Automatic chip thermocompression bonding device and intelligent calibration method

Abstract

The invention discloses an automatic chip thermocompression bonding device and an intelligent calibration method, and belongs to the field of semiconductor packaging and manufacturing. The device comprises a main body frame, an alignment vision system, a side vision system, a bonding arm and a workbench, and micron-order alignment and eutectic bonding of a chip and a substrate can be realized. A bonding arm adopts a vacuum suction nozzle to suck a chip, a substrate is fixed on a workbench through vacuum adsorption, an alignment visual system realizes automatic precision calibration and position feedback of the device by using a calibration sheet, and micron-sized alignment and intelligent matching of chip thermocompression bonding process parameters are realized by collecting image feature recognition. The workbench is finely adjusted to align the marking points of the chip and the substrate, then the bonding arm carries out the thermocompression bonding process, and meanwhile, the side vision system monitors the precision and reliability of chip bonding. The chip alignment precision is high, the method is suitable for the scene that multiple chips are bonded at the same time, the problem that the equipment calibration efficiency is low in the thermocompression bonding process can be solved, and the process reliability can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor packaging, and in particular to an automatic chip thermal compression bonding device and an intelligent calibration method. Background Art

[0002] Driven by the global wave of digitalization, modern electronics manufacturing is accelerating its evolution toward miniaturization and high integration. Advanced packaging technology, a key component in achieving breakthroughs in chip performance, presents unprecedented technical challenges for thermocompression bonding processes as chip integration continues to increase. On the one hand, the demand for high-precision submicron alignment requires bonding equipment to possess micron-level positioning and repeatability. On the other hand, complex three-dimensional packaging structures and multi-chip stacking processes place even higher demands on the control of parameters such as temperature, pressure, and time during the bonding process. Traditional automated thermocompression bonding equipment faces bottlenecks in alignment accuracy, real-time calibration, and process parameter optimization, making it difficult to meet the dual efficiency and quality requirements of large-scale mass production. Overcoming the contradiction between accuracy and efficiency to achieve high-precision, high-efficiency, and high-reliability thermocompression bonding has become a critical technical challenge urgently needed in the electronics manufacturing industry.

[0003] As the core equipment for implementing advanced packaging processes, the chip hot-compression bonding device relies on a high-precision alignment system and uses visual recognition technology to accurately align the chip and substrate at the micron level. At the same time, through the coordinated movement of the workbench and the bonding arm, dynamic adjustment of the multi-dimensional spatial position is achieved to ensure that the chip and substrate precisely fit in three-dimensional space. During the bonding stage, the bonding arm integrates pressure, temperature and position control modules. By precisely adjusting the process parameters, it causes the bonding material to soften at the set temperature, achieving a reliable connection between the chip and the substrate, completing the key step of high-density interconnection. The performance of this device directly determines the accuracy, efficiency and reliability of the hot-compression chip bonding process, and is a key equipment in advanced chip packaging production lines. However, during automated operation, various errors will accumulate while the equipment is running, resulting in a significant decrease in the actual bonding accuracy in automatic control mode, becoming the main bottleneck restricting equipment performance improvements and packaging yield improvements.

[0004] In the US7341877B2 patent, a device calibration method is proposed, which introduces a third recognition system and cooperates with the traditional dual recognition system to achieve high-precision parameter correction through dual-system coarse calibration and third-system fine calibration. However, the complex structure of multiple visual systems not only increases the difficulty of equipment design and maintenance, but also its collaborative operation process is cumbersome. The visual units have inherent differences in installation errors, image acquisition delays and algorithm processing accuracy. Visual deviation coupling often occurs in actual operation, which reduces the overall calibration accuracy of the equipment and affects the stability of the bonding process. In the CN202110908211 patent, an efficient and accurate substrate automatic calibration system is proposed. The coordinate solution of the diagonal mark and the composite posture adjustment mechanism solve the efficiency and accuracy bottlenecks of manual calibration of traditional placement machines. However, when the chips are bonded one by one, the position of the second chip changes, and the chip position cannot be observed and adjusted in real time, which reduces the chip bonding efficiency. Summary of the Invention

[0005] The purpose of the present invention is to provide an automatic chip thermal compression bonding device and an intelligent calibration method, which can be used for automatic calibration of the accuracy of the chip bonding device and automatic calibration of the chip alignment position during chip thermal compression bonding, thereby achieving high-precision chip thermal compression bonding and improving the bonding accuracy and reliability of the chip bonding device.

[0006] To achieve the above objectives, the present invention provides an automatic chip thermal compression bonding device and an intelligent calibration method, comprising:

[0007] The main frame is used to provide force support for the device;

[0008] The alignment vision system is arranged on the main frame and includes a camera lens system, an image chip, an image processor, a lens drive module, a dichroic prism and a rotary motor. The alignment vision system uses a calibration piece to realize automatic calibration of the device accuracy, and realizes intelligent matching of chip types and high-precision alignment by collecting images of the chip and substrate.

[0009] The bonding arm, including a drive motor, a connecting arm, a heating module, a nozzle, and a nozzle rotation platform, is used to pick up the chip and automatically adjust its posture in the θ direction, and realizes the rotation and pressing movement with the help of the motor. The bonding arm applies the corresponding pressure and temperature during the hot pressing bonding process according to the process parameters provided by the alignment vision system;

[0010] The workbench includes a heating stage, a pressure sensor, an X-axis electric translation stage, a Y-axis electric translation stage, and a Z-axis electric translation stage, which is used to adsorb the substrate and provide preheating, pressure feedback, and automatic fine-tuning of the electric translation in the X, Y, and Z directions;

[0011] The side vision system, including a side-view camera lens system and a side-view light source, is used to monitor changes in chip position during the thermal compression bonding process.

[0012] Preferably, the camera lens system has an automatic lens zoom function. By working in coordination with the beam splitter prism and the image chip, it can observe the calibration sheet on the pick-up nozzle located at the pre-bonding position and the calibration sheet on the workbench.

[0013] After receiving the imaging beam, the image chip generates a digital image and transmits the digital image to the image processor. The Z-axis electric displacement stage moves the calibration sheet on the workbench in the Z direction, causing a change in the digital image of the image chip. When the image processor detects that the clarity of a certain frame of the image is low, it will send an instruction to the Z-axis electric displacement stage to achieve autofocus by moving in the Z direction, improving the image clarity. After meeting the clarity requirement, the lens of the camera lens system automatically zooms to obtain a higher-clarity image and performs the autofocus process again until the lens reaches the maximum magnification. The beam splitter prism adjusts its attitude through the rotation motor so that the calibration sheet on the pick-up nozzle also meets the clarity requirement, indicating that the device has achieved automatic calibration.

[0014] Preferably, the alignment vision system uses neural network technology to continuously accumulate data sets, collects various common chip types used by users, and constructs a temperature and pressure recipe curve model library. By identifying key feature points in aspects such as the shape, size, pin arrangement, marked text, and color of different chip types, the chips are classified.

[0015] When the alignment vision system observes the chip, it will perform key feature recognition on the chip, extract the marker features in the target chip and compare them with the model library, automatically match the relevant temperature and pressure curve recipes according to the identified chip type, and feedback the recipes to the bonding arm to meet the temperature and pressure requirements during the thermocompression bonding process.

[0016] Preferably, the bonding arm is driven by the drive motor. After the alignment vision system completes the chip and substrate alignment operation, the lens drive module moves the camera lens system away, and at the same time, the side vision system is turned on to observe the thermocompression bonding process. Then the drive motor drives the bonding arm to rotate and press down from the pre-bonding position until it contacts the substrate to reach the bonding position. When the pressure sensor feeds back the pressure, the heating module starts to heat up to perform the thermocompression bonding process.

[0017] Preferably, the side vision system observes and monitors the chip position in real time, and constructs a closed-loop feedback control system with the position information of the drive motor rotation; during the thermocompression bonding process, when the chip copper pillar melts and collapses downward, the position trend is predicted and timely feedback is given. Based on the feedback information, the bonding arm precisely controls the chip position and automatically performs a displacement compensation operation to prevent the bonding copper pillar from being crushed due to excessive pressure.

[0018] Preferably, the Z-direction electric displacement stage is designed with a wedge mechanism, which can convert the X-direction movement into the Z-direction movement, so as to drive the substrate to move precisely in the Z direction.

[0019] Preferably, after the alignment vision system calculates the position deviation, the workbench can automatically adjust the position information in the X-Y direction through the X-direction electric displacement stage and the Y-direction electric displacement stage; the nozzle rotating stage on the bonding arm is driven by a motor and automatically adjusts the position information in the θ direction, finally realizing the calibration of the chip mounting accuracy.

[0020] Preferably, to achieve the above object, an embodiment of the present invention further provides a method for intelligent calibration of a chip and a substrate, including the following steps:

[0021] S1. The nozzle of the bonding arm sucks the first chip, then returns to the pre-bonding position, and the alignment vision system identifies the position of the first chip and collects and records the position information of the current chip.

[0022] S2. The alignment vision system then identifies the position of the substrate, also collects and records the position information of the current substrate, then matches the recorded position of the first chip with the position of the substrate, and calculates the position deviation (△X1, △Y1, △θ1) between the two.

[0023] S3. The X-direction electric displacement stage and the Y-direction electric displacement stage of the workbench and the nozzle rotating stage on the bonding arm are adjusted through the fed-back position deviation to complete the deviation compensation. The alignment vision system drives the module to move to a position to avoid the bonding arm through the lens, and then the bonding arm moves to the bonding position to start the thermocompression bonding process.

[0024] S4. After the thermocompression bonding of the first chip is completed, record the position of the first chip at this time. The bonding arm sucks the second chip, and the alignment vision system identifies and reads the position information of the second chip, calculates it with the position information of the first chip, and obtains the position deviation (△X 2-1 , △Y 2-1 , △θ 2-1 ) between the two. At the same time, obtain the second bonding position and the first bonding position (△X 2-2 , △Y 2-2 , △θ 2-2 ) on the substrate, add up the above deviations to calculate the total error (△X 2-1 + △X 2-2 , △Y 2-1 + △Y 2-2 , △θ 2-1 + △θ 2-2 ) that needs to be compensated, and then repeat step S3 to start the thermocompression bonding operation of the second chip.

[0025] Preferably, after the first chip is aligned with the substrate, the position of the first chip is set as the initial template to calculate the position deviation of the subsequent chips for matching. The subsequent deviation is as follows:

[0026] ΔX N = ΔX (N-1) - ΔX (N-1)-2 + ΔX N-1 + ΔX N-2 ;

[0027] ΔY N = ΔY (N-1) - ΔY (N-1)-2 + ΔY N-1 + ΔY N-2 ;

[0028] Δθ N = Δθ (N-1) - Δθ (N-1)-2 + Δθ N-1 + Δθ N-2 ;

[0029] Wherein, ΔX n is the abscissa of the deviation of the nth chip, △X n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △X n-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate;

[0030] ΔY n is the abscissa of the deviation of the nth chip, △Y n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △Y n-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate;

[0031] Δθ n is the abscissa of the deviation of the nth chip, △θ n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △θ n-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate.

[0032] An automatic chip thermocompression bonding device and an intelligent calibration method of the present invention have the following beneficial effects:

[0033] (1) The present invention effectively solves the problem that the equipment accuracy is prone to deviation before chip bonding and relies on manual calibration. By adopting the intelligent recognition image clarity calibration method of the image chip and the image sensor, the equipment calibration efficiency is greatly improved, the cumbersome and inefficient manual operation is avoided, and it is ensured that the equipment can quickly and accurately enter the bonding working state.

[0034] (2) The present invention solves the problem that parameters need to be manually adjusted when different chip types are replaced. By automatically identifying the chip type and automatically matching the relevant temperature and pressure curves based on the built-in model library, the operation process is simplified, human intervention is reduced, and the coherence and accuracy of the production process are improved.

[0035] (3) The present invention solves the problem that it is difficult to accurately predict the situation of the chip copper pillar collapsing downward after melting during the bonding process. By introducing a side-view lens system to detect the change of the copper pillar state, the bonding quality problem caused by the copper pillar collapse is prevented in advance, and the stability and reliability of the bonding process are significantly improved.

[0036] (4) In the present invention, a chip calibration method is provided, which solves the problem that the bonding accuracy is affected by the position deviation when multiple chips are bonded. By automatically calculating and compensating the position deviation through the image matching method, it is ensured that each chip can meet the high-precision alignment requirements during the bonding process, and the overall quality and consistency of chip bonding are effectively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a schematic structural diagram of an automatic chip thermocompression bonding device of an automatic chip thermocompression bonding device and an intelligent calibration method of the present invention;

[0038] Figure 2 is a schematic side view structure diagram of an automatic chip thermocompression bonding device of an automatic chip thermocompression bonding device and an intelligent calibration method of the present invention;

[0039] Figure 3 is a block diagram of a chip category image recognition system of an intelligent calibration method of the present invention;

[0040] Figure 4 is a schematic diagram of the bonding process of the bonding arm of the thermocompression bonding device of the present invention;

[0041] Figure 5 is a flow chart of the precision intelligent calibration method of the thermocompression bonding device of the present invention;

[0042] Figure 6 is a schematic structural diagram of the Z-direction electric displacement table of the thermocompression bonding device of the present invention;

[0043] Figure 7 is a schematic diagram of the working process of the thermocompression bonding device of the present invention performing calibration actions;

[0044] Figure 8 Schematic diagram of observing the chip with rear annular light adopted by the thermocompression bonding device of the present invention;

[0045] Figure 9 Schematic diagram of observing the substrate with lower annular light adopted by the thermocompression bonding device of the present invention;

[0046] Figure 10 Schematic diagram of identifying the position deviation between the chip and the substrate by the thermocompression bonding device of the present invention;

[0047] Figure 11 Schematic diagram of the position deviation between the first chip and the second chip by the thermocompression bonding device of the present invention;

[0048] Figure 12 Schematic diagram of observing the chip and the substrate with coaxial light adopted by the thermocompression bonding device of the present invention;

[0049] Figure 13 Schematic diagram of real-time observation and alignment by the thermocompression bonding device of the present invention;

[0050] Description of reference numerals

[0051] 1. Alignment vision system; 2. Main body frame; 3. Side vision system; 4. Bonding arm; 5. Workbench; 6. Substrate; 7. Chip; 101. Camera lens system; 102. Lens driving module; 103. Beam splitter prism; 104. Rotating motor; 1011. Coaxial light; 1012. Rear annular light; 1013. Lower annular light; 301. Side view camera lens system; 401. Driving motor; 402. Connecting arm; 403. Heating module; 404. Suction nozzle; 405. Suction nozzle rotating table; 501. Heating table; 502. Pressure sensor; 504. Z-direction electric displacement table; 505. Y-direction electric displacement table; 506. X-direction electric displacement table; 601. First chip; 602. Second chip; 5041. Motor; 5042. Pair of bevel gears; 5043. Wedge block; 5044. Rising platform; △X. Lateral deviation; △Y. Longitudinal deviation; △θ. Angular deviation. Detailed implementation manners

[0052] The technical solutions of the present invention will be further described below with reference to the drawings and embodiments.

[0053] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the ordinary meanings as understood by those of ordinary skill in the art to which the present invention pertains. The "first", "second" and similar terms used in the present invention do not denote any order, quantity or importance, but are only used to distinguish different components. The terms such as "upper", "lower", "left" and "right" are only used to indicate relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationships may also change accordingly.

[0054] An automatic chip thermocompression bonding device provided by the present invention, as Figure 1 and Figure 2 shown, includes a main body frame 2, which provides force support for the device;

[0055] An alignment vision system 1, which is arranged on the main body frame 2 and includes a camera lens system 101, a lens driving module 102, a beam splitter prism 103 and a rotating motor 104. The camera lens system 101 and the beam splitter prism 103 are used to observe the positions of the chip 6 and the substrate 7. The lens driving module 102 is used to move the camera lens system 101 to avoid movement interference with the bonding arm 4. The rotating motor 104 is used to adjust the beam splitter prism 103 to reach the correct position where alignment can be achieved;

[0056] A bonding arm 4, which includes a driving motor 401, a connecting arm 402, a heating module 403, a nozzle 404 and a nozzle rotating table 405, is used to suck the chip 6 and automatically adjust the attitude in the θ direction, and makes a rotational movement by means of the driving motor 401. The bonding arm 4 applies corresponding pressure and temperature during the thermocompression bonding process according to the process parameters provided by the alignment vision system 1;

[0057] A workbench 5, which includes a heating table 501, a pressure sensor 502, an X-direction electric displacement table 506, a Y-direction electric displacement table 505 and a Z-direction electric displacement table 504, is used to adsorb the substrate 7, and at the same time provides preheating, pressure feedback and automatic fine adjustment of electric displacement in the X - Y - Z directions;

[0058] A side vision system 3, which includes a side-view camera lens system and a side-view light source, is used to monitor the change of the chip position during the thermocompression bonding process.

[0059] Reference Figure 3 , a chip category image recognition method for intelligent calibration, the method includes:

[0060] 1) The alignment vision system continuously accumulates a data set through neural network technology, collects various chip types commonly used by users and constructs a temperature and pressure recipe curve model library, and classifies the chips by identifying key feature points in aspects such as the shape, size, pin arrangement, marked text, color, etc. of different chip types;

[0061] 2) When the alignment vision system observes the chip, it will identify the key features of the chip, extract the mark features in the target chip and compare them with the model library, and automatically match the relevant temperature-pressure curve formula according to the identified chip type.

[0062] 3) The bonding arm receives the temperature-pressure curve formula to meet the temperature and pressure requirements during the thermocompression bonding process.

[0063] Reference Figure 4 and Figure 5 FIG. is a flowchart of a method for intelligent calibration of device accuracy provided by an embodiment of the present invention. The camera lens system 101 has an automatic zoom function, and the focal length of the lens is a + b. By working together with the beam splitter prism 103 and the image chip, it can observe the calibration sheet on the nozzle 404 at the pre-bonding position and the calibration sheet on the workbench.

[0064] After receiving the imaging beam, the image chip generates a digital image and transmits the digital image to the image processor. The Z-axis electric displacement stage 504 moves the calibration sheet on the workbench 5 in the Z direction, causing the digital image of the image chip to change. When the image processor detects that the clarity of a certain frame of image is low, it will send an instruction to the Z-axis electric displacement stage to achieve autofocus by moving in the Z direction and improve the image clarity.

[0065] The image processor detects whether the clarity meets the requirements. After meeting the requirements, the lens of the camera lens system 101 automatically zooms to obtain a higher clarity image, and repeats the above focusing process until the lens reaches the maximum magnification. At this time, the distance between the lens and the calibration sheet on the workbench 5 can be obtained as a + b.

[0066] When the lens reaches the maximum magnification and the image processor detects that the clarity is satisfied, the beam splitter prism 103 adjusts its attitude through the rotation motor 104 so that the calibration sheet on the nozzle also meets the clarity requirements. The distance between the lens and the calibration sheet on the nozzle 404 is also a + b, and the distance from the two calibration sheets to the beam splitter prism forms a square with a side length of b, indicating that the device has achieved automatic calibration.

[0067] Reference Figure 6 FIG. is a schematic structural diagram of the Z-axis electric displacement stage of the thermocompression bonding device of the present invention. In the Z-axis electric displacement stage of the thermocompression bonding device of the present invention, the motor 5041 drives a pair of bevel gears 5042 to convert the rotational motion around the Y axis into rotational motion around the X axis, and drives the wedge block 5043 to linearly move along the X axis through the screw guide rail. Based on the inclined plane cooperation between the wedge block 5043 and the rising platform 5044, the rising platform 5044 is finally driven to vertically lift along the Z axis.

[0068] Embodiment 1

[0069] Reference Figure 7 , which is a schematic diagram of the work process for performing calibration actions of the thermocompression bonding device of the present invention.

[0070] Step 701: The bonding arm 4 sucks the calibration wafer and returns to the pre-bonding position. The workbench 5 places the calibration wafer, and the alignment vision system 1 moves to the pre-bonding position. The position of the beam splitter prism 103 is calibrated through the rotation motor 104, and then the Z-direction distance of the workbench 5 is calibrated through the Z-direction electric displacement stage 504. Finally, the accuracy calibration of the device is completed.

[0071] Step 702: The bonding arm 4 sucks the chip 6, and the workbench 5 places the substrate 7. The alignment vision system 1 moves to the pre-bonding position, identifies the type of the chip 6 to set the temperature and pressure curve and records the chip position information, then identifies the position of the substrate 7 and records the substrate bonding position information, and then the alignment vision system 1 moves to the protection position.

[0072] Step 703: The X-direction electric displacement stage and Y-direction electric displacement stage of the workbench 5 move according to the feedback position deviation obtained by calculation to compensate for the deviation. Then the bonding arm 4 rotates to contact the substrate 7. When the chip 6 contacts the substrate 7 through the feedback pressure value of the pressure sensor 502, the temperature and pressure curve is automatically selected according to the chip type for bonding. The side vision system 3 observes the bonding process and monitors the displacement information of the bonding arm.

[0073] Step 704: The subsequent chips calculate the deviation based on the position of the first chip and calculate the error of the substrate, accumulate the error, and then repeat the previous step for bonding.

[0074] Embodiment 2

[0075] Reference Figure 8-11 , an embodiment of the present invention provides a method for intelligent calibration of a chip and a substrate:

[0076] S1: The nozzle of the bonding arm sucks the first chip and returns to the pre-bonding position. The rear annular light is turned on. At this time, the alignment vision system can only identify the position of the first chip and collect and record the current position information, and then the rear annular light is turned off;

[0077] S2: The lower annular light is turned on. At this time, the alignment vision system can only identify the position of the substrate and collect and record the current position information of the substrate. Then, the recorded position of the first chip is matched with the position of the substrate, and the position deviation (△X1, △Y1, △θ1) between the two is calculated;

[0078] S3: The X-displacement stage and Y-displacement stage of the workbench and the nozzle rotation stage on the bonding arm are adjusted according to the feedback position deviation. After compensating for the deviation, the alignment vision system reaches the protection position through the lens drive module, and the bonding arm reaches the bonding position to start the thermocompression bonding process;

[0079] S4. After completing the thermocompression bonding of the first chip, record the position 601 of the first chip at this time. The bonding arm picks up the second chip 602, the alignment vision system identifies and reads the position information 602 of the second chip, and calculates it with the position information 601 of the first chip to obtain the position deviation between the two (△X 2-1 , △Y 2-1 , △θ 2-1 ). The deviation (△X 2-2 , △Y 2-2 , △θ 2-2 ) between the bonding position 602 of the second chip and the bonding position 601 of the first chip on the substrate. Calculate and accumulate the total deviation to obtain the error to be compensated (△X 2-1 + △X 2-2 , △Y 2-1 + △Y 2-2 , △θ 2-1 + △θ 2-2 ), and repeat step S3 for thermocompression bonding;

[0080] When the position of the first chip is calibrated with the substrate, the first position is set as the initial template to calculate and match the position deviation of the subsequent chips. The subsequent deviation is:

[0081] ΔX N = ΔX (N-1) - ΔX (N-1)-2 + ΔX N-1 + ΔX N-2 ;

[0082] ΔY N = ΔY (N-1) - ΔY (N-1)-2 + ΔY N-1 + ΔY N-2 ;

[0083] Δθ N = Δθ (N-1) - Δθ (N-1)-2 + Δθ N-1 + Δθ N-2 ;

[0084] Among them, ΔX n is the abscissa of the deviation of the nth chip, △X n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △X n-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate;

[0085] ΔY n is the abscissa of the deviation of the nth chip, △Y n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △Yn-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate;

[0086] Δθ n is the abscissa of the deviation of the nth chip, △θ n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △θ n-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate.

[0087] Embodiment III

[0088] Reference Figure 12-13 , an embodiment of the present invention provides a method for intelligent calibration of a chip and a substrate:

[0089] A common-path imaging system is constructed by aligning the beam splitter prism 103 and the coaxial light source 1011 of the alignment vision system 1 to realize real-time image synchronization acquisition of the chip 6 adsorbed by the nozzle 4 and the substrate 7 on the workbench 5. Aiming at the possible image overlap problem in coaxial imaging, the system adopts a feature recognition algorithm based on deep learning to automatically extract the unique marker features such as the edge contour of the chip 6 and the pad array of the substrate 7, and completes the precise segmentation of the dual-target image through the template matching algorithm to identify the field-of-view information of the chip and the substrate.

[0090] The nozzle rotating table 405 of the bonding arm 4 dynamically adjusts the posture of the chip 6 according to Δθ to compensate for the θ-direction angle deviation between the two; the X-axis electric displacement table 505 and the Y-axis electric displacement table 506 carried by the workbench 5 drive the substrate 7 to perform linear displacement to accurately match the X and Y axis coordinates of the chip 6. Through the closed-loop linkage of visual feedback and motion control, the micron-level alignment accuracy between the chip 6 and the substrate 7 is finally achieved, meeting the stringent requirements of high-density interconnection in advanced packaging processes.

[0091] Finally, it should be noted that: the above embodiments are only exemplary descriptions of the technical solutions of the present invention, rather than limitations on the protection scope. Although the present invention has been described in detail in combination with the preferred embodiments, those skilled in the art should understand that: without departing from the core concept of the present invention, any form of modification, equivalent replacement or improvement can be made to the existing technical solutions, such as the combination, division or reuse of features. As long as the above adjustments do not deviate from the technical principle and innovation essence of the present invention, they should be regarded as falling within the protection scope defined by the claims of the present invention.

Claims

1. An automatic chip thermocompression bonding device and an intelligent calibration method, characterized in that, Comprising: A main body frame for providing force support for the device; An alignment vision system disposed on the main body frame, including a camera lens system, an image chip, an image processor, a lens driving module, a beam splitter prism, and a rotating motor. The alignment vision system uses a calibration sheet to automatically calibrate the device accuracy, and realizes intelligent matching of chip types and high-precision alignment by collecting images of the chip and the substrate. A bonding arm, including a driving motor, a connecting arm, a heating module, a nozzle, and a nozzle rotating table, for sucking the chip and automatically adjusting the posture in the θ direction, and realizing the rotation and pressing-down movement by means of the motor. The bonding arm applies corresponding pressure and temperature during the thermocompression bonding process according to the process parameters provided by the alignment vision system. A workbench, including a heating table, a pressure sensor, an X-direction electric displacement table, a Y-direction electric displacement table, and a Z-direction electric displacement table, for adsorbing the substrate, and simultaneously providing preheating, pressure feedback, and automatic fine adjustment of electric displacement in the X-Y-Z directions. A side vision system, including a side-view camera lens system and a side-view light source, for monitoring the change of the chip position during the thermocompression bonding process.

2. The automatic chip thermocompression bonding device and intelligent calibration method according to claim 1, characterized in that: The camera lens system has a function of automatic lens zooming. By working in cooperation with the beam splitter prism and the image chip, it can observe the calibration sheet on the nozzle located at the pre-bonding position and the calibration sheet on the workbench. After receiving the imaging beam, the image chip generates a digital image and transmits the digital image to the image processor. The Z-direction electric displacement table moves the calibration sheet on the workbench in the Z direction, causing the digital image of the image chip to change. When the image processor detects that the clarity of a certain frame of image is low, it will send an instruction to the Z-direction electric displacement table to achieve automatic focusing by moving in the Z direction, improve the image clarity. After meeting the clarity requirement, the lens of the camera lens system automatically zooms to obtain a higher-clarity image and performs the automatic focusing process again until the lens reaches the maximum magnification. The beam splitter prism adjusts its posture through the rotating motor so that the calibration sheet on the nozzle also meets the clarity requirement, indicating that the device has achieved automatic calibration.

3. An automatic chip thermocompression bonding device and intelligent calibration method according to claim 1, characterized in that: The alignment vision system uses neural network technology to continuously accumulate data sets, collects various chip types commonly used by users, and constructs a temperature and pressure recipe curve model library. By identifying key feature points in aspects such as the shape, size, pin arrangement, marked text, and color of different chip types, the chips are classified. When the alignment vision system observes the chip, it will perform key feature recognition on the chip, extract the marker features in the target chip and compare them with the model library, and automatically match the relevant temperature and pressure curve recipes according to the identified chip type, and feedback the recipes to the bonding arm to meet the temperature and pressure requirements during the thermocompression bonding process.

4. An automatic chip thermocompression bonding device and intelligent calibration method according to claim 1, characterized in that: The bonding arm is driven by the driving motor. After the alignment vision system completes the chip and substrate alignment operation, the lens driving module moves the camera lens system away, and at the same time, the side vision system is turned on to observe the thermocompression bonding process. Then, the driving motor drives the bonding arm to rotate and press down from the pre-bonding position until it contacts the substrate to reach the bonding position. When the pressure sensor feeds back the pressure, the heating module starts to heat up to perform the thermocompression bonding process.

5. An automatic chip thermocompression bonding device and intelligent calibration method according to claim 1, characterized in that: The side vision system observes and monitors the chip position in real time, constructs a closed-loop feedback control system with the obtained chip position information and the position information of the driving motor rotation; during the thermocompression bonding process, when the chip copper pillar melts and collapses downward, position trend prediction is performed and timely feedback is given. Based on the feedback information, the bonding arm precisely controls the chip position and automatically performs displacement compensation operations to prevent the bonding copper pillar from being crushed due to excessive pressure.

6. An automatic chip thermocompression bonding device and intelligent calibration method according to claim 1, characterized in that: The Z-axis electric displacement stage is designed with a wedge mechanism, which can convert the X-direction movement into Z-direction movement, so as to drive the substrate to move precisely in the Z direction.

7. An automatic chip thermocompression bonding device and intelligent calibration method according to claim 1, characterized in that: After the alignment vision system calculates the position deviation, the workbench can automatically adjust the position information in the X-Y direction through the X-axis electric displacement stage and the Y-axis electric displacement stage; the nozzle rotating table on the bonding arm is driven by a motor to automatically adjust the position information in the θ direction, and finally the calibration of the chip mounting accuracy is achieved.

8. An automatic chip thermocompression bonding device and intelligent calibration method according to claim 7, provides a chip mounting calibration method, including the following steps: S1. The nozzle of the bonding arm sucks the first chip, then returns to the pre-bonding position. The alignment vision system identifies the position of the first chip, and collects and records the position information of the current chip. S2. The alignment vision system then identifies the substrate position, also collects and records the current substrate position information, then matches the recorded first chip position with the substrate position, and calculates the position deviation (△X1, △Y1, △θ1) between the two. S3. The X-axis electric displacement stage and the Y-axis electric displacement stage of the workbench and the nozzle rotating table on the bonding arm are adjusted through the fed-back position deviation to complete the deviation compensation. The alignment vision system moves to a position to avoid the bonding arm through the lens driving module, and then the bonding arm moves to the bonding position to start the thermocompression bonding process. S4. After completing the thermocompression bonding of the first chip, record the position of the first chip at this time. The bonding arm picks up the second chip, and the alignment vision system identifies and reads the position information of the second chip, and calculates it with the position information of the first chip to obtain the position deviation (△X 2-1 , △Y 2-1 , △θ 2-1 ). At the same time, obtain the second bonding position and the first bonding position (△X 2-2 , △Y 2-2 , △θ 2-2 ) on the substrate, and accumulate and calculate the above deviations to obtain the total error to be compensated (△X 2-1 + △X 2-2 , △Y 2-1 + △Y 2-2 , △θ 2-1 + △θ 2-2 ). Subsequently, repeat step S3 to start the thermocompression bonding operation of the second chip.

9. An automatic chip thermocompression bonding device and intelligent calibration method according to claim 7, characterized in that: When the position of the first chip and the substrate is calibrated, the first chip position is set as the initial template to calculate and match the position deviation of the subsequent chips. The subsequent deviation is: △X N = △X (N-1) - △X (N-1)-2 + △X N-1 + △X N-2 ; △Y N = △Y (N-1) - △Y (N-1)-2 + △Y N-1 + △Y N-2 ; △θ N = △θ (N-1) - △θ (N-1)-2 + △θ N-1 + △θ N-2 ; where, ΔX n is the abscissa of the deviation of the nth chip, △X n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △X n-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate; ΔY n is the abscissa of the deviation of the nth chip, △Y n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △Y n-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate; Δθ n is the abscissa of the deviation of the nth chip, △θ n-1 is the abscissa of the deviation between the position of the nth chip and the position of the first chip, △θ n-2 is the abscissa of the deviation between the bonding position of the nth chip and the bonding position of the first chip on the substrate.

Images