Method and plasma processing system for plasma processing of a substrate in a plasma chamber

By monitoring and adjusting parameter characteristics in plasma processing, especially using current-controlled signals and high sampling rate technology, combined with machine learning to optimize parameters, the cracking problem of glass substrates driven by AC power was solved, achieving higher production stability and product quality.

CN114586129BActive Publication Date: 2025-10-10TRUMPF HUETTINGER SP ZOO
View PDF 15 Cites 0 Cited by

Patent Information

Application Number
CN202080066867.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2019-09-23
Filing Date
2020-09-23
Publication Date
2025-10-10
Estimated Expiration
2040-09-23

AI Technical Summary

Technical Problem

Existing technologies are unable to effectively prevent or reduce cracking defects in glass substrates during large-area sputtering processes driven by AC power, resulting in damage and production interruptions.

Method used

By monitoring relevant parameters in plasma processing, characteristics related to cracking are identified and the power supply signal is adjusted to reduce these characteristics. In particular, without interrupting the processing, current-controlled signals and high sampling rates are used to monitor parameters such as voltage, current, and power, and machine learning and artificial intelligence algorithms are combined to optimize the parameter combination.

Benefits of technology

Significantly reduce or eliminate the occurrence of cracking, improve production continuity and product quality, and reduce production losses.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114586129B_ABST
    Figure CN114586129B_ABST
Patent Text Reader

Abstract

A method of plasma processing a substrate in a plasma chamber, comprising the steps of: a. supplying a power signal to the plasma chamber to form a plasma in the plasma chamber; b. monitoring at least one parameter associated with the plasma processing; c. determining a characteristic associated with the monitored at least one parameter; d. adjusting the power signal to modify the characteristic during the plasma processing.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a method for plasma processing a substrate in a plasma chamber, comprising the steps of supplying a power signal to the plasma chamber to form plasma in the plasma chamber and monitoring at least one parameter related to the plasma processing.

[0002] Furthermore, the present invention particularly relates to a plasma processing system for implementing the method of the present invention. Background Art

[0003] Cracking defects were first discovered in large-area sputtering processes driven by AC power supplies. The defects appear on the substrate and can lead to damage to the substrate, such as architectural glass, as described in more detail in US 2018 / 0040461 A1. The occurrence of cracking depends on the power supply used (medium frequency (MF) power supply or direct current bipolar power supply, also called bipolar power supply, both of which are forms of AC power supply). If cracking is detected, different measures can be taken, such as thorough mechanical cleaning of the vacuum chamber components, grounding / insulation of the rollers, careful treatment of the glass edges, covering the rollers with synthetic fibers, using auxiliary anodes that are biased positive with respect to the ground level, or changing the coating machine geometry. Proper cleanliness of the vacuum chamber is considered to be one of the most important issues.

[0004] For example, system cleanliness deteriorates during deposition, with cracking observed during the deposition of metal layers such as Ag or Ni-Cr. Grounding / insulating the glass rollers and using synthetic fiber-covered rollers keep the glass and all layers deposited on them at a floating potential. This impacts the plasma characteristics and floating potential dynamics within the system. Few glass manufacturers claim to have specially treated glass edges to significantly reduce the likelihood of cracking.

[0005] It has been qualitatively demonstrated that slight rounding of the edges of glass sheets has the potential to eliminate cracking. However, this approach is limited by the thickness of the glass. Processing of glass with thicknesses ranging from 4 to 10 mm has shown that the success of edge rounding in preventing cracking decreases with increasing glass thickness. Other manufacturers have demonstrated that the effect of glass thickness can be eliminated by varying the target-to-glass distance.

[0006] The above methods are not enough to effectively eliminate the occurrence of cracks. Summary of the Invention

[0007] It is an object of the present invention to provide a method and system that eliminates or at least significantly reduces the occurrence of cracking.

[0008] According to a first aspect, the present invention relates to a method for plasma processing a substrate in a plasma chamber, the method comprising the following steps:

[0009] a supply power signal to an electrode disposed within the plasma chamber to form a plasma within the plasma chamber;

[0010] b. monitoring at least one parameter associated with the plasma process;

[0011] c. determining a characteristic associated with at least one monitored parameter;

[0012] d. adjusting the power supply signal during plasma processing to modify, in particular reduce, features,

[0013] e. Modifying features, particularly to eliminate and / or mitigate the formation of cracks on the substrate.

[0014] While known methods only react to the occurrence of cracks and then perform method steps to reduce the cracks, the method of the invention is applied even before the cracks occur. Thus, substrate damage due to cracks can be eliminated or at least significantly reduced.

[0015] The power supply may be designed to deliver AC power greater than 500 W, in particular greater than 5 kW and typically greater than 50 kW.

[0016] The power supply may be designed to deliver AC power in a frequency range between 1 kHz and 200 kHz, in particular between 5 kHz and 100 kHz.

[0017] The output power of the power supply can be connected to two targets in the plasma chamber, so that the two targets can be driven alternately as cathodes and anodes.

[0018] The regulated power supply signal can be a current-, voltage- or power-controlled signal. Advantageously, it is a current-controlled signal.

[0019] The monitored parameters may be, for example, voltage, current, power, reflected waves (at the fundamental frequency or other frequencies), and / or combinations of the foregoing, such as impedance of the plasma process.

[0020] The monitored parameter may be different from the power signal being regulated. For example, if the power signal is a current-controlled signal, the monitored parameter may be voltage. Alternatively, if the power signal is a voltage- or power-controlled signal, the monitored parameter may be current.

[0021] The parameter monitored may be the voltage of one of the targets relative to the potential of the plasma chamber, which may be grounded or earthed.

[0022] The monitored parameter may additionally be the voltage of another target relative to the potential of the plasma chamber, which may be grounded or earthed.

[0023] The monitored parameter may in particular be the voltage between two targets in the plasma chamber.

[0024] The monitored parameter can also be a parameter measured in the plasma chamber or next to the plasma chamber, for example through a window. The monitored or measured parameter can be light, pressure, discharge, electric or magnetic field strength or other signals in the plasma chamber.

[0025] The monitored parameter may be measured at a sampling rate higher than the frequency of the AC power supply, in particular more than ten times the frequency of the AC power supply.

[0026] The monitored parameter can be the derivative of the measured value. Then, the speed or rate of change of the measured value can also be a source for determining the characteristic.

[0027] The monitored parameter can be the filter value. Then, a part of the frequency spectrum in the measurement value can also be the source for determining the characteristic.

[0028] The monitored parameter may be a time frame value. This should mean that a predetermined time frame is placed on the measured value and only a time interval of the measured value will be the source for determining the characteristic.

[0029] In bipolar plasma processing, current-controlled power supplies are often used, meaning the adjustable power supply signal follows a predetermined current setpoint. The resulting voltage, however, depends on the plasma's impedance and can vary very rapidly. Therefore, the voltage typically does not directly follow the current waveform. It has been found that not every disproportion between current and voltage is a cause of cracking. However, typical characteristics that can lead to cracking have been identified.

[0030] A "signature" can be defined as "a characteristic that indicates cracking." This should mean that the "signature" is determinable and related to the likelihood of cracking. Furthermore, the relationship between cracking and the "signature" can also be detected. To detect this relationship, the following steps can be performed:

[0031] A) Determine the likelihood of cracking in a first shape by predefined "features".

[0032] B) The "feature" can then be modified, in particular reduced, compared to the "feature" accompanying the first shape. This can be done by adjusting the power supply signal (method step d.). The second shape of the now detectable "feature" differs from the first shape.

[0033] C) Determine the likelihood of cracking with this modified, particularly reduced, "feature".

[0034] D) If the likelihood of cracking decreases with a modified, in particular reduced, "feature", then this is a "feature indicative of cracking".

[0035] The features may be one or more or a combination of the following:

[0036] Special values

[0037] Waveform characteristics, signal shape or pattern

[0038] And among these:

[0039] ·characteristic,

[0040] Features

[0041] Unique features,

[0042] Temporary characteristics,

[0043] ·abnormal,

[0044] Uneven,

[0045] ·Special disproportion,

[0046] Deviations or changes from expectations,

[0047] Unstable,

[0048] Inconsistency

[0049] Incoordination

[0050] Oscillation, vibration, ripple or ringing,

[0051] Overshoot,

[0052] Afterpulse oscillation,

[0053] Signal drops or decreases,

[0054] decrease and then recover,

[0055] Inconsistency.

[0056] In particular, the characteristic may be a value. The value may reach or exceed a threshold value.

[0057] The feature related to the at least one monitored parameter can be a feature directly derived from the monitored parameter, in particular a maximum or minimum value of the monitored parameter. Alternatively, the feature can be obtained after processing the monitored parameter. Examples for determining a feature can be found in e.g. US8007641B2, DE102011007596B3, US10209294B2, US10181392B2, US7640120B2, US10290477B2, EP2905802B1, EP3234980B1, which are incorporated by reference. But in contrast to those disclosures, the feature for use in the present invention should not be understood as a single event, e.g. a voltage drop in case of a sudden arc in a plasma process. It can be a recurring event. If the power supply signal is periodic, in particular an alternating current signal, it can be a periodic event, in particular in line with the period of the power supply.

[0058] The feature can develop over time, e.g. after one or two or several hours, in particular as a result of contamination or impurity of the plasma chamber or unwanted deposition of components or tools in the plasma chamber.

[0059] It is not always so easy to understand which features or parts of the monitored parameters can be key to reduce the crack potential. Some features seem obvious, some are easy to detect, some are difficult to detect. Especially when a combination of more than one feature can be crucial for the crack, it is difficult to find these features. Advantageously, a method and / or system with a machine learning algorithm and / or artificial intelligence algorithm can be used to identify parameters as essential parameters and / or features as essential features that are related to an increase of the crack potential. For a method and / or system with a machine learning algorithm and / or artificial intelligence algorithm, crack events in the substrate should additionally be monitored and analyzed. By this monitoring and analysis, training and test data sets can be recorded. In this way, the method and system can be trained and in particular checked and in particular continuously improved.

[0060] The monitored parameter can be, but is not limited to:

[0061] a. the potential between electrodes arranged within the plasma chamber, or

[0062] b. the potential between one of the electrodes arranged within the plasma chamber and a reference electrode, or

[0063] c. the potential between two electrodes arranged within the plasma chamber and a reference electrode, wherein the reference electrode can be grounded or floating.

[0064] It is important to note that the method of the present invention, in particular the method steps, can be performed without interrupting the plasma process. In particular, method step d., in which the power supply signal is adjusted during the plasma process to modify, in particular reduce, the characteristic, can be performed without interrupting the plasma process.

[0065] According to a method variant, at least one of an electrode-to-electrode potential and an electrode-to-ground potential is monitored. In particular, if two electrodes are provided in the plasma chamber and each electrode is a target, the target-to-target potential can be monitored. Furthermore, if at least one electrode is provided in the plasma chamber as a target, the target-to-ground potential can be monitored.

[0066] Determining the characteristic may comprise comparing the at least one monitored parameter or a quantity derived therefrom with a threshold limit. The threshold limit may be one, or more, or a combination of:

[0067] Values, such as voltage, current, power, frequency, impedance, etc.,

[0068] The pattern, waveform, or signal shape of voltage, current, power, frequency, impedance, etc.

[0069] The threshold boundary can be derived from one, more, or a combination of the following:

[0070] Measurements with a reference load such as a resistor, capacitor, inductor, "non-cracked plasma", or a combination of these load components,

[0071] The derivation of this measurement,

[0072] ·simulation,

[0073] Statistical analysis,

[0074] Machine learning and / or artificial intelligence.

[0075] The determined feature may be a maximum or minimum value of the monitored parameter. In particular, the feature may be a maximum value of the target-to-target potential or the electrode-to-ground potential.

[0076] Alternatively or additionally, determining the characteristic may include counting the number of times that at least one parameter, or a quantity derived therefrom, exceeds a threshold limit in a given time interval. Thus, determining the characteristic corresponds to determining the number of times a predetermined event occurs in a given time interval. Determining such a characteristic has the effect of observing the monitored parameter over a longer period of time and preventing the power supply signal from being adjusted too quickly or too drastically.

[0077] Adjusting the power signal can include applying multiple parameter sets to generate the power signal, determining the effect of a particular parameter set on the characteristic, and selecting a parameter set that achieves a more preferred level of the characteristic, particularly to its lowest or highest level for further plasma processing. The more preferred level of the characteristic can be a level at which the likelihood of cracking is significantly reduced. A parameter set for generating the power signal can be determined. Furthermore, a test run can be performed to determine a parameter set that changes the characteristic. Furthermore, a parameter set that reduces the characteristic can be obtained from other plasma processes from other plasma supply systems.

[0078] Furthermore, adjusting the power supply signal may include at least one of the following:

[0079] a. modifying the amplitude of at least one power signal component, in particular the amplitude of the current and / or voltage,

[0080] b. modifying the waveform of at least one power signal component, in particular the waveform of the current and / or voltage,

[0081] c. Modifying the frequency of at least one power supply signal component, in particular the frequency of the current and / or voltage.

[0082] If one of the parameters is changed, one or more other parameters may also need to be changed to maintain constant power. For example, if the duty cycle is changed, i.e., the waveform of the power signal component is changed, then the amplitude of the power signal component may also need to be changed. The amplitude, waveform, and frequency of the power signal component may be a parameter set or part of a parameter set used to generate the power signal.

[0083] At least some of the method steps may be performed in response to user demand. For example, in the case of an old power supply, where monitoring and / or detection can only be performed using external equipment to determine characteristics, the method may be triggered by the user.

[0084] Alternatively, at least some of the method steps may be performed in response to detecting that at least one monitored parameter exceeds a threshold limit. Thus, if the monitored parameter exceeds the threshold limit, an improvement in the plasma treatment process may be initiated. The threshold limit may be given, user-defined, or determined by an algorithm, such as a machine learning algorithm and / or an artificial intelligence algorithm.

[0085] Methods and systems using machine learning and / or artificial intelligence algorithms should also monitor cracking events in the substrate. This monitoring and analysis can generate training and test data sets. This allows for training and, in particular, testing and, in particular, continuous improvement of the methods and systems.

[0086] Alternatively, at least some of the method steps can be performed in response to statistical data obtained from a range of power supplies powering a plasma process. In particular, the method of the present invention can be activated as a background application and use statistical data from a range of power supplies available in a central processing system, particularly a cloud computing system, to perform complex modifications and return settings to the power supplies to adjust operating parameters, such as voltage and / or current waveforms, amplitudes, and frequencies. Statistics can be obtained from different power supplies, particularly data collected in the cloud from power supplies in different locations and for different plasma processes. A neural network can be used to determine which parameters and / or features are more likely to be qualified than others, and which of the more qualified parameters and / or features can be more highly correlated in the corresponding method steps.

[0087] According to another aspect, the present invention relates to a plasma processing system comprising:

[0088] a. Plasma chamber,

[0089] b. at least one electrode located within the plasma chamber,

[0090] c. at least one power supply for supplying a power signal to the plasma chamber,

[0091] d. at least one sensor for measuring at least one plasma processing parameter, such as potential,

[0092] e. at least one determining unit configured to determine a characteristic associated with at least one monitored parameter,

[0093] f. A control unit for controlling the power supply in response to the characteristic, wherein the control unit is configured to adjust the power supply signal such that the characteristic is modified, in particular reduced.

[0094] At least one electrode may be a target.The sensor for measuring at least one plasma processing parameter may be a voltage sensor, a current sensor, a directional coupler, or the like.

[0095] The determination unit for determining the feature may be a computing unit or a processor or a software application implemented in a computing unit.

[0096] In particular, the determination unit may comprise a comparator which compares the at least one monitored parameter or a quantity derived therefrom with a threshold limit. For example, a voltage or a current may be compared with a threshold limit.

[0097] The determination unit may further comprise a counter which counts the number of times the monitored parameter or a quantity derived therefrom exceeds a threshold limit. For example, the determination unit, in particular the counter, may count how often the target-to-target voltage rises above a threshold characteristic per unit time.

[0098] The power supply may be a MF power supply.The MF power supply may be designed to deliver an output power wherein the output voltage and / or the output current forms a sinusoidal curve.

[0099] The power supply can also be a bipolar power supply. A bipolar power supply can be designed to deliver output power in which the output voltage and / or output current are formed into a rectangular or stepped rectangular shape, or in a predetermined pattern as described in DE 10 2009 002 684 A1 or DE 10 2014 220 094 A1, which are incorporated herein by reference. The advantage of a bipolar power supply is that the duty cycle, voltage, and frequency can be adjusted without changing the power. In particular, the power can be kept constant. Thus, the power delivered to the plasma process can remain constant, but the parameters used to generate the power can be adjusted. This can prevent cracking.

[0100] The power supply may be designed to balance the power between the two outputs, for example as described in EP 1 593 143 B1 , which is incorporated herein by reference.

[0101] Furthermore, the bipolar power supply can be configured to supply power to two targets simultaneously. In particular, the bipolar power supply can be configured to supply power to a dual magnetron sputtering device.

[0102] The control unit can be integrated into the power supply, or the control unit can be external to the power supply. If the control unit is external to the power supply, it can be used with multiple power supplies.

[0103] The sensor can be located at the power supply output or near the electrodes, particularly near the target. The monitored parameter can be measured directly at the power supply output and / or on the target. In particular, the measurement can be performed at the nearest point accessible for such measurement, for example at the cable connection to the target end block.

[0104] Compared to previous solutions, a control unit configured to perform the method according to the present invention can proactively minimize the root causes of cracking by minimizing characteristics, particularly by minimizing target-to-target potential and / or target-to-ground potential. The method according to the present invention can be used as a continuous adjustment method. The control unit can thus continuously react to changing conditions in the system, such as due to the gradual growth of parasitic coatings on system components or cyclical modifications of the operating conditions of the power supply itself.

[0105] The control unit may comprise a user interface for triggering the adjustment of the power supply signal. In particular, the user may trigger the method of the present invention by using the user interface.

[0106] A plasma processing system may include several power supplies that exchange data with a cloud computing system. Thus, suitable parameters obtained for one power supply can be used in other power supplies to power similar plasma processes.

[0107] When compared with the attached drawings which are not necessarily drawn to scale Figure 1 The foregoing and other objects, features and advantages of the invention, as well as the invention itself, will be more fully understood from the following description when read together. BRIEF DESCRIPTION OF THE DRAWINGS

[0108] Figure 1 A plasma processing system is shown.

[0109] Figure 2 A typical crack defect on the substrate is shown.

[0110] Figure 3 A graph of the signal over time is shown to illustrate the invention.

[0111] Figure 4 Some typical forms of features are shown.

[0112] Figure 5 A flow chart of the method of the present invention is shown.

[0113] Figure 6 An additional flow chart of the method of the present invention is shown. DETAILED DESCRIPTION

[0114] Figure 1 A plasma processing system 1 is shown. This can be a serial coating machine. The plasma processing system 1 includes a plasma chamber 6 for processing a substrate 10, which is disposed on a substrate carrier 15. The plasma processing system 1 includes a power supply 2 connected to electrodes 11 and 12 via power lines 8 and 9. The electrodes 11 and 12 are arranged within the plasma chamber 6. By supplying power to the electrodes 11 and 12, a plasma 7 can be formed and maintained in the plasma chamber 6. The electrodes 11 and 12 can be rotating electrodes. The plasma can be enhanced by magnets; thus, the plasma processing can be a magnetron plasma processing. The plasma processing can be used to sputter material from the electrodes 11 and 12. Such electrodes 11 and 12 in the plasma processing system 1 are also referred to as targets. The plasma processing system 1 can be designed for depositing material on the substrate 10. It can be a PVD, CVD reactive plasma process, etc. The power supply 2 can be an AC power supply, thereby transmitting an AC signal to the two electrodes 11 and 12. The two electrodes 11 and 12 can then operate alternately as cathodes and anodes.

[0115] During the sputtering deposition process, cracks and arcing defects may appear on the surface of insulating materials and may damage the product, resulting in significant waste and economic losses in production. Especially in the case of large area coating (LAC) in architectural glass manufacturing, each crack will hinder production continuity and, in extreme cases, force the production cycle to be stopped before the planned system maintenance.

[0116] In the case of sputter-deposited functional coatings on glass, cracking has been reported to occur at various stages of the multilayer structure deposition. As more glass coaters transition from MF-driven dual-magnetron sputtering to bipolar power supplies, the versatility of frequency and current / voltage output waveform modification available in bipolar power supplies has inadvertently led to an increased likelihood of cracking with certain parameter combinations. In cases of excessive cracking, thorough mechanical cleaning of the vacuum chamber components is often performed to eliminate loss of glass product.

[0117] It has been discovered that cracking can occur due to the accumulation of charge on the surface of the coated glass. The glass undergoing deposition is moved beneath a multi-magnetron arrangement (typically a dual-target magnetron) via rollers made of insulating material. Even with the deposited coating, the glass itself is a dielectric material and, since it has no electrical contact with the chamber walls, can be considered to be at a floating potential. Detailed analysis of the collected data allowed the relationship between cracking and anode potential to be determined. Because the dynamics of plasma species depend on the driving forces (voltage and current waveforms) generated by the plasma discharge, a series of measurements were performed on the anode voltage while varying the bipolar power supply configuration (e.g., current waveform). It was found that the cracking effect was positively influenced by varying the current, voltage, and / or power waveforms. Figure 3 A highly successful parameter setup for the system under investigation is schematically illustrated and will be explained in more detail later. For this particular case, operation in trapezoidal mode with a 1% braking time between pulses yielded the lowest anode voltage levels measured at a pulse frequency of around 10 kHz. In contrast, operation in the default full mode (rectangular voltage and output waveform) at 10 kHz resulted in an anode voltage more than twice as high. Tests conducted on a coating machine production line confirmed that the cracking effect was significantly reduced or even eliminated when the power supply operating point with reduced anode voltage was reached.

[0118] During the normal production cycle of coated glass, the cracking effect is not the only parameter that affects product quality and process efficiency. If the coating stack configuration includes sputtering from a strong arcing target in a reactive atmosphere, indicators such as arc suppression efficiency and target condition must be considered. For these reasons, improvements to the bipolar power supply settings for minimizing anode voltage may include maintaining operation within recommended parameters. For example, if a silicon target is used for reactive sputtering of SiO2 layers, the frequency can be varied within a range that allows operation without the risk of nodule formation. If the arc suppression mechanism is abused or the operating frequency is lower than a critical value, nodules, a type of arc-related localized damage to the target surface, may appear.

[0119] Figure 1The plasma processing system 1 further comprises a monitoring unit 3, which is configured to monitor parameters such as voltage, current, and / or power. The monitoring unit 3 may include a voltage sensor, a current sensor, and / or a power sensor. The monitoring unit 3 may also include an ADC, in particular an ADC with a high sampling rate as described above.

[0120] The plasma processing system 1 also includes a determination unit 4 configured to determine a characteristic associated with at least one of the monitored parameters. This may include a comparator or software to compare the monitored parameter to a threshold value. In this manner, the characteristic may be determined. The determination unit 4 may include a computer, data storage, and program storage configured to operate software having an algorithm to determine the characteristic. Variations of the algorithm have been disclosed in detail above.

[0121] The plasma processing system 1 further comprises a regulating unit 5 configured to regulate the power signal at the output of the power supply 2. As described above, the power supply unit may be a bipolar power supply unit, wherein several regulating parameters are available, such as current, voltage, power, duty cycle, pulse interval time, and waveform shape, such as rectangular, stepped rectangular, trapezoidal, or sinusoidal.

[0122] The monitoring unit 3, the determining unit 4 and the adjusting unit 5 may all be part of the power supply 2a, either together or separately. Figure 1 , wherein all three units 3, 4, 5 are part of the power supply 2a. The control unit 13 comprises the adjustment unit 5. Additionally, the control unit 13 may comprise the determination unit 4. Additionally, the control unit 13 may at least partially comprise the monitoring unit 3.

[0123] Figure 2 A typical crack damage 19 on a glass substrate 10 is shown.

[0124] Figure 3 Five graphs are shown over time t to explain the method steps of the invention.

[0125] Graph 20a shows the power supply signal 21 directed to the electrodes 11, 12 to maintain the plasma 7. This can be a current, voltage or power signal. Advantageously, it is a current signal. It is typically an AC signal because it is provided by a bipolar power supply.

[0126] Graph 20b shows a monitored parameter 22 associated with the plasma process. This can be a current, voltage or power signal. Advantageously, this is a voltage signal measured between the electrode 11 or 12 and the ground potential.

[0127] Graph 20c again shows the monitored parameter 22. In addition, two threshold values ​​23, 23a are shown as dashed lines. It can be seen that the monitored parameter 22 periodically exceeds the threshold values ​​23, 23a. Here, the determination unit 4 determines a characteristic 24 associated with the at least one monitored parameter 22.

[0128] Graph 20d shows the adjusted power supply signal 21a, which now has a trapezoidal shape. Other shapes are possible, but this shape is an example that appears to have successfully improved the plasma supply system and reduced cracking in this or similar ways.

[0129] Graph 20e again shows the monitored parameter 22a. However, this time, it has changed due to changes in power supply signal 21a and the reaction of plasma 7 within plasma chamber 6. It can be seen that characteristic 24 in graph 20c has been modified. Modification of the characteristic can be considered a reduction in its original value. Thus, graph 20c shows that characteristic 24 has decreased, and monitored parameter 22a no longer exceeds threshold value 23.

[0130] Figure 4 Some typical special forms are shown in the four diagrams 30-33.

[0131] Graph 30 shows a periodic parameter drop as a feature 24. Such a feature can be determined by a threshold boundary 23, but the monitored parameter must be part of a time frame of measured values.

[0132] Graph 31 shows ringing, oscillation or ripple as feature 24. To determine such a feature, a filter can be used.

[0133] Graph 32 shows a gradual rise with a sharp kink as feature 24. In order to determine such a feature, the derivative of the measured value can be used as the parameter monitored in order to determine the feature.

[0134] Graph 33 shows a different view of the measured values. Here, the voltage and current are monitored in graphical form. The expected form is likely an ellipse, with one cycle driven per period. Therefore, threshold boundary 23 can be such an ellipse. This also allows for the detection of feature 24.

[0135] Figure 5 A flow chart of a method for plasma processing a substrate in a plasma chamber 6 is shown. In step 51 , a power supply signal is provided to the plasma chamber 6 so as to form a plasma 7 in the plasma chamber 6 .

[0136] In step 52, at least one parameter associated with the plasma process is monitored.

[0137] In step 53 , a characteristic 24 associated with the at least one monitored parameter 22 is determined.

[0138] In step 54, the power supply signal is adjusted during plasma processing to modify, in particular reduce, the characteristics.

[0139] Figure 6 An additional flow chart of the method of the present invention is shown. In this case, a check is made at step 60 to see if cracks can be detected. If cracks are detected, the following steps are performed: Figure 5 In step 63, as in the method according to Figure 5 The parameters monitored are as described in the method for [ 62 ]. Step 62 is repeated for a certain number of different parameter groups. In step 63, a ranked list of parameter groups is created. Specifically, the parameter group with the lowest characteristic is selected. In step 64, the highest ranked parameter group is used to adjust the power supply signal. In step 65, it is again checked whether cracks can be detected. If so, the parameter group is changed in step 66 and monitored again in step 62. The process then proceeds from step 62 to step 63, and so on. If not, the method can end.

Claims

1. A method for plasma processing a substrate (10) in a plasma chamber (6), comprising the following method steps: a. supplying a power signal to electrodes (11, 12) arranged in the plasma chamber (6) to form a plasma (7) in the plasma chamber; b. monitoring at least one parameter associated with the plasma process; c. determining a characteristic associated with the at least one parameter being monitored; d. Adjusting the power supply signal to reduce the signature during plasma processing, e. reducing features to eliminate or reduce the formation of cracks on the substrate (10); in, The characteristic is a recurring event; The steps of confirming the characteristics include: A) Determine the likelihood of cracking through predefined "signatures" in the waveform shape of the parameter; wherein said predefining comprises comparing the monitored at least one parameter or a quantity derived therefrom with a threshold limit; B) confirm that the "feature" can be reduced; C) determine the likelihood of cracking with this reduced "characteristic"; D) If the likelihood of cracking decreases with decreasing "feature", confirm this feature.

2. The method according to claim 1, characterized in that The parameter being monitored is different from the power signal being regulated.

3. The method according to claim 1, characterized in that The feature is a periodic event.

4. The method according to claim 3, characterized in that Machine learning and / or artificial intelligence algorithms are used to identify parameters as essential parameters that correlate with an increased likelihood of cracking.

5. The method according to any one of the preceding claims 1 to 3, characterized in that Parameters monitored include: a. the potential between the electrodes (11, 12) arranged in the plasma chamber, or b. a potential between one of the electrodes (11, 12) arranged in the plasma chamber and a reference electrode, or c. The electrical potential between both electrodes (11, 12) arranged in the plasma chamber and a reference electrode, wherein the reference electrode can be grounded or floating.

6. The method according to claim 3, characterized in that The power signal is periodic, and the periodic event is consistent with the period of the power signal.

7. The method according to claim 1 or 2, characterized in that Method step d. for modifying the feature is performed without interrupting the plasma treatment.

8. The method according to claim 1 or 2, characterized in that Determining the characteristic comprises comparing the at least one monitored parameter or a quantity derived from the at least one parameter to a threshold limit.

9. The method according to claim 1 or 2, characterized in that Determining the characteristic comprises counting the number of times that the at least one parameter or a quantity derived from the at least one parameter exceeds a threshold boundary within a given time interval.

10. The method according to claim 3, characterized in that Adjusting the power signal includes applying a plurality of parameter sets to generate the power signal; determining the effect of a particular parameter set on the characteristic; and selecting a parameter set that achieves a more preferred level of the characteristic for further plasma processing.

11. The method according to claim 9, characterized in that Adjusting the power signal includes at least one of the following: a. modifying the amplitude of at least one power signal component, b. modifying the waveform of at least one power signal component, c. Modifying the frequency of at least one power signal component.

12. The method according to claim 1 or 2, characterized in that At least some of the method steps are performed in response to user demand.

13. The method according to claim 1 or 2, characterized in that At least some of the method steps are performed in response to detecting that the at least one monitored parameter exceeds a threshold limit (23).

14. The method according to claim 1 or 2, characterized in that At least some of the method steps are performed in response to statistical data obtained from a series of power supplies powering the plasma process.

15. A plasma processing system (1), comprising a. Plasma chamber (6), b. at least one electrode (11, 12) located within the plasma chamber (6), c. at least one power supply (2, 2a) for supplying a power signal to the plasma chamber (6), d. at least one sensor for measuring at least one plasma processing parameter, e. at least one determination unit (4) configured to determine a characteristic associated with the at least one parameter being monitored, f. a control unit (13) for controlling the power supply in response to the characteristic, wherein The control unit is configured to adjust the power supply signal so that the characteristic is reduced, wherein said feature is a recurring event; Among them, the determination unit performs: A) Determine the likelihood of cracking through predefined "signatures" in the waveform shape of the parameter; wherein said predefining comprises comparing the monitored at least one parameter or a quantity derived therefrom with a threshold limit; B) confirm that the "feature" can be reduced; C) determine the likelihood of cracking with this reduced "characteristic"; D) If the likelihood of cracking decreases with decreasing "feature", confirm this feature.

16. The plasma processing system according to claim 15, wherein: The determination unit (4) comprises a comparator which compares the at least one monitored parameter or a quantity derived from the at least one parameter with a threshold limit (23).

17. The plasma processing system according to claim 15 or 16, wherein: The determination unit (4) comprises a counter which counts the number of times the at least one monitored parameter or a quantity derived from the at least one parameter exceeds a threshold limit (23).

18. The plasma processing system according to claim 15 or 16, wherein: The power supply (2, 2a) is a bipolar power supply.

19. The plasma processing system according to claim 15 or 16, wherein: The control unit (13) is integrated in the power supply (2, 2a) or is external to the power supply (2, 2a).

20. The plasma processing system according to claim 15 or 16, wherein: The plasma processing system includes several power supplies (2, 2a) that exchange data with a cloud computing system.

Citation Information

Patent Citations

  • Power supplying method for plasma load, involves immediately supplying produced alternating signal to load without intermediate switching of oscillation circuit and extinguishing arc by controlling switching elements during detection of arc

    DE102009002684A1

  • Method for observing status of plasma chamber to control process of coating e.g. glass substrate, during semiconductor manufacturing process, involves supplying measurement signal into plasma chamber with measurement signal frequency

    DE102011007596B3

  • Procedure for operating an MF power generator and MF power generator

    DE102014220094A1

  • Power supply control unit

    EP1593143B1

  • Method of detecting arcs in a plasma process and power supply for supplying an output quantity to a plasma process

    EP2905802B1