TWI600051B - Noise based frequency tuning and identification of plasma characteristics - Google Patents

Description

Frequency adjustment based on noise and identification of plasma characteristics

This disclosure relates generally to the frequency adjustment of the generator. In particular, but not by way of limitation, the present disclosure is directed to systems, methods and apparatus for frequency-modulating generators in plasma processing systems and for identifying changes in plasma characteristics or plasma characteristics.

Automatic frequency adjustment often seeks to match the load impedance presented to the generator to the impedance of the delivered power designed by the generator. In some cases, this can be done by minimizing the magnitude of the load reflection coefficient ρ, which is defined as: ρ = (ZZ ₀ ) / (Z + Z ₀ ^* ) Equation (1) where Z is presented to The load impedance of the generator, Z ₀ is the desired load impedance, and * indicates the conjugate complex number. For many applications, Z ₀ = Z ₀ ^* = 50 ohms.

Auto-tuning algorithms sometimes focus on local optimization, thus missing overall optimization. Figure 8 exemplary performance metrics (e.g., reflected power or the size of the load reflection coefficient) is a graphical function of the frequency, which shows the overall minimum and f ₀ at the local minimum of _{f. 1.} For this demonstration, finding optimization is equivalent to finding the smallest performance metric. We can see that if the automatic frequency adjustment starts at a lower frequency than fa, then the algorithm is likely to settle at the local minimum of f ₁ without knowing that there is an overall minimum at f ₀ .

For a simple load, it is possible to sweep the frequency simply across the entire frequency range to find an overall optimization at f ₀ . For plasma applications, this approach to finding an overall optimization frequency is often not an option. One potential problem is that as the frequency is swept, a frequency such as f _a of Figure 8 may be encountered, where the load impedance is heavily mismatched by the impedance at which the generator can deliver power. If the generator frequency stays for any amount of time at a frequency at which such a generator cannot deliver sufficient power to the load, the plasma may be extinguished. If short-term high reflected power is acceptable, techniques that detect the entire frequency range by varying the frequency for only a short time to detect different frequencies may be an acceptable solution to find the overall optimized frequency. However, in some applications, even these short-term plasma disturbances are not acceptable. It is desirable to have a solution with minimal disturbances in the plasma load to find an overall optimization. In many applications, while finding an optimized operating frequency is equivalent to minimizing load impedance mismatch, other factors such as plasma system stability and system efficiency may be factors in the degree of frequency optimization.

Exemplary aspects of the invention shown in the drawings are summarized below. These and other specific aspects are more fully described in the [Embodiment] section. However, it is to be understood that the invention is not intended to be limited to the forms of the invention or the embodiments. Many modifications, equivalents, and alternatives are possible in the spirit and scope of the invention as expressed by the appended claims.

One aspect may be characterized by a power generation system that is constructed to constitute an automatic frequency adjustment. The power generation system includes a power source to apply a primary power signal at a primary frequency to an output coupled to the plasma load, and the power source is constructed to constitute one or more secondary powers that are less than the power generated at the primary frequency. Power signal to the plasma load. The sensor is arranged to sense one or more attributes of the power delivered to the plasma load to obtain a measure of performance, and the overall optimized identification module analyzes the performance metric and identifies an overall optimization corresponding to the performance metric Total Body optimization frequency. The frequency control module is constructed to direct the primary frequency adjustment toward an overall optimized frequency corresponding to the identification of the overall optimization.

Another aspect may be characterized by a non-transitory, physical processor readable storage medium that encodes machine readable instructions for performing automatic frequency adjustment of the power generation system. The method includes applying a primary power signal at a primary frequency to a plasma load and applying a secondary power signal of one or more secondary frequencies to a plasma load, wherein the power generated at the secondary frequency is lower than at the primary frequency The power produced. A measure of performance is monitored and an optimization frequency corresponding to the overall optimization of the performance metric is obtained. The primary frequency is then adjusted towards the optimized frequency.

100‧‧‧Power Generation System

104, 105‧‧‧ Matching network

106‧‧‧ Plasma

108‧‧‧The plasma chamber

110‧‧‧Power source

112‧‧‧ sensor

114‧‧‧ Circuitry

116‧‧‧Overall Optimized Identification Module

118‧‧‧ frequency control module

122‧‧‧ filter

130, 131‧‧‧ Connection

140‧‧‧External power

150, 151‧‧‧RF or DC generator

200‧‧‧Power Generation System

204‧‧‧matching network

208‧‧‧The plasma chamber

210‧‧‧Power source

212‧‧‧ sensor

214‧‧‧ Circuitry

220‧‧‧ output

300‧‧‧Power Generation System

304‧‧‧matching network

306‧‧‧ Plasma

308‧‧‧The plasma chamber

310‧‧‧Power source

312‧‧‧ sensor

314‧‧‧ Circuitry

316‧‧‧ frequency control module

318‧‧‧Overall Optimized Identification Module

320‧‧‧ Output

322‧‧‧ filter

400‧‧‧Power Generation System

404‧‧‧matching network

406‧‧‧ Plasma

408‧‧‧The plasma chamber

411‧‧‧Low-level source

412, 413‧‧ ‧ sensors

414‧‧‧ Circuitry

416‧‧‧frequency control module

418‧‧‧Overall Optimized Identification Module

420‧‧‧ output

422, 423‧‧‧ filter

424‧‧‧ combiner

500‧‧‧Power Generation System

510‧‧‧ main source of signals

511‧‧‧ secondary signal source

524‧‧‧ combiner

550‧‧‧Power Amplifier

600‧‧‧Power Generation System

610‧‧‧Power source

613‧‧‧Miscellaneous sources

620‧‧‧ output

624‧‧‧ combiner

700‧‧‧Power Generation System

710‧‧‧ main frequency source

713‧‧‧Source of noise

724‧‧‧ combiner

750‧‧‧Power Amplifier

801‧‧‧Real performance metrics

802‧‧‧ Estimated performance metrics

1600‧‧‧Method of the invention

1602~1610‧‧‧ Method steps of the invention

1710‧‧‧ Directional Coupler

1720‧‧‧ Analog-to-digital converter

1730‧‧‧ Filter

1810‧‧‧ demodulator

1820‧‧‧Signal A

1830‧‧‧Signal B

1840‧‧‧ Filter

Multiply cosine and sine functions by 1850‧‧

1900‧‧‧ Implementing the physical components of the overall optimized recognition module and frequency control module

1912‧‧‧ Display section

1920‧‧‧ Non-volatile memory

1922‧‧ ‧ busbar

1924‧‧‧ Random Access Memory (RAM)

1926‧‧‧Processing section

1927‧‧‧ Field Programmable Gate Array (FPGA)

1928‧‧‧Transceiver components

A ₁ ~A _N ‧‧‧ The complex vector representation of signal A

The complex vector representation of B ₁ ~B _N ‧‧‧ signal B

The various objects and advantages of the present invention, as well as the more complete understanding of the invention, may be

Figure 1 illustrates an exemplary power generation system constructed to automatically adjust the frequency of power delivered to the plasma load.

2 illustrates an embodiment of a power generation system in which a sensor resides in a power generation system along with a power source and one or more circuits.

Figure 3 illustrates a specific aspect of a power generation system in which the sensor resides outside of the power generation system.

Figure 4 illustrates a further embodiment of a power generating system.

Figure 5 illustrates a particular aspect of a power generation system in which the primary power signal and the secondary power signal are combined prior to amplification by the power amplifier.

Figure 6 illustrates a specific aspect of a power generation system in which the power source is generated primarily The power signal, and the noise source produces a secondary power signal in the form of a noise.

Figure 7 illustrates a particular aspect of a power generation system in which the secondary signal is noise and the primary power signal and the secondary power signal are combined prior to amplification by the power amplifier.

Figure 8 shows a graph of performance metrics as a function of frequency.

Figure 9A is a graph showing performance metrics, such as reflection coefficients, as a function of frequency.

Figure 9B is a graphical representation showing how the primary power signal frequency can be adjusted to minimize the performance metric shown in Figure 9A.

Figure 9C shows the spectrum of the power production system output at time t ₂ of Figure 9B (power per bandwidth, such as watts per 3 kilohertz bandwidth).

Figure 10A is a graph showing performance metric versus frequency.

Figure 10B is a graph showing how the overall search using the primary power signal can cause the plasma to go out.

Figure 10C is a spectrum diagram showing the output of the power generating system at time t ₂ of Figure 10B.

Figure 11A is a graph showing an optimized frequency estimate using a secondary power signal.

Fig. 11B is a graph showing the adjustment of the main frequency after determining the desired frequency using the secondary power signal.

Figure 11C is a graph showing the power spectral components of the primary and secondary signals of Figure 11B.

Figure 12A is a graph showing an optimized frequency estimate using a secondary power signal.

Figure 12B is a graph showing the adjustment of the primary frequency after determining the desired frequency using the secondary power signal.

Figure 12C is a graph showing the power spectral components of the primary and secondary signals of Figure 12B.

Figure 13A is a graph showing an optimized frequency estimate using a secondary power signal.

Figure 13B is a graph showing noise power as a function of time with noise added to the power generation system output.

Figure 13C is a spectrum diagram showing the output of the power generating system at time t ₂ of Figure 13B.

Figure 14A is a diagram showing aspects of a frequency adjustment method.

Fig. 14B is a diagram showing an additional aspect of the frequency adjustment method shown in Fig. 14A.

Figure 14C is a diagram showing a further aspect of the frequency adjustment method shown in Figures 14A and 14B.

Figure 14D is a graph showing still further aspects of the frequency adjustment method shown in Figures 14A, 14B, and 14C.

Fig. 15A is a diagram showing aspects of a frequency adjustment method.

Fig. 15B is a diagram showing an additional aspect of the frequency adjustment method shown in Fig. 15A.

Figure 15C is a graph showing a further aspect of the frequency adjustment method shown in Figures 15A and 15B.

Figure 16 illustrates a method of adjusting the frequency of a power generating system that can be referenced back and forth to the specific aspects described herein.

Figure 17A is a diagram showing an exemplary sensor.

Figure 17B is a diagram showing another specific aspect of the sensor.

Figure 17C is a diagram showing still another specific aspect of the sensor.

Figure 18 is a diagram showing aspects of an exemplary recognition module.

19 is a block diagram showing components that can be used to implement the specific aspects disclosed herein.

The word "exemplary" is used herein to mean "as an example, example or demonstration." Any specific aspect described herein as "exemplary" is not necessarily interpreted as being better or better than other specific aspects.

For the purposes of this disclosure, a "low level signal" is a signal that is substantially lower than the primary signal being passed to the plasma chamber, for example, at least an order of magnitude smaller.

For the purposes of this disclosure, a "circuit" can include any combination of electrical components that produce an output signal based on an input signal. The circuitry can be digital, analog, or include a processor or a central processing unit (CPU) or portion thereof. The circuitry can include or can read processor readable instructions from a non-transitory, physical computer readable storage medium to perform the methods described below.

For the purposes of this disclosure, components may be in communication, which in some cases includes electrical communication (e.g., capable of transmitting signals therebetween). However, to name two non-limiting examples, those skilled in the art will recognize that communications may also include optical communications and wireless radio communications.

For the purposes of this disclosure, "global optimum" may include a minimum or maximum value of a feature as sampled across a range of frequencies. For example, when the reflected power is characteristic, the overall optimization may be the overall minimum; when the delivered power is characteristic, the overall optimization may be the overall maximum.

The present disclosure relates generally to power generator systems that are constructed and applied (Additional power to the plasma at the primary frequency) Secondary power signals (such as including one or more frequencies) whose power is much lower than the plasma maintenance power. Advantageously, applying a secondary power signal is capable of monitoring one or more aspects of the plasma load without adversely affecting the plasma load itself. Incidentally, when the plasma sustaining power is applied to the plasma load via the matching network, the application of the low level signal can be applied by one or more special frequencies, which leads to the narrow filter band of the matching network. Detectable frequencies (such as mixing and intermodulation frequency). Again, the information obtained regarding the plasma load can be used to control one or more aspects of the generator.

For the control of the generator, for example, automatic frequency adjustment can be performed using information about the plasma load. In particular, an overall optimization of a certain performance metric can be obtained, and the generator can adjust towards this overall optimization frequency without extinguishing the plasma. Two exemplary practices include: (1) processing the noise generated by the generator's primary operating frequency to effectively perform low-power sampling to sweep through the frequency range of interest; or (2) generating additional signal to the primary power signal Low power signal, where low power signals are used to search for overall optimization.

In both cases, the low-power nature of the noise or search signal can detect the frequency range while keeping the generator's main power signal at a sufficient power to pass to the plasma load to maintain the frequency of the plasma (eg, in the performance metric) Local optimization). For example, the primary power signal can be maintained at or near partial optimization, and the search signal or noise (both of which will be referred to hereinafter as "secondary power signal") is sought for overall optimization, thereby allowing The search occurs while allowing substantial power to reach the plasma load.

In the case where the secondary power signal is noise, the noise may be due to the inherent noise caused by the main power signal, or the noise may be added to the main power signal. Noise can occur at multiple secondary frequencies, which are sometimes limited by the frequency controlled by the filter applied to the primary power signal. width. When the secondary power signal is a low level signal, the size of such a signal can be several orders of magnitude lower than the amplitude of the main power signal (for example, -3 dB, -5 dB, -10 dB, -20 dB, -50 dB, - 100 decibels). The low level signal can be sinusoidal or any other type of periodic signal and can be generated at radio frequency (RF) or other frequencies. Signals that begin to singularly and eventually become sinusoidal or periodic are considered sinusoidal or periodic, respectively. The low level signal can sweep over a fixed range of secondary frequencies. Alternatively, the low level signal can be "hopped" between the secondary frequencies based on the adjustment algorithm seeking an overall optimization.

Overall optimization can be found by comparing the degree of optimization of different frequencies and selecting the best frequency. For example, if the measure of the degree of optimization is the minimum load reflection coefficient size, the estimated load reflection coefficient at different frequencies searched by the secondary power signal source is compared, and the load reflection coefficient is selected to be the smallest frequency. As an overall optimization frequency. Measurements and comparisons to find optimizations can occur sequentially; or, for example, when using noise as a secondary power signal, the degree of optimization of different frequencies can be calculated simultaneously, and the best is selected after calculation of different frequencies. frequency.

Once the overall optimization has been found, the primary power signal can be shifted to the overall optimized frequency. Such an offset may involve abrupt switching from one frequency to another, or may involve a power ramp of the secondary power signal and a power slump of the primary power signal such that the secondary power signal becomes the primary power signal.

Further fine adjustments can occur once the primary power signal is operating at an overall optimized frequency. For example, the secondary power signal can come out again to search for overall optimization because the overall optimization of the power level of the primary power signal is different from the overall optimization of the lower power of the secondary power signal, Or because the overall optimization has changed and since The first time the adjustment was repeated, it has changed.

Figure 1 illustrates an exemplary power generation system constructed to automatically adjust the frequency of power delivered to the plasma load. The power generation system 100 is configured to provide RF power to the plasma 106 via a radio frequency (RF) impedance matching circuit (which may be an optional filter 122 internal to the power source 110 and/or a matching network 104 external to the power source 110) Plasma load. Filtering and impedance matching are often performed by the same physical network. Therefore, a filter (such as the optional filter 122) can perform both filtering and impedance matching functions.

Power generation system 100 can include a power source 110 that converts external power 140 to RF power, and power source 110 can be a 13.56 megahertz generator, although this is of course not required. Imagine other frequencies and other power sources 100. The power generation system 100 is constructed to provide sufficient power (e.g., RF voltage) to ignite and maintain the plasma 106 contained in the plasma chamber 108. The plasma 106 is generally used to process a workpiece or substrate (not shown) but is well known to those skilled in the art.

Power source 110 can apply primary power signals to the output primarily at the primary frequency. The output can be configured to couple to an optional matching network 104 and to the plasma chamber 108. In particular, the primary power signal can be delivered to the plasma 106 or to the load of the plasma 106 (also known as a plasma load). The connection(s) 130 from the power source 110 to the optional matching network 104 are often coaxial cables, although other cable types and connection types are also possible. The connection(s) 131 from the matching network 104 to the plasma chamber 108 are often made via custom coaxial connectors, although other cable types and connection types are also possible. For some applications, there is no matching network 104 and the power source 110 is directly connected to the plasma chamber 108. In this case, the RF impedance matching is done inside the power source 110 with an optional filter 122. For some applications, others can An optional RF or direct current (DC) generator 150 can be coupled to the plasma chamber 108 via an optional matching network 104. For some applications, other optional RF or DC generators 151 may be coupled to the plasma chamber 108 via other means, such as other optional matching networks 105. Connecting the other generators to the plasma load via the matching network(s) 104 or by other means (eg, connecting to different electrodes to transfer power to the same plasma) generally complicates the frequency adjustment problem. As described below, although the possibilities of other optional generator(s) 150, 151 and other means of connecting to the plasma (eg, matching network 105(s)) are not excluded, no further demonstration will be made for the sake of brevity. Or discuss.

The sensor 112 can monitor features indicative of the power delivered by the generator or the ability to deliver power, such as only three non-limiting examples: reflected power, delivered power, or impedance mismatch. Further non-limiting examples of features indicative of transfer power or transfer power capability include power delivered to matching network 104, power reflected from matching network 104, power delivered to plasma chamber 108, as seen by power generation system 100 The resulting load impedance, characteristics of the plasma chamber 108 (e.g., plasma density). The sensor 112 can also monitor features indicative of the stability of the plasma system, such as fluctuations in load impedance. The sensor 112 can also monitor features indicative of non-linearity of the plasma load, such as the generation of mixing and intermodulation products.

The additional benefit of using a secondary signal source to implement the frequency adjustment of the generator is that the measurement of the plasma properties can be made from the generator. The optional matching network 104 typically functions as a bandpass filter. This property of the matching network 104 makes it difficult to reliably measure the plasma at the harmonics of the generator output frequency, although such information may be useful. However, the characteristics of the plasma impedance modulation can be determined by observing the mixed and interactive modulation products produced by the secondary signal source. For example, if the main source of the signal is at 13.56 megahertz and the secondary news The source is at 13.57 megahertz, then we expect the mixed product to be at 13.55 megahertz and the interactive modulation product at 13.56 megahertz plus multiple 10 kHz, for example at 13.53, 13.54, 13.58 ... megahertz. Measuring the amplitude and phase relationships of the mixed and interactive modulation products and deriving, for example, the amount of amplitude and phase modulation present, can provide information about the properties of the plasma. The processing of information can be done in a number of ways: from simply analyzing the time series measurements from the sensor and making higher-order statistics on the time series, to using a dedicated receiver that adjusts to the mixed and interactive modulation product frequencies to capture the amplitude And phase relationships, to the use of any number of mathematical transformations (including but not limited to discrete Fourier transforms). It may be useful to monitor the mixing and intermodulation products and to detect changes in plasma characteristics as indicated by, for example, a phase modulation (for example, a property), such as for endpoint detection in an etch operation to fabricate a semiconductor.

The sensor 112 can be a directional coupler, a current voltage sensor, or other multi-turn network, and can monitor between the power source 110 and the matching network 104 or between the matching network 104 and the plasma chamber 108. Current and voltage or a combination of voltage and current (eg, incident and reflected signals). In another non-limiting example, the sensor 112 can optically detect or direct the plasma chamber 108 to optically measure the density of the plasma 106. These examples are by no means a description of the range or limitation of the position of the sensor 112 or the position of the sensor 112. Instead, the exemplary sensor 112 can take a wide variety of forms and can be coupled in a variety of ways. System (see the various non-limiting examples of Figures 2-7). Incidentally, the sensor 112 can be one or more sensors that have resided in the optional matching network(s) 104 or plasma chamber 108.

Signals from the sensor 112 or sensors that have resided in the matching network 104 and the plasma chamber 108 may be provided to one or more circuits 114 that are also controlled in communication with the power source 110. It. One or more circuits 114 may be used from sensor 112 and/or have been Information about the sensors resident in the matching network 104 and the plasma chamber 108 is adjusted to adjust the primary and/or secondary frequencies at which the power source 110 operates to optimize the power delivered to the plasma 106. Optimize or optimize the measure of another degree of optimization (eg, plasma stability).

In some cases, because such adjustments result in locally optimized operations (such as just two examples of local minimization of reflected power or local maximum of transmitted power), some adjustment algorithms can further adjust the primary frequency so that Seek overall optimization (for example, through a series of fast "frequency hopping"). However, such a search can be made by poor impedance matching of the power spectral region (for example FIG. 8 in the vicinity of f _a), so that the transmission power can fall significantly, and in some cases can cause plasma 106 is turned off (for example, in FIG. 8 f _a).

In order to avoid this, such seeking overall optimization can be performed by one or more secondary signals, thereby enabling the high power primary power signal to be maintained at a sufficient power to be delivered to the plasma 106 while seeking the overall overall The frequency of optimization (such as local optimization). Figures 11-13 show a graph of the monitored features as a function of frequency and how the secondary power signals showing amplitudes substantially lower than the primary power signal can be used to search for overall optimization. Once the relevant systems and devices have been described, these graphics will be discussed in depth later.

Figure 1 illustrates an exemplary power generation system for automatically adjusting the frequency of power delivered to a plasma load. The power source 110 can provide a primary power signal to the plasma load of the plasma 106 in the plasma processing chamber 108, wherein the impedance seen by the power source 110 is arranged between the power source 110 and the plasma chamber 108. Matching network 104 and impedance matching by power source 110 are impedance matched. While power source 110 can be frequency adjusted to find an optimized frequency, the delivered power is typically optimized here, but other metrics of optimization can be used. Such adjustments can sometimes cause the primary power signal from power source 110 to be adjusted to Local optimization rather than overall optimization. In such a case, one or more secondary signals may be generated by power source 110 and processed to identify overall optimization without having to use the primary power signal to find an overall optimization.

In other cases, the secondary power source can provide a secondary power signal (see, for example, Figures 4 and 6). One or more secondary power signals may be provided at an amplitude or power level below the primary power signal (either substantially below the primary power signal, a fraction of the primary power signal, or at a substantially lower power level) The effect on the plasma 106 is negligible compared to the main power signal). The one or more secondary power signals may include a plurality of secondary frequencies that are all generated at the same time (see Figures 11-13). Alternatively, one or more secondary power signals can be adjusted to two or more different frequencies at different times (as shown in Figures 11-13).

One or more secondary power signals can be used to sample power delivered at frequencies other than the primary power signal without applying a lot of power to the secondary frequencies to affect the plasma. In other words, the primary power signal can be maintained at a frequency at which plasma can be maintained (eg, at or near local optimization) while one or more secondary power signals are used to seek overall optimization.

In particular, the sensor 112, or two or more sensors, and/or sensors already present in other components of the power generation system 100 can monitor performance at the frequency of the primary power signal as well as at the secondary frequency. Metrics. One or more sensors (e.g., sensor 112) may also be measured at the frequency of the expected blending and intermodulation products to extract information about the nonlinear characteristics of the plasma 108. For example, changes in mixing and intermodulation products can be used to sense plasma ignition or endpoint detection of a plasma process. Injecting one or more secondary frequency components and measuring the properties of the mixed and intermodulation products can sense the nonlinear characteristics of the plasma 108 under the harmonics of the primary power signal, even with the matching network 104 and filtering The 122 may not allow direct measurement of harmonics.

For example, sensor 112 can be a reflected power sensor or a transmitted power sensor, and the features can be reflected power or transmitted power, respectively. Other features may also be monitored and used to identify local and overall optimizations (such as to name a few non-limiting examples: load impedance seen by power source 110, supply cable 130 to matching network 104(s)) The voltage and current on, the density of the plasma 106). The sensor 112 and/or other sensors may provide information describing the feature(s) to one or more circuits 114 (eg, logic circuits, digital circuits, analog circuits, non-transitory computer readable media) And combinations of the above). One or more circuits 114 may be in communication (e.g., electrical communication) with the sensor 112 and the power source 110. The one or more circuits 114 may adjust the primary frequency of the power source 110 to adjust the power source 110 to optimize the power delivered by the plasma load.

In some embodiments, the optimization of performance metrics includes controlling the feedback loop, which uses secondary power signals to search or seek overall optimization. In this case, based on feedback from the performance metrics from the sensor 112 (or two or more sensors and/or sensors already present in other components of the power generation system 100), More circuits 114 can control the secondary power signal and its one or more secondary frequencies. For example, the frequency of the secondary power signal can sweep over a fixed frequency range that encompasses the primary frequency of the primary power signal, and one or more circuits 114 can monitor the performance metric as a function of the frequency of the secondary power signal. Based on this sweep, one or more circuits 114 can identify the overall optimization and then instruct the power source 110 to adjust its primary frequency, such as to move the primary power signal to the identified overall optimization. Frequency hopping or other adjustment schemes can be used to find overall optimization via one or more secondary power signals.

Secondary power signals can take many different forms. In one situation, one or more The plurality of circuits 114 may instruct the power source 110 to apply a secondary power signal in the form of a low level signal at a secondary frequency (as shown in FIG. 11) or a plurality of secondary frequencies (as shown in FIG. 12). The order is to apply a low level signal at those secondary frequencies (see Figure 11), or to optimize the performance metric based on the algorithm (Figure 12). In another scenario, the one or more circuits 114 can instruct the power source 110 to apply a secondary power signal in the form of a noise. The noise may be inherent to the primary power signal, in which case one or more of the circuits 114 are not necessarily supplying instructions to the power source 110; or may be non-inherent noise added to the output of the power source 110 (see figure 6 and 7).

Regardless of the form in which the secondary power signal is presented, in many instances, the amplitude is one or more orders of magnitude lower than the primary power signal. For example, the secondary power signal can be 1 to 100 decibels lower than the primary power signal. In other embodiments, the secondary power signal can be 1 dB, 5 dB, 10 dB, 20 dB, 50 dB, or 100 dB lower than the primary power signal.

As shown, the one or more circuits 114 can include an overall optimization identification module 116 and a frequency control module 118. The overall optimization recognition module 116 can analyze information from the sensor 112 for each of the one or more secondary frequencies and identify frequencies corresponding to the overall optimization. This frequency may be referred to as the identified overall optimization frequency, and it corresponds to an overall optimization of the characteristics of the generator delivery power. The frequency control module 118 may adjust the primary frequency toward the identification during initial adjustment of the primary power signal (which may result in identifying local optimizations) and once the overall optimization is identified by the overall optimization identification module 116 The main frequency of the main power signal is adjusted during the period of the overall optimization frequency.

In particular, once the identified overall optimization frequency is identified, the frequency control module 118 can instruct the power source 110 to adjust the primary frequency to jump to the identified overall optimized frequency, or to jump to the lower amplitude primary frequency while increasing The amplitude of the secondary frequency at the identified overall optimized frequency is such that the roles of the primary and secondary frequencies are reversed. In this way, the primary frequency can be shifted to a frequency corresponding to the overall optimization of the power characteristics (such as low reflected power or low oscillating level) without suppressing or extinguishing the spectral region of the plasma (see Figures 8-13). The vicinity of f _a ) to apply power.

The overall optimization recognition module 116 and the frequency control module 118 can be cycled to repeatedly improve the accuracy of optimizing the primary frequency adjustment toward the overall. For example, where the feature being monitored (such as plasma impedance) is non-linear, the overall minimum of the feature can be found when a low level secondary power signal is applied; but when a much larger main power signal is applied at the same frequency Higher power signals can have different overall optimization frequencies. As such, the secondary power signal can be used again to further explore the overall optimization of the primary power signal, and this can be repeated in a round-robin fashion. Optimizing the frequency adjustment toward the overall may include changing the frequency to a frequency associated with the overall optimization, or simply changing the frequency to a frequency that is closer to the overall optimization than the original frequency.

In some embodiments, one or more secondary frequencies begin to fall/rise above a sufficiently steep portion of the frequency curve (eg, between f _a and f ₀ in Figures 8-14), and the primary frequency can be switched immediately One of one or more secondary frequencies. When a steep portion of such a curve is identified, the overall optimization recognition module 116 can determine that it is approaching the overall optimization, thereby instructing the power source 110 to switch the primary frequency to a frequency near the secondary power signal, thereby enabling The main power signal is skipped and avoided or the frequency curve region of the plasma can be suppressed (eg, near f _a ). Once the primary power signal switches frequency, one or more secondary power signals can then be explored for overall optimization, or the primary power signal can be used to further explore overall optimization.

In many aspects, the supply connection(s) 130 can be implemented by a pair of conductors or by a collection of two conductor coaxial cables connecting the power source 110 to the matching network 104. In other specific aspects, the cable 130 is implemented by one or more twisted pairs of cables. In still other specific aspects, cable 130 can be implemented by any cable network including, but not limited to, simple conductor wiring and quadrupole connections. Although the connection(s) 131 are often implemented with connectors, they can take a wide variety of forms, including simple conductor wiring.

Matching network 104 can be implemented by a wide variety of matching network architectures. It will be appreciated by one of ordinary skill in the art that the matching network 104 can be used to match the load of the plasma 106 to the power source 110. By properly designing the matching network 104 or 105, it is possible to convert the load impedance of the plasma 106 to a desired load impedance value close to the power source 110. The correct design of the matching network 104 or 105 may include a matching network within the power source 110 (e.g., via the filter 122) or a matching network external to the power source 110, as seen in Figures 1-7.

One or more of the circuits 114 may be the original equipment of the power generation system 100; and in other specific aspects, the one or more circuits 114 may be refurbished components that may be added to frequencies that would otherwise not be described herein Adjusted power generation system.

In a particular aspect, power generation system 100 can include an optional filter 122. Filter 122 can be constructed to attenuate portions of the main power signal outside of the selected bandwidth and make additional impedance matching. For example, because 50 ohms is the master impedance for the cable and connector 130, the desired impedance seen at the output of power source 110 is typically 50 ohms or some other convenient impedance. . The impedance at the input of filter 122 (on the opposite side of the output of power source 110) is the desired impedance provided by the active component of the power source, such as a metal oxide semiconductor field effect transistor (MOSFET), and typically Very different from 50 ohms, for example Typical 5+j6 ohms as a single MOSFET amplifier. For such a system, filter 122 will be designed to match 50 ohms at the output to 5+j6 ohms at the input. In addition to impedance matching, filters are typically designed to limit the harmonics produced by the active components. For example, the filter can be designed to match the 50 ohms at the output to a value close to 5+j6 and suppress the second or third at the output in the frequency range in which the generator is expected to operate (eg, from 12.882 to 14.238 megahertz). Signals with harmonics above the 25 megahertz frequency amount to a certain amount, typically at least 20 decibels.

The sensor 112 can be arranged in a wide variety of locations, including the location of a portion of the power generation system 100 and the location externally. The location of the sensor 112 to monitor the feature may also vary with different specific aspects, as will be seen in Figures 2-7.

2 illustrates a specific aspect of power generation system 200 in which sensor 212 resides in power generation system 200 along with power source 210 and one or more circuits 214. The power generation system 200 includes an output 220 that is configured to be coupled to the optional matching network(s) 204 or directly coupled to the plasma chamber 208 if the matching network(s) 204 are not present. Thus, the primary power signal and one or more secondary power signals can be provided to the output 220 and thus constructed to be passed to the matching network(s) 204.

FIG. 3 illustrates a specific aspect of power generation system 300 in which sensor 312 resides outside of power generation system 300. Here, power generation system 300 includes a power source 310, one or more circuits 314, an optional filter 322, and an output 320 of power generation system 300. Sensor 312 is coupled to one or more circuits 314 and provides information describing performance metrics (eg, load reflection coefficient magnitude or plasma density). The sensor 312 monitors between the power generation system 300 and the optional matching network(s) 304, between the matching network(s) 304 and the plasma chamber 308, in the plasma chamber 308, or power Between system 300 and plasma chamber 308 (if (multiple) match The characteristics of the network 304 do not exist. Sensor 312 may also be monitored within or within matching network(s) 304.

Although Figures 1-3 demonstrate a single power source 110, 210, 310, those skilled in the art will recognize that the power source 110, 210, 310 can simultaneously generate both primary and secondary power signals. For example, for two non-limiting examples, the power sources 110, 210, 310 can originate both the high power primary power signal and the low level secondary power signal, or the power sources 110, 210, 310 can generate high power. The main power signal and the noise inherent to the main power signal are used as the secondary power signal. Alternatively, the power source 110, 210, 310 can generate a primary power signal and combine it with the noise generated or amplified. Although these examples demonstrate how a single power source 110, 210, 310 can generate both a primary power signal and a secondary power signal, the specific aspects that will be exemplified in Figures 4-7 are that the power source produces a primary power signal and the low level The source of the signal produces a secondary power signal.

4 illustrates a particular aspect of power generation system 400 having a power source 410, a low level signal source 411, one or more circuits 414, an optional sensor 412 that may be arranged in power generation system 400, or may An optional sensor 413 disposed outside of the power generation system 400, a combiner 424 that combines outputs from the power source 410 and the low level signal source 411. It will be appreciated by those of ordinary skill in the art that the combiner can be implemented by a coupler known in the art.

FIG. 5 illustrates a particular aspect of power generation system 500 in which the primary and secondary signals are combined prior to amplification by power amplifier 550.

6 illustrates a particular aspect of power generation system 600 in which power source 610 generates a primary power signal and noise source 613 produces a secondary power signal in the form of a noise. The primary power signal and the secondary power signal or noise may be combined in the power generation system 600, and the combined signals may be provided to the output 620 of the power generation system 600. It will be appreciated by those of ordinary skill in the art that the noise source 613 can be implemented by a wide variety of different types of devices, including noise diodes. Advantageously, the noise source 613 can generate a continuous secondary frequency, and the secondary frequency response can be processed in parallel at a plurality of different frequencies (eg, by multiple demodulation channels or fast Fourier transform modules) . For example, the reflection coefficients at multiple frequencies can arrive in parallel to identify a frequency that provides a low reflection coefficient, a stable frequency, or a balance between stability and low reflection coefficient.

FIG. 7 illustrates a particular aspect of power generation system 700 in which the primary and secondary signals are combined prior to amplification by power amplifier 750. In this specific aspect, the secondary signal is generated by the noise source 713.

The system illustrated in Figures 1-7 can be more easily understood with reference to the figures seen in Figures 8-15.

Figure 8 shows a graph of performance metrics as a function of frequency. The solid line 801 shows the true performance metric (such as the magnitude of the load reflection coefficient) as a function of frequency, which is produced if the primary power signal is adjusted to each frequency and measurements are taken. Dotted line 802 shows the estimated performance metric, using one or more of the main secondary power signal and the power signal at a constant frequency (for example f ₁₎ is obtained.

As discussed, the power level of the primary frequency affects the measure of performance (such as the load reflection coefficient); therefore, the performance metric estimated using the low level power signal will be different from the performance metric at the higher power of the primary signal. However, as discussed further herein, the low level signal can closely estimate the desired primary frequency (e.g., it produces low reflection coefficients and/or low instability). Main signal The frequency can then be fine-tuned at a higher power level without testing the frequency that can cause the plasma to go out.

Figure 9 shows an aspect in which the initial dominant frequency can be applied between f ₁ and f _a and shows how the frequency adjustment algorithm (which depends on the frequency of sweeping and testing the main power) can become a local best trapped in the performance metric. In the middle, there is no information provided by the low-power secondary signal. More specifically, the adjustment algorithm can adjust the primary frequency to a frequency f ₁ that is believed to be optimized. In particular, Figure 9A shows a measure of performance (such as the reflection coefficient) as a function of frequency; the solid line of Figure 9B shows the algorithm using only the primary power or how the primary power signal frequency can be adjusted to minimize the measure of performance; 9C shows the spectrum of the power generation system output 220, 320, 420, 520, 620 or 720 at time t ₂ of Figure 9B (power per bandwidth, such as watts per 3 kilohertz bandwidth). As indicated by the dashed line in Figure 9B, the overall optimized frequency may be identified using a low level secondary signal.

However, as shown by the solid line, when local optimization at f ₁ is reached, if the primary frequency is used to search for overall optimization, such an attempt may result in the application of power near the frequency f _a , which may result in a plasma It is extinguished as seen in Figures 10A and 10B. Figure 10A shows the measure of performance as a function of frequency. The solid line of Figure 10A shows the performance metric for the plasma point, and the dashed line shows the performance metric for the plasma to extinguish. Figure 10B shows how an overall search using primary power signals can cause plasma to extinguish, as insufficient power can be delivered near f _a to maintain the plasma. Figure 10C shows the spectrum of the power production system output at time t ₂ of Figure 10B.

One or more secondary power signals can instead be used to search for overall optimization, as shown in Figure 11 (showing a secondary power signal) and Figure 12 (showing multiple secondary power signals), while the primary power signal remains At a fixed frequency (such as at or near local optimization). Figure 11, shown The frequency adjustment is performed using a secondary power signal in the form of a low level signal and in a special order at a single frequency. The frequency adjustment shown in Figure 12 uses a secondary power signal in the form of a low level signal whose spectral components are optimized at a plurality of secondary frequencies adjusted according to the algorithm to optimize performance.

As shown, one or more secondary power signals can be applied at a power level well below the primary power signal and one or more secondary frequencies can be applied. The secondary frequencies may be fixed frequencies with equal or unequal spacing, or may be variable frequencies, as shown in FIG. In addition, primary and secondary power signals can be applied simultaneously.

As exemplified in FIG. 11, the secondary signal may be applied continuously, or as illustrated in FIG. 12, only during the period in which overall optimization is sought. In addition, although a single feature is shown in the graphs of Figures 8-13, in other embodiments, multiple features can be monitored simultaneously, such as the magnitude of the load reflection coefficient and the measured plasma stability (such as fluctuations in load impedance). And analysis of all monitoring features (or multiple monitoring features) can be used to identify overall optimization. In this way, overall optimization is identified without applying a full power main signal near f _a or any frequency at or quenching the plasma.

In some modes of operation, the amplitude of one or more secondary power signals applied to one or more secondary frequencies is so small that it can be considered negligible compared to the primary power signal, thus The slurry has no significant effect. For other applications, if the goal is simply to seek overall optimization without extinguishing the plasma, the amplitude of one or more secondary power signals can be significant compared to the primary power signal. In this case, care must be taken not to exceed the voltage and current ratings of the plasma system because of the high gain amplitude at the beat frequency.

Figure 11 shows a specific aspect in which the frequency is continuously swept through the frequency range. The range swept by the secondary frequency (typically) will typically be the frequency range in which the power generation system is expected to operate (eg, 12.882 to 14.238 megahertz), but this need not be the case. Examples of other frequency ranges that may be considered include information on plasma conditions when using secondary power signals, for example by analyzing mixed and interactive modulation products. In other scenarios as exemplified in FIG. 12, one or more secondary frequencies may be adjusted according to an algorithm to find an optimized frequency instead of being swept in a predetermined pattern as shown in FIG. As also shown in FIG. 12, once the overall optimization has been identified, the secondary power signal can be turned off instead of being continuously applied as shown in FIG.

As exemplified in Figures 11A and 12A, the optimized frequency estimate made using one or more secondary power signals may not exactly correspond to the true optimization. Typically, this divergence is due to the nonlinearity of the plasma load. As exemplified in Figures 11B and 12B, after the secondary power signal is used to determine the optimized frequency, the primary frequency can be adjusted to further optimize performance. Figures 11C and 12C show the spectral components of the primary and secondary frequencies of Figures 11B and 11C, respectively.

Figure 13 shows the case where the secondary power signal is a noise. Fig. 13C shows the spectrum of the power generation system output at time t ₂ of Fig. 13B. The noise can be inherent to the main power signal or can be added to the power generation system output (see Figures 6 and 7 for example). Figure 13B shows the noise power as a function of time, and the hypothetical situation is that noise is added to the power generation system output.

Once the overall optimization has been identified, the primary power signal can be adjusted or switched (or directed) to the frequency corresponding to the overall optimization without passing the primary power signal or suppressing the spectral region of the plasma (eg, near f) _a ). For example, in Figure 14, the amplitude of the primary power signal is ramped down, while the amplitude of the secondary frequency that is optimized overall is ramping up. In this way, the primary power signal and the secondary power signal exchange positions. Figure 15 shows another variation of the primary frequency switching towards overall optimization, where the frequency of the primary power signal suddenly changes to the identified overall optimized frequency.

In some specific aspects, although the overall optimized frequency of the identification can be selected from a certain frequency, this is not necessary. For example, the identified overall optimization frequency can be between two or more secondary frequencies. For example, interpolation between multiple secondary frequencies can be used to identify the overall optimized frequency of the identification.

16 illustrates a method of adjusting the frequency of a power generation system that uses a secondary power signal to find an overall optimization to explore an overall optimization of performance metrics. The method 1600 primarily applies a primary power signal to the plasma system at a primary frequency (e.g., to the matching network 104(s) connected to the plasma chamber 108) (block 1602). At the same time, method 1600 applies a low level signal to the plasma system in one or more or consecutive (eg, the case of a noise) secondary frequency (block 1604).

The low level signal may be the sum of periodic or periodic signals, may be noise inherent to the primary power signal, or may be noise added to the primary power signal. The one or more secondary frequencies may have equal frequency intervals or may have varying intervals. One or more secondary frequencies may be applied all at once or at separate times and may be adjusted at any time. One or more secondary frequencies can be swept across a fixed frequency range. Alternatively, one or more secondary frequencies may be adjusted via feedback to detect and explore overall optimization. One or more secondary or continuous secondary frequencies may be applied or applied only as needed.

The feature monitored by method 1600 is a measure of the performance of the frequency function (such as the magnitude of the load reflection coefficient), especially one or more or consecutive secondary frequencies and/or expectations at the primary frequency and/or at the primary and secondary frequencies. The blending and interactive modulation products (block 1606). The method 1600 then identifies an optimization frequency corresponding to the overall optimization of the feature (block 1608). This can be done by minimizing or maximizing the algorithm familiar to those skilled in the art. Finally, method 1600 adjusts the primary frequency of the primary power signal to the optimized frequency identified in the identification operation (block 1610). This adjustment can be done in a variety of ways. For example, the adjustment may have to avoid applying the main power for only an extended period of time in the region where the reflected power approaches 100% (譬 near f _a in Figure 8), as this may extinguish the plasma (unless, for example, the plasma is by another Power source 150 or 151 is maintained). Thus, for two non-limiting examples, the primary power signal can be switched to the optimized frequency, or the power levels of the primary and secondary power signals can be gradually reversed, such that the power signal exchanges positions.

In some embodiments, method 1600 ends when the primary power signal has been moved to a frequency that is identified as being generally optimized using one or more secondary power signals. However, in other examples, method 1600 may cycle to further fine tune optimization or be responsible for overall optimization changes due to, for example, non-linearity of the plasma load or parameters that may be changed at any time, such as plasma chamber gas pressure. .

The identification of the optimized frequency (block 1608) may occur immediately as the sample is obtained from monitoring (block 1606), or may occur after the frequency range has been sampled. The movement of the primary frequency may occur only once the overall optimization has been identified (block 1608) (block 1610), or it may occur once more optimized frequencies than the current primary frequency are identified.

The method of using secondary power signals to monitor features can also be used to identify changes in plasma characteristics or plasma characteristics. Instead of identifying the optimization frequency and overall optimization of the primary frequency adjustment toward the identification, the output or monitoring feature (block 1608) can instead be used to identify changes in the plasma or plasma characteristics. Monitoring mixed and interactive modulation products can be used to monitor the nonlinear behavior of the plasma, or simply to detect if the plasma is lit. Instead of observing special blending and intermodulation products, higher-order statistics (such as dual-spectrum) can be used to identify changes in plasma characteristics or plasma characteristics.

Figure 17 shows three exemplary implementations of a sensor (e.g., sensor 112 or 412) example. The sensor can be, for example, a directional coupler 1710 (as shown in Figure 17A) or a voltage and current (VI) sensor (as shown in Figure 17B), and any of these embodiments can include a filter 1730 and Analog to digital converter 1720 (as shown in Figure 17C).

Figure 18 shows an exemplary embodiment of an overall optimization recognition module (e.g., 116 or 418). Part of the functionality shown in Figure 18 can also be part of the sensor. The embodiment shown in Figure 18 uses a plurality of demodulators 1810 that allow multiple frequency components to be processed simultaneously. Signals 1820 (labeled A) and 1830 (labeled B) may be, for example, forward and reflected power, voltage and current, or some other interesting measurement. After multiplying the cosine and sine functions 1850 with the filter 1840, the complex vectors of A and B at different frequencies are denoted as A ₁ , B ₁ to A _N , B _N , which are used to calculate power and load reflections at multiple frequencies. coefficient. Typically, one channel will be reversed for the primary frequency. Other channels can be set to one or more secondary frequencies, or set to the expected blending and intermodulation products. As noted previously, this is only one embodiment, and by way of example, many other embodiments may use, for example, discrete Fourier transforms rather than dedicated demodulation channels.

The exemplary arrangement of the components shown in Figures 1-7 is logical, the connections between the various components are merely exemplary, and the display of these specific aspects is not intended to be a true hardware diagram; therefore, in the real embodiment The components can be combined or further separated, and the components can be connected in a variety of ways without changing the basic operation of the system.

Rather than a single secondary power source (as seen in Figures 4-7), two, three, four or more secondary power sources may instead be used to generate two or more secondary power signals.

For the purposes of this disclosure, the secondary power signal can be periodic, such as an RF signal. However, in other specific aspects, aperiodic power signals (such as noise) can be used.

Although the present disclosure has repeatedly shown adjustments for local and overall minima, those skilled in the art will appreciate that it is also tuned for local and overall maximums, and that the present disclosure can be readily applied to the primary frequencies at which power is delivered. The feature is monitored in the case where the overall maximum value of the feature is monitored and optimized.

The methods described in connection with the specific aspects disclosed herein can be directly implemented in hardware, processor-executable instructions encoded in non-transitory processor readable media, or a combination of both. For example, referring to FIG. 19, a block diagram is shown that shows the physical components that can be used to implement the overall optimization recognition module 116 and the frequency control module 118 in accordance with an exemplary embodiment. As shown, in this particular aspect, display portion 1912 and non-volatile memory 1920 are coupled to busbar 1922, which is also coupled to a random access memory (RAM) 1924, processing portion (which includes N). Processing components) 1926, field programmable gate array (FPGA) 1927, transceiver component 1928 (which includes N transceivers). Although the components shown in FIG. 19 represent solid components, FIG. 19 is not intended to be a detailed hardware diagram; thus many of the components shown in FIG. 19 may be implemented by a common architecture or dispersed among additional physical components. Furthermore, it is contemplated that other physical components and architectures that are either existing or yet to be developed can be used to implement the functional components described with reference to FIG.

The display portion 1912 is generally operative to provide a user interface for the user; in several embodiments, the display is implemented by a touch screen display. In general, non-volatile memory 1920 is a non-transitory memory that functions to store (e.g., persistently store) data and code executable by the processor (including executable code associated with methods that implement the methods described herein). In some specific aspects, for example, the non-volatile memory 1920 includes a boot load code, a job system code, a file system code, a non-transitory processor executable code to facilitate execution of the method described with reference to FIG. Methods and other methods described herein.

In many embodiments, although the non-volatile memory 1920 is implemented by flash memory (eg, NAND or ONENAND memory), it is contemplated that other memory types may be utilized as well. While it is possible to execute code from non-volatile memory 1920, executable code in non-volatile memory is typically loaded into RAM 1924 and is processed by one or more of the N processing elements in processing portion 1926. More to implement. Non-volatile memory 1920 or RAM 1924 can be used to store the overall optimized frequency as described in Figures 8-14.

The N processing components, along with the RAM 1924, generally operate to execute instructions stored in the non-volatile memory 1920 to enable the source impedance of the generator to be modified to achieve one or more purposes. For example, non-transitory processor-executable instructions that implement the methods described with reference to Figures 16 and 18 can be persistently stored in non-volatile memory 1920 and executed by N processing components associated with RAM 1924. As will be appreciated by one of ordinary skill in the art, processing portion 1926 can include a video processor, a digital signal processor (DSP), a graphics processing unit (GPU), and other processing components. The DSP can be used, for example, to employ discrete Fourier transforms to analyze specific aspects of various aspects of the generator output power indicative of the plasma load.

Additionally or alternatively, FPGA 1927 can be constructed to implement one or more aspects of the methods described herein (such as the methods described with reference to Figures 16 and 18). For example, non-transitory FPGA configuration instructions can be permanently stored in non-volatile memory 1920 and accessed by FPGA 1927 (eg, during power-on) to build FPGA 1927 into an overall optimized recognition mode. The functions of group 116 and frequency control module 118.

The input member is operative to receive signals (eg, from sensors 112, 312, 412, An output signal of 413) indicating one or more aspects of output power and/or plasma loading. The signals received at the input member may include, for example, voltage, current, forward power, reflected power, and plasma load impedance. It is contemplated that the input member can include both digital and analog input, and can include an analog to digital conversion component to convert the analog signal to a digital signal. The output member is generally operative to provide one or more analog or digital signals to achieve an operational aspect of the generator. For example, the output portion can provide a frequency control signal to the oscillator shown and described herein. It is also contemplated that a signal that controls the amplitude and phase of the applied power can also be output from the output member.

The illustrated transceiver component 1928 includes N transceiver chains that can be used to communicate with external devices via a wireless or wired network. Each of the N transceiver chains can represent a transceiver associated with a particular communication scheme (e.g., wireless fax (WiFi), Ethernet, Profibus, ...). The transceiver components can be used, for example, to communicate with one or more other devices associated with the plasma processing tool.

In this specification, the same reference characters are used to refer to terminals, signal lines, wires, and their corresponding signals. In this respect, the words "signal", "wire", "connection", "terminal", "pin" (pin) are always interchangeable in this manual. use. It should also be appreciated that "signals", "wires" or similar words may represent one or more signals, such as a single bit transmitted over a single wire or multiple parallel bits transmitted through multiple parallel wires. In addition, each wire or signal can represent a two-way communication between two or more components to which the signal or wire is connected, as can be the case.

The foregoing description of the specific aspects of the invention is intended to enable any of those skilled in the art to make and use the invention. Those skilled in the art will readily appreciate that there are numerous modifications to these specific aspects, and that the general principles defined herein can be applied to other specific aspects without departing from the present invention. The spirit and scope of the Ming. Therefore, the invention is not intended to be limited to the details of the details disclosed herein,

100‧‧‧Power Generation System

104, 105‧‧‧ Matching network

106‧‧‧ Plasma

108‧‧‧The plasma chamber

110‧‧‧Power source

112‧‧‧ sensor

114‧‧‧ Circuitry

116‧‧‧Overall Optimized Identification Module

118‧‧‧ frequency control module

122‧‧‧ filter

130, 131‧‧‧ Connection

140‧‧‧External power

150, 151‧‧‧RF or DC generator

Claims (23)

A power generation system constructed to automatically adjust a frequency of power delivered to a plasma load, the power generation system comprising: a power source that applies a primary power signal at a primary frequency to an output, the output being constructed directly or through a matching network The circuit is coupled to a plasma load, the power source being configured to apply one or more secondary power signals to the plasma load at one or more frequencies, wherein the power generated at the secondary frequencies is lower than The power generated at the primary frequency; a sensor arranged to sense one or more attributes of power delivered to the plasma load to obtain a measure of performance; an overall optimization identification module that analyzes the a measure of performance and identifying the overall optimized frequency as an overall optimization of the metric corresponding to the performance; a frequency control module that adjusts the primary frequency toward the overall identity of the identification corresponding to the overall optimization The frequency of optimization. A system as claimed in claim 1, wherein the power generated at the secondary frequencies is 1 to 100 decibels lower than the power at the primary frequency. A system as claimed in claim 1, wherein the sensor is a sensor selected from the group consisting of: a voltage and current sensor, a directional coupler. The system of claim 1, wherein the overall optimization of the measure of performance is a measure of the performance, which is less than or greater than all other samples in a bandwidth at which the frequency of the primary power signal can be adjusted. The metric value. The system of claim 1, wherein the measure of performance is a performance metric selected from the group consisting of: reflected power calculated relative to a desired reference impedance, the power The source sees how far the load impedance deviates from the desired impedance, and a measure of the magnitude of the load reflection coefficient. The system of claim 1, wherein the one or more secondary power signals are noise. A system of claim 6 wherein the source of power comprises a single oscillator amplifier combination to generate the primary power signal and the noise is inherent to the single oscillator amplifier combination. A system as claimed in claim 1, comprising a secondary oscillator; wherein the source of power comprises a primary oscillator that generates the primary power signal and the secondary oscillator that generates the secondary power signals. A non-transitory and physical computer readable storage medium encoding a machine readable command for performing automatic frequency adjustment of a power generation system, the method comprising: applying the primary frequency directly or through a matching network a primary power signal to a plasma load; applying a secondary power signal of one or more secondary frequencies to the plasma load, wherein the power generated at the secondary frequencies is lower than the power generated at the primary frequency Metricing a measure of performance; identifying an optimized frequency corresponding to the overall optimization of the performance metric; and adjusting the primary frequency toward the optimized frequency. A non-transitory and physical computer readable storage medium as claimed in claim 9 wherein the method is cyclically applied to repeatedly improve the accuracy of the primary frequency adjustment towards the overall optimization. Non-transitory and physical computer readable storage medium as claimed in claim 9 The secondary power signal is the sum of periodic or periodic signals. A non-transitory and physical computer readable storage medium as claimed in claim 9 wherein simultaneously applying the one or more secondary power signals comprises: sweeping the one or more secondary frequencies through a fixed Frequency Range. A non-transitory and physical computer readable storage medium as claimed in claim 9, wherein simultaneously applying the one or more secondary power signals comprises: adjusting a single power signal to a plurality of times at different times Want different people in the frequency. A non-transitory and physical computer readable storage medium as claimed in claim 9 wherein the primary frequency adjustment toward the optimized frequency comprises: reducing an amplitude of power applied to the primary frequency, and increasing the applied The amplitude of the power of the target frequency. A non-transitory and physical computer readable storage medium as claimed in claim 9 wherein the primary frequency adjustment toward the optimized frequency comprises: changing the primary frequency to a target frequency and near the target frequency Make additional adjustments. The non-transitory and physical computer readable storage medium of claim 9 wherein the overall optimization of the performance metric is a value of the performance metric that is less than or greater than a primary power signal that can be adjusted. The performance metric value of all other samples in the frequency bandwidth. A non-transitory and physical computer readable storage medium as claimed in claim 9 wherein the measure of performance is a performance metric selected from the group consisting of: a reflection calculated relative to a desired reference impedance Power, the source of power sees how far the load impedance deviates from the desired impedance, and a measure of the magnitude of the load reflection coefficient. A non-transitory and physical computer readable storage medium as claimed in claim 9 wherein the simultaneous application of one or more secondary power signals comprises: applying noise. A non-transitory and physical computer readable storage medium as claimed in claim 9 wherein the primary frequency adjustment toward the optimized frequency comprises: reducing an amplitude of power applied to the primary frequency, and increasing the applied The amplitude of the power of the target frequency. A non-transitory and physical computer readable storage medium as claimed in claim 9 wherein the primary frequency adjustment toward the optimized frequency comprises: changing the primary frequency to a target frequency and near the target frequency Make additional adjustments. A power generation system configured to automatically adjust a frequency of power delivered to a plasma load, the power generation system comprising: a power source that applies a primary power signal to an output, the output being configured to be coupled directly or through a matching network a plasma load that is constructed to form one or more secondary power signals applied to one or more or successive secondary frequencies to the plasma load, wherein the power generated at the secondary frequencies is low The power generated at the primary frequency; a sensor arranged to sense one or more attributes of power delivered to the plasma load to obtain a measure of performance; an overall optimization recognition module comprising The non-transitory and physical processor can read a storage medium that encodes instructions read by the processor or the field programmable gate array for automatic frequency adjustment of the power generation system, the method comprising: The power source directly or through the matching network to apply the main power signal at the primary frequency to the plasma load; simultaneously applying one or more or consecutive secondary frequencies Or more secondary plasma power signal to the load; monitoring performance metrics; Identifying an optimized frequency corresponding to the overall optimization of the performance metric; and adjusting the primary frequency toward the optimized frequency. A power generation system according to claim 21, comprising: at least one demodulation channel coupled to the sensor to obtain the performance metric at at least one particular frequency. A power generation system as in claim 21, comprising: a discrete Fourier transform module coupled to the sensor to obtain the performance metric at one or more particular frequencies.