having the same outer diameter while corresponding to different wafer specifications.

[0034] For example, a system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells may be provided, wherein the laser sensor includes a non-contact sensor that recognizes the wafer size by matching it with a plurality of holes of the cassette; a recognition unit that identifies wafers of different sizes; and a control unit that automatically synchronizes the edge ring of the etching chamber and the alignment module according to the recognition result.

[0035] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the alignment module further comprises: an error calculation unit that detects normal loading of the wafer using a linear moving sensor without rotating the wafer; a retry control unit that retryes the alignment process if the calculated error value exceeds an allowable range; and a signal transmission unit that transmits an alignment completion signal to the transfer module if the error value is within an allowable range.

[0036] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the alignment module further comprises: an error calculation unit that calculates positional error and rotational error by moving a linear moving sensor in a straight line without rotating the wafer; a retry control unit that retryes the alignment process if the calculated error value exceeds an allowable range; and a signal transmission unit that transmits an alignment completion signal to the transfer module if the error value is within an allowable range.

[0037] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the linear moving sensor comprises: an optical sensor unit including at least one of a line scan camera or a laser displacement sensor; a sensor head mounted on a linear guide rail movable in the X-axis direction; and a driving motor for moving the sensor head.

[0038] For example, a system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells may be provided, wherein the above allowable range is such that the rotation error is 0.05 degrees or more and 0.15 degrees or less, and the position error is 3 μm or more and 7 μm or less.

[0039] For example, a system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells may be provided, wherein the transfer module comprises a robot arm with a two-axis structure including a central axis and an articulation axis; an end effector including a vacuum suction pad; and a motion control unit that dynamically adjusts an acceleration / deceleration profile according to the size and thickness of the wafer.

[0040] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for solar cells may be provided, wherein the motion control unit includes an acceleration limiting unit that limits the acceleration of the robot arm when the wafer thickness is less than or equal to a reference value; and a jerk control unit that adjusts the jerk value according to substrate characteristics using an S-curve profile.

[0041] For example, the above system may be provided for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells, which is a process for large-area solar cell wafers.

[0042] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the etching chamber further comprises: a process control unit that adjusts process conditions in a feed-forward manner according to an error value received from the alignment module; a gas supply system that supplies etching gas; and an RF power source that supplies RF power.

[0043] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the linear moving sensor further comprises: an ultrasonic sensor that operates in a predetermined frequency range to detect the edge of the wafer; and a signal analysis unit that analyzes the acoustic reflection pattern of the wafer to detect the edge position.

[0044] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for solar cells may be provided, wherein the alignment module further comprises: a piezoelectric vibration sensor for measuring the natural vibration frequency of the wafer; a defect detection unit for detecting thickness non-uniformity or internal defects of the wafer by analyzing the measured natural vibration frequency; and a signal processing unit for converting the vibration signal into the frequency domain.

[0045] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the alignment module further comprises an infrared thermal imaging sensor; a temperature measuring unit for measuring the temperature distribution of the wafer surface; and a thermal expansion correction unit for predicting and correcting positional changes due to thermal expansion during etching based on the measured temperature distribution.

[0046] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the alignment module further comprises a data fusion unit that fuses measurement data of the ultrasonic sensor and the piezoelectric vibration sensor to calculate comprehensive state information of the wafer; and a state estimation unit that performs optimal estimation considering the measurement uncertainty of each sensor.

[0047] For example, a system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells may be provided, wherein the etching gas further includes CF3I gas along with CF4 and O2.

[0048] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the process control unit includes a zone-specific control unit that differentially adjusts the flow rate of the CF3I gas by zone according to an alignment error value; and a non-uniformity compensation unit that increases or decreases the flow rate of the CF3I in a specific area within a predetermined range relative to the reference value to compensate for non-uniformity when the rotation error is greater than or equal to a reference value.

[0049] For example, a system for transporting a square solar cell wafer under vacuum conditions including dry etching for a solar cell may be provided, wherein the alignment module further comprises: a multi-sensor fusion unit that calculates a comprehensive correction value based on multi-sensor fusion by integrating measurement data from the ultrasonic sensor, the piezoelectric vibration sensor, and the infrared thermal imaging sensor; and a correction signal transmission unit that transmits a correction signal composed of three components, position correction, shape correction, and thermal correction, to the etching chamber according to the comprehensive correction value. Effects of the invention

[0050] According to one embodiment of the present invention, safe processing of large-area thin wafers is possible. By using a linear scan alignment method that does not rotate the wafer, high measurement precision can be secured while minimizing mechanical stress.

[0051] According to one embodiment of the present invention, through non-contact automatic wafer specification recognition using a laser sensor, specifications can be automatically determined and the entire system synchronized without manual intervention by an operator. This can prevent human error and improve productivity.

[0052] According to one embodiment of the present invention, various wafer sizes can be processed with a single piece of equipment through a common chuck and a replaceable edge ring structure, thereby eliminating the need to purchase separate equipment for each size.

[0053] According to one embodiment of the present invention, uniform etching can be achieved across the entire wafer surface by reflecting alignment results in the etching conditions in a feedforward manner.

[0054] According to one embodiment of the present invention, the defect rate can be minimized and consistent quality maintained through an automatic retry and rework branching function when the allowable range is exceeded.

[0055] According to one embodiment of the present invention, throughput can be maximized through a vacuum continuous processing structure and pipeline operation from the loader to the etching chamber.

[0056] According to one embodiment of the present invention, it is possible to contribute to strengthening the competitiveness of the solar cell industry by increasing the throughput of a solar cell production line, lowering the defect rate, and reducing operating costs. Brief explanation of the drawing

[0057] FIG. 1 is a schematic diagram showing the overall configuration of a transport system for vacuum manufacturing equipment including dry etching of a square solar cell wafer according to one embodiment of the present invention. FIG. 2 is a drawing showing the detailed structure of a loader elevator according to an embodiment of the present invention, comprising a lower lifting structure, an elevator plate, a lifting guide, a cassette, and stacked wafers, and at the bottom, the specific shape of a cassette according to M6, M10, and M12 specifications is shown. FIG. 3 is a drawing showing an automatic wafer size recognition system using a laser sensor according to an embodiment of the present invention, illustrating the configuration of a laser sensor, a non-contact sensor, a recognition unit, a control unit, and a cassette hole, and a scanning principle through a three-hole matching method. FIG. 4 is a drawing showing the structure of an alignment module according to an embodiment of the present invention, including a roller guide, a linear guide, a driving unit, and a control unit, and illustrates a diagonal alignment method of a square wafer. FIG. 5 is a drawing showing the detailed structure of a replaceable edge ring according to one embodiment of the present invention. FIG. 6 is a drawing showing the structure of a square common chuck installed in an etching chamber according to one embodiment of the present invention. FIG. 7 is an overall process flow diagram of a transport system according to an embodiment of the present invention, showing a control algorithm including processing steps at each stage from start to end and retry or feedforward correction logic based on the determination of an allowable range. Specific details for implementing the invention

[0058] Hereinafter, preferred embodiments of the present invention will be described with reference to the attached drawings. However, embodiments of the present invention may be modified in various other forms, and the scope of the present invention is not limited to the embodiments described below.

[0059] In one aspect of the present invention, a transport system for vacuum manufacturing equipment including dry etching of square solar cell wafers can overcome the limitations of conventional circular semiconductor wafer reference equipment and enable safe processing of large-area thin wafers. As illustrated in FIG. 1, the system can have a loader elevator (100), a load lock chamber (110), a transfer module (200), an alignment module (300), and an etching chamber (400) connected in a vacuum continuous processing structure.

[0060] FIG. 1 is an overall configuration diagram of a transport system for vacuum manufacturing equipment including dry etching of a square solar cell wafer according to one embodiment of the present invention, and is a schematic diagram of a system in which a loader elevator (100), a load lock chamber (110), a transfer module (200), an alignment module (300), and an etching chamber (400) are connected in a vacuum continuous processing structure.

[0061] FIG. 2 is a drawing showing the detailed structure of a loader elevator according to one embodiment of the present invention, and includes a lower lifting structure (120), an elevator plate (121), a lifting guide (122), a cassette (123), stacked wafers (124), and a chamber door (125), and at the bottom, the specific shape of a cassette according to M6, M10, and M12 specifications is shown.

[0062] The loader elevator (100) may have a structure for stacking multiple wafers in multiple layers. For example, 20 to 30 wafers may be stacked vertically, and each wafer may be separated at a constant pitch interval to enable individual pickup. As shown in FIG. 2, the lower lifting structure (120) may include an elevator plate (121) and a lifting guide (122) to sequentially transport the stacked wafers (124) to a pickup position. The lifting method may ensure sub-millimeter level positional accuracy through a precision-machined guide and a central lifting screw. The loader elevator (100) may be provided in a load lock chamber (110). The load lock chamber (110) may be responsible for switching between atmospheric pressure and a vacuum state. A cassette (123) may be loaded through the chamber door (125) of FIG. 2, and after vacuuming the multilayer wafers, they may be transferred to a transfer module. Deformation or breakage of thin wafers can be prevented by gradually controlling the pressure change applied to the wafer during the vacuum transition process.

[0063] The transfer module (200) can transfer a wafer from the load lock chamber (110) including a rotary robot arm. The robot arm may have a two-axis structure including a central axis and an articulation axis, and may be capable of 360-degree rotation and bending angle adjustment from 0 to 180 degrees. A large-area wafer can be stably supported through an end effector including a vacuum suction pad, and the suction force can be adjusted in the range of -80 kPa or more and -95 kPa or less depending on the wafer thickness.

[0064] An end effector including a vacuum suction pad may not facilitate suction due to the vacuum when the chamber is maintained in a vacuum state. In this case, the end effector may use a fixing jig, an adhesive pad, or a fixing pad utilizing electrostatic force instead of a vacuum suction pad. For example, the end effector may use an adhesive pad made of a polymer material, and in this case, the end effector can be adjusted so as not to affect the object to be transported.

[0065] The alignment module (300) can receive a wafer from the transfer module (200) and perform precise alignment. As shown in FIG. 4, the roller guide (310) is positioned diagonally to contact the corners of a square wafer and can move along a linear guide (311) to accommodate various wafer sizes. The roller guide can be selectively operated according to the size of the wafer through the driving unit (312) and the control unit (313).

[0066] The etching chamber (400) can receive an aligned wafer from the alignment module (300) and perform dry etching. As shown in FIGS. 5 and 6, it may include a square common chuck and a plurality of interchangeable edge rings (410). The common chuck may have a square shape with dimensions of 225 mm x 225 mm and may detachably connect edge rings corresponding to different wafer specifications. Through this structure, square wafers of various specifications such as M6, M10, and M12 can be processed with a single common chuck, thereby reducing equipment investment costs and ensuring production flexibility.

[0067] In one aspect of the present invention, the loader elevator includes a laser sensor that automatically recognizes the wafer size, a lower lifting structure, and a plurality of cassettes, and can process square wafers of various sizes in an automated manner.

[0068] FIG. 3 is a drawing showing a wafer size automatic recognition system using a laser sensor according to an embodiment of the present invention, illustrating the configuration of a laser sensor (130), a non-contact sensor (131), a recognition unit (132), a control unit (133), and a cassette hole (134), and a scanning principle through a three-hole matching method.

[0069] A laser sensor that automatically recognizes wafer size can determine wafer specifications in a non-contact manner, as illustrated in FIG. 3. The laser sensor (130) can be positioned on the top of the cassette and operate by matching with the holes of each cassette, and can perform scanning through a red laser line. The recognition unit (132) can distinguish M6, M10, and M12 specifications, respectively, where M6 may refer to a wafer of 166mm x 166mm, M10 to a wafer of 182mm x 182mm, and M12 to a wafer of 210mm x 210mm. These specification classifications are standard specifications based on cell size in the solar cell industry and may each have different power output and efficiency characteristics.

[0070] The lower lifting structure can control the wafer loading position according to the output of a laser sensor. As shown in FIG. 2, the lower lifting structure (120) may include an elevator plate (121) and a lifting guide (122), and can maintain a constant pitch spacing through a precision-machined guide. The lifting operation can be performed through a dual load guide and a central lifting screw, and the positional precision can be maintained at ±0.1 mm or less. It can automatically move to the optimal pickup position corresponding to each wafer size according to the sensor recognition result, thereby eliminating the manual setting process.

[0071] Multiple cassettes can each accommodate wafers of different specifications. As shown at the bottom of FIG. 2, M6, M10, and M12 cassettes (123) can each be designed with a dedicated structure, and each cassette can have a support structure optimized for the corresponding wafer specification. The cassettes can stack 20 to 30 wafers, and the spacing between wafers is maintained at 3 mm to 7 mm to facilitate individual pickup. When replacing cassettes, a laser sensor automatically recognizes the new specification, and the entire system can be automatically adjusted to the corresponding specification, thereby ensuring flexibility of the production line.

[0072] FIG. 4 is a diagram showing the structure of an alignment module according to an embodiment of the present invention, including a roller guide (310), a linear guide (311), a driving unit (312), a control unit (313), an error calculation unit (314), a retry control unit (315), and a signal transmission unit (316), and illustrates a diagonal alignment method of a square wafer.

[0073] In one embodiment of the present invention, the alignment module can overcome the limitations of conventional rotary alignment for circular wafers through a diagonal alignment method optimized for the characteristics of a square wafer.

[0074] A plurality of roller guides arranged diagonally on a square wafer and contacting the corners can provide a key alignment mechanism as illustrated in FIG. 4. The roller guides (310) can be arranged at a 45-degree angle relative to the diagonal of the wafer, and two or more rollers can simultaneously contact the diagonal corners of the wafer to guide it to a reference position. The diameter of each roller can be 10 mm or more and 20 mm or less, and the surface can be treated with a Teflon coating or ceramic material to prevent wafer damage. The rotational speed of the rollers is controlled to be 10 rpm or more and 50 rpm or less to smoothly move the wafer to an alignment position.

[0075] A structure for moving a roller guide on a linear guide can provide versatility for various wafer sizes. The linear guide (311) can use a precision ball bearing method, and the movement precision can be maintained at ±1 μm or less. The movement range can be set to 50 mm or more to cover all specifications from M6 to M12, and an optimal position corresponding to each wafer specification can be pre-set. The linear guide can be configured as a fixed or sliding type, and can simultaneously support cells of various specifications through a multi-stage structure.

[0076] The drive unit can be responsible for precise position control of the roller guide. The drive unit (312) may include a servo motor and a ball screw, and the position resolution may be 0.1 μm or more and 1 μm or less. The drive speed may be adjustable in the range of 100 mm / s or more and 300 mm / s or less, and it may operate at a low speed during the alignment process to ensure precision, and operate at a high speed during the movement process to improve productivity.

[0077] The control unit can selectively operate the roller guides according to the size of the wafer. The control unit (313) can receive wafer specification information from the laser sensor of the loader elevator and activate only the corresponding roller guides. For M6 wafers, the inner roller guide can be selectively used, and for M10 and M12 wafers, the outer roller guide can be selectively used, and unused rollers can be moved to a waiting position to be removed from the wafer path. Through this selective operation, alignment performance optimized for each wafer specification can be achieved, and the alignment time can be reduced to 5 seconds or less.

[0078] FIG. 6 is a diagram showing the structure of a square common chuck of an etching chamber according to one embodiment of the present invention, illustrating the coupling relationship between a driving hole for a hole lifter formed in the common chuck and an edge ring.

[0079] In one embodiment of the present invention, the etching chamber can provide versatility to process various wafer sizes in a single device through an innovative combined structure of a common chuck and a replaceable edge ring.

[0080] As illustrated in FIG. 6, the square common chuck can provide a base structure that is commonly used for all wafer sizes. The common chuck may have a square shape of 225 mm x 225 mm, which may be designed with a margin larger than the maximum size M12 wafer (210 mm x 210 mm). The chuck may be made of aluminum alloy or stainless steel, and the surface may be anodic or electropolished to ensure corrosion resistance and flatness. The chuck may be designed with a thickness of 20 mm or more and 30 mm or less to provide sufficient rigidity, and may include multiple holes for vacuum suction and through holes for driving lift pins.

[0081] A common chuck including vacuum suction holes may not facilitate suction under vacuum when the chamber is maintained in a vacuum state. In this case, the common chuck may use a fixing jig, an adhesive pad, or a fixing pad utilizing electrostatic force instead of vacuum suction holes. For example, the common chuck may use an adhesive pad made of a polymer material, and in this case, the common chuck can be adjusted so as not to affect the object to be transported.

[0082] Multiple interchangeable edge rings can correspond to different wafer specifications while maintaining compatibility with a common chuck. As illustrated in FIG. 5, the edge ring (410) may have an internal opening corresponding to the wafer sizes of M6, M10, and M12, respectively. The edge ring for M6 may have a square opening of 168 mm x 168 mm, allowing for a margin of 1 mm compared to the actual M6 wafer (166 mm x 166 mm), and the edge ring for M10 may have an opening of 184 mm x 184 mm to correspond to the M10 wafer (182 mm x 182 mm). The edge ring may be made of quartz, ceramic, or fluorine resin to ensure chemical resistance to etching gas.

[0083] The drive hole for the drive hole lifter is formed in a common chuck, allowing different edge rings to be detachably connected. The drive hole can be precisely fixed to the edge ring. The depth of the drive hole is designed to be between 5mm and 10mm to ensure stable fixation of the edge ring, and a 2.5mm wide round slope is formed around the outer edge of the drive hole to facilitate smooth wafer placement. When replacing the edge ring, it can be easily replaced by releasing the vacuum and lifting the edge ring from the top, reducing the replacement time to less than 5 minutes. Through this structure, rapid response is possible when wafer specifications change on the production line without replacing the entire equipment, thereby significantly reducing equipment investment costs.

[0084] In one embodiment of the present invention, the precise shape design of the edge ring may be a key element for stable wafer placement and uniform etching.

[0085] The curved surface formed at the edge can provide smooth contact between the wafer and the edge ring, as shown in FIG. 5. The curved surface may have a radius of curvature of 5T mm, where T represents the basic thickness unit of the edge ring. For example, if T=1 mm, the radius of curvature can be 5 mm, which can prevent stress concentration at the wafer edge and enable even load distribution. The machining precision of the curved surface can be maintained at Ra 0.1 μm or less to minimize wafer surface damage.

[0086] A cross-sectional structure having an inclined surface on the top and a vertical surface on the bottom can ensure stable seating due to the self-weight of the wafer. The upper inclined surface corresponds to the 2T portion of the total thickness of 3T mm and has an inclination angle of 52 degrees, allowing the wafer to slide and naturally settle in the correct position. This inclination angle can be an optimized value considering the wafer's sliding angle and angle of repose, preventing wafer detachment due to excessive inclination or incomplete seating due to insufficient inclination. The lower vertical surface corresponds to the remaining 1T portion and can serve as the final support surface for the wafer.

[0087] A structure in which multiple edge rings have the same outer diameter can be a key design element for ensuring compatibility with a common chuck. The outer diameter of all edge rings can be standardized to 211.9 mm, and tolerances are managed to ±0.05 mm or less, allowing for precise coupling with the drive hole for the hole lifter of the common chuck. Through this standardized outer diameter design, edge rings with different inner diameters, such as the M6 ​​edge ring (inner diameter 168 mm) and the M10 edge ring (inner diameter 184 mm), can be interchangeably used on the same common chuck. This standardized outer diameter structure can be achieved through the standardization of edge ring thickness and outer flange design, which simultaneously ensures reduced manufacturing costs and quality control efficiency.

[0088] In one embodiment of the present invention, a precise recognition system of a laser sensor can enable fully automated production through automatic determination of wafer specifications and synchronization of the entire system.

[0089] A non-contact sensor that recognizes the wafer size by matching it with multiple holes formed in a cassette can provide an innovative recognition method as illustrated in FIG. 3. The non-contact sensor (131) can generate three laser lines that exactly match the cassette holes (134), and can determine the wafer specifications through the position and spacing of each hole. In the case of the M6 ​​cassette, the hole spacing is set to 166 mm, for the M10 cassette to 182 mm, and for the M12 cassette to 210 mm, so the specifications can be identified by analyzing the pattern of the laser passing through the holes. The wavelength of the sensor can be a red laser in the range of 400 nm to 780 nm, and safety can be ensured with an output of 5 mW or less.

[0090] The recognition unit, which identifies wafers of different sizes, can perform real-time specification identification through a high-precision algorithm. The recognition unit (132) can determine the wafer specifications within 0.1 seconds by analyzing the blocking pattern of the laser signal, and the recognition accuracy can be 99.9% or higher. During the identification process, tilting of the cassette or positional error can be automatically corrected, and positional deviations within ±2mm can be automatically compensated.

[0091] A control unit that automatically synchronizes the edge ring and alignment module of the etching chamber according to the recognition result can be responsible for integrated control of the entire system. The control unit (133) can simultaneously transmit wafer specification information to all modules of the system, and each module can be automatically switched to a setting that matches the corresponding specification. For example, when an M10 wafer is detected, the roller guide of the alignment module moves to a position dedicated to M10, and the etching chamber can notify the operator that replacement with an edge ring for M10 is required.

[0092] In one embodiment of the present invention, the optimized stacking structure of the cassette and the sensor-linked design can simultaneously achieve efficient wafer supply and accurate specification recognition.

[0093] A stacking structure accommodating 20 to 30 wafers can be designed to balance productivity and stability, as illustrated in FIG. 2. The stacking structure can maintain a spacing of 3 mm to 7 mm between wafers, which ensures safe pickup of individual wafers and sufficient air circulation. Based on 25 wafers, the total cassette height can be 120 mm to 170 mm, ensuring structural rigidity of the cassette itself while maintaining accessibility for a robot arm. The stacked wafers (124) can be individually separated by supports made of anti-static material to prevent damage caused by mutual contact.

[0094] Multiple holes formed for size recognition of the laser sensor can provide perfect matching with the sensor through precise machining. The diameter of the holes can be machined to be between 5mm and 8mm, ensuring the complete passage of the laser beam while maintaining the structural strength of the cassette. The position of the holes can be manufactured using CNC machining to precisely align with the centerline of each wafer specification, and the positional tolerance can be managed to ±0.1mm or less.

[0095] In one embodiment of the present invention, the three-hole matching method of the non-contact sensor and the wafer specification-specific discrimination algorithm can implement an automatic recognition system with high accuracy and reliability.

[0096] The scanning method for recognizing wafer size by matching three holes in a cassette can utilize a unique pattern recognition technique as illustrated in Fig. 3. The three-hole matching method can use the principle of triangulating wafer size by simultaneously scanning holes located at the center and both ends of the cassette. During the scanning process, the laser passes through each hole sequentially, and the wafer specifications can be determined by analyzing the distance between holes and the pattern. The scanning speed is set to between 10 mm / s and 50 mm / s, allowing for both accurate signal acquisition and fast processing time.

[0097] The discrimination algorithm for distinguishing wafers of M6, M10, and M12 specifications may include advanced pattern analysis techniques. The algorithm is based on differences in hole spacing, but can clearly distinguish between 16mm and 28mm differences for M6 (166mm spacing), M10 (182mm spacing), and M12 (210mm spacing). Multiple scans may be performed to remove noise during the discrimination process, and the result may be finalized when the same result is obtained three consecutive times. The processing time of the algorithm is less than 1 second, and the misidentification rate can be maintained at 0.01% or less.

[0098] FIG. 5 is a drawing showing the detailed structure of a replaceable ceramic edge ring according to one embodiment of the present invention.

[0099] In one embodiment of the present invention, the precision cross-sectional structure of the edge ring may be a key design element that ensures perfect wafer placement and uniformity of the etching process.

[0100] A curved surface having a radius of curvature of 5T mm at the edge can effectively disperse stress concentration that may occur at the wafer edge. Based on T=1 mm, a radius of curvature of 5 mm is sufficiently larger than the wafer edge radius (typically between 0.1 mm and 0.3 mm), allowing for stable support through surface contact rather than point contact. The surface roughness of the curved surface is controlled to Ra 0.05 μm or less to prevent micro-scratches on the wafer surface, and the surface machining can be performed by precision cutting using a diamond tool.

[0101] Of the upper thickness of 3T mm, 2T features an inclined surface with a 52-degree inclination angle that can induce natural wafer seating. The 52-degree inclination angle is optimized by considering the static friction coefficient of the silicon wafer (0.1 or greater and 0.3 or less), allowing the wafer to stably reach its final position while sliding. The surface of the inclined surface can secure appropriate friction through fine texture processing, thereby preventing excessive sliding or stopping.

[0102] The lower 1T structure, which forms a vertical plane, can be responsible for the final support and position fixation of the wafer. The vertical plane is machined at a precise angle of 90 degrees to allow the wafer to be seated perfectly parallel to the horizontal plane. The height of the vertical plane, 1T mm, is sufficiently higher than the wafer thickness (typically 180 μm or more and 200 μm or less) to completely enclose the wafer and prevent the wafer from detaching during the etching process.

[0103] In one embodiment of the present invention, the driving hole for the lifter is formed in a common chuck and can perform the function of driving the lifters together. The driving hole can be machined in a circular shape and can be formed in four parts to precisely fix all edge rings. The depth of the driving hole is designed to be 5 mm or more and 10 mm or less to ensure stable fixation of the edge ring, and a round slope with a width of 2.5 mm is formed around the outer edge of the driving hole to help the wafer to be smoothly seated.

[0104] The inner wall surface of the driving hole of the driving hole lifter is machined vertically to ensure tight contact with the edge ring, and the surface roughness is managed to Ra 1.6μm or less to improve vacuum sealing performance. An O-ring groove may be additionally formed on the bottom surface of the driving hole to further ensure airtightness in a vacuum environment.

[0105] When replacing the edge ring, it can be easily replaced by releasing the vacuum and lifting the edge ring from the top, and the replacement time can be reduced to less than 5 minutes. Through this structure, rapid response is possible when wafer specifications change on the production line without replacing the entire equipment, thereby significantly reducing equipment investment costs.

[0106] FIG. 7 is an overall process flow diagram of a transport system according to an embodiment of the present invention, showing a control algorithm including processing steps at each stage from start to end and retry or feedforward correction logic based on the determination of an allowable range.

[0107] In one embodiment of the present invention, the rotation-free linear scanning method of the alignment module and the conditional sequence control can simultaneously achieve safe alignment of large-area thin wafers and high process reliability.

[0108] The error calculation unit, which calculates position error and rotation error using a linear moving sensor without rotating the wafer, can overcome the limitations of conventional rotary alignment as shown in FIG. 4. The error calculation unit (314) can slide detect whether the wafer is loaded normally using a linear moving sensor.

[0109] The error calculation unit, which calculates position error and rotation error by moving a linear moving sensor in a straight line without rotating the wafer, can overcome the limitations of conventional rotational alignment as illustrated in FIG. 4. The error calculation unit (314) can calculate the position deviation relative to the reference coordinate system by sequentially scanning the four corners of the wafer while the sensor head moves linearly along the X-axis. The scan speed can be set to between 100 mm / s and 300 mm / s, and the total scan time can be reduced to 5 seconds or less. Position error βx, βy and rotation error The calculation can be performed based on the least squares method, and the measurement uncertainty can be maintained at ±0.5 μm or less.

[0110] A retry control unit that retryes the alignment process when the calculated error value exceeds the allowable range can provide an automated quality control system. The retry control unit (315) can automatically perform one to three re-alignments on the same wafer when the allowable range is exceeded, and the retry interval can be set to 5 seconds to 15 seconds to secure system stabilization time. During the retry, the sensor scan speed can be reduced by 50% and measurement precision can be improved, and the scan area can be extended to ±5mm from the edge of the substrate to perform more accurate measurements.

[0111] A signal transmission unit that transmits an alignment completion signal to a transfer module when the error value is within the allowable range can ensure smooth interoperability between systems. When alignment is successful, the signal transmission unit (316) can transmit an "ALIGN OK" signal to the transfer module to permit the proceeding to the next step, and when the retry fails, it can activate a rework path that returns the wafer to the loader along with an "ALIGN FAIL" signal. The signal transmission delay time can be maintained at 100ms or less, enabling real-time control.

[0112] In one embodiment of the present invention, the multi-sensor configuration and precision movement mechanism of the linear moving sensor can provide high accuracy and reliability under various measurement conditions.

[0113] An optical sensor unit comprising at least one of a line scan camera or a laser displacement sensor may enable optimal measurement according to wafer characteristics. The line scan camera may include a linear image sensor of the CCD or CMOS type, and the number of pixels may be 2,048 or more and 8,192 or less. The pixel size may be designed to be 5 μm or more and 15 μm or less, enabling edge detection with sub-pixel precision, and the linearity and angle of the edge may be measured with a precision of 0.01 degrees or less. The scan speed may be adjustable within a range of 10 kHz or more and 100 kHz or less, and multi-sampling and averaging processing for noise removal may be performed even during high-speed scanning. The laser displacement sensor may use a semiconductor laser diode with a wavelength of 400 nm or more and 780 nm or less as a light source, and through triangulation or confocal methods, ultra-precision measurement with a resolution of 0.1 μm or less may be possible at a measurement distance of 10 mm or more and 100 mm or less. The selection of the sensor can be automatically switched according to the wafer surface characteristics, and a line scan camera can be applied first to wafers with high surface reflectivity, while a laser displacement sensor can be applied first to rough or transparent wafers. In some embodiments, the two sensors can be used simultaneously to perform complementary measurements, and measurement reliability can be improved by more than 95% compared to a single sensor.

[0114] A sensor head mounted on a linear guide rail movable along the X-axis can ensure nanometer-level precise position control and stable, vibration-free scanning. The linear guide rail may utilize precision ball bearings or air bearings, and movement precision can be maintained at ±0.5μm or less. The straightness of the guide rail is managed to be 10μm / m or less, minimizing measurement errors during the scanning process. The sensor head may feature an integrated structure with the camera and laser sensor precisely positioned at 5mm intervals, and it is manufactured from aluminum alloy or stainless steel to minimize thermal deformation and vibration. The center of gravity of the sensor head is designed to align with the central axis of the guide rail to prevent moment generation during movement, and the travel range is set to be at least 50mm larger than the wafer size, allowing it to accommodate wafers of all specifications. The sensor head is equipped with a vibration detection sensor that temporarily suspends measurement if external vibration exceeds an allowable threshold (0.1G or higher) and resumes it after the vibration stabilizes. Through the temperature compensation function, positional accuracy can be maintained even with a temperature change of ±5℃, and guide wear due to long-term use is monitored in real time to enable preventive maintenance.

[0115] The drive motor that moves the sensor head can provide ultra-precision control and adaptive speed regulation considering the characteristics of thin wafers. The drive motor may utilize a direct-drive linear servo motor or a ball screw-coupled rotary servo motor, and can achieve an ultra-precision level of position resolution of 0.05μm or less. The motor controller uses a 32-bit high-speed processor to perform real-time position feedback and speed control, and the gain values ​​of the PID control algorithm can be automatically adjusted according to wafer characteristics. The motor torque can be selected to be between 0.3Nm and 3Nm, taking into account the weight of the sensor head (0.5kg or more and 2kg or less) and acceleration / deceleration requirements, and the maximum travel speed can be varied up to 10mm / s for precision measurement and 500mm / s for high-speed movement. A 23-bit absolute encoder is installed in the drive unit to provide real-time position feedback with a resolution of 0.1μm, and position accuracy can be maintained at ±0.3μm or less through temperature compensation and backlash correction. In the special mode for thin wafer measurement, the acceleration is 0.5 m / s to minimize vibration. 2 It is limited to the following level, and if minute vibrations are detected during measurement, it can be stopped immediately and stabilized. Electromagnetic noise generated during motor operation can be controlled through shielded cables and filters so as not to affect the sensor signal, and the driving sound can be maintained at 40dB or less to ensure a precise measurement environment.

[0116] In one embodiment of the present invention, the precise setting of the allowable range can provide an optimized standard that balances the quality requirements of the solar cell process and the precision of the measurement system.

[0117] An allowable rotational error range of 0.05 degrees to 0.15 degrees can reflect the precision requirements for solar cell pattern formation. The lower limit of 0.05 degrees represents the minimum detectable range considering sensor measurement uncertainty and mechanical vibration; errors smaller than this are difficult to distinguish from measurement noise and may cause unnecessary correction operations. The upper limit of 0.15 degrees represents the maximum allowable rotational error in the post-processing stage; exceeding this limit may cause positional deviations during pattern alignment to exceed acceptable limits. For example, on a 210mm wafer, a rotational error of 0.15 degrees can result in a positional deviation of approximately 0.55mm at the edge.

[0118] An allowable positional error range of 3μm to 7μm can be a practical criterion considering the mask alignment precision and mechanical backlash of the etching process. The lower limit of 3μm is the minimum meaningful error level considering mechanical backlash and thermal expansion; for errors smaller than this, the correction effect is negligible and may instead cause unnecessary system load. The upper limit of 7μm is the maximum allowable error that satisfies the mask alignment precision requirements of the etching process and can provide a sufficient margin when considering the minimum linewidth (50μm or more) of typical solar cell patterns.

[0119] The tolerance range can be dynamically adjusted based on wafer size, and stricter standards may be applied to larger wafers. Additionally, a setting interface may be provided to allow users to adjust the tolerance range according to process conditions and quality requirements.

[0120] In one embodiment of the present invention, the 2-axis robot structure and adaptive motion control of the transfer module can ensure safe and precise transfer of large-area thin wafers.

[0121] A robot arm with a two-axis structure including a central axis and an articulation axis can provide an optimal transfer trajectory in a confined space. The central axis can rotate 360 ​​degrees, and the backlash of the bearings and drive motors can be managed to be 0.01 degrees or less. The articulation axis can adjust the bending angle from 0 to 180 degrees, and the robot arm can be converted from a fully extended straight form to a fully folded form. Servo motors can be used for driving each axis, and positional accuracy can be maintained at ±0.1 degrees or less.

[0122] An end effector including vacuum suction pads can have a design optimized for stable support of large-area wafers. The number of suction pads can be set from 4 to 16, and the number of activated pads can be adjusted according to the wafer size. The vacuum level can be controlled within a range of -80kPa to -95kPa, and the suction force can be automatically adjusted according to the wafer thickness and material. The suction pads can be made of NBR or silicon to prevent damage to the wafer surface.

[0123] The motion control unit, which dynamically adjusts acceleration and deceleration profiles according to wafer size and thickness, can perform transfer optimized for each wafer characteristic. The motion control unit can receive wafer information in real time and automatically set transfer conditions, with a standard acceleration of 2.0 m / s for a 156 mm x 156 mm wafer. 2 for a 210mm x 210mm wafer, 1.5m / s 2 The transfer speed can be adjusted in the range of 200 mm / s to 300 mm / s depending on the wafer size, and different profiles can be applied for each transfer step.

[0124] In one embodiment of the present invention, the thin wafer-specific control algorithm of the motion control unit can optimize productivity while preventing wafer breakage.

[0125] The acceleration limiting unit, which limits the acceleration of the robot arm to 70% or less of the standard acceleration when the wafer thickness is 200 μm or less, can apply an innovative protection algorithm specialized for the mechanical vulnerability of thin wafers. The acceleration limiting unit can calculate dynamic limit values ​​by considering composite factors such as material, surface treatment status, temperature, and humidity in real time, in addition to wafer thickness information. Detailed limit values ​​by thickness can be applied stepwise as follows: 70% of the standard acceleration at 200 μm, 60% at 180 μm, 45% at 150 μm, 30% at 120 μm, and 20% at 100 μm. When transitioning between each step, a gradual adaptation time of at least 5 seconds is provided to prevent abrupt changes, thereby ensuring the stability of the robot control system. In environmental condition compensation, acceleration limits can be further tightened by 10% to account for increased wafer brittleness when the temperature is 25°C or lower or the humidity is 30% or lower, and the transfer speed can be reduced by 20% under conditions with a high risk of electrostatic discharge. For real-time wafer status monitoring, wafer vibration and deformation can be continuously monitored through a 6-axis accelerometer and strain sensor installed on the robot arm, and wafer breakage can be prevented by immediately performing emergency deceleration if the allowable threshold (vibration 0.05G, strain 10με) is exceeded. Through machine learning-based adaptive control, accumulated transfer data is analyzed to learn and continuously improve optimal limit values ​​for each wafer, and a prediction accuracy of over 95% can be achieved based on more than 1,000 training data iterations.

[0126] The jerk control unit, which adjusts jerk values ​​according to substrate characteristics using an S-curve profile, can minimize impact on the wafer through smooth acceleration and deceleration. By controlling the rate of change of acceleration, the jerk control unit can prevent wafer vibration caused by sudden changes in acceleration. For standard thickness wafers, the jerk is 10 m / s 3 More than 20 m / s 3It can be set to less than or equal to 3 m / s on thin wafers. 3 8 m / s or more 3 It may be further restricted to the following. The S-curve profile may consist of 7 stages: acceleration start, acceleration hold, acceleration end, constant speed, deceleration start, deceleration hold, and deceleration end.

[0127] In one embodiment of the present invention, the large-area solar cell wafer processing capability of the system can provide versatility that encompasses all major standards of the current solar cell industry.

[0128] Large-area solar cell wafers ranging from 156mm to 210mm are the mainstream specifications in the current solar cell industry and can have different power output and efficiency characteristics. The 156mm wafer is a specification derived from the traditional 6-inch silicon wafer and corresponds to an M6 cell; while its output per unit area is relatively low, it is easy to handle and offers high compatibility with existing production lines. The 210mm wafer is a large specification used in the latest high-efficiency solar cells and corresponds to an M12 cell; because its area is approximately 1.8 times larger than that of the 156mm wafer, it offers higher output per unit cell, but the difficulty of mechanical handling may increase.

[0129] The key advantage of this system in large-area wafer processing is its ability to effectively resolve mechanical stress and deformation issues associated with increasing wafer size. Compared to 156mm wafers, 210mm wafers can experience more than three times the sagging due to their self-weight per area, and inertial forces during transport can also increase significantly. The system's rotationless linear alignment method and adaptive transport control are optimized for these characteristics of large-area wafers, ensuring safe processing. Furthermore, the common chuck and interchangeable edge ring structure allow a single piece of equipment to handle all sizes, thereby significantly reducing equipment investment costs.

[0130] In one embodiment of the present invention, the feedforward control system of the etching chamber and the optimized gas supply configuration can achieve uniform etching quality through fine-tuning of process conditions using alignment information.

[0131] The process control unit, which adjusts process conditions in a feedforward manner based on error values ​​received from the alignment module, can compensate in advance for etching non-uniformity caused by wafer position errors through real-time adaptive control. The process control unit covers position errors βx, βy and rotation error Based on the information, differential etching conditions can be applied to each region. For example, If a rotational error of +0.05 degrees is detected, the gas flow rate in the upper region can be increased by +3% relative to the reference value, and the RF bias can be adjusted by +5W. The response time of the feedforward control is maintained at 1 second or less, so that correction can be applied simultaneously with the start of the process.

[0133] In one embodiment of the present invention, the innovative acoustic-based edge detection system of the ultrasonic sensor can implement a differentiated technology that provides stable and precise measurements even under surface conditions that are difficult for optical sensors to overcome.

[0134] The precisely selected frequency range of 40 kHz to 60 kHz is the result of comprehensively optimizing attenuation, resolution, and penetration in air, enabling adaptive measurements utilizing the characteristics of each frequency. At 40 kHz, the air wavelength of 8.5 mm corresponds to 42.5 times the wafer thickness of 0.2 mm, allowing for deep penetration into the wafer; with the beam spread angle set to 15 degrees, a measurement area with a diameter of 15 mm can be secured. At 60 kHz, the wavelength of 5.7 mm provides higher spatial resolution, enabling the detection of even fine geometric features at the edges, and with the beam focused at a spread angle of 10 degrees, precise measurements with a diameter of 10 mm are possible. Frequency selection can be automatically switched in real-time according to wafer characteristics; 40-45 kHz can be applied to thick substrates or porous surfaces, while 55-60 kHz can be applied to thin and smooth substrates. The sensor is based on a piezoelectric ceramic (PZT) transducer, and the transmitter and receiver are integrated to minimize acoustic interference. The transducer is designed with a Q factor of 200 or higher to provide high sensitivity and frequency selectivity, and has a temperature coefficient of -0.04% / ℃, making it robust against environmental changes.

[0135] The signal analysis unit, which detects edge locations by analyzing the acoustic reflection patterns of the wafer, can achieve sub-micron level precision through advanced digital signal processing techniques. The signal analysis unit utilizes a 32-bit floating-point DSP to perform real-time signal processing, and with a sampling frequency set to 500 kHz, it can secure sufficient temporal resolution while satisfying the Nyquist theorem. In acoustic impedance difference analysis, a clear reflection signal can be obtained because the impedance ratio at the wafer-air interface is very large, approximately 105:1, and distance information can be extracted by analyzing the time delay of the echo signal. In signal intensity analysis, the 50% point of the maximum value is defined as the edge, but the average value of five consecutive measurements can be used for noise removal. Through adaptive threshold settings, detection criteria can be dynamically adjusted according to environmental noise levels, and measurement conditions can be automatically optimized if the SNR drops below 20 dB. The distance measurement error can be maintained at 0.05% or less through real-time correction for changes in the speed of sound (343-346 m / s) in the temperature range of 20-30℃, and an accuracy of ±0.1% can be guaranteed even in the humidity range of 30-70%.

[0136] The unique advantage of the ultrasonic sensor is that it provides consistent performance regardless of wafer surface treatment conditions or contamination, enabling stable edge detection even under conditions where measurement is difficult with optical sensors, such as anti-reflective coatings (reflectance <5%), textured surfaces (roughness Ra >1μm), and transparent substrates. Additionally, it can provide a double benefit by simultaneously acquiring wafer thickness and internal structure information, and with a thickness measurement precision of ±3μm or less, it can also be utilized for quality control.

[0137] In one embodiment of the present invention, a natural vibration frequency analysis system utilizing a piezoelectric vibration sensor can provide an innovative approach to non-destructively evaluate the mechanical properties of a wafer and predict process quality in advance.

[0138] Piezoelectric vibration sensors, which measure the natural vibration frequency of a wafer, can generate a unique vibration fingerprint that comprehensively reflects the wafer's size, thickness, material, internal stress, and defect status. The piezoelectric sensors can be fabricated based on lead zirconium titanate (PZT-5H) ceramic or polyvinylidene fluoride (PVDF) film, and their size is optimized to range from 3mm x 3mm to 8mm x 8mm, allowing them to be directly attached to or installed in close proximity to the wafer support. The sensor possesses a piezoelectric constant of d33 = 374 pC / N or greater, enabling it to detect even minute vibrations, with a sensitivity of 2mV / m / s. 2 It is configured to measure a wide range of vibrations, from natural to forced vibrations of the wafer. The frequency response range is extended from 0.1 Hz to 15 kHz, covering from the fundamental mode to higher-order modes of the wafer, and an amplification factor of 40 dB or more at the resonant frequency can be achieved. As for the excitation method, sinusoidal excitation using a piezoelectric actuator (0.1-1000 Hz sweep), impulse excitation using an impact hammer, or non-contact excitation using an air jet can be selected, and the excitation intensity can be limited to 0.5 N or less, a level that does not damage the wafer.

[0139] The defect detection unit, which detects wafer thickness non-uniformity or internal defects by analyzing measured natural vibration frequencies, can identify even minute anomalies through advanced pattern recognition and statistical analysis. The defect detection unit can quantitatively analyze characteristic changes appearing in defective wafers in contrast to the single, sharp natural frequency peaks (half-width <2Hz, Q factor >150) observed in normal wafers. In thickness non-uniformity detection, it can comprehensively evaluate the appearance of multiple frequency peaks (volume peaks within -10dB relative to the main peak observed with ±5μm deviation), peak dispersion (enlargement of half-width by more than 5Hz), and frequency distribution asymmetry (skewness >0.5). In internal defect detection, it can assess a sharp drop in Q factor (below 50) and an increase in damping rate (normal 0.015s -1 -> Defect 0.035s -1 Quantitative indicators such as abnormalities and mode shape distortion can be used. In the case of microcracks, the presence and approximate location of the crack can be estimated through phenomena where specific modes are suppressed or new modes emerge depending on the location, and positional precision of ±8mm or less can be achieved through multi-point measurement using triangulation.

[0140] The signal processing unit, which converts vibration signals into the frequency domain using an FFT algorithm, can perform real-time high-precision analysis through state-of-the-art digital signal processing techniques. The signal processing unit can utilize a multi-core DSP supporting 64-bit floating-point operations and achieve a frequency resolution of 0.02 Hz or less through 16,384-point FFT operations. Spectral leakage can be minimized by using a Hanning window as the window function, and the SNR can be improved by more than 30 dB through overlapping averaging. The automatic peak detection algorithm can select only true resonant peaks using first and second derivatives and effectively remove noise or harmonic components. By applying a pipeline structure for real-time processing, FFT operations can be performed simultaneously with data acquisition, reducing the total analysis time to less than 3 seconds.

[0141] This is a diagram showing the overall process flow diagram and control algorithm of a system for transporting a square solar cell wafer under vacuum conditions, including dry etching for a solar cell according to the present invention. It includes processing steps for each stage from start to end and conditional sequence control logic based on tolerance determination.

[0142] The process starts at "Start" and proceeds to the first step, "Substrate Loading (1->2)". In this step, a wafer is transferred from the loader elevator (100) to the transfer module (200). Simultaneously with automatic specification recognition by a laser sensor, a robot arm picks up the wafer through vacuum suction.

[0143] In the second step, "Alignment Module Transfer (2->3)," the wafer is transferred from the transfer module (200) to the alignment module (300). The robot arm is converted into a folded form to pass through a confined space and place the wafer onto the alignment chuck.

[0144] In the third step, "Perform Linear Scan Alignment," a linear moving sensor is used to verify whether the wafer is loaded in the correct position without rotating the wafer.

[0145] Alternatively, in the third step, "Performing Linear Scan Alignment," a linear moving sensor measures positional and rotational errors by moving in a straight line without rotating the wafer. A diagonal alignment roller guide guides the wafer to a reference position, while a line scan camera or laser displacement sensor performs precise measurements.

[0146] In the fourth step, "Calculation of Error Values," based on the linear scan results, βx, βy, The value is calculated. An algorithm based on the least squares method is applied to quantify the position error and rotation error.

[0147] The fifth step is "tolerance determination," where the measured error value is compared with a preset tolerance range. Rotational error is determined to be within the range of 0.05 degrees or more and 0.15 degrees or less, and positional error is determined to be within the range of 3 μm or more and 7 μm or less. Based on this determination result, the process branches into two paths.

[0148] If the tolerance determination is "FAIL," the path branches to the "Retry" path. During the retry process, one to three repeated measurements are performed on the same substrate. During the retry, the sensor scan speed is reduced and the measurement area is expanded to improve precision. If the retry is successful, the process returns to the main flow; if failure persists, the path branches to the rework path and the substrate is returned to the loader.

[0149] If the tolerance range judgment is "PASS," the process proceeds to the "etching chamber transfer" step. A transfer robot picks up the wafer with a corrected posture and trajectory and transfers it to the etching chamber (400). During this process, the error value measured in the alignment is transmitted together and utilized in subsequent processes.

[0150] In the "feedforward correction" step, process conditions are adjusted in real time based on the received error values. At least one of the gas flow rate, RF power, or chamber pressure is adjusted in a feedforward manner to compensate for positional and rotational errors of the substrate. For example, if a rotational error is detected, the gas flow rate and RF bias in the corresponding area are differentially adjusted.

[0151] In the "dry etching" step, plasma etching is performed under adjusted process conditions. Anti-reflective film removal or surface texturing is performed using CF4 and O2 gases, and process stability is ensured through real-time monitoring.

[0152] In the final stage, the "unloading (4->2->1)" process, the etched wafer is returned from the etching chamber to the loader via the transfer module. The robot arm is converted back into an extended form to safely transport the wafer and load it into a designated location within the cassette.

[0153] When all processes are completed, the system reaches "termination" and returns to its initial state for the next wafer processing. During continuous processing, a pipeline method is applied so that multiple wafers can be processed simultaneously at different stages.

[0154] The core of this process flow diagram is conditional sequence control, which automatically selects a retry or feedforward correction path based on alignment results, thereby simultaneously achieving quality assurance and productivity improvement. Furthermore, integrated control is possible through real-time data sharing between modules, enabling fully automated unmanned operation.

[0155] In one embodiment of the present invention, an innovative three-way fusion system of ultrasonic, piezoelectric vibration, and infrared thermal imaging sensors can synergistically combine the unique characteristics of each sensor to present a new paradigm for wafer state analysis.

[0156] The data fusion unit, which calculates comprehensive wafer state information by fusing measurement data from ultrasonic and piezoelectric vibration sensors, can fundamentally overcome the limitations of a single sensor through the innovative integration of complementary measurements. The data fusion unit can utilize a hybrid state estimation algorithm combining an improved Kalman filter and Bayesian inference, and can integrate edge detection information from the ultrasonic sensor (precision ±1.0μm) and structural integrity information from the piezoelectric sensor (natural frequency resolution 0.02Hz) in real time. During the fusion process, the measurement uncertainty and correlation of each sensor can be dynamically considered, and the weights of the ultrasonic data can be adaptively adjusted within a range of 0.5–0.7 depending on surface conditions, while the weights of the piezoelectric data can be adjusted within a range of 0.3–0.5 depending on wafer thickness. The fusion algorithm can utilize a deep learning model trained on over 5,000 training data points, including neural network-based pattern recognition, and can achieve a prediction accuracy of over 97.8%. Through the mutual verification function, the agreement between measurement results of two sensors can be evaluated in real time, and the fusion result can be determined as valid only if an agreement is maintained within ±2μm in position information and ±3% in structure information.

[0157] The multi-sensor fusion unit, which calculates comprehensive correction values ​​based on multi-sensor fusion including an infrared thermal imaging sensor with a resolution of 320x240 pixels or higher, can achieve unprecedented precision through optimal estimation in 3D sensor space. The multi-sensor fusion unit can adopt a three-stage hierarchical fusion structure; in the first stage, it performs mechanical state vectors through ultrasonic-piezoelectric fusion, in the second stage, thermal-mechanical coupled states through thermal imaging data integration, and in the third stage, dynamic state prediction through temporal change tracking. The thermal imaging sensor enables ultra-precise temperature measurements of 0.05°C or less in a high-resolution mode of 640x480 pixels and can detect minute temperature gradients (0.1°C / cm) on the wafer surface. Thermal expansion coefficient database (Silicon: 2.6x10 -6 / ℃, Germanium: 5.8x10 -6 Optimal correction for each material can be performed in real-time linked with (e.g., / ℃), and internal stress distribution can be estimated through the analysis of the second moment of the temperature distribution. The comprehensive correction value can be expressed as a 12-dimensional state vector, and position correction (X, Y, It can be composed of four categories: shape correction (thickness uniformity, flatness, stress), thermal correction (temperature distribution, thermal expansion, thermal history), and quality correction (defect probability, reliability, prediction error).

[0158] The performance of the multi-sensor fusion system surpasses that of existing single or binary sensor systems, demonstrating innovative achievements such as an etching uniformity of 1.8% (a 49% improvement over the existing 3.5%), a positional precision of ±0.6μm (a 50% improvement), and a wafer defect rate of 0.03% (an 85% reduction). The processing time of the fusion system has been reduced to less than 2.5 seconds through parallel processing and hardware acceleration, which can provide sufficient responsiveness for real-time process control. In terms of long-term stability, ternary sensor fusion significantly improves robustness against drift or failure of individual sensors, allowing for safe mode operation with over 90% performance even in the event of a single sensor failure and 70% performance even in the event of two sensor failures.

[0159] This is a diagram showing the overall process flow diagram and control algorithm of a system for transporting a square solar cell wafer under vacuum conditions, including dry etching for a solar cell according to the present invention. It includes processing steps for each stage from start to end and conditional sequence control logic based on tolerance determination.

[0160] The process starts at "Start" and proceeds to the first step, "Substrate Loading (1->2)". In this step, a wafer is transferred from the loader elevator (100) to the transfer module (200). Simultaneously with automatic specification recognition by a laser sensor, a robot arm picks up the wafer through vacuum suction.

[0161] In the second step, "Alignment Module Transfer (2->3)," the wafer is transferred from the transfer module (200) to the alignment module (300). The robot arm is converted into a folded form to pass through a confined space and place the wafer onto the alignment chuck.

[0162] In the third step, "Performing Linear Scan Alignment," a linear moving sensor measures positional and rotational errors by moving in a straight line without rotating the wafer. A diagonal alignment roller guide guides the wafer to a reference position, while a line scan camera or laser displacement sensor performs precise measurements.

[0163] Alternatively, in the third step, "Perform Linear Scan Alignment," a linear moving sensor can be used to verify that the wafer is loaded in the correct position without rotating the wafer.

[0164] In the fourth step, "Calculation of Error Values," based on the linear scan results, βx, βy, The value is calculated. An algorithm based on the least squares method is applied to quantify the position error and rotation error.

[0165] The fifth step is "tolerance determination," where the measured error value is compared with a preset tolerance range. Rotational error is determined to be within the range of 0.05 degrees or more and 0.15 degrees or less, and positional error is determined to be within the range of 3 μm or more and 7 μm or less. Based on this determination result, the process branches into two paths.

[0166] If the tolerance determination is "FAIL," the path branches to the "Retry" path. During the retry process, one to three repeated measurements are performed on the same substrate. During the retry, the sensor scan speed is reduced and the measurement area is expanded to improve precision. If the retry is successful, the process returns to the main flow; if failure persists, the path branches to the rework path and the substrate is returned to the loader.

[0167] If the tolerance range judgment is "PASS," the process proceeds to the "etching chamber transfer" step. A transfer robot picks up the wafer with a corrected posture and trajectory and transfers it to the etching chamber (400). During this process, the error value measured in the alignment is transmitted together and utilized in subsequent processes.

[0168] In the "feedforward correction" step, process conditions are adjusted in real time based on the received error values. At least one of the gas flow rate, RF power, or chamber pressure is adjusted in a feedforward manner to compensate for positional and rotational errors of the substrate. For example, if a rotational error is detected, the gas flow rate and RF bias in the corresponding area are differentially adjusted.

[0169] In the "dry etching" step, plasma etching is performed under adjusted process conditions. Anti-reflective film removal or surface texturing is performed using CF4 and O2 gases, and process stability is ensured through real-time monitoring.

[0170] In the final stage, the "unloading (4->2->1)" process, the etched wafer is returned from the etching chamber to the loader via the transfer module. The robot arm is converted back into an extended form to safely transport the wafer and load it into a designated location within the cassette.

[0171] When all processes are completed, the system reaches "termination" and returns to its initial state for the next wafer processing. During continuous processing, a pipeline method is applied so that multiple wafers can be processed simultaneously at different stages.

[0172] The core of this process flow diagram is conditional sequence control, which automatically selects a retry or feedforward correction path based on alignment results, thereby simultaneously achieving quality assurance and productivity improvement. Furthermore, integrated control is possible through real-time data sharing between modules, enabling fully automated unmanned operation.

[0173] The present invention will be explained in more detail below using examples. However, the following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0174] Example 1

[0175] In this embodiment, the process from automatic specification recognition of an M10 wafer (182mm x 182mm) to overall system synchronization is specifically described using the laser sensor system shown in FIGS. 2 and 3.

[0176] Twenty wafers were stacked at 4mm intervals in the M10 dedicated cassette (123). Three holes (134) were precision machined at 182mm intervals on the top of the cassette, and the diameter of the holes was 6mm. As the cassette was placed on the lower lifting structure (120) of the load lock chamber (110), the chamber door (125) was automatically closed.

[0177] The laser sensor (130) started operating at a height of 150 mm above the cassette. A red laser with a wavelength of 650 nm was positioned to generate three parallel lines that matched exactly with the cassette holes. A non-contact sensor (131) monitored the change in intensity of the laser signal passing through each hole in real time. During the scanning process, the distance between the central hole and the two side holes was measured to be 91 mm (half of 182 mm) each. The total scanning time was 0.8 seconds, and measurement accuracy was ensured through three consecutive measurements.

[0178] The recognition unit (132) compared and analyzed the measured hole spacing data with a stored database. Among M6 (166mm), M10 (182mm), and M12 (210mm), it was finally determined to be M10, which showed a 99.8% agreement with the 182mm pattern. During the determination process, a position error of ±1mm was automatically corrected, and the recognition accuracy was 99.9%.

[0179] The control unit (133) simultaneously transmitted the M10 recognition result to all modules of the system within 0.1 seconds. In the alignment module, the roller guide (310) automatically moved to the M10 dedicated position and was adjusted to a position corresponding to a size of 182 mm on the linear guide (311). The stroke range of the drive unit (312) was set to match the M10 standard. In the etching chamber, the control unit detected that the currently installed edge ring was for M6 and generated an "M10 edge ring replacement required" alarm to the operator. After the replacement of the M10 edge ring (inner diameter 184 mm x 184 mm) was completed, a "Ready" signal was received.

[0180] Based on the recognition result of the laser sensor, the elevator plate (121) was automatically adjusted to a pickup position optimized for the M10 wafer. The position was set with a precision of ±0.05 mm along the lifting guide (122), and the pitch correction of the central lifting screw was made to accurately correspond to a wafer spacing of 4 mm.

[0181] When the synchronization of all modules was completed, the system control unit generated a "M10 mode ready" signal. During this process, the adsorption pads of the transfer module were activated with 6 pads to match the M10 size, the vacuum adsorption pressure was set to -88kPa, which is optimized for an M10 wafer thickness of 200μm, and the acceleration / deceleration profile of the transfer robot was limited to 1.2m / s² to reflect the characteristics of the large-area wafer.

[0182] The effectiveness of the laser sensor-based automatic recognition system was demonstrated through this embodiment. From cassette loading to system synchronization, the entire process was completed within 3 seconds, and the M10 specification was identified with a high accuracy of over 99.9%. Compared to conventional manual setup, the processing time was reduced by 90%, and process errors caused by the use of incorrect edge rings could be prevented in advance. As a result of continuously processing 20 M10 wafers, consistent automatic recognition performance was observed on all wafers, and no system synchronization failures occurred. Due to the non-contact method of the laser sensor, no damage to the cassette or wafer was observed.

[0183] Example 2

[0184] In this embodiment, the process of diagonal alignment and error correction of an M12 wafer (210mm x 210mm, thickness 180μm) using the alignment module shown in FIG. 4 is described in detail.

[0185] The M12 wafer received from the transfer module was placed on the quartz chuck of the alignment module (300). Since the system was already set to M12 mode, the outer roller guide (310) was automatically activated. The roller guide was positioned at a 45-degree angle to correspond to a diagonal length of 297 mm (210 x √2).

[0186] A total of four roller guides simultaneously contacted the wafer from the four diagonal corners. Each roller had a diameter of 18 mm to accommodate large-area wafers, and its surface was treated with a ceramic coating to minimize friction. The rollers rotated at a safe speed of 15 rpm, preventing wafer damage with a contact pressure of 0.5 N. The rollers gently contacted the diagonal corners of the wafer and guided it to the center position; during this process, the wafer was aligned solely by linear movement without rotation, thereby minimizing mechanical stress on the large-area thin wafer.

[0187] The roller guides on the linear guide (311) were positioned to the M12 standard. The precision ball bearing type linear guide provided a movement precision of ±0.5μm. The servo motor of the drive unit (312) moved each roller to an optimal position via a ball screw. The X-axis roller and the Y-axis roller were each positioned at ±148.5mm (210mm x cos 45°) from the center and moved at a precision control speed of 50mm / s. A position resolution of 0.1μm was achieved.

[0188] The control unit (313) activated the error calculation unit (314) to operate the linear moving sensor. The sensor head slid to detect whether the wafer was loaded normally.

[0189] The retry control unit (315) compared the measured error value with the allowable range. Based on the M12 large area, the allowable range for rotational error was set to 0.12 degrees or less, and the allowable range for positional error was set to 5 μm or less. It was confirmed that all measurement results were within the allowable range. If the allowable range was exceeded, a retry was performed up to 2 times, the retry interval was 8 seconds (large area wafer stabilization time), and the scan speed during the retry was reduced to 75 mm / s, thereby improving precision.

[0190] The signal transmission unit (316) transmitted an "ALIGN OK" signal to the transfer module. At the same time, the measured error value information was also transmitted so that it could be utilized in a subsequent process. The transmitted data consisted of a status of "Alignment complete", a position error [+4.2, -3.8] μm, a rotation error of +0.08 degrees, a wafer size of "M12", and an alignment time of 4.2 seconds.

[0191] Several advantages over conventional rotary alignment were confirmed during this alignment process. In terms of mechanical stability, the absence of wafer rotation resulted in zero centrifugal force, preventing deformation caused by the inertial force of large-area wafers. The wafer breakage rate during alignment was 0%, completely eliminating the previous rate of 0.2%. Regarding measurement precision, the full-edge scanning achieved an averaging effect, and the impact of local defects was minimized. Repeatability improved significantly to ±0.5μm compared to the previous ±1.2μm. Processing speed was also reduced to 4.2 seconds from 8.5 seconds, and the retry rate decreased from 5% to 1%.

[0192] As a result of processing 50 M12 wafers in succession, the alignment success rate was 99% (including retries), and the average alignment time was 4.1 ± 0.3 seconds. Positional precision was ±0.8 μm, a 60% improvement over the target, and there were zero wafer damages. Safe and precise alignment of large-area thin wafers was successfully achieved through a combination of diagonal alignment and linear scanning.

[0193] Example 3

[0194] In this embodiment, the process of diagonal alignment and error correction of an M12 wafer (210mm x 210mm, thickness 180μm) using the alignment module shown in FIG. 4 is described in detail.

[0195] The M12 wafer received from the transfer module was placed on the quartz chuck of the alignment module (300). Since the system was already set to M12 mode, the outer roller guide (310) was automatically activated. The roller guide was positioned at a 45-degree angle to correspond to a diagonal length of 297 mm (210 x √2).

[0196] A total of four roller guides simultaneously contacted the wafer from the four diagonal corners. Each roller had a diameter of 18 mm to accommodate large-area wafers, and its surface was treated with a ceramic coating to minimize friction. The rollers rotated at a safe speed of 15 rpm, preventing wafer damage with a contact pressure of 0.5 N. The rollers gently contacted the diagonal corners of the wafer and guided it to the center position; during this process, the wafer was aligned solely by linear movement without rotation, thereby minimizing mechanical stress on the large-area thin wafer.

[0197] The roller guides on the linear guide (311) were positioned to the M12 standard. The precision ball bearing type linear guide provided a movement precision of ±0.5μm. The servo motor of the drive unit (312) moved each roller to an optimal position via a ball screw. The X-axis roller and the Y-axis roller were each positioned at ±148.5mm (210mm x cos 45°) from the center and moved at a precision control speed of 50mm / s. A position resolution of 0.1μm was achieved.

[0198] The control unit (313) activated the error calculation unit (314) to operate the linear moving sensor. The sensor head sequentially scanned the four corners of the wafer while moving linearly along the X-axis for a distance of 250 mm. The scan speed was 150 mm / s, the measurement interval was 0.1 mm, and the total scan time was 4.2 seconds. By using a combination of a line scan camera and a laser displacement sensor, the measurement results showed a position error of βx = +4.2 μm, βy = -3.8 μm, and a rotation error of = +0.08 degrees, and the measurement uncertainty was achieved at ±0.3 μm.

[0199] The retry control unit (315) compared the measured error value with the allowable range. Based on the M12 large area, the allowable range for rotational error was set to 0.12 degrees or less, and the allowable range for positional error was set to 5 μm or less. It was confirmed that all measurement results were within the allowable range. If the allowable range was exceeded, a retry was performed up to 2 times, the retry interval was 8 seconds (large area wafer stabilization time), and the scan speed during the retry was reduced to 75 mm / s, thereby improving precision.

[0200] The signal transmission unit (316) transmitted an "ALIGN OK" signal to the transfer module. At the same time, the measured error value information was also transmitted so that it could be utilized in a subsequent process. The transmitted data consisted of a status of "Alignment complete", a position error [+4.2, -3.8] μm, a rotation error of +0.08 degrees, a wafer size of "M12", and an alignment time of 4.2 seconds.

[0201] Several advantages over conventional rotary alignment were confirmed during this alignment process. In terms of mechanical stability, the absence of wafer rotation resulted in zero centrifugal force, preventing deformation caused by the inertial force of large-area wafers. The wafer breakage rate during alignment was 0%, completely eliminating the previous rate of 0.2%. Regarding measurement precision, the full-edge scanning achieved an averaging effect, and the impact of local defects was minimized. Repeatability improved significantly to ±0.5μm compared to the previous ±1.2μm. Processing speed was also reduced to 4.2 seconds from 8.5 seconds, and the retry rate decreased from 5% to 1%.

[0202] As a result of processing 50 M12 wafers in succession, the alignment success rate was 99% (including retries), and the average alignment time was 4.1 ± 0.3 seconds. Positional precision was ±0.8 μm, a 60% improvement over the target, and there were zero wafer damages. Safe and precise alignment of large-area thin wafers was successfully achieved through a combination of diagonal alignment and linear scanning.

[0203] Example 4

[0204] In this embodiment, the process of switching from M6 to M10 specifications and the wafer processing process according to each specification are explained using the common chuck and interchangeable edge ring (410) structure shown in FIGS. 5 and 6.

[0205] After the processing of the M6 ​​wafer batch was completed, the edge ring replacement process for the next batch, the M10 wafer processing, began. The vacuum inside the etching chamber (400) was returned to atmospheric pressure, and the upper port of the chamber was opened. The existing M6 edge ring had an internal opening that was a 168mm x 168mm square, an outer diameter of 211.9mm (common standard), a thickness of 15mm, and a weight of 1.2kg. The M6 ​​edge ring was removed by lifting it vertically from the drive hole of the drive lifter of the common chuck. Due to the 0.1mm gap between the inner diameter of the drive hole (212mm) and the outer diameter of the edge ring (211.9mm), it was easily separated without a separate tool.

[0206] After removing the edge ring, the condition of the common chuck was inspected. The common chuck was a 225mm x 225mm square and was made of an aluminum alloy with anodizing treatment. The surface of the common chuck and the inside of the drive hole were cleaned with a nitrogen blower to remove residual particles. The roundness and surface roughness of the drive hole were checked, and the result showed that it was in good condition with an Ra of 1.2μm.

[0207] An edge ring dedicated to the M10 was prepared and installed on a common chuck. The specifications of the M10 edge ring were that the inner opening was a 184mm x 184mm square, providing a 1mm margin over the 182mm M10 wafer, and the outer diameter was 211.9mm, which was the same common standard as the M6. The material was quartz with excellent chemical resistance, and the radius of curvature of the edge was 5mm (5T, based on T=1mm). After aligning the edge ring over the drive hole, it was lowered vertically to engage. The outer diameter of 211.9mm precisely interlocked with the inner diameter of 212mm groove, fixing it in a concentric state.

[0208] An M10 wafer (182mm x 182mm, thickness 200μm) was placed on the edge ring. The special cross-sectional structure of the edge ring induced stable placement due to the wafer's own weight. Based on a thickness of 3T = 3mm, the cross-sectional structure of the edge ring features an upper 2T (2mm) section with a 52-degree inclined surface to induce wafer sliding, while the lower 1T (1mm) section is a vertical surface for final support and position fixation. The curved edge surface has a radius of curvature of 5mm to prevent stress concentration. The wafer slid naturally along the inclined surface of the edge ring and was stably placed on the lower vertical surface. Perfect centering was achieved due to a 1mm clearance between the 184mm internal opening and the 182mm wafer.

[0209] After the wafer was seated, the fixing member operated to secure the wafer to the surface of the common chuck. The chucking voltage was set to 1.5 kV to match the characteristics of the M10 wafer. Vacuum suction holes positioned in the common chuck operated auxiliaryly to ensure the flatness of the wafer. It was confirmed that safe separation was possible during wafer unloading as the lift pin rose through the lift pin drive hole. Due to the combined structure of the common chuck and the edge ring, M10 processing was possible simply by changing process conditions without the need for a separate chuck replacement.

[0210] Dry etching was performed on M10 wafers. The 225mm x 225mm size of the common chuck covered up to a 210mm diagonal, providing sufficient margin for processing M10 (182mm) wafers. As a result of the process, the etching uniformity was 3.8%, which increased by 0.6% compared to M6, but this was determined to be due to the effect of the increased area. A stable margin was secured with an edge exclusion width of 2mm, there were zero wafer breakages, and the process reproducibility was excellent at ±2.1%.

[0211] The time required for the transition from M6 to M10 was analyzed. A total of 5 minutes and 45 seconds was taken, consisting of 90 seconds for vacuum release and chamber opening, 30 seconds for M6 edge ring removal, 60 seconds for common chuck cleaning and inspection, 45 seconds for M10 edge ring installation, and 120 seconds for vacuuming and process condition setting. This demonstrated a 70% reduction in time compared to the conventional method of replacing the entire chuck.

[0212] The advantages of the common chuck and replaceable edge ring structure were clearly demonstrated. In terms of versatility, a single common chuck could handle all M6 to M12 specifications, and in terms of cost-effectiveness, costs could be reduced by replacing only the edge ring without the need to purchase a separate chuck. In terms of efficiency, production flexibility was secured through rapid specification switching, and in terms of quality, consistent process quality was maintained with the same chuck structure. As a result of continuous operation over 30 days while repeatedly switching between M6, M10, and M12 specifications, the stability of the edge ring replacement process and the consistency of process quality were confirmed.

[0213] Example 5

[0214] In this embodiment, the integrated operation process of the entire system processing M10 wafer batches (25 wafers) according to the process flow diagram shown in FIG. 7 is described in detail in chronological order.

[0215] System initialization has started. The load lock chamber (110) is 1x10 -6 The vacuum level of Torr was maintained, and the transfer module (200) robot arm was waiting in the HOME position. The alignment module (300) was in a standby state, but the previous batch M6 setting remained, and the etching chamber (400) had the M6 ​​edge ring installed. M10 batch information was entered, and the wafer quantity was set to 25 wafers, wafer size to 182mm x 182mm, thickness to 200μm, and the process recipe to M10_ARC_ETCH_STD.

[0216] The board loading and automatic recognition process proceeded from 30 to 60 seconds. After the load lock chamber door was opened for 5 seconds, it took 10 seconds to place the M10 cassette onto the lower lifting structure. An additional 15 seconds were required to seal the door and initiate vacuuming. During the laser sensor automatic recognition, a 3-hole scan was completed to detect a pattern with an 182mm spacing, and after it was determined to be an M10 specification (99.8% confidence), a command to switch the entire system mode was sent.

[0217] System synchronization and setting changes were performed from 60 to 180 seconds. Reconfiguration of the alignment module took 60 seconds, and the roller guide position was changed from M6 to the M10 corresponding position. The tolerance settings were updated to adjust the rotational error to ±0.12° and the positional error to ±5μm, and sensor calibration was completed. Replacing the etching chamber edge ring took 120 seconds; after releasing the vacuum and opening the top of the chamber, the M6 ​​edge ring was removed and the M10 edge ring was installed. Vacuuming and process condition optimization were completed, and the RF power was set to 750W, CF4105sccm, O25sccm, and the chamber pressure to 770mTorr.

[0218] The first wafer processing cycle proceeded from 180 seconds to 320 seconds. In the 1->2->3 path transferring from the loader to the alignment module, it took 3 seconds for the elevator to ascend to the first wafer position, 5 seconds for the six vacuum pads to adsorb at -88kPa at the transfer robot pickup, and 8 seconds for the alignment module to be transferred and placed. During the alignment process, diagonal alignment began, and the four roller guides operated simultaneously. The linear scan error measurements showed △x=+2.1μm, △y=-1.8μm, and Θ=+0.04°; upon determining the tolerance range, all values ​​were within the criteria and were processed as PASS. The alignment completion time was 4.8 seconds.

[0219] Transfer from alignment to the etch chamber proceeded via path 3->2->4. It took 3 seconds to re-pick up the wafer in the corrected position, and 7 seconds for transfer to the etch chamber and chucking. In the feedforward adjustment of process conditions, the top CF4 was increased by +1.5 sccm to compensate for Θ=+0.04°, and the RF bias was increased by +3W to compensate for positional error. The dry etching process was performed for 70 seconds; real-time monitoring confirmed a normal torque range of 1.02 Nm, and the plasma emission exhibited uniform blue emission. After the completion of etching, unloading proceeded via path 4->2->1; it took 5 seconds for chucking release and wafer pickup, and 8 seconds for returning to the load lock chamber and loading the cassette, resulting in a total time of 140 seconds per wafer.

[0220] Continuous processing optimization began at 320 seconds. Pipeline processing was implemented, and starting from the second wafer, simultaneous processing was performed in a pipeline manner. Wafer A was processed in the etching chamber for 70 seconds, Wafer B was aligned in the alignment module for 5 seconds, and Wafer C waited in the loader, thereby reducing the actual throughput to 75 seconds / wafer.

[0221] An example of an abnormal situation handling occurred on the 13th wafer. Due to an exceedance of the tolerance, the alignment result of the 13th wafer was △x=+8.2μm, which exceeded the tolerance of 5μm, activating an automatic retry sequence. During the retry process, the first retry was performed with a scan speed reduced to 75mm / s and an expanded area. The retry result was △x=+4.1μm, △y=-2.9μm, and Θ=+0.07°, passing the tolerance determination, and the retry time was +15 seconds. Feedforward correction was enhanced, and based on the error value of the wafer that succeeded in the retry, the RF bias was additionally adjusted by +8W and the CF4 flow rate by +2.5sccm.

[0222] Batch completion and statistical analysis were completed at 3100 seconds. The total batch processing results showed a total of 25 wafers processed, an average processing time of 75 seconds / wafer (pipeline effect), 2 retries (8%), 0 reworks, and 0 wafer breakage. In the quality statistics, the etching uniformity was 3.6±0.4%, the positional accuracy was ±1.2μm (target achieved), and the process reproducibility was CV 2.8% (excellent). In the system efficiency analysis, the overall utilization rate was 96.2%, the total waiting time was 90 seconds (waiting for edge ring replacement), and the throughput was 48 wafers / hour, exceeding the target of 45 wafers / hour.

[0223] In the next batch preparation, a batch-to-batch transition was prepared for the M12 transition. The next batch information consisted of 20 M12 wafers, and an edge ring replacement was scheduled from M10 to M12. The alignment setup was prepared in advance, and the estimated transition time was 4 minutes.

[0224] In terms of the key performance indicators of the integrated control system, the level of automation achieved continuous unmanned processing of 25 sheets without operator intervention, automatic detection and recovery of abnormal situations, real-time process optimization, and quality assurance. In terms of system interoperability, perfect synchronization among five modules, real-time data sharing and feedback, and overall control through a single integrated interface were implemented. Regarding productivity improvement, throughput increased by 300% compared to conventional manual processing, the defect rate improved by 80% from 0.5% to 0.1%, and the equipment utilization rate increased from 85% to 96%.

[0225] Through this embodiment, it was demonstrated that an integrated transport system for dry etching of square solar cell wafers is suitable for fully automated mass production.

[0226] Although embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and it will be obvious to those skilled in the art that various modifications and variations are possible within the scope of the technical concept of the present invention as described in the claims. Explanation of the symbols delete

Claims

Claim 1 A loader elevator for placing a cassette in which multiple wafers are stacked in multiple layers; a load lock chamber provided in the loader elevator for switching between atmospheric pressure and vacuum states; a transfer module including a rotary robot arm for transporting wafers from the load lock chamber; an alignment module for receiving wafers from the transfer module and aligning them; and an etching chamber for receiving aligned wafers from the alignment module and performing dry etching, wherein the loader elevator includes a laser sensor for automatically recognizing wafer size; a lower lifting structure for controlling the wafer loading position according to the output of the laser sensor; and multiple cassettes for stacking multiple wafers, wherein the alignment module includes a plurality of roller guides arranged diagonally to a square wafer and contacting the corners; a driving unit for moving the roller guides on a linear guide; and a control unit for selectively operating the roller guides according to the size of the wafer, wherein the alignment module includes an error calculation unit for calculating position error and rotation error by moving a linear moving sensor in a straight line without rotating the wafer; The apparatus further includes a retry control unit that retryes the alignment process when the calculated error value exceeds an allowable range; and a signal transmission unit that transmits an alignment completion signal to the transfer module when the error value is within an allowable range, wherein the etching chamber includes a square common chuck; a plurality of replaceable edge rings detachably coupled to the common chuck and corresponding to different wafer specifications; and a lifter driving hole formed in the common chuck to fix the edge rings, and the laser sensor includes a non-contact sensor that recognizes the wafer size by matching with a plurality of holes formed in the cassette; and a recognition unit that distinguishes wafers of different sizes.A system for transporting a square solar cell wafer under vacuum conditions including dry etching for solar cells, characterized by including a control unit that automatically synchronizes the edge ring of the etching chamber and the alignment module according to the recognition result. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 A system for transporting square solar cell wafers under vacuum conditions including dry etching for solar cells, wherein, in claim 1, the edge ring comprises: a curved surface formed at the corner; a cross-sectional structure having an inclined surface at the top and a vertical surface at the bottom; and a structure in which a plurality of edge rings have the same outer diameter. Claim 6 delete

Images