Vacuum robot station automatic teaching method and system, and storage medium

Abstract

The application discloses a kind of vacuum robot station automatic teaching method and system, storage medium.The method is placed in the measuring device integrated with displacement sensor and linear CCD sensor in the station to be taught, and the center of the two coincides;Robot wafer stretches into device and drops, and the contact station is judged based on the distance change rate of displacement sensor, the Z-axis coordinate when contact is recorded to determine the height of wafer stretching in and leaving, and the height correction coefficient of fitting Z-axis position and each sensor reading during dropping is recorded;At the height of wafer stretching in, the first offset is obtained by height coupling correction based on each sensor reading, current height and correction coefficient, and the horizontal position of robot is adjusted to meet the accuracy requirement;After wafer is placed, second offset is obtained by correction again according to current height, and the station coordinate is updated based on the relationship between offset before and after placement and iterated to meet the accuracy requirement.The application realizes single-device X / Y / Z three-axis full-automatic teaching, without manual input, and teaching accuracy and efficiency are high.

Description

Technical Field

[0001] This invention relates to the field of semiconductor vacuum robot technology, and in particular to an automatic teaching method and system for vacuum robot workstations based on an integrated measurement device, as well as a storage medium. Background Technology

[0002] In the semiconductor manufacturing industry, vacuum robots are responsible for the precise handling of wafers between various process chambers and transfer chambers. The robot needs to accurately know the coordinate parameters of each workstation (including R / T direction position and Z-axis height) to ensure the accuracy and safety of wafer handling. Therefore, the teaching (i.e., calibration) accuracy of the workstation coordinates directly affects the reliability of wafer transfer and chip yield.

[0003] Currently, the teaching of workstation coordinates mainly relies on manual completion, which has problems in both the Z-axis height and X / Y directions.

[0004] Regarding Z-axis height teaching: the mainstream teaching method is manual operation. For example, after the robot finger carries the wafer into the workstation, a safe distance of approximately 3mm must be maintained between the lower surface of the wafer and the contact surface of the workstation. If the distance is too small, the wafer vibration during movement can easily cause it to scratch the contact surface; if the distance is too large, it will affect the wafer placement time efficiency. On-site operators need to repeatedly measure whether the safe distance meets the standard with a feeler gauge after the robot enters the workstation. If it does not meet the standard, the robot is controlled to move up and down with a teach pendant to adjust it, and the measurement is repeated until the requirement is met. Similarly, after the wafer is placed, a safe distance of approximately 3mm must also be maintained between the lower surface of the wafer and the upper surface of the finger, which also requires feeler gauge measurement and repeated adjustment. The whole process is time-consuming, inefficient, relies on human experience, and the accuracy is greatly affected by human factors.

[0005] Regarding X / Y direction teaching: traditional methods mainly involve manual visual alignment or the fabrication of specialized fixtures to align the center of the finger with the center of the workstation. However, the accuracy of visual alignment is only about ±0.5mm, far below the micron-level precision requirements of semiconductor processes. Furthermore, the fixture-based teaching method has a fundamental flaw: during the previous process, the center of the wafer and the center of the finger may not necessarily coincide (depending on the actual placement of the wafer on the finger). This results in the wafer and the workstation center not aligning, meaning the fixture can only guarantee alignment between the finger and the workstation, not between the wafer and the workstation.

[0006] To address the automation of teaching, existing technologies have proposed several solutions. For example, patent CN113228246 discloses a substrate handling robot and an automatic teaching method, which installs two optical sensors on the robot's manipulator to obtain the teaching position by detecting the intersection of light beams and scanning a target placed at the workstation. However, this solution requires the sensors to be mounted on the manipulator, necessitating the placement of optical fibers and amplifiers, which undoubtedly increases the complexity and weight of the manipulator. Furthermore, the sensors are in motion and are highly susceptible to vibration. In addition, this solution only addresses teaching in the X / Y plane; the Z-axis direction requires additional up-and-down movement to detect the upper end of the target, resulting in limited accuracy.

[0007] On the other hand, patent CN120422205 discloses an automated teaching system and method for a wafer handling robot. A sensor is fixedly installed directly above the target workstation. After the robot grips the wafer, it rotates along its axis to bring the wafer edge past the sensor's optical axis. The edge tangent point is detected by signal abrupt change, and the center position is then fitted using the least squares method. This solution relies on the sensor to detect the signal abrupt change (step signal) when the wafer edge passes by. However, detecting the edge signal involves both the sensor's own response delay and the system's data acquisition delay. The displacement error caused by this delay varies at different robot speeds, with the error increasing at higher speeds. While a speed-decreasing strategy gradually reduces the error, it cannot completely eliminate the impact of the delay on accuracy.

[0008] It is evident that existing technologies not only fail to detect the actual position of the wafer after it is placed on the workstation and are unable to eliminate the minute displacement errors generated during the wafer placement and handover process, but also lack integrated means to automatically complete high-precision three-axis teaching. Therefore, new solutions urgently need to be proposed by those skilled in the art. Summary of the Invention

[0009] Therefore, the main objective of this invention is to provide an automatic teaching method and system for vacuum robot workstations, as well as a storage medium, to solve the problems of low efficiency and insufficient accuracy in three-axis teaching caused by the reliance on manual operation in existing vacuum robot workstation teaching methods, which cannot automatically detect and correct the junction errors generated after wafer placement.

[0010] To achieve the above objectives, according to one aspect of the present invention, an automatic teaching method for a vacuum robot workstation is provided, based on an integrated measurement device, the steps of which include:

[0011] Place the integrated measuring device at the teaching station so that the centers of the two coincide; the integrated measuring device integrates a displacement sensor and a linear CCD sensor with at least two intersecting measuring lines that pass through the center of the device.

[0012] The robot finger, carrying the wafer, is inserted into the integrated measuring device and lowered. The displacement sensor measures the distance in real time. Based on the distance change rate characteristics, the robot determines the contact position and records the robot's Z-axis coordinate at the moment of contact. The insertion height and exit height of the wafer are then determined accordingly. During the descent, the Z-axis position and the readings of each linear CCD sensor are recorded simultaneously, and a height correction coefficient is obtained by fitting the data.

[0013] At the tape insertion height, height coupling correction is performed based on the readings of each linear CCD sensor, the height measured by the current displacement sensor, and the height correction coefficient to obtain the first offset of the wafer, and the robot's horizontal position is adjusted accordingly to meet the accuracy requirements.

[0014] After placing the wafer, the readings of each linear CCD sensor are read again, and height coupling correction is performed again based on the current height to obtain the second offset of the wafer after placement. The target station coordinates are updated based on the relationship between the offsets before and after placement, and the robot is controlled to re-fetch wafers iteratively until the offsets before and after placement meet the accuracy requirements.

[0015] In a possible preferred embodiment, the step of determining the contact station based on the distance change rate feature includes: calculating the change rate of displacement sensor readings in adjacent steps; when the change rate stabilizes within a first preset range during the free fall phase, and the change rate of subsequent consecutive steps is lower than a second preset threshold, it is determined that the wafer has contacted the station.

[0016] In a possible preferred embodiment, the method further includes the step of: if the rate of change briefly falls below a second preset threshold during the descent but then recovers, it is determined to be a false contact, and the robot finger continues to descend.

[0017] In a possible preferred embodiment, the step of fitting to obtain the height correction coefficient includes: during the descent of the vacuum robot, synchronously recording the Z-axis position and the readings of each of the linear CCD sensors to establish a data queue; based on the readings corresponding to each descent height in the data queue, using a linear fitting method to obtain the height correction coefficient of each of the linear CCD sensors caused by the unit height change.

[0018] In a possible preferred embodiment, the calculation step of the first offset includes: obtaining the current height of the wafer measured by the displacement sensor, calculating the height difference between the current height and the calibration reference height; and performing height coupling correction on the original offset determined based on the readings of each of the linear CCD sensors according to the height difference and the height correction coefficient to obtain the first offset.

[0019] In a possible preferred embodiment, the linear CCD sensor includes a first linear CCD sensor and a second linear CCD sensor, with measurement lines respectively arranged along the X and Y directions, and the high coupling correction is performed according to the following formula:

[0020]

[0021]

[0022] in, , These are the current readings of the first linear CCD sensor and the second linear CCD sensor, respectively, where K is the CCD pixel equivalent. and This is the factory calibration reference reading. The current height is measured in real time by the displacement sensor. This is the factory-calibrated reference height. and These are the height correction coefficients corresponding to the X and Y directions, respectively.

[0023] In a possible preferred embodiment, the step of updating the target workstation coordinates based on the relationship between the offsets before and after placement, and controlling the robot to re-pick the wafer iteratively, includes: comparing whether the difference between the first offset and the second offset is greater than a preset threshold; if it is greater, adding the difference to the current workstation coordinate offset to update the workstation coordinates; using the updated workstation coordinates as the target position, controlling the robot to re-pick up the wafer, and repeating the above steps of horizontal position adjustment and offset detection after placement until the difference is not greater than the preset threshold.

[0024] In a possible preferred embodiment, the method further includes: repeating the step of lowering the robot finger with the wafer to the determination contact station and determining the wafer insertion height and exit height thereon at least twice, and taking the average value of the wafer insertion height and exit height obtained each time as the final Z-axis teaching parameter.

[0025] In a possible preferred embodiment, when the number of linear CCD sensors is three or more, the wafer center position is fitted by least squares using the readings of all the linear CCD sensors to obtain the offset of the wafer center relative to the center of the integrated measurement device.

[0026] To achieve the above objectives, according to another aspect of the present invention, an automated teaching system for a vacuum robot workstation is also provided, comprising:

[0027] An integrated measurement device, configured to be placed at a teaching station, includes: a device body with a positioning structure at its bottom for engaging with a positioning element on the station to align the device center with the station center; a displacement sensor mounted at the top center of the device body, with a vertically downward measuring direction, for measuring the distance between the wafer surface above it and the displacement sensor; at least two linear CCD sensors mounted inside the device body, the measuring lines of each linear CCD sensor intersecting each other and passing through the center of the device; a vacuum robot for moving the wafer; and a control unit communicatively connected to both the vacuum robot and the integrated measurement device, and configured to perform the method described in any one of the above descriptions.

[0028] In a possible preferred embodiment, the control unit includes: a communication module for receiving measurement data from the displacement sensor and each linear CCD sensor in real time; a parameter module for reading and writing workstation teaching parameters; a motion module for controlling the motion of the vacuum robot end effector; and a teaching calculation module for performing contact determination, high coupling correction, and workstation coordinate correction calculations.

[0029] To achieve the above objectives, according to another aspect of the present invention, a computer-readable storage medium is also provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the automatic teaching method for a vacuum robot workstation as described in any of the preceding claims.

[0030] This invention discloses an automated teaching method and system for a vacuum robot workstation, along with a storage medium. It integrates a displacement sensor and a linear CCD sensor into a device that can be precisely placed on the workstation. The robot carries the wafer through a conventional pick-and-place process, inserting, lowering, placing, and removing it. This allows for simultaneous and automated teaching of the X / Y coordinates and Z-axis height parameters of the workstation on a single device, achieving fully automated closed-loop calibration. Furthermore, due to differences between the wafer's height during teaching and the sensor's factory calibration height, factors such as light source divergence angle, wafer thickness and chamfering, and edge penumbra can cause systematic deviations in the linear CCD sensor's detection results that vary with height. Existing technologies do not recognize or address this issue. Therefore, this invention eliminates the systematic deviations introduced by height differences by simultaneously recording and fitting a height correction coefficient during the descent process, performing height-coupled correction on the detection offset. By re-detecting after wafer placement, updating the workstation coordinates based on the relationship between the offsets before and after placement, and performing closed-loop iteration, the invention eliminates minor displacement errors generated during wafer placement and handover, thereby further improving the accuracy of workstation teaching. Attached Figure Description

[0031] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0032] Figures 1 to 2 This is a schematic diagram of the integrated measurement device architecture in the system of the present invention;

[0033] Figure 3 This is a schematic diagram of the integrated measuring device in the system of the present invention;

[0034] Figure 4 This is a schematic diagram of the system integration of the present invention into a vacuum transfer device;

[0035] Figure 5 This is a schematic diagram of the wafer position and the linear CCD sensor position in the system of the present invention, showing the obstruction state of the linear CCD sensor when the wafer is offset relative to the center of the workstation and when it is directly facing the center;

[0036] Figure 6 This is a schematic diagram of the system structure of the present invention;

[0037] Figure 7 This is a schematic diagram of the control unit in the system of the present invention.

[0038] Explanation of reference numerals in the attached figures

[0039] 1. Transmission cavity; 2. Vacuum robot; 3. Integrated measuring device; 31. Displacement sensor; 32. Linear CCD sensor; 33. Base; 34. Positioning pin; 35. Sensor bracket; 36. Communication interface; 37. Positioning hole; 9. Wafer. Detailed Implementation

[0040] To enable those skilled in the art to better understand the technical solutions of the present invention, the specific technical solutions of the present invention will be clearly and completely described below in conjunction with embodiments, so as to help those skilled in the art further understand the present invention. Obviously, the embodiments described in this application are merely some embodiments of the present invention, and not all embodiments. It should be noted that, for those skilled in the art, the embodiments and features in the embodiments of this application can be combined with each other without departing from the concept of the present invention and without conflict. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the disclosure and protection scope of the present invention.

[0041] Furthermore, the terms "first," "second," "S1," "S2," etc., used in the specification, claims, and drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such features can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those described herein. At the same time, the stages described in each step are not necessarily to be implemented in the same step; it should be understood that the implementation order of the contents of each step stage can be adjusted and interchanged without violating the inventive concept, so that embodiments of the invention described herein can be implemented in orders other than those described herein. Additionally, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. Unless otherwise expressly specified and limited, the terms "set," "arrange," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; a mechanical connection or an electrical connection; a direct connection or an indirect connection through an intermediate medium; or a connection within two elements. Those skilled in the art can understand the specific meaning of the above terms in this case based on the specific circumstances and in conjunction with existing technology.

[0042] To address the issue that existing vacuum robot 2-station teaching relies on manual operation, cannot automatically detect and correct the junction error generated after wafer 9 placement, resulting in low efficiency and insufficient accuracy in three-axis teaching. For example... Figures 1 to 7 As shown, this invention provides an automated teaching system for a vacuum robot at two workstations, an example of which includes: a transfer chamber 1, a vacuum robot 2, and multiple workstations (such as...). Figure 4 The diagram shows workstations 2 and 1, an integrated measurement device 3, and a host computer. The vacuum robot 2 is fixed inside the transfer cavity 1, and its end effector (i.e., fingers) can move between workstations to drive the wafer 9 to complete pick-and-place operations. The integrated measurement device 3 is placed at the teaching workstation (…). Figure 4 (Taking workstation 1 as an example)

[0043] like Figure 7As shown, the host computer, acting as a control unit, includes, for example, a communication module, a parameter module, a motion module, and a teaching calculation module. The communication module receives measurement data from the displacement sensor 31 and each linear CCD sensor 32 in real time; the parameter module reads and writes the workstation teaching parameters; the motion module controls the motion of the end effector of the vacuum robot 2; and the teaching calculation module performs contact determination, high coupling correction, and workstation coordinate correction calculations. The communication connections are as follows: the communication interface 36 of the integrated measuring device 3 connects to the robot controller via an RS485 bus, outputting the measurement data from the displacement sensor 31 and the linear CCD sensor 32 in real time; the host computer communicates with the robot controller via Ethernet, issuing motion commands to the vacuum robot 2 and performing teaching calculations.

[0044] Furthermore, such as Figures 1 to 3 As shown, the example of the integrated measurement device 3 includes: a device body, a displacement sensor 31, at least two linear CCD sensors 32, and a communication interface 36.

[0045] The device body consists of a circular base 33 and a sensor bracket 35. Its shape is adapted to the wafer 9, and its placement at the workstation is consistent with the placement method of the wafer 9, facilitating installation by operators. The outer diameter of the device body is adapted to the size of a 300mm wafer 9. The bottom has three positioning holes 37, which cooperate with the positioning pins 34 on the workstation to ensure that the center of the device and the center of the workstation are precisely aligned when the device is placed on the workstation. The positioning accuracy is preferably ≤±2μm; the center of the device is the projection point of the center of the workstation onto the device.

[0046] The displacement sensor 31 is mounted at the top center of the device body, with the measurement direction vertically downward, and is used to measure the distance between the surface of wafer 9 and the displacement sensor 31. In this embodiment, the displacement sensor 31 is a laser displacement sensor 31, preferably with a resolution of 5μm, a range of 0-50mm, and a response time ≤1ms. In other embodiments, the displacement sensor 31 can also be replaced with a laser rangefinder or an ultrasonic rangefinder, as long as non-contact distance measurement can be achieved.

[0047] In a preferred embodiment, this example uses two linear CCD sensors 32, denoted as CCD-A (first linear CCD sensor) and CCD-B (second linear CCD sensor). Both sensors have 2048 pixels and a pixel pitch of 2μm (i.e., CCD pixel equivalent K = 2μm / pixel). The receiving end of each linear CCD sensor 32 is installed inside the device body, and the transmitting end is suspended directly above the receiving end by a bracket, forming a through-beam detection structure. The measurement line of CCD-A passes through the device center O along the X-axis, and the measurement line of CCD-B passes through the device center O along the Y-axis, with the two measurement lines forming a 90° angle. When the wafer 9 is inserted into the device, the edge of the wafer 9 blocks the linear CCD sensor 32. Each linear CCD sensor 32 outputs the center position value of the blocked segment, such as... Figure 5 As shown, this value directly reflects the offset of the center of wafer 9 in the direction of the sensor's measurement line.

[0048] It should be noted that at least two linear CCD sensors 32 are sufficient for X / Y direction teaching, with two orthogonally arranged sensors being the minimum configuration. In other embodiments, the number can be increased to three, four (e.g., evenly distributed along 0° / 90° / 180° / 270° directions) or more. The distribution angle of each measurement line is not limited to orthogonality and can be arbitrary, but all pass through the center of the device. When there are three or more linear CCD sensors 32, each sensor forms a blockage on the edge of the wafer 9 from different directions. Using the readings of all linear CCD sensors 32, the center position of each blockage segment is projected along its measurement line direction, and the position of the center of the wafer 9 is fitted using the least squares method to obtain the offset of the center of the wafer 9 relative to the center of the device. Therefore, the more sensors there are and the richer the angle distribution, the higher the center fitting accuracy, and the more redundant the capability to resist single sensor failure and local blockage interference.

[0049] The integrated measuring device 3 has undergone benchmark calibration at the factory, and the benchmark parameters are stored internally. An example of the stored benchmark parameters in this embodiment is: calibration benchmark height. =15.000mm (the reference height measured by displacement sensor 31 when the standard template is placed on the device base 33 during factory calibration); reference CCD reading Pixels. It should be noted that the height correction factor... , It is not a factory-fixed value, but is obtained in real time by linearly fitting the data queue established during the Z-axis teaching process to adapt to the actual conditions of different workstations.

[0050] The teaching principle of this embodiment will be explained below:

[0051] In the Z-axis direction, the robot finger slowly lowers the wafer 9. The sudden drop in the rate of change of the displacement sensor 31 reading determines the moment when the wafer 9 contacts the work station (i.e., the wafer 9 lands on the surface of the measuring device corresponding to the contact surface of the work station). This automatically obtains the contact surface height, and then automatically calculates the wafer insertion height and departure height in combination with the process safety distance. In the X / Y direction, the offset of the wafer 9's center relative to the device center can be directly obtained by using the center position of the obstruction segment of the linear CCD sensor 32 that is blocked by the wafer 9. This allows the robot's horizontal position to be adjusted to achieve alignment between the wafer 9 and the work station.

[0052] Because the standard sample is placed on the device base 33 during factory calibration (e.g., the calibration reference height is about 15mm), while the wafer 9 is at the height corresponding to the process station during actual teaching (the detection height can reach about 52mm), the difference is about 37mm. In actual measurement, when the wafer 9 is at this initial detection height, the CCD reading deviates from the reference value by 1024 pixels, reaching 1036~1042 pixels. Affected by factors such as the light source divergence angle, the thickness of the wafer 9 (about 775μm), and the chamfer and edge penumbra effects, the center of the CCD obstruction section will have a systematic shift of tens of micrometers at different detection heights at the same center position. Therefore, in this embodiment, during the descent process of Z-axis teaching, the CCD readings at different heights are recorded simultaneously, and a height correction coefficient is fitted to perform height coupling correction on the offset to eliminate this systematic deviation.

[0053] In addition, during the process of placing the wafer 9 from the finger onto the surface of the measuring device, a slight displacement (interface error) may occur. Therefore, this embodiment eliminates this by detecting the offset again after placement and performing closed-loop iterative correction.

[0054] Based on the principles described above, the teaching process is illustrated in detail below:

[0055] (I) Device Installation and Initialization

[0056] First, the operator places the integrated measuring device 3 on the teaching station (such as station 1), and the positioning hole 37 cooperates with the station positioning pin 34 to ensure accurate positioning. Then, the communication line is connected to read the factory reference parameters stored in the device.

[0057] (II) Z-axis height teaching

[0058] Step S1: The host computer controls the robot to pick up a wafer 9 (300mm wafer 9, 775μm thick) from station 2. The robot's finger, carrying the wafer 9, extends into station 1, where the integrated measuring device 3 is placed, according to the currently stored initial station coordinates (R, T) and the initial Z-axis height. It should be noted that this step uses the same initial station coordinates (R, T) as the robot's extension in the subsequent X / Y direction teaching steps to ensure that the data queue established during the descent process and the X / Y detection are at the same horizontal position reference.

[0059] Step S2: The host computer controls the robot to slowly lower the finger at a step speed of 0.1 mm / step. After each step, the Z-axis position, displacement sensor 31 reading, rate of change of adjacent steps, and CCD-A and CCD-B readings are recorded simultaneously to establish a data queue coupled with height. This data queue contains a complete correspondence from the initial height to the contact plane, naturally recording the systematic deviation of CCD occlusion at different detection heights, which is used for fitting height correction coefficients in subsequent steps.

[0060] The contact determination adopts the rate of change knee point detection method: Let the rate of change of displacement sensor 31 readings in adjacent steps be Ri = Di - D(i-1), where Di is the reading of displacement sensor 31 in the i-th step, then:

[0061] 1) Free fall phase: The displacement reading changes linearly with the decrease of Z, and the rate of change is stable within the first preset range of 80-120μm / step;

[0062] 2) Approaching contact stage: The rate of change gradually decreases;

[0063] 3) Contact determination: When the rate of change for 5 consecutive steps is lower than the second preset threshold of 10μm / step, it is determined that wafer 9 has contacted the workstation (i.e., wafer 9 is stably placed on the surface of the measuring device), and the Z-axis coordinate of the robot at this time is recorded as the contact surface height. .

[0064] If the rate of change briefly falls below the second preset threshold during the descent but then recovers to the first preset range, it is determined to be a false contact (caused by adhesion or relative horizontality), and the robot finger is controlled to continue descending.

[0065] The measured data (excerpt) from this embodiment is shown in the table below:

[0066] 1 22.3 52185 37185 - 1045.6 1040.7 -- 5 21.9 51770 36770 104 1045.5 1040.5 Freefall 10 21.4 51280 36280 98 1045.3 1040.2 Freefall 15 20.9 50790 35790 98 1045.1 1039.9 Freefall 20 20.4 50285 35285 101 1044.9 1039.8 Freefall 25 19.9 49795 34795 98 1044.4 1039.7 Freefall 28 19.6 49490 34490 102 1044.2 1039.2 Freefall 30 19.4 49280 34280 105 1044.1 1039.2 Freefall 32 19.2 49095 34095 93 1044 1039.1 Close contact 34 19 48930 33930 83 1043.9 1039 Close contact 36 18.8 48810 33810 60 1043.6 1038.9 Close contact 38 18.6 48720 33720 45 1044.1 1038.9 Close contact 40 18.4 48665 33665 28 1043.8 1039.1 Close contact 41 18.3 48660 33660 5 1043.8 1039.1 Contact determination 42 18.2 48665 33665 -5 1043.8 1039.1 Contact determination 43 18.1 48665 33665 0 1043.8 1039.1 Contact determination 44 18 48660 33660 5 1043.8 1039.1 Contact determination 45 17.9 48660 33660 0 1043.8 1039.1 Contact determination

[0067] As shown in the table above, the rate of change during the free descent phase from step 5 to 25 stabilizes at 98–104 μm / step; the rate of change gradually decreases from step 34 onwards; and from step 41 onwards, the change in each of the last 5 steps is less than 10 μm / step, indicating that wafer 9 has contacted the workstation, and the contact surface height is recorded. =18.3mm.

[0068] In addition, the table shows "height difference" "Listed as the reading and calibration reference height of displacement sensor 31" The difference is (h - ) in the high-coupling correction formula. ); the descent steps in the data queue ( The data refers to the high-correction coefficients obtained from linear fitting in step S3. , The sample data is as follows. The table also shows that as the detection height decreases, the CCD-A reading gradually changes from 1045.5 pixels to 1043.8 pixels, and the CCD-B reading gradually changes from 1040.5 pixels to 1039.1 pixels. This clearly demonstrates the systematic shift of the CCD occlusion center with increasing detection height. This data queue naturally records the relationship between this deviation and height, providing a data basis for the linear fitting in step S3.

[0069] Step S3: Perform linear fitting on the (height difference, CCD-A offset, CCD-B offset) data corresponding to each descent height in the descent data queue to obtain the height correction coefficient of each linear CCD sensor 32 caused by the unit height change. This embodiment obtains the following fitting results: =0.94μm / mm =0.75μm / mm, meaning that for every 1mm deviation of the detection height corresponding to the occlusion center of the CCD-A from the calibration reference height, a systematic deviation of 0.94μm is introduced in the X-direction, and the corresponding deviation for the CCD-B is 0.75μm / mm. This correction coefficient is used in subsequent X / Y direction teaching.

[0070] Step S4: Based on the safety distance required by the process Automatically calculates the tape insertion height Z_in and the finger removal height after wafer 9 is placed. :

[0071]

[0072]

[0073] Among them, the strip extension height To ensure a 3mm safe distance between the lower surface of the wafer and the contact surface of the workstation when the robot inserts the wafer, balancing scratch prevention and wafer placement efficiency; [Alternative height] After placing wafer 9 down, maintain a 3mm safe distance between the lower surface of wafer 9 and the upper surface of the finger to facilitate safe finger removal.

[0074] Step S5: The robot lifts the finger away from the device and repeats steps S1 to S4 three times. The average of the insertion height and exit height of the strip obtained each time is taken as the final Z-axis teaching parameter to eliminate single measurement error.

[0075] (III) Teaching in the X / Y directions

[0076] Step S6: The robot finger, carrying wafer 9, inserts the wafer at a normal speed and according to the wafer insertion height. The robot extends into the device with the initial workstation coordinates (R, T), and the accumulated adjustment is cleared to zero. The current position in the Cartesian coordinate system is then converted and recorded. Since the occlusion amount output by the linear CCD sensor 32 is a continuous analog quantity, there is no response delay issue as with through-beam sensors. Teaching in the X / Y directions does not require deceleration, and the robot can complete the extension at normal speed. After reaching the preset R / T position, the initial workstation polar coordinates are converted to the robot's Cartesian coordinate system position and recorded.

[0077]

[0078]

[0079] in, , The initial polar coordinates currently stored at workstation 1. The data is stored in units of one-thousandth of a degree, and during conversion, it is divided by 1000 and converted to radians. At this time, wafer 9 blocks two linear CCD sensors 32, and the measured CCD-A = 1045.2 pixels and CCD-B = 1040.3 pixels are obtained.

[0080] Step S7: Obtain the current height h = 51.66 mm of wafer 9 as measured by displacement sensor 31, and calculate the height difference between the current height and the calibration reference height. The first offset is obtained by performing high coupling correction according to the following formula:

[0081]

[0082]

[0083] Substituting the data from this step: the uncorrected original offset is:

[0084] ;

[0085] The high coupling correction amount is:

[0086] ;

[0087] First offset after correction , It is evident that without high-coupling correction, the difference between the detected height and the calibration reference height will introduce a systematic deviation of 30μm, far exceeding the accuracy requirement of ±5μm.

[0088] In other alternative embodiments, when the number of linear CCD sensors 32 is three or more, this step uses the readings of all linear CCD sensors 32 to fit the position of the center of the wafer 9 using the least squares method to obtain the offset of the center of the wafer 9 relative to the center of the device, and then obtains the first offset by the same high-coupling correction.

[0089] Step S8: Determine if the first offset meets the accuracy requirement of ±5μm. In this step, the X-direction offset of 8.0μm exceeds the threshold (Y-direction offset of 5.0μm is at the boundary), which does not meet the requirement. Therefore, the robot controls the finger to move in the direction of reducing the offset, i.e., -8μm in the X-direction and -5μm in the Y-direction. The correction distance is recorded and accumulated to the cumulative adjustment amount.

[0090]

[0091]

[0092] Reread the CCD value and repeat step S7 (e.g., iterate up to 10 times).

[0093] The second measurement data were: CCD-A = 1042.2 pixels (a decrease of 3.0 pixels from the first measurement, corresponding to 6.0 μm), CCD-B = 1038.3 pixels (a decrease of 2.0 pixels from the first measurement, corresponding to 4.0 μm), with the original offset... After the same high-coupling correction ( After that, the first offset converges to Meeting the ±5μm accuracy threshold, the center of wafer 9 is now aligned with the center of the workstation, with a cumulative adjustment of [amount missing]. .

[0094] (iv) Secondary testing and closed-loop correction after placement

[0095] Step S9: The robot controls the finger to extend from the tape to a certain height. Descend to contact surface height The wafer 9 is placed on the surface of the measuring device. Two linear CCD sensors 32 detect the position of the wafer 9 again and calculate the second offset of the wafer 9 after it is placed.

[0096] It should be emphasized that after placing wafer 9, the current height of wafer 9 measured by displacement sensor 31 becomes h = 48.66 mm, corresponding to a height difference Δh = 33.66 mm, which is smaller than Δh = 36.66 mm in step S7. This step involves re-acquiring the current height and performing height coupling correction again; if the Δh=36.66mm from step S7 is mistakenly used for correction, spurious errors will be introduced.

[0097] X-direction pseudo-error = ;

[0098] Y-direction pseudo-error = ;

[0099] This pseudo-error can be mistakenly included in the handover error, contaminating the iteration results. Real-time recalibration based on the current height ensures the comparability of detection results at different measurement heights, which is a prerequisite for handover error detection.

[0100] Step S10: In the first iteration, the first offset measured before placement is (2.0, 1.0) μm; after placement, CCD-A is detected to be 1045.8 pixels and CCD-B to be 1040.6 pixels, the original offset is... After height coupling correction of Δh=33.66mm (correction amount) After that, the second offset (12.0, 8.0) μm is obtained. The joint piece error is the difference between the first offset and the second offset, i.e., (12.0-2.0, 8.0-1.0) = (10.0, 7.0) μm, which exceeds the preset threshold of ±5 μm, and is therefore determined not to meet the accuracy requirements. This difference is accumulated into the current station coordinate offset to update the station coordinates, and this second offset is recorded as the correction compensation value for this station.

[0101] = -10; = -7;

[0102] The Cartesian coordinates of the workstation are then corrected based on the updated cumulative offset.

[0103]

[0104]

[0105] This is then converted to polar coordinates of the workstation:

[0106] ;

[0107] .

[0108] The robot picks up wafer 9 again, uses the updated workstation coordinates as the target position, adjusts the insertion position according to the correction compensation value, and repeats the above steps of horizontal position adjustment and offset detection after placement (up to 3 complete iterations).

[0109] In the second iteration, the first offset measured before placement was (2.0, 1.0) μm; after placement, CCD-A was measured to be 1042.3 pixels and CCD-B to be 1038.4 pixels. The original offset was:

[0110]

[0111]

[0112] After correction, the second offset is (5.0, 3.6) μm, and the crossover error is (5.0-2.0, 3.6-1.0) = (3.0, 2.6) μm, which is no greater than the preset threshold of ±5 μm, and the iteration terminates.

[0113] Step S11: After the teaching is completed, record the complete teaching parameters for workstation 1. .

[0114] (v) Multi-station teaching and device removal

[0115] After completing the teaching at station 1, the operator moves the integrated measurement device 3 to the next station to be taught (station 2), repeating the above teaching process to complete the teaching at all stations. After all teaching is completed, the communication line is disconnected, the integrated measurement device 3 is removed from the station, and the robot can then perform normal wafer 9 handling operations according to the taught parameters.

[0116] Through the above embodiments, the Z-axis teaching time is approximately 30 seconds (including 3 repeated measurements), the X / Y teaching time is approximately 20 seconds (including iterative adjustments), and the complete teaching time for a single station is less than 1 minute. Compared to the traditional manual teaching method that requires 15-30 minutes per station, the efficiency is improved by more than 15 times. In terms of teaching accuracy, the positioning accuracy in the X / Y direction reaches ±5μm, and the Z-axis height accuracy reaches ±10μm, both meeting the requirements of semiconductor processes. At the same time, through the height coupling correction in the above example, the deviation of the CCD measurement system at the tens of micrometer level introduced by the approximately 37mm difference between the detection height and the calibration reference height is eliminated, as well as the 2~3μm pseudo-transfer error introduced by the 3mm difference between the X / Y teaching height and the detection height after placement is eliminated. Adhesion and relative horizontal interference are automatically eliminated through the knee point detection of the rate of change. Through secondary detection after placement and closed-loop iterative correction, the transfer piece error can be controlled within ±5μm.

[0117] In summary, the automatic teaching method and system for the vacuum robot 2 provided by this invention integrates a displacement sensor 31 and at least two linear CCD sensors 32 into a measuring device that can be precisely placed on the workstation to be taught, so that the center of the device coincides with the center of the workstation. The vacuum robot 2 only needs to carry the wafer 9 according to the conventional pick-and-place process to complete the insertion, descent, placement and removal actions, and can automatically complete the teaching of the X / Y / Z three-axis parameters of the workstation on a single device. This transforms the traditional manual teaching process that relies on feeler gauge measurement and visual alignment into a fully automated closed-loop calibration process.

[0118] Based on this concept, the teaching method of this invention uses rate of change knee point detection in the Z-axis direction to automatically determine the moment when wafer 9 contacts the workstation, and accordingly automatically determines the wafer insertion height and departure height, eliminating the need for manual measurement with feeler gauges and automatically eliminating false contact caused by adhesion and relative horizontality. In the X / Y direction, the occlusion amount of the linear CCD sensor 32 is used to directly obtain the offset of the wafer 9 center relative to the device center. This avoids the defects of insufficient alignment accuracy by the naked eye and the inability of tooling alignment to guarantee the alignment of wafer 9 with the workstation. Also, because the occlusion amount is a continuous analog quantity, there is no speed-related error caused by the response delay of the through-beam sensor. More importantly, the data queue synchronously established during the Z-axis teaching descent is used to fit the height correction coefficient and perform height coupling correction on the offset, eliminating the systematic deviation of tens of micrometers introduced by the difference between the wafer 9 detection height and the calibration reference height. Furthermore, the secondary detection after lowering is corrected again based on the current height, ensuring the comparability of detection results at different measurement heights.

[0119] Furthermore, this invention, through a closed-loop iteration of "placement-detection-correction," updates the workstation coordinates based on the relationship between the offsets before and after placement, eliminating placement and handover errors that cannot be detected and corrected by existing technologies. This forms a complete technical chain of "Z-axis height setting—XY correction—secondary verification—closed-loop iteration," with each link coupled and working synergistically to achieve fully automated three-axis teaching on a single device. The teaching accuracy in the X / Y directions reaches ±5μm, and the Z-axis height accuracy reaches ±10μm. The teaching time for a single workstation is reduced to less than 1 minute, improving efficiency by more than 15 times compared to traditional manual teaching. Moreover, it requires no manual parameter input, is simple to operate, and has strong versatility, showing promising prospects for industrial applications.

[0120] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The present invention is limited only by the claims and their full scope and equivalents. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the protection scope of the invention.

[0121] Those skilled in the art will understand that, besides implementing the system, apparatus, unit, and its modules provided by this invention in purely computer-readable program code, the same program can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system, apparatus, and its modules provided by this invention can be considered a hardware component, and the modules included therein for implementing various programs can also be considered structures within the hardware component; alternatively, modules for implementing various functions can be considered both software programs implementing the method and structures within the hardware component.

[0122] Furthermore, all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0123] Furthermore, various different implementations of the present invention can be combined arbitrarily, as long as they do not violate the spirit of the present invention, they should also be regarded as the content disclosed in the present invention.

Claims

1. An automatic teaching method for a vacuum robot workstation, based on an integrated measurement device, comprising the following steps: Place the integrated measuring device at the teaching station so that the centers of the two coincide; the integrated measuring device integrates a displacement sensor and a linear CCD sensor with at least two intersecting measuring lines that pass through the center of the device. The robot finger, carrying the wafer, is inserted into the integrated measuring device and lowered. The displacement sensor measures the distance in real time. Based on the distance change rate characteristics, the robot determines the contact position and records the robot's Z-axis coordinate at the moment of contact. The insertion height and exit height of the wafer are then determined accordingly. During the descent, the Z-axis position and the readings of each linear CCD sensor are recorded simultaneously, and a height correction coefficient is obtained by fitting the data. At the tape insertion height, height coupling correction is performed based on the readings of each linear CCD sensor, the height measured by the current displacement sensor, and the height correction coefficient to obtain the first offset of the wafer, and the robot's horizontal position is adjusted accordingly to meet the accuracy requirements. After placing the wafer, the readings of each linear CCD sensor are read again, and height coupling correction is performed again based on the current height to obtain the second offset of the wafer after placement. The target station coordinates are updated based on the relationship between the offsets before and after placement, and the robot is controlled to re-fetch wafers iteratively until the offsets before and after placement meet the accuracy requirements.

2. The method according to claim 1, wherein, The step of determining the contact station based on the distance change rate feature includes: Calculate the rate of change of displacement sensor readings in adjacent steps. When the rate of change stabilizes within a first preset range during the free fall phase, and the rate of change in subsequent consecutive steps is lower than a second preset threshold, it is determined that the wafer has contacted the workstation.

3. The method according to claim 2, wherein, The steps also include: If the rate of change briefly falls below the second preset threshold during the descent but then recovers, it is determined to be a false contact, and the robot finger continues to descend.

4. The method according to claim 1, wherein, The step of obtaining the height correction coefficient through fitting includes: During the descent of the vacuum robot, the Z-axis position and the readings of each of the linear CCD sensors are recorded simultaneously to establish a data queue; Based on the readings corresponding to each descent height in the data queue, the height correction coefficient of each linear CCD sensor caused by the unit height change is obtained by using a linear fitting method.

5. The method according to claim 4, wherein, The calculation steps for the first offset include: Obtain the current height of the wafer measured by the displacement sensor, and calculate the height difference between the current height and the calibration reference height; The first offset is obtained by performing height coupling correction on the original offset determined based on the readings of each of the linear CCD sensors, according to the height difference and the height correction coefficient.

6. The method according to any one of claims 4 or 5, wherein, The linear CCD sensor includes a first linear CCD sensor and a second linear CCD sensor, with measurement lines respectively arranged along the X and Y directions. The high coupling correction is performed according to the following formula: in, , These are the current readings of the first linear CCD sensor and the second linear CCD sensor, respectively, where K is the CCD pixel equivalent. and This is the factory calibration reference reading. The current height is measured in real time by the displacement sensor. This is the factory-calibrated reference height. and These are the height correction coefficients corresponding to the X and Y directions, respectively.

7. The method according to claim 1, wherein, The steps of updating the target workstation coordinates based on the relationship between the offsets before and after placement, and controlling the robot to re-collect the piece iteratively, include: Compare whether the difference between the first offset and the second offset is greater than a preset threshold; If the difference is greater than the current workstation coordinate offset, the difference is added to the current workstation coordinate offset to update the workstation coordinate. Using the updated workstation coordinates as the target position, the robot is controlled to pick up the wafer again, and the above steps of horizontal position adjustment and offset detection after placement are repeated until the difference is not greater than the preset threshold.

8. The method according to claim 1, wherein, The steps also include: Repeat the steps of lowering the robot finger with the wafer to the contact point and determining the wafer insertion and exit heights at least twice, and take the average value of the wafer insertion and exit heights obtained each time as the final Z-axis teaching parameter.

9. The method according to claim 1, wherein, When the number of linear CCD sensors is three or more, the wafer center position is fitted using the least squares method by using the readings of all the linear CCD sensors to obtain the offset of the wafer center relative to the center of the integrated measurement device.

10. An automated teaching system for a vacuum robot workstation, comprising: An integrated measurement device, configured to be placed at a teaching station, includes: a device body with a positioning structure at the bottom for engaging with a positioning component on the station to align the device center with the station center; a displacement sensor mounted at the top center of the device body, with a vertically downward measuring direction, for measuring the distance between the wafer surface above it and the displacement sensor; and at least two linear CCD sensors mounted inside the device body, wherein the measuring lines of each linear CCD sensor intersect each other and all pass through the center of the device. Vacuum robots are used to move wafers. The control unit is communicatively connected to both the vacuum robot and the integrated measuring device, and is configured to perform the method as described in any one of claims 1 to 9.

11. The system according to claim 10, wherein, The control unit includes: a communication module for receiving measurement data from the displacement sensor and each linear CCD sensor in real time; a parameter module for reading and writing workstation teaching parameters; a motion module for controlling the motion of the vacuum robot's end effector; and a teaching calculation module for performing contact determination, high coupling correction, and workstation coordinate correction calculations.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the automatic teaching method for the vacuum robot workstation as described in any one of claims 1 to 9.

Images