Vacuum evaporation coating method for semiconductor laser chip

Abstract

This invention relates to the field of semiconductor thin film deposition technology, and discloses a vacuum evaporation coating method for semiconductor laser chips, comprising: acquiring spatial distribution information of evaporation clouds in real time, and simultaneously calculating focus characteristics, spatial drift characteristics, and transient severe disturbances based on this information; decoupling and adjusting the amplitude and center point of electron beam scanning according to the focus characteristics and spatial drift characteristics, and pausing the adjustment when a transient severe disturbance is detected. This invention transforms film uniformity control from passive averaging downstream to active, real-time stabilization at the evaporation source; by coordinating the control of the position and morphology of the evaporation cloud and actively avoiding the influence of transient disturbances, the stability and repeatability of the process are improved.

Description

Technical Field

[0001] This invention relates to a vacuum evaporation coating method for semiconductor laser chips, belonging to the field of semiconductor thin film deposition technology. Background Technology

[0002] In the current fabrication of optical films for semiconductor laser chips, electron beam evaporation is a key thin film deposition technique. To ensure uniform heating of the target material and avoid local overheating that could lead to material splashing or localized burn-through, existing processes generally employ a pre-set, fixed scanning mode to control the reciprocating motion of the electron beam on the target surface. This has become the recognized standard operating procedure in the field. The effectiveness of this fixed scanning method is based on the assumption that the physical state of the molten pool is relatively stable. However, in actual physical processes, the melting, evaporation, and even gas release of the target material can cause rapid and unpredictable dynamic changes in the morphology and hot spot distribution of the molten pool surface. The existing fixed scanning mode is an open-loop execution that does not perceive the instantaneous real state of the molten pool during operation. When the molten pool spontaneously forms hot spots or changes its morphology due to physical processes, the fixed scanning action cannot actively avoid this. On the contrary, the fixed application of energy may exacerbate this local instability, which directly leads to random fluctuations in the spatial distribution of the evaporation cloud.

[0003] This limitation is prevalent in existing technologies. Many technical solutions in this field, even those employing electron beam evaporation technology, often focus on optimizing the film structure or static process parameters. The electron beam scanning control method they rely on is, in principle, still open-loop and fixed. For example, Chinese invention patent CN113699488B discloses a coating method for the cavity surface of a semiconductor laser chip. Although this solution uses E-Gun evaporation coating, its key technical point is to form a specific film system by first depositing an Al layer and then passing oxygen. Its limitations on process conditions are only static parameters such as an Al deposition rate of 3A / s-5A / s or an oxidation treatment time of 100s. This method essentially still relies on a preset, fixed electron beam scanning mode. Its control logic cannot perceive or respond to the instantaneous physical state of the molten pool during evaporation (such as hot spot distribution or morphological changes). Similarly, it cannot fundamentally solve the aforementioned problem of random fluctuations in the spatial distribution of the evaporation cloud. The guarantee of film uniformity ultimately still passively depends on the mechanical averaging of substrate rotation.

[0004] Therefore, the technical problem to be solved by this invention is how to establish a control method that can sense the dynamic characteristics of the spatial distribution of evaporation clouds in real time and adjust the electron beam scanning mode parameters in reverse and dynamically based on these characteristics, so that the energy input can actively adapt to the physical changes of the molten pool and suppress the random drift of evaporation clouds from the source. Summary of the Invention

[0005] To address the problems mentioned in the background section, the technical solution of this invention is as follows:

[0006] A vacuum evaporation coating method for a semiconductor laser chip is applied in a vacuum chamber containing an electron beam evaporation source and a substrate. The electron beam evaporation source has an electron beam controller with an adjustable scanning mode. The method includes:

[0007] Step a: During the evaporation coating process, at least two spatially separated sensors deployed in the vacuum chamber are used to acquire in real time the spatial distribution information of the deposition rate of the evaporation cloud generated by the evaporation source.

[0008] Step b: Based on spatial distribution information, calculate the focusing characteristics that characterize the spatial distribution morphology of evaporation clouds;

[0009] Step c: Based on spatial distribution information, calculate the spatial drift characteristics that characterize the spatial distribution location of the evaporation cloud; analyze the time series characteristics of the spatial distribution information in real time to determine whether there are transient and severe disturbances.

[0010] Step d: Based on the focusing characteristics, and only when it is determined that there is no transient violent disturbance, dynamically adjust the amplitude parameter of the electron beam scanning mode; based on the spatial drift characteristics, and only when it is determined that there is no transient violent disturbance, dynamically adjust the center point coordinate parameter of the electron beam scanning mode.

[0011] Step e, and when a transient severe disturbance is determined, the dynamic adjustment of the amplitude parameter and the center point coordinate parameter is suspended for a preset suspension time.

[0012] Preferably, the step of determining whether a transient severe disturbance exists includes: calculating the first derivative of the original deposition rate signal corresponding to the spatial distribution information; and determining that a transient severe disturbance exists when the absolute value of the first derivative exceeds a preset first threshold; and pausing the dynamic adjustment of the amplitude parameter and the center point coordinate parameter includes: freezing the amplitude parameter and the center point coordinate parameter to the values ​​at the moment before the transient severe disturbance is determined within a preset suspension time.

[0013] Preferably, in the step of calculating the focusing characteristic characterizing the spatial distribution morphology of the evaporation cloud, the focusing characteristic is a dispersion statistic calculated from the deposition rate signals of at least two spatially separated sensors, wherein the dispersion statistic is the variance of the deposition rate signal. , For the first Deposition rate signals from individual sensors, The average deposition rate signal from all sensors. The total number of sensors, variance Defined as: ;

[0014] Preferably, the step of dynamically adjusting the amplitude parameter of the electron beam scanning mode based on the focus characteristics includes: adjusting the variance... Compared with the preset target variance range, when the variance When the variance exceeds the upper limit of the preset target variance range, increase the amplitude parameter, and when the variance... When the variance falls below the lower limit of the preset target variance range, the amplitude parameter is reduced.

[0015] Preferably, in the step of calculating the spatial drift characteristics characterizing the spatial distribution location of the evaporation cloud, the spatial drift characteristics are the amount of drift of the centroid of the evaporation cloud from a predetermined center position.

[0016] Preferably, the step of dynamically adjusting the center point coordinate parameters of the electron beam scanning mode according to the spatial drift characteristics includes: when the drift indicates a deviation in the first direction, shifting the center point coordinate parameters to a region on the electron beam evaporation source target opposite to the first direction.

[0017] Preferably, the method further includes: real-time monitoring of the temporal stability of spatial drift characteristics; when the value of the spatial drift characteristics exceeds a preset second threshold, or when the spatial drift characteristics remain unchanged at a non-zero value for a period of time exceeding a preset static duration, determining that at least two spatially separated sensors have malfunctioned; when a malfunction is determined, suspending the steps of dynamically adjusting the amplitude parameter and dynamically adjusting the center point coordinate parameter, and resetting the amplitude parameter and the center point coordinate parameter to their respective initial set values.

[0018] Preferably, the method further includes: calculating a total flux signal characterizing the total evaporation rate based on spatial distribution information, and dynamically adjusting a preset target variance range according to the total flux signal.

[0019] Preferably, at least two spatially separated sensors comprise a quartz crystal microbalance.

[0020] Preferably, in the step of calculating the focusing characteristic that characterizes the spatial distribution morphology of the evaporation cloud, the focusing characteristic is the ratio of the deposition rate of the center sensor to that of the edge sensor.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] 1. By establishing a real-time feedback path between the spatial distribution of the evaporation cloud and the electron beam scanning mode parameters, the control of film uniformity is transformed from relying on substrate rotation for hysteretic and passive error averaging to actively and instantly stabilizing the physical state at the evaporation source, thereby improving the stability of the evaporation cloud, which is the fundamental condition of the coating process.

[0023] 2. By utilizing the spatial distribution information obtained from the same set of spatially separated sensors, and through parallel processing, the drift characteristics representing the centroid position of the evaporation cloud and the focusing characteristics representing its distribution width were extracted simultaneously. Furthermore, these two physically related characteristics were mapped to the center point coordinates and scanning amplitude of the electron beam scanning mode, which are two independently adjustable parameters, in the control logic, thus constructing two non-interfering adjustment loops and achieving stable control of the position and morphology of the evaporation cloud.

[0024] 3. This method further identifies transient disturbances such as material splashing inherent in the physical process by analyzing the time-series characteristics of the deposition rate signal. When such disturbances are identified, the adjustment of the electron beam scanning mode is temporarily suspended, avoiding over-adjustment of the control system based on erroneous transient signals and maintaining the physical stability of the molten pool during the disturbance period. The provided method, through real-time perception and multi-dimensional adjustment of the physical state of the evaporation source, enables the entire coating process to actively adapt to the morphological changes of the target material during consumption, reducing the dependence of the process results on the initial state of the target material and random fluctuations of the molten pool, and improving the batch-to-batch repeatability and stability of the semiconductor laser chip optical film preparation process. Attached Figure Description

[0025] Figure 1 This is the logic diagram of the multi-feature decoupling feedback control for evaporation clouds in this invention;

[0026] Figure 2 This is a timing diagram of the transient disturbance identification and control suspension response of the present invention;

[0027] Figure 3 This is a diagram illustrating the root cause analysis of the uniformity problem in the existing open-loop process of this invention.

[0028] Figure 4 This is a schematic diagram of the spatial layout of the evaporation source and multiple sensors in this invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below; obviously, the described embodiments are some embodiments of the present invention, but 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.

[0030] This invention discloses a vacuum evaporation deposition method for semiconductor laser chips, applied in a vacuum chamber containing an electron beam evaporation source and a substrate. The electron beam evaporation source is equipped with an electron beam controller with an adjustable scanning mode. The method constructs a real-time feedback control system, the core logic of which includes: acquiring, in real-time, the spatial distribution information of the deposition rate of the evaporation cloud generated by the evaporation source through at least two spatially separated sensors, preferably a quartz crystal microbalance, deployed within the vacuum chamber; based on this spatial distribution information, the system calculates in parallel the focus intensity characteristics characterizing the spatial distribution morphology of the evaporation cloud and the spatial drift characteristics characterizing the spatial distribution position of the evaporation cloud, and simultaneously analyzes the time series characteristics of this information in real time to determine whether there are transient and severe disturbances; furthermore, at the control execution level, the method establishes two decoupled adjustment loops: dynamically adjusting the amplitude parameter of the electron beam scanning mode according to the focus intensity characteristics, and dynamically adjusting the center point coordinate parameter of the electron beam scanning mode according to the spatial drift characteristics. Simultaneously, this method establishes a high-priority arbitration mechanism, namely, when a transient and severe disturbance is determined to exist, the dynamic adjustment of the amplitude parameter and the center point coordinate parameter is suspended within a preset suspension time. In the specific implementation of this method, the acquisition of spatial distribution information relies on at least two, preferably three or four symmetrically arranged quartz crystal microbalances. The step of calculating the spatial drift characteristics characterizing the spatial distribution position of the evaporation cloud may specifically include: comparing the deposition rate signals of the quartz crystal microbalances at different positions in real time to calculate a drift vector characterizing the deviation of the centroid of the evaporation cloud from the predetermined center position. In an example of a four-quadrant layout, this drift vector can be characterized by the rate difference between two pairs of diagonal sensors. At the same time, the step of calculating the focusing characteristics characterizing the spatial distribution morphology of the evaporation cloud may be specifically implemented by calculating the discreteness statistics constituted by the deposition rate signals of at least two spatially separated sensors. In a preferred implementation, the focusing characteristic is the variance of the deposition rate signal. Its definition is: ,in For the first Deposition rate signals from individual sensors, For all The average deposition rate signal of each sensor, and the focus characteristic can also be characterized as the ratio of the deposition rate of the center sensor to that of the edge sensors.

[0031] The two decoupled adjustment loops constructed in this method have the following specific execution logic: The step of dynamically adjusting the center point coordinate parameters of the electron beam scanning mode based on spatial drift characteristics is achieved by establishing a reverse compensation mechanism. That is, when the drift vector points to a first direction, the control system translates the center point coordinate parameters to a region on the electron beam evaporation source target opposite to the first direction to compensate for the drift. The step of dynamically adjusting the amplitude parameters of the electron beam scanning mode based on focusing characteristics is achieved through an interval control logic. In a preferred embodiment, the variance... Compared with a preset target variance range, when the variance When the variance exceeds the upper limit of the preset target variance range, it indicates that the evaporation cloud is over-focused. In this case, the system should increase the amplitude parameter to make the energy input more dispersed. Conversely, when the variance is below the upper limit, it indicates that the evaporation cloud is over-focused. When the variance falls below the lower limit of the preset target variance range, it indicates that the evaporation cloud is excessively dispersed. At this time, the system reduces the amplitude parameter to make the energy input more concentrated. This method also includes a mechanism for identifying and avoiding transient severe disturbances. The step of determining whether a transient severe disturbance exists may specifically include: calculating the first derivative of the original deposition rate signal corresponding to the spatial distribution information in real time. This first derivative characterizes the instantaneous rate of change of the deposition rate. The system compares the absolute value of the first derivative with a preset first threshold, which is calibrated according to the upper limit of the normal rate of change of the steady drift in the process. When the absolute value of the first derivative exceeds the preset first threshold, the system determines that a transient severe disturbance has occurred, rather than a steady drift. At this time, the system immediately triggers the step of pausing the dynamic adjustment of the amplitude parameter and the center point coordinate parameter. In a specific implementation, the pausing step includes: within a preset suspension time, adjusting the amplitude... The parameters and center point coordinate parameters are frozen to their values ​​just before the transient drastic disturbance is determined. This method may also include monitoring and adjusting the sensor status and the overall process status to further improve the robustness and adaptability of the method. On one hand, the system can monitor the temporal stability of the spatial drift characteristics in real time. When the amplitude of the spatial drift characteristics exceeds a preset second threshold, which is greater than the normal process fluctuation range, or when the spatial drift characteristics remain stationary at a non-zero value for a period exceeding a preset stationary duration, the system determines that at least two spatially separated sensors have failed. At this time, the system will suspend the steps of dynamically adjusting the amplitude parameters and dynamically adjusting the center point coordinate parameters, and reset the amplitude parameters and center point coordinate parameters to their respective initial settings. On the other hand, this method may also include: calculating the total flux signal characterizing the total evaporation rate based on spatial distribution information; an example is the average deposition rate of all sensors. Furthermore, based on the total throughput signal, the preset target variance range is dynamically adjusted.

[0032] Example 1: In a vacuum evaporation process, a multilayer semiconductor laser chip DBR reflector film is continuously fabricated using electron beam evaporation. When the deposition reaches the tenth layer, the surface morphology of the molten pool slowly changes due to the long-term consumption of the target material. The spatial distribution information obtained in real time from the four quadrant-distributed quartz crystal microbalances deployed in the vacuum chamber shows a continuous imbalance. Based on this spatial distribution information, the system calculates in parallel the spatial drift characteristics representing the position of the evaporation cloud and the focusing characteristics representing the morphology of the evaporation cloud. The calculation results show that the spatial drift characteristic is that the centroid of the evaporation cloud has shifted towards quadrant A, and the calculated value of the focusing characteristic, i.e., the variance... The value has exceeded the upper limit of the preset target variance range, indicating that the evaporation cloud is in an over-focused state.

[0033] The system then executes two decoupled adjustment loops: First, based on the spatial drift characteristics, it dynamically adjusts the center point coordinate parameters of the electron beam scanning mode, shifting them towards the C quadrant region on the target, opposite to quadrant A, to compensate for the spatial drift; second, simultaneously based on the focusing characteristics, the system dynamically adjusts the amplitude parameters of the electron beam scanning mode, increasing the value of this amplitude parameter to make the molten pool energy input more dispersed, thus broadening the evaporation cloud morphology until its variance... After falling back to within the target variance range, during the aforementioned period of stable operation under coordinated regulation, a material splash occurred at a certain point on the target material due to localized overheating, causing the original deposition rate signal of the quartz crystal microbalance located in quadrant B to drop. A transient, large-amplitude positive spike is generated. When analyzing the time-series characteristics of this signal, the system calculates that the absolute value of its first derivative has exceeded a preset first threshold. The system immediately determines that a transient and severe disturbance exists. The high-priority arbitration mechanism is triggered, and the system immediately suspends the dynamic adjustment of the amplitude parameter and the center point coordinate parameter, freezing these two parameters to their values ​​at the moment before the disturbance was determined, and maintaining a preset suspension time of 500 milliseconds. During this suspension time, the electron beam scanning mode remains unchanged, avoiding over-adjustment of the control system due to receiving erroneous spike signals, and the physical state of the molten pool is kept stable. After 500 milliseconds, the signal from the B-quadrant sensor has returned to normal, the system determines that the transient and severe disturbance has ended, automatically releases the suspension, and the two decoupled adjustment loops smoothly resume operation to continue compensating for the position and morphology of the evaporation cloud. The entire preparation process of the multilayer DBR film system is carried out stably under the compensation of slow drift and the avoidance of transient disturbances, and finally a film layer with the required uniformity and batch repeatability is obtained.

[0034] Example 2: To objectively verify the effectiveness of the synergistic effect of the various technical features in the method of the present invention, a comparative experiment was conducted in a standard electron beam evaporation chamber. This chamber is equipped with an electron beam evaporation source, whose electron beam controller can receive external commands to adjust the center point coordinate parameters and amplitude parameters of the scanning mode. Four quartz crystal microbalances are symmetrically arranged within the chamber, and a 4-inch silicon substrate is mounted on the substrate station. Five sets of comparative conditions were set up for the experiment: Control group 1, using the fixed scanning mode of the prior art without any feedback adjustment; Control group 2, only applying feedback adjustment of the spatial drift characteristics, i.e., only adjusting the center point coordinate parameters; Control group 3, only applying feedback adjustment of the focus characteristics, i.e., only adjusting the amplitude parameters. Control group 4 applied feedback adjustment to both spatial drift and focus characteristics, but did not enable transient severe disturbance identification and suspension mechanisms. The present invention sample group employed the complete method of the present invention, i.e., simultaneously applying synergistic adjustment to spatial drift and focus characteristics, and enabling transient severe disturbance identification and suspension mechanisms. All groups operated under the same basic process parameters: the evaporation material was titanium dioxide, the target deposition thickness was 500 nm, and the substrate was rotated at a uniform speed of 20 rpm. During the evaporation coating process, the intermediate process parameters calculated by the control system of each group were monitored and recorded in real time. After coating, the silicon substrate was removed, and the film thickness was measured at five locations—the center point and four quadrants at the edge—using an ellipsometry. Calculate the film uniformity, where For maximum thickness, Minimum thickness, The average thickness at 5 points is shown in Table 1.

[0035] Table 1: Comparison of intermediate and final film uniformity under different control methods

[0036]

[0037] Analysis of the data in Table 1 shows that in control group 1, without any feedback, the position and morphology of the evaporation cloud were in an uncontrolled drift state, with both the mean drift vector magnitude and mean variance being high, resulting in the worst final film uniformity of 3.1%. Control groups 2 and 3 controlled position and morphology individually, respectively. The corresponding controlled parameters, namely the mean drift vector magnitude or mean variance, were reduced, and the film uniformity improved accordingly to 1.8% and 2.0%. However, the uncontrolled parameters remained high, limiting further improvement in uniformity. The data in control group 4 showed nonlinear results; when the center point and amplitude were controlled simultaneously, but without feedback... When using transient severe disturbance identification and suspension mechanisms, the uniformity result (4.3%) was worse than all other groups. This was due to material splashing that occurred during the process, which was misinterpreted by the system as a process drift signal, causing the control system to execute drastic and erroneous adjustment actions, which in turn damaged the stability of the molten pool. In contrast, the sample of this invention, under the same operating conditions, recorded 4 control suspension events in its transient disturbance response system, indicating that the system successfully identified these disturbances and suspended the adjustment actions, avoiding the occurrence of erroneous adjustments. At the same time, its spatial drift characteristics and focus characteristics remained at a low level, ultimately achieving a film uniformity of 0.7%.

[0038] Example 3: This example combines Figures 1 to 4 A vacuum evaporation coating method for a semiconductor laser chip is described, such as... Figure 1 As shown, the system uses the evaporation cloud generated by the electron beam evaporation source as the control object. It collects information in real time through a spatial distribution information acquisition module. This information is sent in parallel to three processing units: a focus feature calculation unit, a spatial drift feature calculation unit, and a transient severe disturbance identification unit. The focus feature is used to characterize the spatial morphology of the evaporation cloud, the spatial drift feature is used to characterize the spatial position of the evaporation cloud, and the transient severe disturbance identification is used to analyze time series features. The calculation results of these three units are jointly input to an adjustment arbitration mechanism. When no transient severe disturbance is detected, the mechanism allows adjustment to be performed, feeding back the focus feature to the amplitude parameter dynamic adjustment module and the spatial drift feature to the center point coordinate parameter dynamic adjustment module. The adjustment instructions of these two modules are fed back to the electron beam evaporation source. However, when a transient severe disturbance is determined to exist, the arbitration mechanism will suspend the execution of the above two dynamic adjustment modules.

[0039] like Figure 2As shown in the figure, this graph illustrates the relationship between the deposition rate signal and its corresponding first derivative over time (in seconds). At 3.8 seconds, the deposition rate signal begins to rise sharply, causing the absolute value of its first derivative to exceed the preset threshold at 4.2 seconds. At this point, the system determines that there is a transient and severe disturbance, and the control state immediately switches from 0 (normal) to 1 (suspended), pausing adjustment until 5.2 seconds, when the first derivative recovers to within the threshold, at which point the control state switches back to 0 (normal), thus avoiding erroneous adjustment caused by signal spikes. Figure 3 As shown, the fundamental reasons for the poor uniformity and batch repeatability of the final film include: the electron beam scanning uses a fixed open-loop mode, lacking real-time feedback correlation; the physical state of the molten pool itself is unstable, with transient disturbances such as target melting, evaporation, gas release, and even material splashing; and the evaporation cloud distribution is subject to random fluctuations due to the above reasons, including changes in the distribution width (i.e., focus) and drift in the centroid position of the evaporation cloud. Existing compensation methods, such as downstream passive substrate rotation averaging, cannot suppress these errors at their root due to their slow mechanical response speed. Figure 4 As shown, at the bottom of a vacuum chamber, an electron beam evaporation source is set, which acts on the target material above to form a molten pool. At the top of the chamber, directly opposite the molten pool, a substrate is placed. Between the molten pool and the substrate, an evaporation cloud generated from the molten pool diffuses upward. At the same time, in the diffusion path of the evaporation cloud, below the substrate, four sensors are spatially separated and symmetrically distributed around the periphery of the evaporation cloud, and are labeled as sensor A, sensor B, sensor C and sensor D, respectively.

[0040] Example 4: This example is an engineering procedure for standardizing and setting key parameters. This procedure is executed after the vacuum evaporation chamber is first put into production or after the target material is replaced. It determines the basis for setting parameters such as the preset first threshold, preset hang-up time, preset target variance range, preset second threshold, and preset static time. The first stage of this calibration procedure is used to determine the preset target variance range and establish its dynamic adjustment relationship with the total flux signal. In the vacuum chamber, three independent calibration runs are performed at different total evaporation rates. The total flux signal of the three runs is stabilized at 5 Å / s, 10 Å / s, and 15 Å / s by controlling the total electron beam power, respectively. In each run, the control system turns off the automatic adjustment function and instead executes a programmed amplitude parameter scan, so that the amplitude parameter of the electron beam scanning mode linearly increases from a small value to a large value. During this process, the system records the amplitude parameter and the corresponding focus characteristics, i.e., variance, in real time. Stable readings were obtained while a sample was deposited on the substrate. After the calibration run, the uniformity of the sample film was analyzed to determine the amplitude parameters A1, A2, and A3 corresponding to the optimal uniformity at three rates of 5 Å / s, 10 Å / s, and 15 Å / s, respectively, as well as the corresponding focus characteristic variance. V1, V2, and V3 are respectively; thus, the preset target variance range is determined, and a lookup table based on the three-point interpolation of (5, V1), (10, V2), and (15, V3) is established in the system controller. This lookup table is the specific implementation logic for dynamically adjusting the preset target variance range according to the total throughput signal.

[0041] The second stage of the calibration procedure is used to determine the threshold related to transient disturbances. During the calibration run at 10 Å / s, the system captured a material splash event caused by target degassing and recorded the original deposition rate signal from the sensor. Time-domain analysis of this signal determined that during normal stable fluctuations, the absolute value of the first derivative of the signal did not exceed 45 Å / s, while during this splash event, the instantaneous peak value of the absolute value of the first derivative reached 910 Å / s. Based on this data, the preset first threshold was set to 150 Å / s, a value greater than normal fluctuations while still capturing disturbances. Further analysis of the disturbance signal showed that the total time from the peak appearance to complete recovery to the stable fluctuation range was 380 ms. To ensure that the control system resumes adjustment after the disturbance completely subsides, the preset hangup time was set to 500 ms. The third stage is used to determine the threshold related to sensor failure. During calibration operation, a fault simulation test is performed, that is, when the system is running smoothly, the signal connection of one of the quartz crystal microbalances is actively interrupted. The spatial drift characteristic calculated by the system at this time immediately jumps from the normal fluctuation level of 0.3 to 4.8 and remains unchanged. In all non-fault calibration operation data, the maximum magnitude of the spatial drift characteristic caused by the process never exceeds 0.9. Based on this, the preset second threshold is set to 2.5, which is sufficient to avoid all process fluctuations and can accurately identify signal anomalies caused by sensor failure. At the same time, by analyzing all calibration data, it is determined that in the non-fault state, the longest time that the spatial drift characteristic remains stationary at a non-zero value is 1.8 seconds. Therefore, the preset stationary time is set to 3.0 seconds.

[0042] Example 5: An alternative implementation of the focusing characteristic is applied to a vacuum chamber with a sensor layout including a central sensor and one or more edge sensors. The focusing characteristic is determined as the deposition rate ratio of the central sensor to the edge sensors. Its calibration procedure includes: In an offline test run, the system shuts down all automatic adjustment loops and performs a programmed amplitude parameter scan, linearly increasing the amplitude parameter of the electron beam scanning mode from its minimum to its maximum value. During this process, the system records the amplitude parameter and the corresponding stable rate ratio in real time and forms a gradient film on the test substrate. After the test, by analyzing the film uniformity of the test substrate, the amplitude parameter setpoint with optimal uniformity is determined, and the rate ratio corresponding to this setpoint, 1.45, is set as the target ratio for this process. In the subsequent formal coating process, the control loop compares the real-time calculated rate ratio with this target ratio. When the real-time ratio is higher than 1.45, it indicates that the evaporation cloud is over-focused, and the system increases the amplitude parameter accordingly. When the real-time ratio is lower than 1.45, it indicates that the evaporation cloud is over-divergent, and the system decreases the amplitude parameter.

[0043] A baseline bias calibration procedure for spatial drift characteristics is performed before the process in the vacuum evaporation chamber to eliminate measurement errors introduced by non-ideal symmetry of sensor mounting positions or inherent asymmetry of the evaporation source. The procedure includes: before the formal coating begins, the system performs a 60-second stable evaporation process with all control loops closed; during this period, the system acquires spatial distribution information from at least two spatially separated sensors and calculates a time-averaged drift vector characterizing the system's inherent bias, which is stored as a baseline bias vector; during the formal coating process, after calculating the original spatial drift characteristics in real time at each step, the system subtracts this baseline bias vector to obtain a calibrated spatial drift characteristic characterizing the actual physical drift, and dynamically adjusts the center point coordinate parameters based on this calibrated characteristic.

[0044] Example 6: In a vacuum chamber of a quartz crystal microbalance employing a four-quadrant layout (A, B, C, D), to implement the control logic in this specific embodiment, sensors A and C are symmetrically distributed along the first axis (X-axis), and sensors B and D are symmetrically distributed along the second axis (Y-axis); the system controller receives four initial deposition rate signals at time t. , , and Based on this, the system calculates the adjustment instructions for the center point coordinate parameters. Based on this spatial distribution information, the system calculates the spatial drift characteristics, which in this implementation are defined as a two-dimensional drift error vector. Its first axis component Calculated as Its second axis component Calculated as The system uses a proportional-integral controller to control the drift error vector. The compensation offset, converted to center point coordinate parameters, is decoupled into two components. and .

[0045] X-axis compensation offset Calculated as Y-axis compensation offset Calculated as ,in, and The drift error component at the current moment. This is the proportional gain coefficient. This is the integral gain coefficient. and The coefficients are calibrated according to the calibration procedure in Example 4. The overshoot and settling time of the step signal response are adjusted by analyzing the system. The system then calculates the compensation offset and uses an initial scan center reference point. Superposition, that is and This generates and applies dynamically adjusted center point coordinate parameters to the electron beam controller. .

[0046] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions 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 solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A vacuum evaporation coating method for a semiconductor laser chip, applied in a vacuum chamber comprising an electron beam evaporation source and a substrate, wherein the electron beam evaporation source has an electron beam controller with an adjustable scanning mode, characterized in that, The method includes: Step a: During the evaporation coating process, at least two spatially separated sensors deployed in the vacuum chamber are used to acquire in real time the spatial distribution information of the deposition rate of the evaporation cloud generated by the evaporation source. Step b: Based on spatial distribution information, calculate the focusing characteristics that characterize the spatial distribution morphology of evaporation clouds; Step c: Based on spatial distribution information, calculate the spatial drift characteristics that characterize the spatial distribution location of the evaporation cloud; analyze the time series characteristics of the spatial distribution information in real time to determine whether there are transient and severe disturbances. Step d: Based on the focusing characteristics, and only when it is determined that there is no transient violent disturbance, dynamically adjust the amplitude parameter of the electron beam scanning mode; based on the spatial drift characteristics, and only when it is determined that there is no transient violent disturbance, dynamically adjust the center point coordinate parameter of the electron beam scanning mode. Step e, and when a transient severe disturbance is determined, the dynamic adjustment of the amplitude parameter and the center point coordinate parameter is suspended for a preset suspension time; The step of determining whether a transient severe disturbance exists includes: calculating the first derivative of the original deposition rate signal corresponding to the spatial distribution information; and determining that a transient severe disturbance exists when the absolute value of the first derivative exceeds a preset first threshold; and pausing the dynamic adjustment of the amplitude parameter and the center point coordinate parameter includes: freezing the amplitude parameter and the center point coordinate parameter to the values ​​at the moment before the transient severe disturbance is determined within a preset suspension time. In the step of calculating the focus characteristics characterizing the spatial distribution morphology of evaporation clouds, the focus characteristics are discrete statistics calculated from deposition rate signals from at least two spatially separated sensors, where the discrete statistics are the variance of the deposition rate signals. , For the first Deposition rate signals from individual sensors, The average deposition rate signal from all sensors. The total number of sensors, variance Defined as: Alternatively, the focus characteristic can be the ratio of the deposition rate of the center sensor to that of the edge sensors. In the step of calculating the spatial drift characteristics that characterize the spatial distribution location of the evaporation cloud, the spatial drift characteristics are the amount of drift of the centroid of the evaporation cloud from the predetermined center position.

2. The vacuum evaporation coating method for a semiconductor laser chip according to claim 1, characterized in that, The steps for dynamically adjusting the amplitude parameters of the electron beam scanning mode based on the focusing characteristics include: adjusting the variance... Compared with the preset target variance range, when the variance When the variance exceeds the upper limit of the preset target variance range, increase the amplitude parameter, and when the variance... When the variance falls below the lower limit of the preset target variance range, the amplitude parameter is reduced.

3. The vacuum evaporation coating method for a semiconductor laser chip according to claim 1, characterized in that, The steps of dynamically adjusting the center point coordinate parameters of the electron beam scanning mode based on spatial drift characteristics include: when the drift indicates a deviation in the first direction, shifting the center point coordinate parameters to a region on the electron beam evaporation source target opposite to the first direction.

4. The vacuum evaporation coating method for a semiconductor laser chip according to claim 1, characterized in that, The method further includes: real-time monitoring of the temporal stability of spatial drift characteristics; when the value of the spatial drift characteristics exceeds a preset second threshold, or when the spatial drift characteristics remain unchanged at a non-zero value for a period of time exceeding a preset static duration, determining that at least two spatially separated sensors have malfunctioned; when a malfunction is determined, suspending the steps of dynamically adjusting the amplitude parameter and dynamically adjusting the center point coordinate parameter, and resetting the amplitude parameter and the center point coordinate parameter to their respective initial settings.

5. The vacuum evaporation coating method for a semiconductor laser chip according to claim 2, characterized in that, The method also includes: calculating the total flux signal characterizing the total evaporation rate based on spatial distribution information, and dynamically adjusting the preset target variance range according to the total flux signal.

6. The vacuum evaporation coating method for a semiconductor laser chip according to claim 1, characterized in that, At least two spatially separated sensors include a quartz crystal microbalance.

Images