Bonding end point judgment method and system based on gas deflation rate monitoring

Abstract

The invention discloses a bonding end point judgment method and system based on gas deflation rate monitoring, relates to the technical field of microelectronic manufacturing, semiconductor packaging and MEMS manufacturing, and solves the technical problems that end point judgment mainly depends on indirect methods such as fixed time, specific temperature or pressure / displacement parameters and the like, and judgment is inaccurate. According to the method, a traditional method depending on indirect parameters such as fixed time and temperature is abandoned, end point judgment is carried out by monitoring the intrinsic parameter, namely the deflation rate of the target gas released by the bonding interface in real time, which directly reflects physical and chemical changes of the interface, and the real process of interface fusion and sealing can be sensitively captured; therefore, the strength, the sealing performance and the product yield of the bonding interface are improved, the excellent batch-to-batch consistency is ensured, the composite judgment logic combining the dynamic threshold value condition and the trend stability condition is adopted, intelligent and self-adaptive judgment of the process end point is achieved, and the available process window is widened.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of microelectronic manufacturing, semiconductor packaging and MEMS manufacturing, in particular to a bonding endpoint judgment method and system based on gas outgassing rate monitoring. BACKGROUND

[0002] Wafer bonding or chip bonding technology is a key process in modern microelectronic manufacturing and micro-electro-mechanical system (MEMS) field. Its core is to form a firm, sealed and permanent combination on atomic or molecular scale through specific physical or chemical action between two surface treated substrates. This technology is widely used in three-dimensional integration, device packaging, sensor manufacturing and formation of various functional structures, and is an indispensable process basis for realizing device miniaturization, multifunctionalization and high performance. According to the different mechanisms, the main bonding technologies mainly include anodic bonding, eutectic bonding, hot-press bonding and direct fusion bonding.

[0003] In the existing bonding process, the endpoint judgment mainly relies on indirect methods such as fixed time, specific temperature or pressure / displacement parameters. However, these methods have inherent limitations: first, the simple time control method cannot adapt to the slight fluctuations of different batches of materials and equipment states, and is easy to cause under-bonding or over-bonding; second, the temperature control method can monitor the process environment temperature, but cannot truly reflect the physical and chemical changes at the bonding interface, and the temperature is only a necessary but insufficient condition; in addition, the pressure or displacement control method is not sensitive to the key phenomena at the interface. In summary, the existing technologies are indirect measurement, which cannot directly, real-time and accurately represent the real reaction state of the bonding interface, thereby causing problems such as narrow process window, poor product consistency and unstable yield.

[0004] In the bonding process, the water, solvent, gas and other substances adsorbed in the bonding layer or on the surface will be desorbed or chemically reacted due to the action of the process conditions, and released into the bonding chamber in the form of gas. The strength of this outgassing process is closely related to the fusion and sealing process of the bonding interface: as the bonding interface gradually forms and seals, the channel for gas release is reduced, and the outgassing rate will show a certain decay law until it reaches a stable low level. The change curve of the outgassing rate directly and sensitively reflects the completion degree of the bonding process. Based on this innovative understanding, the present application aims to provide a method and system for accurately representing the endpoint of the bonding process by directly and real-time monitoring the gas outgassing rate, and there is no published technology using this direct indicator for endpoint judgment. SUMMARY

[0005] In view of the deficiencies of the prior art, the application provides a bonding end point judgment method and system based on gas release rate monitoring, which solves the problem of inaccurate judgment of the end point mainly depending on indirect methods such as fixed time, specific temperature or pressure / displacement parameters.

[0006] To achieve the above object, the application is implemented by the following technical scheme: a bonding end point judgment method based on gas release rate monitoring, which specifically comprises the following steps: S1, placing a first substrate and a second substrate to be bonded in a sample chamber, starting a vacuum molecular pump group to vacuumize the sample chamber and apply a process condition; S2, after the chamber reaches a high vacuum stable state, being communicated with a test chamber, and monitoring a target gas concentration signal released by a bonding layer by using a residual gas analyzer; S3, receiving the gas concentration signal by a data processing unit, and calculating a gas release rate per unit time; S4, continuously tracking a change curve of the release rate, and when it is monitored that the release rate decreases from a peak value and reaches a stable state, it is determined that the bonding process reaches an end point.

[0007] As a further scheme of the application, the condition of the high vacuum stable state is: continuous acquisition of pressure data P(t), then setting a target high vacuum range P target , and simultaneously introducing a bistable judgment condition, when the pressure P(t)≤1×10 -3 , and in the past time window, the pressure fluctuation amplitude P max -P min ≤k×P(t), wherein k is a constant.

[0008] As a further scheme of the application, the process of being communicated with the test chamber is: the first stage valve is opened to 10%-20% opening, maintained for 30-60 seconds, if the test chamber pressure suddenly rises and exceeds the limit, the valve is automatically closed and an alarm is given, otherwise the test chamber pressure rising rate is continuously monitored, if the pressure rising rate<0.5Pa / s, the valve is completely opened, after the two chambers are completely communicated, the test chamber pressure is continuously monitored until the constraint condition is met, then the RGA enters the target gas monitoring mode, the constraint condition includes: the pressure rising rate<0.015Pa / min, and the RGA background signal fluctuation<5%.

[0009] As a further scheme of the application, the way of monitoring the target gas concentration signal released by the bonding layer by using the residual gas analyzer is: Start RGA and select the ion monitoring working mode to obtain the mass number corresponding to the preset target gas. Map each gas to one or more characteristic mass-to-charge ratios and output the target mass number set M. target Simultaneously, background spectra are acquired using a full-spectrum scanning mode with a step size of 0.5 amu and a sampling time ≥ 100 ms / point. N consecutive acquisitions are performed, and the average value is saved as the background reference vector B0 (m / z). Then, the target mass number set M is obtained. target The system performs a scan, acquires the raw signal in real time, and performs dynamic background subtraction.

[0010] As a further aspect of the present invention, the method of real-time acquisition of the original signal and dynamic background subtraction is as follows: At different chamber temperatures T i The full-spectrum background signal under high vacuum steady state was collected and a background library was constructed. B i (m / z) represents the background spectrum under the i-th operating condition. Then, the chamber temperature T is obtained through a sensor. t And calculate the change in chamber temperature |T t -T t-1 | and simultaneously compare it with a temperature threshold; if the change value |T t -T t-1 If the temperature exceeds the threshold, dynamic background update is initiated to find two adjacent operating conditions T. low <T t <T high And calculate the temperature normalization factor. Generate dynamic background spectrum Then, for each mass-to-charge ratio m / z and time t, the net ion current is calculated. Output net signal, where I raw (m / z, t) represents the full-spectrum background signal.

[0011] As a further aspect of the present invention, the gas release rate is calculated as follows: Based on the obtained net signal, preprocessing and noise reduction are performed. Then, based on the obtained preprocessed concentration-time series signal, a dynamic mass balance equation is established, including the test chamber volume, temperature, and pumping speed parameters of the pumping system. This equation is then applied according to the formula... Calculate the venting rate Q, where V is the effective volume of the chamber, C is the gas concentration, and S is the effective pumping speed of the pumping system under the corresponding gas composition.

[0012] As a further aspect of the present invention, processing the concentration signal includes denoising the original signal using a moving average filter with an adaptive window length or a Savitzky-Golay convolutional smoothing algorithm.

[0013] As a further aspect of the present invention, the method for determining whether the bonding process has reached its end point is as follows: Current venting rate Q t ≤a×Q max , where a is a preset ratio; In the most recent consecutive time windows Within the range, the rate of change of the venting rate between any two sampling points satisfies: Where b is the stability critical value, t i and t j ∈[t- ,t].

[0014] As a further aspect of the present invention, the sample chamber is a chamber that can be independently temperature-controlled and pressurized, and can apply temperature and pressure conditions to the sample according to a preset process curve. The target gas includes at least one of water vapor, hydrogen, oxygen, nitrogen oxides or organic solvent vapor.

[0015] A bonding endpoint determination system based on gas outgassing rate monitoring, comprising: Test chamber, sample chamber, second gate valve connecting the two chambers, residual gas analyzer, full-range vacuum gauge, turbomolecular pump, backing mechanical pump, vacuum molecular pump assembly and data processing unit; The residual gas analyzer is installed in the test chamber to monitor the concentration of the target gas; the data processing unit is used to receive the concentration signal, calculate the venting rate, and perform endpoint determination.

[0016] This invention provides a method and system for determining bonding endpoints based on gas outgassing rate monitoring. Compared with existing technologies, it has the following advantages: This invention abandons traditional methods that rely on indirect parameters such as fixed time and temperature. Instead, it uses an intrinsic parameter—the rate of release of the target gas from the bonding interface itself—in real time to determine the endpoint. This parameter directly reflects the physicochemical changes at the interface. This method can sensitively capture the actual process of interface fusion and sealing, effectively avoiding underbonding or overbonding caused by process fluctuations. This improves the strength, sealing performance, and product yield of the bonding interface, ensuring excellent batch-to-batch consistency.

[0017] This invention employs a composite judgment logic combining dynamic threshold conditions and trend stability conditions, along with optional dynamic background subtraction and adaptive signal filtering. This allows the system to automatically adapt to process fluctuations under different materials, batches, and equipment conditions. This overcomes the poor adaptability of existing fixed-parameter methods, achieving intelligent and adaptive judgment of process endpoints and broadening the available process window.

[0018] This invention is not only used for endpoint determination, but its gas release spectrum monitoring throughout the process provides valuable online diagnostic data for process development and optimization. By analyzing the release timing and rate characteristics of different gases, the reaction mechanism of the bonding interface can be inferred, the pretreatment effect can be evaluated, or the source of pollution can be identified, thereby guiding the optimization of process parameters and material formulations and shortening the development cycle. Attached Figure Description

[0019] Figure 1 This is a flowchart of the bonding endpoint determination method of the present invention; Figure 2 This is a schematic diagram of the bonding endpoint monitoring system described in this invention; Figure 3 This is a three-dimensional structural model diagram of the bonding endpoint monitoring system described in this invention; Figure 4 This is a flowchart of the method of the present invention; Figure 5 This is a schematic diagram of the gas outgassing rate versus time during a typical bonding process.

[0020] In the figure: 1. Residual gas analyzer; 2. First full-range vacuum gauge; 3. Test chamber; 4. First gate valve; 5. Turbomolecular pump; 6. Foreboard mechanical pump; 7. Second gate valve; 8. Quick-release flange; 9. Sample chamber; 10. Angle valve; 11. Second full-range vacuum gauge; 12. Vacuum molecular pump assembly. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] First Embodiment Please see Figure 1 and Figure 5 This application provides a method for determining the bonding endpoint based on gas outgassing rate monitoring, which specifically includes the following steps: S1. Perform surface activation, cleaning and alignment on the first substrate and the second substrate to be bonded to form a bonding layer in a pre-bonded state. Place the assembled substrate in a sample chamber that can be independently temperature-controlled and pressurized. After sealing the chamber, turn on the vacuum molecular pump group to pre-evacuate the sample chamber. At the same time, the initial temperature and pressure conditions can be applied to the sample according to the preset process curve.

[0023] S2. Continuously collect pressure data P(t) using a full-range vacuum gauge installed in the sample chamber, and then set the target high vacuum range P. target ,and At the same time, a bistable condition is introduced: when the pressure P(t) ≤ 1×10 -3 And in the past Within the time window, the pressure fluctuation amplitude P max -P min When k ≤ k×P(t), where k is a constant, it is determined to be a high vacuum stable state; Based on the high vacuum stable state, the second gate valve is opened slowly in stages. In the first stage, the valve is opened to 10%-20% and maintained for 30-60 seconds. If the pressure in the test chamber suddenly rises beyond the limit, the valve is automatically closed and an alarm is triggered. Otherwise, the pressure rise rate in the test chamber is continuously monitored. If the pressure rise rate is <0.5Pa / s, the valve is fully opened. After the two chambers are fully connected, the pressure in the test chamber is continuously monitored until the constraint conditions are met. Only then is the RGA triggered to enter the target gas monitoring mode. The target gas includes at least one of water vapor, hydrogen, oxygen, nitrogen oxides, or organic solvent vapor. The constraint conditions include: pressure rise rate <0.015Pa / min, and RGA background signal fluctuation <5%. Start RGA and select the ion monitoring working mode to obtain the mass number corresponding to the preset target gas. Map each gas to one or more characteristic mass-to-charge ratios and output the target mass number set M. target Simultaneously, background spectra are acquired using a full-spectrum scanning mode with a step size of 0.5 amu and a sampling time ≥ 100 ms / point. N consecutive acquisitions are performed, and the average value is saved as the background reference vector B0 (m / z). Then, the target mass number set M is obtained. target A scan is performed to acquire the raw signal in real time and dynamic background subtraction is executed. The specific processing method is as follows: At different chamber temperatures T i The full-spectrum background signal under high vacuum steady state was collected and a background library was constructed. B i (m / z) represents the background spectrum under the i-th operating condition. Then, the chamber temperature T is obtained through a sensor. t And calculate the change in chamber temperature |T t -T t-1 | and simultaneously compare it with a temperature threshold; if the change value |T t -T t-1 If the temperature exceeds the threshold, dynamic background update is initiated to find two adjacent operating conditions T. low <T t <T high And calculate the temperature normalization factor. Generate dynamic background spectrum Then, for each mass-to-charge ratio m / z and time t, the net ion current is calculated. Output net signal, where I raw (m / z, t) represents the full-spectrum background signal.

[0024] S3. Based on the obtained net signal, preprocessing and noise reduction are performed. Noise reduction involves using an adaptive window length moving average filter or a Savitzky-Golay convolution smoothing algorithm to suppress high-frequency noise while preserving signal trend characteristics. Then, based on the obtained preprocessed concentration-time series signal, a dynamic mass balance equation is established, incorporating the test chamber volume, temperature, and pumping system speed parameters. This equation is then applied according to the formula... Calculate the venting rate Q, where V is the effective volume of the chamber, C is the gas concentration, and S is the effective pumping speed of the pumping system under the corresponding gas composition. Simultaneously, feature parameters of the venting rate curve are extracted, including instantaneous venting rate value, venting rate change rate, local curvature of the venting rate curve, and venting rate percentage relative to peak value. A feature-time series is then constructed based on the obtained feature parameters.

[0025] S4. Continuously track and analyze the gas release rate change curve, and record its maximum value Q. max and the corresponding time t peak Based on the obtained parameters, a multi-condition composite endpoint judgment is initiated, and the specific judgment conditions include: rate threshold condition and stability condition; The rate threshold condition is: the current venting rate Q. t ≤a×Q max , where a is a preset ratio; The stability condition is: within the most recent consecutive time window Within the range, the rate of change of the venting rate between any two sampling points satisfies: Where b is the stability critical value, t i and t j ∈[t- ,t]; When both the rate threshold condition and the stability condition are met, the system determines that the bonding process has reached its end point, automatically sends an end point arrival signal, closes the second gate valve, and isolates the test chamber.

[0026] Second Embodiment This embodiment discloses a method for determining the endpoint of hydrogen-oxygen catalytic bonding in fused silica glass based on the detection of water vapor release using a residual gas analyzer. The specific technical solution is as follows: Test equipment such as Figure 2 or Figure 3As shown, it includes a test chamber, a sample chamber, a first gate valve, a second gate valve, an all-metal angle valve, a quick-release flange, a first full-range vacuum gauge, a second full-range vacuum gauge, a residual gas analyzer, a turbomolecular pump, a backing mechanical pump, and a vacuum molecular pump assembly.

[0027] The test samples included 12 quartz samples prebonded with 1:6 Na2SiO3 solution, with a sample size of 10×10×10mm.

[0028] The steps for detecting water vapor release include: Open the first gate valve, close the second gate valve and the angle valve, open the first full-range vacuum gauge, start the fore-stage mechanical pump to pre-evacuate the test chamber, observe the reading of the first full-range vacuum gauge, when the reading is less than 10 Pa, start the turbomolecular pump to evacuate the test chamber to the ultimate pressure. Turn on the residual gas analyzer and set the residual gas analyzer detection mass number to 18 (H2O).

[0029] Open the quick-release flange, load the test sample into the sample chamber, and close the quick-release flange. Open the angle valve, open the second full-range vacuum gauge, and start the vacuum molecular pump assembly to pre-evacuate the sample chamber.

[0030] Observe the second full-range vacuum gauge; when the reading is less than 10... -2 When Pa is reached, open the second gate valve to connect the sample chamber and the test chamber, and close the angle valve. Use a residual gas analyzer to monitor the water pressure in the test chamber in real time, and use the data processing unit to record the water vapor partial pressure and calculate the venting rate.

[0031] When the venting rate drops to 5% of the peak rate and remains constant for 30 seconds, the bonding process is considered complete. Close the second gate valve, close the second full-range vacuum gauge, open the quick-release flange, and remove the test sample.

[0032] Figure 5 This is the curve showing the change in moisture partial pressure over time during the hydrogen-oxygen catalytic bonding process, as measured by the residual gas analyzer in this embodiment.

[0033] Depend on Figure 5 The curves showing the change in the partial pressure of water vapor during bonding reveal the following: First, the bonding process releases a large amount of water vapor, resulting in a high partial pressure of water vapor inside the test chamber. Simultaneously, the released water vapor is rapidly removed by the turbomolecular pump, causing the water pressure inside the test chamber to drop rapidly in the early stages. Second, as the bonding process continues, the bonding interface gradually seals, reducing the venting channels, and the water pressure inside the test chamber changes from a rapid decrease to a slow decrease. Finally, as the bonding process ends, the bonding interface essentially stops releasing water vapor, and the water pressure inside the test chamber reaches a plateau, decreasing extremely slowly.

[0034] Depend onFigure 5 The partial pressure change curve of bonded moisture shows that the hydrogen-oxygen catalytic bonding of quartz samples completed the bonding process within four days. General research experience suggests that hydrogen-oxygen catalytic bonding of quartz samples requires four weeks of room temperature curing to fully reach a dehydrated and cured state. Subsequent tensile strength tests on quartz samples that underwent four days of vacuum treatment after pre-bonding revealed that the bond strength of all 12 tested samples was greater than 12 MPa, which is not lower than the bond strength after four weeks of room temperature curing.

[0035] Third Embodiment This application provides a bonding endpoint determination system based on gas outgassing rate monitoring, including a test chamber, a sample chamber, a second gate valve connecting the two chambers, a residual gas analyzer, a full-range vacuum gauge, a turbomolecular pump, a forepump, a vacuum molecular pump assembly, and a data processing unit; the residual gas analyzer is installed in the test chamber to monitor the target gas concentration; the data processing unit is used to receive the concentration signal, calculate the outgassing rate, and perform endpoint determination.

[0036] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0037] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for determining a bonding endpoint based on a gas outgassing rate monitoring, characterized by, The method specifically comprises the following steps: S1, placing the first substrate and the second substrate to be bonded in a sample chamber, starting a vacuum molecular pump group to vacuumize the sample chamber and apply process conditions; S2, after the chamber reaches a high vacuum stable state, connecting with a test chamber, and monitoring the target gas concentration signal released by the bonding layer by using a residual gas analyzer; S3, receiving the gas concentration signal by a data processing unit, and calculating the gas outgassing rate per unit time; S4, continuously tracking the change curve of the outgassing rate, and when the outgassing rate is monitored to decrease from the peak value and reach a stable state, it is determined that the bonding process reaches the end point.

2. The method of claim 1, wherein the method is characterized by: The conditions of the high vacuum stable state are: Collecting pressure data P(t) continuously, then set the target high vacuum range P target , and At the same time, introduce a bistable determination condition, when the pressure P(t)≤1x10 -3 , and in the past time window, the pressure fluctuation amplitude P max -P min =kxP(t), where k is a constant.

3. The method of claim 1, wherein the method is characterized by: The process of connecting with the test chamber is: The first stage valve is opened to 10%-20% opening, maintained for 30-60 seconds, if the test chamber pressure suddenly rises and exceeds the limit, the valve is automatically closed and an alarm is given, otherwise the test chamber pressure rising rate is continuously monitored, if the pressure rising rate <0.5 Pa / s, the valve is fully opened, after the two chambers are completely connected, the test chamber pressure is continuously monitored until the constraint condition is met, then the RGA enters the target gas monitoring mode, the constraint condition includes: pressure rising rate <0.015 Pa / min, RGA background signal fluctuation <5%.

4. The method of claim 1, wherein the method is characterized by: The way of monitoring the target gas concentration signal released by the bonding layer by using the residual gas analyzer is: Start the RGA and select the ion monitoring mode of operation, get the preset target gas corresponding to the mass number, each gas is mapped to one or more characteristic mass-to-charge ratio output target mass number set M target Synchronously collect the background base spectrum, specifically, use the full spectrum scanning mode, step 0.5amu, sampling time≥100ms / point, continuously collect N times and take the average value to save as the background reference vector B0(m / z), then scan the obtained target mass number set M target Collect the original signal in real time and perform dynamic background subtraction.

5. The method of claim 4, wherein the method is characterized by: The way of real-time collecting the original signal and performing dynamic background subtraction is: At different chamber temperatures T i Next, collect full spectrum background signals at high vacuum steady state, and build a background library , where B i (m / z) represents the background spectrum under the i-th working condition, then obtain the chamber temperature T t through the sensor, and calculate the change value of the chamber temperature |T t -T t-1 |, and compare it with the temperature threshold value, if the change value |T t -T t-1 | is greater than the temperature threshold value, then start dynamic background updating, find two adjacent working conditions T low <T t <T high , and calculate the temperature normalization factor , generate a dynamic background spectrum , then calculate the net ion flow for each mass-to-charge ratio m / z and time t, and output the net signal, where I raw (m / z, t) represents the full spectrum background signal.

6. The method of claim 1, wherein the method is characterized by: The calculation way of the gas outgassing rate is: The obtained net signal is pre-processed and de-noised, then based on the obtained pre-processed concentration-time sequence signal, a dynamic mass balance equation containing test chamber volume, temperature, and pumping system pumping speed parameters is established, and according to the formula The outgassing rate Q is calculated, wherein V is the effective volume of the chamber, C is the gas concentration, and S is the effective pumping speed of the pumping system under the corresponding gas component.

7. The method of claim 6, wherein the method is characterized by: The concentration signal processing includes adopting a moving average filter with an adaptive window length or a Savitzky-Golay convolution smoothing algorithm to denoise the original signal.

8. The method of claim 1, wherein the method is characterized by: The way of determining that the bonding process reaches the end point is: Current deflation rate Q t ≤ a x Q max where a is a preset ratio; In a recent consecutive time window , the rate of change of the outgassing rate between any two sampling points satisfies: , where b is a stability threshold, t i and t j ∈ [t- , t].

9. The method of claim 1, wherein the method is characterized by: The sample chamber is a chamber that can be independently controlled in temperature and pressure, which can apply temperature and pressure conditions to the sample according to a preset process curve, and the target gas includes at least one of water vapor, hydrogen, oxygen, nitrogen oxide or organic solvent vapor.

10. A system for determining a bonding endpoint based on a gas outgassing rate monitoring, for carrying out the method for determining a bonding endpoint according to any one of claims 1 to 9, characterized in that It comprises: The test chamber, the sample chamber, the second plug valve connecting the two chambers, the residual gas analyzer, the full-range vacuum gauge, the turbo molecular pump, the forepump, the vacuum molecular pump group and the data processing unit; The residual gas analyzer is installed in the test chamber for monitoring the target gas concentration, and the data processing unit is used for receiving the concentration signal, calculating the outgassing rate and performing end point judgment.

Images