Method and apparatus for processing substrates

Abstract

Methods and apparatuses for processing a substrate are provided herein, for example, a matching network configured for use with a plasma processing chamber, comprising an input configured to receive one or more radio frequency (RF) signals, an output configured to deliver one or more RF signals to the processing chamber, a first sensor operably connected to the input and a second sensor operably connected to the output and configured to measure impedance during operation, at least one variable capacitor connected to the first sensor and the second sensor, and a controller configured to tune the at least one variable capacitor of the matching network to a first target position based on weighted output impedance values ​​measured during a pulsed state of a voltage waveform and tune the at least one variable capacitor to a second target position based on weighted input impedance values ​​measured during a pulsed state of a voltage waveform based on the measured impedance.

Description

Technical Field

[0001] Embodiments of the present disclosure generally relate to methods and apparatuses for processing substrates, for example, in an RF matching network configured for use with a radio frequency (RF) processing chamber, using weighted average tuning to process substrates.

Background Art

[0002] Methods and apparatuses for processing substrates in a vacuum processing chamber using one or more of RF power supplies are known (for example, one or more RF power supplies can be configured for single level pulsing, dual level pulsing or multi-level pulsing). For example, in single level pulsing (for example, pulsing between an on state and an off state), there is only one state to adjust (for example, the on state). However, in dual level pulsing, the RF power supply is switched between a high state and a low state (the low state is not, for example, an off state). In multi-level pulsing, the RF power supply can be switched between multiple states.

[0003] RF matching networks are often connected between an RF power supply and a vacuum processing chamber and are configured to ensure that the output of the RF power supply is efficiently coupled to the plasma to maximize the amount of energy coupled to the plasma (for example, known as RF power supply tuning). For example, in dual-level pulsing, there are two or more impedance states that require impedance matching. Current RF matching networks are configured to use motorized capacitors (e.g., in series or shunt) to tune to one state over time average and to frequency tune to the other state in real time. However, frequency tuning is limited in impedance matching by single-axis tuning, which can result in process capability limitations and increased reflected power. Furthermore, plasma load impedance states can fluctuate due to pulsed power levels, e.g., bias power on and off, or pulsed voltage waveforms. The motor in RF matching cannot keep up with rapidly changing impedance states. For example, conventional RF matching networks are configured to match to the first state in multi-level pulsing. For example, with ultrafast pulsed signals or pulsed voltage waveforms on a microsecond timescale, frequency tuning may not keep up with plasma impedance fluctuations and may even fail to tune well within the pulse cycle.

[0004] Accordingly, the inventors herein provide an improved method and apparatus for processing a substrate using weighted average tuning in a radio frequency (RF) matched network configured for use with an RF processing chamber. [Overview of the project]

[0005] Methods and apparatus for processing substrates are provided herein. For example, in some embodiments, a matching network configured for use with a plasma processing chamber comprises an input configured to receive one or more radio frequency (RF) signals; an output configured to send one or more RF signals to the processing chamber; a first sensor operably connected to the input; a second sensor operably connected to the output and configured to measure impedance during operation; at least one variable capacitor connected to the first and second sensors; and a controller configured to tune at least one variable capacitor of the matching network to a first target position based on a weighted output impedance value measured in a pulsed state, and at least one variable capacitor to a second target position based on a weighted input impedance value measured in a pulsed state, based on the measured impedance.

[0006] According to at least some embodiments, a plasma processing chamber comprises a chamber body and a chamber lid, an RF source power connected to the chamber lid and configured to create plasma from a gas placed in the processing area of ​​the chamber body, one or more RF bias power supplies configured to sustain a plasma discharge, and a matching network, the matching network comprising an input configured to receive one or more radio frequency (RF) signals, an output configured to send one or more RF signals to the processing chamber, a first sensor operably connected to the input, and a second sensor operably connected to the output and configured to measure impedance during operation, at least one variable capacitor connected to the first and second sensors, and a controller configured to tune at least one variable capacitor of the matching network to a first target position based on a weighted output impedance value measured in a pulsed state, and at least one variable capacitor to a second target position based on a weighted input impedance value measured in a pulsed state, based on the measured impedance.

[0007] According to at least some embodiments, a method for processing a substrate includes measuring impedance at the input of a matched network configured to receive one or more radio frequency (RF) signals and at the output of a matched network configured to send one or more RF signals to a processing chamber; and, based on the measured impedance, tuning at least one variable capacitor of the matched network to a first target position based on a weighted output impedance value measured in a pulsed state, and tuning at least one variable capacitor to a second target position based on a weighted input impedance value measured in a pulsed state.

[0008] Other and further embodiments of this disclosure are described below.

[0009] Embodiments of the present disclosure, briefly summarized above and described in more detail below, can be understood by referring to exemplary embodiments of the present disclosure shown in the accompanying drawings. However, since the present disclosure may allow for other equally valid embodiments, the accompanying drawings are merely illustrative of general embodiments of the present disclosure and should not be considered limiting to the scope. [Brief explanation of the drawing]

[0010] [Figure 1] This is a cross-sectional view of a processing chamber according to at least some embodiments of the present disclosure. [Figure 2] This is a diagram of a system according to at least some embodiments of the present disclosure. [Figure 3] This is a diagram of a harmonized network according to some embodiments of the present disclosure. [Figure 4] This is a graph of sampling impedance according to at least some embodiments of the present disclosure. [Figure 5] This is a diagram of a system according to at least some embodiments of the present disclosure. [Figure 6]This is a diagram of internal synchronization for dual-level pulsing according to at least some embodiments of the present disclosure. [Figure 7] This is a flowchart of a method for processing a substrate according to at least some embodiments of the present disclosure.

[0011] For ease of understanding, the same reference numerals are used to designate identical elements common to the figures where possible. The figures are not drawn to a fixed scale and may be simplified for clarity. Elements and features of one embodiment may be usefully incorporated into other embodiments without further description. [Modes for carrying out the invention]

[0012] Embodiments of methods and apparatus for processing substrates are provided herein. For example, a matching network configured for use with a plasma processing chamber may have an input configured to receive one or more radio frequency (RF) signals, and an output configured to send one or more RF signals to the processing chamber. A first sensor may be operably connected to the input, and a second sensor may be operably connected to the output and configured to measure impedance during operation. At least one variable capacitor may be connected to the first and second sensors. A controller may be configured to tune at least one variable capacitor of the matching network to a first target position based on a weighted output impedance value measured in a pulsed state, and at least one variable capacitor to a second target position based on a weighted input impedance value measured in a pulsed state, based on the measured impedance. The advantages of the apparatus and methods described herein include, but are not limited to, tuning to optimized weighted plasma impedance values ​​to achieve minimum total reflected power for all states during multilevel pulsing, spatial power distribution and uniformity, and flexibility for defining tuning targets based on different process and pulsing conditions.

[0013] Figure 1 is a cross-sectional view of an example of a suitable processing chamber 100 for carrying out the etching process according to this disclosure. Suitable processing chambers that may be adapted for use with the teachings disclosed herein include, for example, one or more etching processing chambers available from Applied Materials, Inc. in Santa Clara, California. Other processing chambers may be adapted to benefit from one or more of the methods of this disclosure.

[0014] The processing chamber 100 includes a chamber body 102 and a chamber lid 104 that seals the internal volume 106. The chamber body 102 is generally made of aluminum, stainless steel or other preferred material. The chamber body 102 generally includes side walls 108 and a bottom 110. A substrate support pedestal access port (not shown) is generally defined in the side wall 108 and selectively sealed by a slit valve to facilitate the entry and exit of substrates 103 from the processing chamber 100. An exhaust port 126 is defined in the chamber body 102 and connects the internal volume 106 to a pump system 128. The pump system 128 generally includes one or more pumps and throttle valves used to relieve and regulate the pressure of the internal volume 106 of the processing chamber 100. In this embodiment, the pump system 128 maintains the pressure in the internal volume 106 at operating pressures generally between about 1 mTorr and about 500 mTorr, between about 5 mTorr and about 100 mTorr, or between about 5 mTorr and about 50 mTorr, as required by the process.

[0015] In one embodiment, the chamber lid 104 may be sealed and supported on the side wall 108 of the chamber body 102. The chamber lid 104 may be opened to allow an overflow relative to the internal volume 106 of the processing chamber 100. The chamber lid 104 includes a window 142 to facilitate optical process monitoring. In one embodiment, the window 142 is made of quartz or other suitable material that is transparent to signals utilized by an optical monitoring system 140 mounted on the outside of the processing chamber 100.

[0016] The optical monitoring system 140 is positioned to observe, through a window 142, at least one of the internal volume 106 of the chamber body 102 and / or the substrate 103 placed on the substrate support pedestal assembly 148. In one embodiment, the optical monitoring system 140 is coupled to the chamber lid 104 to facilitate the deposition process, which provides information that allows the deposition process to adjust, if necessary, to compensate for inconsistencies (such as thickness) in the incoming substrate pattern features using optical measurements, and provides process state monitoring (plasma monitoring, temperature monitoring, etc.).

[0017] Process gases and / or cleaning gases can be introduced into the internal volume 106 of the chamber body 102 from a gas panel 158 coupled to the processing chamber 100 through a showerhead assembly 130. A vacuum pump system, such as a pump system 128, maintains the pressure inside the chamber body 102 while removing accumulated by-products.

[0018] In the example shown in Figure 1, inlet ports 132', 132'' are provided in the chamber lid 104 to allow gas to be delivered from the gas panel 158 into the internal volume 106 of the processing chamber 100. In some embodiments, the gas panel 158 is adapted to supply oxygen and an inert gas such as argon, or an oxygen and helium process gas or mixed gas, to the internal volume 106 of the processing chamber 100 through the inlet ports 132', 132''. In one embodiment, the process gas supplied from the gas panel 158 includes at least an oxidizing agent, such as oxygen gas. In some embodiments, the oxidizing agent-containing process gas may further include an inert gas such as argon or helium. In some embodiments, the process gas includes a reducing agent such as hydrogen and may be mixed with an inert gas such as argon, or other gases such as nitrogen or helium. In some embodiments, chlorine gas may be supplied alone or in combination with at least one of the inert gases such as nitrogen, helium, or argon. Non-limiting examples of oxygen-containing gases include one or more of O2, CO2, N2O, NO2, O3, H2O, etc. Non-limiting examples of nitrogen-containing gases include N2, NH3, etc. Non-limiting examples of chlorine-containing gases include HCl, Cl2, CCl4, etc. In the embodiment, the showerhead assembly 130 is coupled to the inner surface 114 of the chamber lid 104. The showerhead assembly 130 includes a plurality of openings that allow the gas to flow from the inlet ports 132', 132'' through the showerhead assembly 130 into the inner volume 106 of the processing chamber 100 in a predefined distribution across the surface of the substrate 103 being processed in the processing chamber 100.

[0019] In some embodiments, the processing chamber 100 may utilize capacitively coupled RF energy for plasma processing, or in some embodiments, the processing chamber 100 may use inductively coupled RF energy for plasma processing. In some embodiments, a remote plasma source 177 may optionally be coupled to a gas panel 158 to facilitate the separation of a mixed gas from the remote plasma before it enters the internal volume 106 for processing. In some embodiments, RF source power 143 is coupled to the showerhead assembly 130 through a matching network 141. The RF source power 143 can generally be generated at a maximum of about 5000 W, for example, between about 200 W and about 5000 W, or between 1000 W and 3000 W, or about 1500 W, and optionally at tunable frequencies in the range of about 50 kHz to about 200 MHz.

[0020] The showerhead assembly 130 further includes a region that is transparent to optical measurement signals. The optically transparent region or passage 138 is suitable for allowing the optical monitoring system 140 to observe the internal volume 106 and / or the substrate 103 placed on the substrate support pedestal assembly 148. The passage 138 may be one or more openings, material formed or disposed in the showerhead assembly 130, which is substantially transparent to wavelengths of energy provided by and reflected to the optical monitoring system 140. In one embodiment, the passage 138 includes a window 142 to prevent gas leakage through the passage 138. The window 142 may be a sapphire plate, a quartz plate or other suitable material. The window 142 may alternatively be disposed in the chamber lid 104.

[0021] The showerhead assembly 130 can be composed of multiple zones that enable separate control of the gas flowing in the internal volume 106 of the processing chamber 100. In the example shown in FIG. 1, the showerhead assembly 130 is configured as an inner zone 134 and an outer zone 136 separately coupled to the gas panel 158 through the inlet ports 132’, 132”.

[0022] The substrate support pedestal assembly 148 is disposed in the internal volume 106 of the processing chamber 100 below a gas distribution assembly such as the showerhead assembly 130. The substrate support pedestal assembly 148 holds the substrate 103 during processing. The substrate support pedestal assembly 148 generally includes a plurality of lift pins (not shown) disposed through the substrate support pedestal assembly 148 configured to lift the substrate 103 from the substrate support pedestal assembly 148 and facilitate the replacement of the substrate 103 by a robot (not shown) in a conventional manner. The inner liner 118 can closely circumscribe the periphery of the substrate support pedestal assembly 148.

[0023] The substrate support pedestal assembly 148 includes a mounting plate 162, a base 164, and an electrostatic chuck 166. The mounting plate 162 is coupled to the bottom 110 of the chamber body 102 and includes passages for routing utilities such as fluids, power lines, and sensor leads to the base 164 and the electrostatic chuck 166. The electrostatic chuck 166 includes an electrode 180 (e.g., a clamp electrode) for holding the substrate 103 below the showerhead assembly 130. The electrostatic chuck 166 is driven by a chuck power supply 182 to generate an electrostatic force for holding the substrate 103 against the chuck surface, as is conventionally known. Alternatively, the substrate 103 can be held against the substrate support pedestal assembly 148 by a clamp, vacuum, or gravity.

[0024] The base 164 or the electrostatic chuck 166 may include a heater 176, at least one optional embedded isolator 174, and a plurality of conduits 168, 170 to control the lateral temperature profile of the substrate support pedestal assembly 148. The conduits 168, 170 are fluidly coupled to a fluid source 172, which circulates temperature-adjusted fluid through those conduits. The heater 176 is adjusted by a power source 178. The conduits 168, 170 and the heater 176 are utilized to control the temperature of the base 164, to control heating and / or cooling of the electrostatic chuck 166, and ultimately to control the temperature profile of the substrate 103 disposed on the electrostatic chuck 166. The temperatures of the electrostatic chuck 166 and the base 164 may be monitored using a plurality of temperature sensors 190, 192. The electrostatic chuck 166 may further include a plurality of gas passages (not shown), such as grooves, formed in the substrate support pedestal support surface of the electrostatic chuck 166 and fluidly coupled to a source of heat transfer (or backside) gas, such as helium (He). During operation, the backside gas is provided in the gas passages at a controlled pressure to improve heat transfer between the electrostatic chuck 166 and the substrate 103. In embodiments, the temperature of the substrate may be maintained at between 20 degrees Celsius and 450 degrees Celsius, such as between 100 degrees Celsius and 300 degrees Celsius, or between 150 degrees Celsius and 250 degrees Celsius.

[0025] The substrate support pedestal assembly 148 may be configured as a cathode assembly and includes an electrode 180 coupled to a plurality of RF bias power sources 184, 186. The RF bias power sources 184, 186 are coupled between an electrode 180 disposed in the substrate support pedestal assembly 148 and another electrode, such as the showerhead assembly 130 (or chamber lid 104) of the chamber body 102. The RF bias power excites and sustains a plasma discharge formed from a gas disposed in the processing region of the chamber body 102.

[0026] Referring again to Figure 1, in some embodiments, the dual RF bias power supplies 184, 186 are coupled to electrodes 180 located in a substrate support pedestal assembly 148 via a matching network 188. The signals provided by the RF bias power supplies 184, 186 are fed through a single feed via the matching network 188 to the substrate support pedestal assembly 148 to ionize the mixed gas provided in a plasma processing chamber such as the processing chamber 100, and thus provide the ionic energy necessary to carry out etching deposition or other plasma enhancement processes. The RF bias power supplies 184, 186 are generally capable of generating RF signals with frequencies from about 50 kHz to about 200 MHz (e.g., about 13.56 MHz + / - 5%) and powers from about 0 watts to about 10,000 watts (e.g., about 50 W for low-power operation to about 10,000 W for high-power operation), from 1 watt (W) to about 100 W, or from about 1 W to about 30 W. Additional bias power may be coupled to electrode 180 to control the plasma characteristics.

[0027] In at least some embodiments, the impedances at the input and output ports of the matched network 188 and / or matched network 141 may be measured in all states of multilevel pulsing. The impedances at the input and output ports of the matched network may be used to determine the weighted input and output impedances for tuning. For example, the apparatus and method described herein uses weighted average tuning in multilevel pulsing. In at least some embodiments, a weighted combination of measured output impedances may be selected for feedforward tuning, and the weighted impedance may be defined from the input impedance measured in the multilevel pulsing state. Furthermore, in at least some embodiments, frequency tuning may be used in conjunction with weighted average tuning for hybrid tuning. The matched network described herein may receive a TTL synchronization signal from the RF generator and / or advanced waveform generator 202, as described in more detail below. Alternatively or additionally, the matched network may receive a TTL synchronization signal internally triggered on the detected pulse rising edge.

[0028] A controller 150 is coupled to the processing chamber 100 to control its operation. The controller 150 includes a central processing unit 152, memory 154, and support circuitry 156, which are used to control the process sequence and adjust the gas flow from the gas panel 158. The central processing unit 152 may be any form of general-purpose computer processor that may be used in an industrial setting. Software routines may be stored in memory 154, such as random access memory, read-only memory, floppy disks, or hard disk drives, or other forms of digital storage. Support circuitry 156 is conventionally coupled to the central processing unit 152 and may include a cache, clock circuitry, input / output systems, power supply, etc. Bidirectional communication between the controller 150 and various components of the processing chamber 100 is handled through a number of signal cables.

[0029] Figure 2 is a diagram of System 200 according to at least some embodiments of the present disclosure.

[0030] For example, in at least some embodiments, one or more RF power supplies (e.g., RF bias power supply 184 and / or RF source power supply 143) may be configured to supply RF power for plasma generation to the RF base plate of the cathode assembly (e.g., electrostatic chuck 166). In such embodiments, the upper electrode (e.g., showerhead assembly 130 (or chamber lid 104)) may be grounded. The frequency of one or more RF power supplies may range from 13.56 MHz to an ultra-high frequency band such as 60 MHz, 120 MHz, or 162 MHz. In at least some embodiments, one or more RF power supplies may also be supplied through the upper electrode. One or more RF power supplies may operate in continuous mode or pulsed mode. For example, in pulsed mode, the pulse frequency may be 100 Hz to about 10 kHz, and the duty cycle may be from about 5% to about 95%.

[0031] RF impedance matching networks, for example, matching network 188 and / or matching network 141, are connected between one or more RF power supplies and the processing chamber 100 to optimize power supply efficiency. The matching networks are configured for use with plasma processing chambers, such as physical vapor deposition chambers, chemical vapor deposition chambers, atomic layer deposition chambers, etching chambers, or other processing chambers that use matching networks. For illustrative purposes, matching networks (e.g., matching network 141 and / or matching network 188) are described herein in relation to etching chambers, such as processing chamber 100.

[0032] The matching network includes an input stage 201 configured to connect to one or more RF power supplies (e.g., RF bias power supplies 184, 186) of the plasma processing chamber and to receive one or more radio frequency (RF) signals. The matching network also includes an output stage 203 configured to connect to a substrate support pedestal assembly (e.g., substrate support pedestal assembly 148) of the processing chamber and to send one or more RF signals to the processing chamber.

[0033] The matching network includes one or more variable (tunable) capacitors, such as a first variable capacitor 205 and a second variable capacitor 207 (e.g., a shunt variable capacitor), which may be connected in series or in parallel. The first variable capacitor 205 and the second variable capacitor 207 have variable capacitances that allow the first variable capacitor 205 and the second variable capacitor 207 to be tuned to one or more frequencies. For example, in at least some embodiments, the first variable capacitor 205 and the second variable capacitor 207 may have capacitances ranging from about 3 pF to about 2500 pF. In at least some embodiments, such as when the processing chamber is operating in a high-power or low-power state, the first variable capacitor 205 and the second variable capacitor 207 may be tuned to one or more of the frequencies described above, for example, target frequencies ±10%, and target frequencies from 100 kHz to about 250 MHz.

[0034] In at least some embodiments, one or more additional capacitors, inductors, transistors, etc. (not shown) are also provided, which may be connected in parallel and / or in series with the first variable capacitor 205 and the second variable capacitor 207.

[0035] The first variable capacitor 205 and the second variable capacitor 207 may be the same as or different from each other. In at least some embodiments, the first variable capacitor 205 may be coupled to the output stage 203, and the second variable capacitor 207 may be connected to the input stage 201, or vice versa.

[0036] In at least some embodiments, one or more RF filters may be connected to a matching network to enable power in a selected frequency range and to isolate the RF power supplies from one another. For example, in at least some embodiments, RF filter 204 is connected to a matching network 188 and to a substrate support pedestal assembly 148.

[0037] In at least some embodiments, an advanced waveform generator 202 may be used to supply one or more waveforms (e.g., pulsed voltage waveforms and / or adapted voltage waveforms, which may be the sum of harmonic frequencies associated with the adapted voltage waveform). One or more voltage waveforms may be coupled to bias electrodes (e.g., electrodes 180 of a substrate-supported pedestal assembly 148) through one or more filter assemblies. For example, in at least some embodiments, an RF filter 206 is connected to the advanced waveform generator 202 and electrodes 180. The advanced waveform generator 202 may output a synchronization signal to a matching network 188. For example, in at least some embodiments, the synchronization signal may be a transistor-to-transistor logic (TTL) 209 signal, as described in more detail below. Alternatively or additionally, the RF source power may be configured to output a synchronization signal to the matching network 188. Alternatively or additionally, the matching network may be configured to provide an internal synchronization signal, as described in more detail below.

[0038] Figure 3 shows a harmonized network 188 configured for use with a processing chamber 100, according to at least some embodiments of the present disclosure. In at least some embodiments, the harmonized network 188 may be an L-type or π-type harmonized network.

[0039] The harmonized network 188 comprises a local controller, one or more sensors, and one or more electric capacitors, all of which are connected via EtherCAT (indicated by dashed line 301). EtherCAT is a real-time industrial Ethernet protocol, and with its short cycle time and low jitter, EtherCAT provides fast and accurate synchronization during plasma processing. One or more other interfaces may be used to connect the components of the harmonized network 188 to each other, and / or to connect the RF generator and the plasma processing chamber to the harmonized network 188. For example, transmission line 303 (indicated by solid line) may be used to connect the RF generator to the harmonized network 188 and to connect the harmonized network 188 to the plasma processing chamber, for example, to supply RF power to the plasma processing chamber.

[0040] In at least some embodiments, the local controller 300 functions as a local EtherCAT master, and all harmonized network components, such as sensors and electric capacitors, are EtherCAT slave devices controlled by the local controller 300. For example, a command sent by the local controller 300 (e.g., the EtherCAT master controller) is passed to all EtherCAT slave devices. A first electric capacitor 302 (vacuum capacitor) with an EtherCAT interface may be connected to the local controller 300, and a second electric capacitor 304 (vacuum capacitor) with an EtherCAT interface may be connected. The first electric capacitor 302 may be connected to the second electric capacitor 304 in a series or parallel configuration. For example, in the illustrated embodiment, the first electric capacitor 302 (e.g., a shunt variable capacitor) is connected in parallel with the second electric capacitor 304 (e.g., a series variable capacitor). The first electric capacitor 302 and the second electric capacitor 304 are electric variable capacitors and are configured to be adjusted during operation. For example, the local controller 300 may be configured to adjust the first electric capacitor 302 and the second electric capacitor 304 to minimize reflected power during plasma processing.

[0041] The local controller 300 may be connected (directly or indirectly) to a first sensor 306 located at the input of the matching network 188 and a second sensor 308 (when used) located at the output of the matching network 188, respectively, to acquire inline RF voltage, current, phase, harmonic, and impedance data. In at least some embodiments, the first sensor 306 and the second sensor 308 may be multi-frequency voltage / current probes. The measured data may be used for automatic impedance tuning, load impedance monitoring, etc.

[0042] In at least some embodiments, the interlock circuit 307 may be connected to the local controller 300 and configured to prevent RF generator failure. For example, the interlock circuit 307 may include a fault protection circuit configured to shut off the RF power output from the RF generator when the reflected RF power exceeds a certain percentage of the forward power (e.g., >20%), which is the RF power sent through the matching network 188 to the load, for example, the plasma in the processing chamber.

[0043] As described above, the EtherCAT communication interface connects the local controller 300 to the first electric capacitor 302, the second electric capacitor 304, the first sensor 306, and the second sensor 308. The EtherCAT communication interface connects the RF generator directly to each of the first sensor 306 and the second sensor 308, for example, to transmit a TTL signal 305 from the RF generator (e.g., RF bias power supplies 184, 186 (and / or bias power supply 189)) to each of the first sensor 306 and the second sensor 308 for fast response and short adjustment time.

[0044] In at least some embodiments, when connected to an RF generator and a plasma processing chamber, the local controller 300 is configured as an EtherCAT master device that controls and monitors local EtherCAT slave devices, such as sensors and stepping motors. The local controller 300 is also integrated with an EtherCAT slave controller, so that the local controller 300 can act as an EtherCAT slave device and the controller 150 acts as an EtherCAT master device. That is, the local controller 300 is configured to perform master-slave conversion with the controller 150. The tool controller may be implemented on an industrial computer and embedded with the necessary drivers. In such embodiments, the local controller 300 can receive feedback requests from the controller 150 and provide feedback to the controller 150 during plasma processing. For example, the local controller 300 can receive inline RF voltage, current, phase, harmonic, and impedance data acquired via the first sensor 306 and the second sensor 308. Sensor data and variable capacitor positions are transmitted to the controller 150 and combined with other system processing data, such as forward power data and reflected power data from RF bias power supplies 184 and 186, thereby enabling cooperative intelligent real-time control during operation.

[0045] The harmonized network 188 may comprise at least one of a first network port 310 (e.g., dual RJ45 type ports) configured to connect to the controller 150, and a second series port configured to connect to an external computing device (e.g., a laptop or other suitable computing device) for manual control of the harmonized network 188. For example, in at least some embodiments, the controller 150 may connect to the first network port 310 of the harmonized network 188 for plasma process control. The local controller 300 may receive inline RF voltage, current, phase, harmonic, and impedance data acquired via the first sensor 306 and the second sensor 308. Sensor data and variable capacitor positions are transmitted to the controller 150 and combined with other system processing data, such as forward power data and reflected power data from the RF bias power supplies 184 and 186, to thus create cooperative intelligent real-time control during operation. In at least some embodiments, the harmonized network 188 may include a second series port 312 configured to connect to a computing device 314 for algorithm uploading and for manual control of the harmonized network, for example, by using external software and an application programming interface (API). In at least some embodiments, the external software and API may be uploaded, stored in memory 154, accessed by the controller 150, and / or in the memory (not shown) of the local controller 300. In at least some embodiments, sensor data that can be obtained from the first sensor 306 and the second sensor 308 may be accessed by the computing device 314. Furthermore, when connected to the second series port 312, the computing device 314 may be configured to control the first electric capacitor 302 and the second electric capacitor 304.The provision of a first network port 310 and a second serial port 312 provides the harmonized network 188 with greater flexibility compared to conventional harmonized networks. For example, advanced process relation control algorithms can be deployed in real time, and the harmonized network 188 can operate fully autonomously, in cooperation with the controller 150, and / or manually controlled via the computing device 314. During processing, if necessary, the EtherCAT-based distributed RF impedance harmonized network described herein allows a user using the computing device 314 to have complete control over the harmonized network 188 and its associated components.

[0046] Figure 4 is a graph 400 of sampling impedance according to at least some embodiments of the present disclosure. For example, in at least some embodiments, the voltage waveform or RF power pulse applied to the substrate (e.g., substrate 103) in the processing chamber 100 may include two stages. The plasma sheath impedance varies with the supplied pulse voltage waveform and RF power pulse, and the matching network 188 monitors the TTL synchronization signal from the waveform generator 202 or RF power supply. For example, in a pulse cycle, two or more pulse data points may be collected by the matching network 188 to collect impedance at different stages. In at least some embodiments, a first data sample may be collected at a first stage for impedance Z1 (e.g., at 404, which may correspond to a sheath collapse stage), and a second data sample may be collected at a second stage for impedance Z2 (e.g., at 402, which may correspond to an ion current stage). The collected data samples may be used to obtain a weighted impedance value. In at least some embodiments, data samples need to be collected after a time delay defined, for example, based on the pulse frequency, duty cycle, and / or rising edge of the TTL synchronization signal.

[0047] In at least some embodiments, a high-voltage DC power supply 208 may be used to power the electrode 180 to chuck the substrate (e.g., wafer) during processing for temperature control. In at least some embodiments, a third electrode (not shown) may be provided at the edge of the cathode assembly for edge uniformity control. In such embodiments, a third low-frequency RF power supply in the frequency range of 50 kHz to 2 MHz may be supplied to the edge electrode and operated in continuous mode.

[0048] Figure 5 is a diagram of System 500 according to at least some embodiments of the present disclosure. System 500 is substantially equivalent to System 200. Therefore, only features specific to System 500 are described herein.

[0049] For example, one or more RF power supplies (e.g., RF source power 143) are connected to the upper electrode for plasma generation. The frequency of one or more RF power supplies can operate in frequencies from about 13.56 MHz to about 200 MHz, such as 60 MHz, 120 MHz, or 162 MHz, as needed. One or more RF power supplies can operate in continuous mode or pulsed mode. The pulsed frequency can be from 100 Hz to 10 kHz, and the duty cycle can be from about 5% to about 95%. RF bias power (e.g., RF bias power supply 184) is connected to the bottom electrode with a frequency range of about 100 kHz to about 15 MHz. The RF bias power can operate in either continuous mode or pulsed mode. The pulsed frequency can be from 100 Hz to 10 kHz, and the duty cycle can be from about 5% to about 95%. Either or both of the RF source power 143 and the RF bias power may be configured to send synchronous TTL signals to matching network 141 and matching network 188, respectively. As described above, a third electrode may be used at the edge of the cathode assembly for edge uniformity control. In such embodiments, a third low-frequency RF power supply in the frequency range of 50 kHz to 2 MHz may be supplied to the edge electrode and operated in continuous mode. Similar to system 200, the RF filter may be connected to a matching network 188 (not shown), a matching network 141, and a high-voltage DC power supply 208. In at least some embodiments, the RF filter may be connected to the matching network 188.

[0050] Figure 6 is a figure 600 of internal synchronization for dual-level pulsing according to at least some embodiments of the present disclosure. For example, as described above, a trigger signal may be generated externally (e.g., via one or more RF power supplies or advanced waveform generators) or internally (e.g., via a matched network) to obtain impedance values ​​for determining a weighted average. In the latter embodiment, voltage and current sensors in the matched network (e.g., a first sensor 306 and a second sensor 308) are configured to internally detect the start of the pulse signal. The voltage and current sensors may sense multiple impedance samples during the pulse cycle, for example, at a first sample time defined with respect to the start of the trigger signal, generally the rising edge of the pulse, and at a second sample time. Multiple impedances may be measured at the same pulse level or different levels and may be used in a weighted average tuning algorithm. Thus, the matched network uses the weighted impedance from all measured samples at the same pulse level or different levels.

[0051] The collected data samples may be measured in a pulse or averaged from multiple pulses. For example, a threshold 601 may be defined for pulse detection. The threshold 601 may be set between two pulse state levels. The start of a pulse is detected when a voltage measured by a sensor (not shown) exceeds the threshold 601. In at least some embodiments, sample 1 and sample 2 may be taken from different pulses. In at least some embodiments, sample 1 and sample 2 may be the average of many pulse data points collected from multiple pulses. For example, in at least some embodiments for calculating an averaged value, a pulse high state may be measured 10 times in 10 pulses to obtain averaged sample 1, and a pulse low state may be measured 10 times in the same or different 10 pulses to obtain averaged sample 2. The data point measurement is triggered after a delay of a first sample time and a second sample time relative to the start of the pulse. The first and second sample times determine when the measurement is taken in the pulse.

[0052] For example, the first data sample 602 and the second data sample 604 can be collected in high and low states, respectively, by defining a first pulse time and a second pulse time with respect to a trigger signal, such as the rising edge 606 of a pulse.

[0053] The inventors discovered that the plasma load impedance can be varied by multilevel pulse conditions due to RF power sources with different power levels or combinations (e.g., RF bias power supply 184 and RF source power supply 143). Furthermore, in conventional matching networks, the vacuum capacitor motors used generally cannot respond quickly enough (e.g., move) in the pulse cycle, and conventional matching networks cannot adjust both impedances simultaneously.

[0054] Therefore, the inventors have provided weighted combinations of impedances (e.g., various impedance samples acquired via either an external or internal trigger) for single-level pulsing or multi-level pulsing. For example, in dual-level pulsing, using either or both an external or internal trigger, a first impedance Z1 may be measured at a first pulse level and a first sample time, and a second impedance Z2 may be measured at a second pulse level and a second sample time. The weighted target impedance can then be calculated using equation (1). Z w =Z1*w+Z2*(1-w), where w ranges from 0 to 1... (1) Here, w is a weight value between 0 and 1. In at least some embodiments, a multi-state weighting algorithm may also be used if more weight values ​​are required. For example, w1 and w2 may be weight values ​​in a triple-level pulsing situation. During operation, the weighted target impedance of the matched network described herein changes with the weight value during a dual-impedance state. For example, when w is equal to 0, the matched network is tuned to an impedance equal to the second level in the pulse, but when w is equal to 1, the matched network is tuned to an impedance equal to the first level.

[0055] In at least some embodiments, the optimized w value may be determined based on the minimum total reflected power for both states. For example, the reflected power used varies with the weight value in dual-level pulsing. For instance, the first level may have the minimum reflected power at w=1, and the second level may have the minimum reflected power at w=0. For example, when w=0.8, the total reflected power for both states is at its minimum. Other criteria may also be used to select appropriate weight values.

[0056] Similarly, the matched network provides measured and weighted output impedances in dual-level pulsing. For example, the plasma load impedance can vary with pulsed power levels, bias power on and off, or pulse voltage waveforms. Therefore, Z1 and Z2 can be the measured impedances at the output of the matched network, in this case either externally or internally synchronized. As mentioned above, since motors in conventional matched networks cannot keep up with rapidly changing impedance states, a weighted output impedance can be calculated and used as the target plasma load impedance. For example, in at least some embodiments, the weighted impedance at the output of the matched network can be used as a tuning target in a feedforward tuning algorithm.

[0057] In at least some embodiments, the weighted input impedance and weighted output impedance can be stored in a lookup table (for example, in memory 154), or a circuit model can be used to directly move the variable capacitor in the RF matching to the target position. In some embodiments, a learning-based tuning algorithm can be employed to find the appropriate weighted target impedance at the RF matching input and / or output.

[0058] Figure 7 is a flowchart of a method 700 for processing a substrate according to at least some embodiments of the present disclosure. For illustrative purposes, the method 700 is described herein using a processing chamber 100 for etching a substrate 103.

[0059] In 702, the method 700 includes measuring impedance at the input of a matched network configured to receive one or more radio frequency (RF) signals and at the output of a matched network configured to send one or more RF signals to a processing chamber. For example, in at least some embodiments, in 702, a first sensor 306 may be configured to measure impedance at the input of a matched network 188, and a second sensor 308 may be configured to measure impedance at the output of the matched network 188.

[0060] Next, in 704, method 700 includes adjusting at least one capacitor of a matching network to a first target position based on a measured impedance, for example, based on a weighted output impedance value measured in a pulsed state of the voltage waveform, and adjusting at least one variable capacitor to a second target position based on a weighted input impedance value measured in a pulsed state of the voltage waveform, for example.

[0061] For example, a variable capacitor (e.g., a second electric capacitor 304) may be tuned based on a weighted output impedance value measured in a pulsed state and a weighted input impedance value measured in a pulsed state. For example, as described above, the weighted output impedance value and the weighted input impedance value are calculated using equation (1). In at least some embodiments, at least one capacitor may comprise two variable capacitors (e.g., a first electric capacitor 302 and a second electric capacitor 304). In such embodiments, at least two variable capacitors may be tuned simultaneously or at different times.

[0062] In at least some embodiments, such as when the second sensor 308 is not used, the variable capacitor may be adjusted based on a weighted output impedance stored in a lookup table or circuit model (for example, in memory 154).

[0063] In at least some embodiments (for example, in a single-level pulsed signal configuration comprising an RF signal provided by an RF bias power supply, an RF signal provided by an RF source power supply, and / or a pulsed voltage waveform), the impedance value used for equation (1) can be obtained from the pulse-on state. Furthermore, the first sample time and the second sample time may be the same or different, based on at least one of the pulse frequency, duty cycle, or rising edge of the transistor-transistor logic (TTL) synchronous signal.

[0064] Similarly, in at least some embodiments (for example, in a dual-level pulsed signal configuration comprising at least one of an RF signal provided by an RF bias power supply, an RF signal provided by an RF source power supply, and / or a pulsed voltage waveform), the impedance value used for equation (1) may be acquired in a high-level pulse stage and a low-level pulse stage. In such embodiments, the high-level pulse stage is acquired at a first sample time, and the low-level pulse stage is acquired at a second sample time different from the first sample time. Furthermore, the first and second sample times are triggered after a delay from the start of the pulse detected when the measured voltage of the pulse is equal to or greater than a threshold.

[0065] The weighted impedance value can be calculated using equation (1) and impedance data obtained for a single-level pulse signal configuration, a dual-level pulse signal configuration, or a multi-level pulse signal configuration. The calculated weighted impedance value is stored in memory 154 and is automatically accessed by controller 150 during operation in 704 to tune the first variable capacitor and / or the second variable capacitor to the weighted target impedance value.

[0066] While the above applies to embodiments of the present disclosure, other and further embodiments of the present disclosure can be conceived without departing from its basic scope.

Claims

1. A matching network configured for use with a plasma processing chamber, An input configured to receive one or more radio frequency (RF) signals, An output configured to send one or more RF signals to the plasma processing chamber, A first sensor operably connected to the input and configured to measure impedance during operation, and a second sensor operably connected to the output and configured to measure impedance during operation, At least one variable capacitor connected to the first sensor and the second sensor, A controller configured to adjust at least one variable capacitor of the matching network to a first target position based on a weighted output impedance value obtained by weighting the impedances measured using the second sensor in multiple different sections of the pulse based on one or more RF signals, and to adjust at least one variable capacitor to a second target position based on a weighted input impedance value obtained by weighting the impedances measured using the first sensor in multiple different sections of the pulse, A consistent network equipped with these features.

2. The weighted output impedance value and the weighted input impedance value are calculated using the following formula: Z ｗ =Z １ *w+Z ２ *(1-w)、 Here, Z １ However, it was measured at time 1 and pulse level 1, Z ２ The matched network according to claim 1, wherein the values ​​are measured at time 2 and pulse level 2, where w is a weight value between 0 and 1.

3. A matching network according to claim 1, comprising a single-level pulse signal configuration having an RF signal provided by an RF bias power supply, wherein the plurality of different segments of the pulse correspond to pulse data points collected to obtain impedance in a first data sample and a second data sample of the pulse.

4. The first data sample is taken at a first time, and the second data sample is taken at a second time different from the first time. The matching network according to claim 3, wherein the first time and the second time are based on at least one of the pulse frequency, duty cycle, or rising edge of a transistor-to-transistor logic (TTL) synchronization signal.

5. A matching network according to claim 1, comprising at least one of an RF signal provided by an RF bias power supply or an RF signal provided by an RF source power supply, wherein the plurality of different segments of the pulse correspond to pulse data points collected to obtain impedances in high-level pulse stages and low-level pulse stages.

6. The high-level pulse stage is taken in a first time period, and the low-level pulse stage is taken in a second time period different from the first time period. The matching network according to claim 5, wherein the first time and the second time are triggered after a delay from the start of the pulse detected when the measured voltage is equal to or greater than a threshold.

7. The matching network according to claim 1, wherein the matching network is connected to an RF bias power supply capable of operating at frequencies of approximately 100 kHz, 13.56 MHz, 15 MHz, 60 MHz, 120 MHz, or 162 MHz, and in at least one of continuous mode or pulse mode, wherein in pulse mode, the pulse frequency is from approximately 100 Hz to approximately 10 kHz and the duty cycle is from approximately 5% to approximately 95%.

8. The matching network according to claim 1, wherein the matching network is connected to an RF source power capable of operating at frequencies of approximately 13.56 MHz, 60 MHz, 120 MHz, 162 MHz, or 200 MHz, and in at least one of continuous mode or pulse mode, wherein in pulse mode, the pulse frequency is from approximately 100 Hz to approximately 10 kHz and the duty cycle is from approximately 5% to approximately 95%.

9. The matching network according to claim 1, wherein the at least one variable capacitor comprises a series variable capacitor and a shunt variable capacitor.

10. A plasma processing chamber, Chamber body and chamber lid, An RF power source connected to the chamber lid and configured to create plasma from a gas placed in the processing area of ​​the chamber body, One or more RF bias power supplies configured to sustain a plasma discharge, Unified network and The matching network is equipped with, An input configured to receive one or more radio frequency (RF) signals output from the one or more RF bias power supplies, An output configured to send one or more RF signals to the plasma processing chamber, A first sensor operably connected to the input and configured to measure impedance during operation, and a second sensor operably connected to the output and configured to measure impedance during operation, At least one variable capacitor connected to the first sensor and the second sensor, A controller configured to adjust at least one variable capacitor of the matching network to a first target position based on a weighted output impedance value obtained by weighting the impedances measured using the second sensor in multiple different sections of the pulse based on one or more RF signals, and to adjust at least one variable capacitor to a second target position based on a weighted input impedance value obtained by weighting the impedances measured using the first sensor in multiple different sections of the pulse, A plasma processing chamber equipped with the following features.

11. The weighted output impedance value and the weighted input impedance value are calculated using the following formula: Z ｗ =Z １ *w+Z ２ *(1-w)、 Here, Z １ is measured at time 1 and pulse level 1, and Z ２ is measured at time 2 and pulse level 2, where w is a weight value between 0 and 1, the plasma processing chamber according to claim 10.

12. A plasma processing chamber according to claim 10, comprising a single-level pulsed signal configuration having an RF signal provided by an RF bias power supply, wherein a plurality of different segments of the pulse correspond to pulse data points collected to obtain impedances in a first data sample and a second data sample of the pulse.

13. The first data sample is taken at a first time, and the second data sample is taken at a second time different from the first time. The plasma processing chamber according to claim 12, wherein the first time and the second time are based on at least one of the pulse frequency, duty cycle, or rising edge of a transistor-to-transistor logic (TTL) synchronization signal.

14. A plasma processing chamber according to claim 10, comprising at least one of an RF signal provided by an RF bias power supply or an RF signal provided by an RF source power supply, wherein a plurality of different segments of the pulse correspond to pulse data points collected to obtain impedances in high-level pulse stages and low-level pulse stages.

15. The high-level pulse stage is taken in a first time period, and the low-level pulse stage is taken in a second time period different from the first time period. The plasma processing chamber according to claim 14, wherein the first time and the second time are triggered after a delay from the start of a pulse detected when the measured voltage is equal to or greater than a threshold.

16. The plasma processing chamber according to claim 10, wherein the matching network is connected to an RF bias power supply capable of operating at frequencies of approximately 100 kHz, 13.56 MHz, 15 MHz, 60 MHz, 120 MHz, or 162 MHz, and in at least one of continuous mode or pulsed mode, wherein in pulsed mode, the pulse frequency is from approximately 100 Hz to approximately 10 kHz and the duty cycle is from approximately 5% to approximately 95%.

17. The plasma processing chamber according to claim 10, wherein the matching network is connected to an RF source power capable of operating at frequencies of approximately 13.56 MHz, 60 MHz, 120 MHz, 162 MHz, or 200 MHz, and in at least one of continuous mode or pulsed mode, wherein in pulsed mode, the pulse frequency is from approximately 100 Hz to approximately 10 kHz and the duty cycle is from approximately 5% to approximately 95%.

18. The plasma processing chamber according to claim 10, wherein the at least one variable capacitor comprises a series variable capacitor and a shunt variable capacitor.

19. A method for processing a substrate, Measuring impedance at the input of a matched network configured to receive one or more radio frequency (RF) signals, and at the output of the matched network configured to send the one or more RF signals to a processing chamber, The at least one variable capacitor of the matching network is adjusted to a first target position based on a weighted output impedance value obtained by weighting the impedances measured in multiple different sections of the pulse based on one or more RF signals, and the at least one variable capacitor is adjusted to a second target position based on a weighted input impedance value obtained by weighting the impedances measured in multiple different sections of the pulse. Methods that include...

20. The weighted output impedance value and the weighted input impedance value are calculated using the following formula: Z ｗ =Z １ *w+Z ２ *(1-w)、 Here, Z １ However, it was measured at time 1 and pulse level 1, Z ２ The method according to claim 19, wherein the measurement is taken at time 2 and pulse level 2, where w is a weight value between 0 and 1.

Images