CN110945624A - Plasma chamber with electrode assembly - Google Patents

Abstract

A processing tool for plasma processing includes a chamber body having an interior volume providing a plasma chamber, a workpiece support having a ceiling and an opening on a side opposite the ceiling, an actuator, a gas distributor, an electrode assembly, and a first RF power source, the workpiece support holding a workpiece, such that at least a portion of the front surface of the workpiece faces the opening, the actuator producing relative motion between the chamber body and the workpiece support such that the opening moves laterally across the workpiece, the gas distributor delivering a process gas to the plasma chamber, the electrode assembly comprising a plurality of coplanar filaments, the plurality of coplanar filaments extend laterally through the plasma chamber between the workpiece support and the top plate, each filament of the plurality of filaments includes a conductor, the first RF power source providing a first RF power to the conductor of the electrode assembly to form a plasma.

Description

Plasma chamber with electrode assembly

Technical Field

The present disclosure relates to a processing tool that includes a plasma chamber, such as a plasma chamber for depositing a film on a workpiece (e.g., a semiconductor wafer), etching a workpiece, or processing a workpiece.

Background

Plasmas are typically generated using Capacitively Coupled Plasma (CCP) sources or Inductively Coupled Plasma (ICP) sources. The basic CCP source contains two metal electrodes similar to a parallel plate capacitor, separated by a small distance in a gaseous environment. One of the two metal electrodes is driven by a fixed frequency Radio Frequency (RF) power supply, while the other electrode is connected to RF ground, creating an RF electric field between the two electrodes. The generated electric field ionizes the gas atoms, releasing electrons. Electrons in the gas are accelerated by the RF electric field and directly or indirectly ionize the gas by collisions, creating a plasma.

Basic ICP sources typically comprise a spiral or coil-shaped conductor. When an RF current flows through the conductor, an RF magnetic field is formed around the conductor. The RF magnetic field, along with the RF electric field, ionizes the gas atoms and generates a plasma.

Plasmas of various process gases are widely used in the manufacture of integrated circuits. For example, plasma may be used for thin film deposition, etching, and surface treatment.

Atomic Layer Deposition (ALD) is a thin film deposition technique based on the sequential use of gas phase chemical processes. Some ALD processes use plasma to provide the necessary activation energy for chemical reactions. Plasma enhanced ALD processes may be performed at lower temperatures than non-plasma enhanced (e.g., "thermal") ALD processes.

Disclosure of Invention

In one aspect, a processing tool for plasma processing includes a chamber body having an interior space providing a plasma chamber, a chamber body having a ceiling and an opening on a side opposite the ceiling, a workpiece support holding a workpiece such that at least a portion of a front surface of the workpiece faces the opening, an actuator producing relative motion between the chamber body and the workpiece support such that the opening moves laterally across the workpiece, a gas distributor delivering a process gas to the plasma chamber, an electrode assembly including a plurality of coplanar filaments (coplanar films) extending laterally through the plasma chamber between the workpiece support and the ceiling, each filament of the plurality of filaments includes a conductor, the first RF power source providing a first RF power to the conductor of the electrode assembly to form a plasma.

Implementations may include one or more of the following features.

The workpiece support is rotatable about an axis of rotation, and the actuator can rotate the workpiece support such that rotation of the support carries the workpiece across (across) the opening.

A plurality of coplanar filaments may extend across the wedge-shaped region. The workpiece may be completely engaged (fit) within the wedge region such that in operation the entire front surface of the workpiece is exposed to the plasma. The workpiece may be larger than the wedge region such that in operation a wedge portion of the front surface of the workpiece is exposed to the plasma. The opening may be wedge-shaped.

The plurality of coplanar wires may be linear wires and different wires may have different lengths to define the wedge-shaped region. A plurality of coplanar filaments may extend in parallel. The plurality of coplanar filaments may be evenly spaced apart. Different filaments may be oriented at different angles. The plurality of coplanar filaments may be oriented such that a plasma density generated in the wedge-shaped region is lower at an apex (apex) of the wedge-shaped region than at a base (base) of the wedge-shaped region. The plurality of coplanar wires may be oriented to have longitudinal axes at non-zero angles with respect to a direction of motion of a portion of the substrate below the opening. The non-zero angle may be greater than 10 °.

The spacing between the coplanar filaments may be sufficient to avoid plasma region pinch (ping) between the region above and the region below the electrode assembly within the chamber. The bottom of the chamber may be open. The tool may include a top electrode on a top plate of the chamber.

The ends of the conductors of the plurality of coplanar filaments may be connected to a first RF power source through a recursive RF feed structure. The opposite ends of the conductors of the plurality of coplanar wires may be connected to a shared bus. The bus may be connected to the first RF power supply at two opposing locations.

The first plurality of conductors of the plurality of coplanar wires may be connected to a first RF power source, and the second plurality of conductors of the plurality of coplanar wires may be floating or grounded. First ends of the conductors of the plurality of coplanar wires may be coupled to a first RF power source through a shared bus. The conductors of the first set and the conductors of the second set may be arranged to alternate in a direction perpendicular to the longitudinal axis of the filament.

In another aspect, a plasma reactor includes a chamber body having an interior space that provides a plasma chamber, a gas distributor that delivers a process gas to the plasma chamber, a workpiece support that holds a workpiece, an electrode assembly including a plurality of conductors spaced apart from the workpiece support in a parallel coplanar array and extending laterally across the workpiece support, a first RF power source that provides a first RF power to the electrode assembly, and a dielectric baseplate between the electrode assembly and the workpiece support that provides an RF window between the electrode assembly and the plasma chamber.

Implementations may include one or more of the following features.

A plurality of conductors may be positioned between the dielectric top plate and the dielectric window. The dielectric top plate may be a ceramic body and the dielectric bottom plate may be quartz or silicon nitride.

The lower surface of the bottom plate may have a plurality of parallel slots, and a plurality of parallel coplanar conductors may be positioned in the plurality of parallel slots. A plurality of wires may be positioned in the plurality of slots. Each filament may include a conductor and a non-metallic shell surrounding the conductor. The shell may form a conduit and the conductor may be suspended in and extend through the conduit. The conductor may comprise a hollow conduit.

A plurality of conductors may be coated on the dielectric top plate. A plurality of conductors may be embedded in the dielectric top plate.

The plurality of conductors may be evenly spaced apart. The spacing between the workpiece support and the plurality of conductors may be 2mm to 50 cm.

The plurality of conductors may include a first plurality of conductors and a second plurality of conductors arranged in an alternating pattern with the first plurality of conductors. The RF power source may be configured to apply a first RF input signal to the first multiple conductor and a second RF input signal to the second multiple conductor. The RF power source may be configured to generate the first RF signal and the second RF signal at the same frequency. The RF power supply may be configured to generate the first RF signal and the second RF signal such that a phase difference between the first RF signal and the second RF signal is 180 °. The RF power source may be configured to provide an adjustable phase difference between the first RF signal and the second RF signal.

The plurality of conductors may have a plurality of first ends at a first side of the plasma chamber and a plurality of second ends at an opposite second side of the plasma chamber. The RF power supply can be configured to apply a first RF input signal to a first end of the first multiconductor and a second RF input signal to a second end of the second multiconductor. The second end of the first multiple conductor may be floating and the first end of the second multiple conductor may be floating. A first end of the first multiconductor can be connected to a first shared bus and a second end of the second multiconductor can be connected to a second shared bus. The first plurality of filaments may be grounded and the first end of the second plurality of filaments may be grounded.

A first end of the first multiconductor can be connected to a first shared bus located outside the plasma chamber on a first side of the chamber, and a second end of the second multiconductor can be connected to a second shared bus located outside the plasma chamber on a second side of the chamber. The second end of the first multiple conductor may be connected to a third shared bus located outside the plasma chamber on the second side of the chamber, and the first end of the second multiple conductor may be connected to a fourth shared bus located outside the plasma chamber on the first side of the chamber.

In another aspect, a plasma reactor includes a chamber body having an interior space providing a plasma chamber, a gas distributor delivering a process gas to the plasma chamber, an RF power source, and at least one RF switch, a gas distributor comprising a plurality of filaments extending laterally through the plasma chamber between a ceiling of the plasma chamber and the workpiece support, each filament comprising a conductor surrounded by a cylindrical insulating shell, wherein the plurality of filaments comprises a first plurality of filaments and a second plurality of filaments arranged in an alternating pattern with the first plurality of filaments, the first bus is coupled to the first multiplex wire, the second bus is coupled to the second multiplex wire, the RF power source applies an RF signal to the intra-chamber electrode assembly, the at least one RF switch is configured to controllably electrically couple and decouple the first bus to one of: i) ground, ii) an RF power source, or iii) the second bus.

Implementations may include one or more of the following features.

The at least one RF switch may include a plurality of RF switches connected in parallel between the first bus and one of: i) ground, ii) an RF power source, or iii) a second bus.

The at least one RF switch may be configured to controllably electrically couple and decouple the first bus to the second bus. The at least one RF switch may include a plurality of switches connected in parallel between different pairs of locations of the first bus and the second bus to controllably electrically couple and decouple the first bus to the second bus.

The at least one RF switch may include a first switch configured to controllably couple and decouple the first bus to ground and at least one second RF switch configured to controllably couple and decouple the second bus to ground. The at least one RF switch may include a first plurality of switches connected in parallel between different locations of the first bus and ground, and the at least one second switch may include a second plurality of switches connected in parallel between different locations of the second bus and ground. The different locations of the first bus may comprise opposite ends of the first bus and the different locations of the second bus may comprise opposite ends of the second bus.

The at least one RF switch may include a first plurality of switches connected in parallel between different positions of the first bus and the RF power source, and the at least one second switch may include a second plurality of switches connected in parallel between different positions of the second bus and the RF power source. The different locations of the first bus may comprise opposite ends of the first bus and the different locations of the second bus may comprise opposite ends of the second bus. The at least one RF switch may include a first plurality of switches connected in parallel between different locations of the first bus and the RF power source, and the at least one second switch may include a second plurality of switches connected in parallel between different locations of the second bus and ground. The different locations of the first bus may comprise opposite ends of the first bus and the different locations of the second bus may comprise opposite ends of the second bus.

The at least one RF switch includes a first switch configured to controllably electrically couple and decouple the first bus to the RF power source, and includes at least one second switch configured to controllably electrically couple and decouple the second bus to the RF power source.

Some implementations can include a third bus coupled to a first plurality of filaments and a fourth bus coupled to a second plurality of filaments, wherein the plurality of filaments have a plurality of first ends and a plurality of second ends, and the first end of each respective filament is closer to a first sidewall of the plasma chamber than the second end of the respective filament, and wherein the first bus is coupled to the first end of the first plurality of filaments, the second bus is coupled to the first end of the second plurality of filaments, the third bus is coupled to the second end of the first plurality of filaments, and the fourth bus is coupled to the second end of the second plurality of filaments.

The at least one RF switch may be configured to controllably electrically couple and decouple the first bus to and from the second bus, and may include at least one second RF switch configured to controllably electrically couple and decouple the third bus to and from the fourth bus.

The at least one RF switch may include a first switch configured to controllably couple and decouple the first bus to ground, and may include at least one second RF switch configured to controllably couple and decouple the third bus to ground.

The RF source may be coupled to the fourth bus through a first tap (tap) and to the second bus through a second tap.

Some implementations may include at least one third RF switch configured to controllably couple and decouple the third bus to ground, and at least one fourth RF switch configured to controllably couple and decouple the fourth bus to ground. The at least one RF switch may include a first switch configured to controllably couple and decouple the first bus to ground, and at least one second RF switch configured to controllably couple and decouple the second bus to the RF source, at least one third RF switch configured to controllably couple and decouple the third bus to ground, and at least one fourth RF switch configured to controllably couple and decouple the fourth bus to the RF source.

The at least one RF switch includes a first switch configured to controllably electrically couple and decouple the first bus to the RF source, and includes at least one second RF switch configured to controllably electrically couple and decouple the second bus to the RF source, at least one third RF switch configured to controllably electrically couple and decouple the third bus to the RF source, and at least one fourth RF switch configured to controllably electrically couple and decouple the fourth bus to the RF source.

In another aspect, a plasma reactor includes a chamber body having an interior space that provides a plasma chamber, a gas distributor that delivers process gas to the plasma chamber, a pump coupled to the plasma chamber to evacuate the chamber, a workpiece support that holds a workpiece, a chamber inner electrode assembly including a plurality of filaments extending laterally through the plasma chamber between a ceiling of the plasma chamber and the workpiece support, each filament including a conductor surrounded by a cylindrical insulating shell, a pump, a workpiece support, a chamber inner electrode assembly including a plurality of filaments extending laterally through the plasma chamber between the ceiling of the plasma chamber and the workpiece support, a bus outside the chamber and coupled to opposing ends of the plurality of filaments, the RF power source applying an RF signal to the chamber inner electrode assembly, the plurality of RF switches are configured to controllably electrically couple and decouple a plurality of different locations of the bus to one of: i) ground or ii) the RF power supply.

Certain implementations may have one or more of the following advantages. Plasma uniformity can be improved. The repeatability of the plasma process can be improved. The metal pollution can be reduced. Particle generation can be reduced. Plasma charging damage may be reduced. Plasma uniformity can be maintained under different process operating conditions. The plasma power coupling efficiency can be improved. The plasma region size can be reduced for a given size workpiece. Plasma process throughput can be improved. The workpieces may be carried through the plurality of chambers in succession while remaining on the support. The effect of the relative velocity during exposure to the plasma can be compensated for and hence within-wafer uniformity can be improved. The effect of local non-uniformity in the plasma region can be reduced by switching, and thus within-wafer uniformity can be improved. A low impedance RF ground may be provided. Particle generation can be reduced. Plasma charging damage may be reduced. Plasma uniformity can be maintained under different process operating conditions. The plasma power coupling efficiency can be improved. A grounded top electrode integrated with the gas distribution showerhead may be implemented for introducing gas in a uniform manner without creating unnecessary gas decomposition in the showerhead holes.

The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims.

Drawings

Fig. 1 is a schematic side view of an example of a processing tool including a plasma chamber.

Fig. 2A is a schematic top view of an example of a processing tool including a plasma chamber.

Fig. 2B-2C are cross-sectional side views of the processing tool of fig. 2A along section lines B-B and C-C, respectively.

Figures 3A-3C are schematic cross-sectional perspective views of various examples of filaments of an electrode assembly within a chamber.

Figure 4A is a schematic top view of a portion of an electrode assembly within a chamber.

Figures 4B-C are cross-sectional schematic side views of an electrode assembly in a chamber having different plasma region states.

Figures 5A-D are schematic top views of various examples of electrode assembly configurations within a chamber.

Fig. 6A is a schematic top view of an example of a processing tool.

Fig. 6B is a schematic top view of an example of a wedge-shaped electrode assembly.

Fig. 6C is a schematic top view of an example of a frame of a wedge-shaped electrode assembly.

Fig. 6D is a cross-sectional side view of an example of a frame of a wedge-shaped electrode assembly.

Fig. 6E is a schematic top view of an example of a wedge-shaped electrode assembly.

Fig. 7A-7D are conceptual diagrams of examples of electrical configurations of wedge-shaped electrode assemblies.

Fig. 8A is a schematic top view of an example of an electrode assembly.

Fig. 8B-8F are conceptual diagrams of examples of electrical configurations of a switching electrode assembly.

Fig. 9A-9B are conceptual illustrations of examples of mode-selectable switched electrode systems.

FIG. 10 is a conceptual schematic of an example of a switched wedge electrode system.

Fig. 11A is a schematic diagram of an example of a PIN diode switch.

Fig. 11B is a schematic diagram of an example of a saturable inductive switch.

Fig. 12A is a schematic diagram of an example of frequency-based switching.

12B-C are conceptual illustrations of examples of electrical configurations of frequency-switched electrode systems.

Fig. 13 is a schematic side view of an example of a plasma reactor.

Fig. 14A is a schematic top view of another example of a plasma reactor.

Fig. 14B and 14C are schematic side views of the plasma reactor of fig. 14A taken along lines 14B-14B and 14C-14C, respectively.

Fig. 15A-15C are schematic cross-sectional views of the electrode assembly.

Like reference symbols in the various drawings indicate like elements.

Detailed Description

In conventional plasma reactors, the workpiece remains stationary within the reaction chamber. A plasma region is generated above the stationary workpiece, which then treats the workpiece surface. However, certain plasma processing applications may benefit from moving the workpiece through the plasma region, i.e., relative motion between the plasma region and the workpiece. In addition, for some tools, the substrate is moved between different chambers to perform a series of processing steps.

One way to achieve relative motion between the workpiece and the plasma region is by placing the workpiece on a workpiece support that moves along a linear path, such as a conveyor belt. In such a configuration, the workpiece may make a single pass in one direction through the plasma region and exit on the other side of the chamber. This may be advantageous for certain sequential processes in which a workpiece travels through multiple chambers of different types as part of a manufacturing process.

Another way to achieve relative motion between the workpiece and the plasma region is by placing the workpiece on a rotating workpiece support. Rotating the workpiece support enables multiple passes through the plasma region without changing the direction of travel, which can improve throughput because the workpiece support does not need to continuously change the direction of travel of the workpiece. However, if the support is rotated, different regions of the workpiece may move at different speeds relative to the regional plasma.

In conventional CCP sources, plasma uniformity is generally determined by the electrode size and distance between electrodes, as well as the gas pressure, gas composition, and applied RF power. At higher radio frequencies, additional effects (additive effects) may become significant or even dominate the non-uniformity due to the presence of standing waves or skin effects (skin effects). This additional effect becomes more pronounced at higher frequencies and plasma densities.

Plasma uniformity in conventional ICP sources is generally determined by the configuration of the ICP coil, including its size, geometry, distance to the workpiece and associated RF window position, as well as gas pressure, gas composition and power. In the case of multiple coils or coil segments, the current or power distribution and the relative phase of the current or power distribution may also be significant factors if driven at the same frequency. Power deposition tends to occur within a few centimeters below or near the ICP coil due to the skin effect, and such localized power deposition often results in process non-uniformities that reflect the coil geometry. This plasma non-uniformity results in a potential difference across the workpiece, which may also lead to plasma charging damage (e.g., transistor gate dielectric breakdown).

A large diffusion distance is typically required to improve the uniformity of the ICP source. However, conventional ICP sources with thick RF windows are generally inefficient at high gas pressures due to low power coupling, which results in high drive currents, resulting in high resistive power losses. In contrast, the electrode assembly within the chamber need not have an RF window, but rather only a cylindrical shell. This may provide better power coupling and higher efficiency.

In a plasma chamber with a moving workpiece support, the moving workpiece support may be DC grounded by, for example, a rotating mercury coupler, a brush, or a slip ring. However, the moving workpiece support may not be sufficiently grounded at radio frequencies. The RF ground path should have a much lower impedance than the plasma so that the RF ground path becomes a sufficient RF ground. The lack of a sufficient RF grounding path can make it difficult to control the ion energy at the workpiece and reduce the repeatability of the process.

Therefore, there is a need for a plasma source having the following properties: the plasma source can efficiently produce a uniform plasma with desired characteristics (plasma density, electron temperature, ion energy, dissociation, etc.) over the workpiece size; the plasma source may adjust uniformity (e.g., pressure, power, gas composition) over the operating window; it has stable and repeatable electrical performance even if the workpiece is moving; and it does not produce excessive metal contamination or particles. The electrode assembly within the chamber may better provide one or more of these properties.

FIG. 1 is a schematic side view of an example of a processing tool. The processing tool 100 has a chamber body 102 enclosing an interior volume 104. The interior space 104 may be cylindrical, such as for receiving a circular workpiece support. At least a portion of the interior space serves as a plasma chamber or plasma reactor. The chamber body 102 has supports 106 for providing mechanical support for various components within the interior volume 104. For example, the support 106 may provide support for the top electrode 108. The top electrode may be suspended within the interior space 104 and may be spaced apart from, abut, or form a portion of the top plate. Portions of the sidewalls of the chamber body 102 may be grounded independent of the top electrode 108.

The gas distributor 110 is located near the ceiling of the plasma reactor section of the processing tool 100. In some implementations, the gas distributor 110 is integrated with the top electrode 108 as a single component. The gas distributor 110 is connected to a gas supply 112. The gas supply 112 delivers one or more process gases to the gas distributor 110, the composition of which may depend on the process to be performed, such as deposition or etching.

A vacuum pump 113 is coupled to the interior space 104 to evacuate the process tool. For some processes, the chamber is operated in the Torr range, and the gas distributor 110 supplies argon, nitrogen, oxygen, and/or other gases.

A workpiece support 114 for supporting a workpiece 115 is positioned in the processing tool 100. The workpiece support 114 has a workpiece support surface 114a facing the ceiling of the processing tool 100. For example, the workpiece support surface 114a may face the top electrode 108. The workpiece support 114 is operable to rotate about an axis 150. For example, the actuator 152 may turn the drive shaft 154 to rotate the workpiece support 114. In some implementations, the axis 150 coincides with the center of the workpiece support 114 (coincident).

In some implementations, the workpiece support 114 includes a workpiece support electrode 116 inside the workpiece support 114. In some implementations, the workpiece support electrode 116 may be grounded or connected to a grounded impedance or circuit. In some implementations, the RF bias power generator 142 is coupled to the workpiece support electrode 116 through an impedance match 144. The workpiece support electrode 116 may additionally comprise an electrostatic chuck, and a workpiece bias voltage supply 118 may be connected to the workpiece support electrode 116. The RF bias power generator 142 may be used to generate a plasma, control the electrode voltage or electrode sheath (sheath) voltage, or control the ion energy of the plasma.

In addition, the workpiece support 114 may have an internal passage 119 for heating or cooling the workpiece 115. In some implementations, an embedded resistive heater may be disposed inside the internal channel 119.

In some implementations, the workpiece support 114 is heated by radiation, convection, or conduction from a heating element located within the bottom interior space 133.

The electrode assembly 120 is positioned in the interior space 104 between the top electrode 108 and the workpiece support 114. The electrode assembly 120 includes one or more coplanar wires 300, the one or more coplanar wires 300 extending laterally in the chamber and above the support surface 114a of the workpiece support 114. At least a portion of the coplanar filaments of the electrode assembly 120 above the workpiece support 114 extend parallel to the support surface 114 a. Although the left side of fig. 1 depicts the wire 300 parallel to the direction of motion (into and out of the page) of the workpiece 115, the wire 300 may be at a non-zero angle relative to the direction of motion, such as substantially perpendicular to the direction of motion.

A top gap 130 is formed between the top electrode 108 and the electrode assembly 120 within the chamber. A bottom gap 132 is formed between the workpiece support 114 and the electrode assembly 120 within the chamber.

The inner space 104 may be divided by a barrier into one or more areas 101a, 101b, at least one of which serves as a plasma chamber. The obstruction defines one or more openings 123 above the workpiece support. In some implementations, the electrode assembly 120 is positioned within the opening 123. In some implementations, an electrode assembly is placed over the opening 123. In some implementations, the barrier is integrally formed from the support 106, and the opening 123 is formed on the support 106. In some implementations, the opening 123 formed on the support 106 is configured to support the electrode assembly 120.

The electrode assembly 120 is driven by an RF power source 122. The RF power source 122 can supply power to one or more coplanar filaments of the electrode assembly 120 at a frequency of 1MHz to over 300 MHz. For some processes, the RF power source 120 provides a total RF power of 100W to greater than 2kW at a frequency of 60 MHz.

In some implementations, it may be desirable to select the bottom gap 132 to allow plasma generated radicals, ions, or electrons to interact with the workpiece surface. The choice of gap depends on the application and on the method of operation. For some applications where a flux of free radicals (but a very low flux of ions/electrons) needs to be delivered to the workpiece surface, operation at larger gaps and/or higher pressures may be selected. For other applications where a flux of radicals and a substantial flux of plasma ions/electrons need to be delivered to the workpiece surface, operation at a smaller gap and/or lower pressure may be selected. For example, in some low temperature plasma enhanced ALD processes, radicals of the process gas are necessary for the deposition or processing of ALD films. A radical is an atom or molecule with an unpaired valence electron. Free radicals are generally highly chemically reactive towards other species. The reaction of radicals with other chemicals often plays a significant role in film deposition. However, free radicals are generally short-lived due to their high chemical activity and therefore cannot be transported far during their life cycle. Placing the radical source (i.e., the in-chamber electrode assembly 120 as a plasma source) near the surface of the workpiece 115 may increase the supply of radicals to the surface to improve the deposition process.

The life cycle of the free radicals is generally dependent on the pressure of the surrounding environment. Thus, the height of the bottom gap 132 that provides a satisfactory concentration of radicals may vary depending on the desired chamber pressure during operation. In some implementations, the bottom gap 132 is less than 1cm if the chamber is operated at a pressure in the range of 1 to 10 Torr. In other (lower) temperature plasma enhanced ALD processes, exposure to plasma ion flux (and accompanying electron flux) as well as radical flux may be necessary to deposit and process ALD films. In some implementations, the bottom gap 132 is less than 0.5cm if the chamber is operated at a pressure in the range of 1-10 Torr. Lower operating pressures can operate at larger gaps due to lower volumetric recombination rates with respect to distance. In other applications (e.g., etching), lower operating pressures (less than 100mTorr) are typically used and the gap may be increased.

In such applications where the bottom gap 132 is small, the plasma generated by the electrode assembly 120 may have significant non-uniformity between filaments, which may be detrimental to the processing uniformity of the workpiece. By moving the workpiece through a plasma having spatial non-uniformity, the effect of the spatial non-uniformity of the plasma on the process can be mitigated by a time-averaging effect (i.e., the cumulative plasma dose received by any given region of the workpiece after a single pass through the plasma is substantially similar).

The top gap may be selected to be large enough to allow plasma to develop between the electrode assembly and the top electrode (or top of the chamber) within the chamber. In some implementations, if the chamber is operated at a pressure in the range of 1-10Torr, the top gap 130 can be between 0.5-2cm, such as 1.25 cm.

The top electrode 108 may be configured in various ways. In some implementations, the top electrode is connected to RF ground 140. In some implementations, the top electrode is electrically isolated ("floating"). In some implementations, the top electrode 108 is biased to a bias voltage. The bias voltage may be used to control characteristics of the generated plasma, including ion energy. In some implementations, the top electrode 108 is driven with an RF signal. For example, driving the top electrode 108 relative to the already grounded workpiece support electrode 116 may increase the plasma potential at the workpiece 115. The increased plasma potential may cause the ion energy to increase to a desired value.

The top electrode 108 may be formed of different process compatible materials. Various criteria for process calculability include the resistance of the material to etching by process gases and to sputtering from ion bombardment. Furthermore, where material quantities are etched, process compatible materials preferentially form volatile or gaseous compounds (which can be evacuated by vacuum pump 113) and do not form particles that can contaminate workpiece 115. Thus, in some implementations, the top electrode is made of silicon. In some implementations, the top electrode is made of silicon carbide.

In some implementations, the top electrode 108 may be omitted. In such implementations, the RF ground path may be provided by a subset of the workpiece support electrode, the coplanar wires of the electrode assembly 120, or by the chamber wall or other ground-referenced (ground-referenced) surface in communication with the plasma.

In some implementations, the fluid supply 146 circulates fluid through channels in the electrode assembly 120 within the chamber. In some implementations, a heat exchanger 148 is coupled to the fluid supply 146 to remove or supply heat to the fluid.

Depending on the chamber configuration and the process gases supplied, the plasma reactor in the process tool 100 may provide an ALD apparatus, an etch apparatus, a plasma processing apparatus, a plasma enhanced chemical vapor deposition apparatus, a plasma doping apparatus, or a plasma surface cleaning apparatus.

Fig. 2A is a schematic top view of an example of a processing tool 200. The processing tool 200 is similar to the processing tool 100 except as described. The processing tool 200 has a cylindrical chamber body 202, an interior volume 204 having a cylindrical shape, a support 206, an electrode assembly 220, and a precursor station 260. The support 206 is centrally located within the processing tool 200 and forms a plurality of radial partitions 270 that divide the interior space 204 into multiple processing regions. For example, the plurality of treatment regions may be configured to have a wedge shape (e.g., a circular cross-section or an equilateral triangle), or may be cut away at the apex. The processing regions can be configured in various ways to achieve various functions as desired for the operation of the processing tool 200.

The precursor processing zone is configured to process workpiece 115 with one or more precursors (e.g., for an ALD process). For example, first precursor station 260a positioned within precursor processing region 280a may be configured to flow or pump chemical precursor a such that workpiece 115 is processed as workpiece 115 moves under precursor station 260 a. Precursor station 260a may then treat workpiece 115 with chemical precursor B in preparation for surface treatment of workpiece 115, such as surface treatment for ALD film formation plasma.

In some implementations, precursor processing region 280 includes a plurality of sub-regions with respective precursor stations 260 for respective chemical precursors. In some implementations, the sub-regions are arranged sequentially along the path of the workpiece 115. In some implementations, movement of the workpiece 115 is stopped during precursor surface treatment. In some implementations, workpiece 115 is continuously moved through precursor processing zone 280.

The gas isolation zone 281 is configured to provide spatial isolation of the respective processing environments of a plurality of processing regions, such as a first processing region and a second processing region. The gas isolation zone 281 may include a first pumping zone 282, a purge zone 283, and a second pumping zone 284, each separated by a respective radial partition 270. In conventional systems, isolation of the processing environment may be provided by a hermetic seal between the first and second processing regions. However, due to the rotating workpiece support 114, it may not be practical to provide such a seal. Conversely, a level of isolation sufficient for plasma processing applications (e.g., ALD) can be provided by interposing a gas isolation region 281 between the first and second processing regions.

Referring to FIG. 2B, a cross-sectional view of a portion of the processing tool 200 along section line B is illustrated. During operation, the first pumping zone 282 adjacent to a first processing region (e.g., precursor processing region 280a) creates a negative pressure differential relative to the first processing region. For example, a vacuum pump may be used to create the negative pressure differential. The negative pressure differential causes process gas that leaks from the first process region to be pumped out through the first pumping zone 282, as indicated by the arrows. Similarly, a second pumping zone 284 adjacent to the second processing region provides a negative pressure differential with respect to the second processing region (e.g., plasma processing region 285 a).

A purge zone 283 located between the first pumping zone 282 and the second pumping zone 284 supplies a purge gas. Examples of purge gases include non-reactive gases such as argon and nitrogen. The purge gas supplied by the purge zone 283 is pumped to the first and second pumping zones as indicated by the arrows due to the negative pressure differential created by the first and second pumping zones. The presence of the purge gas may prevent the respective process gases of the first and second processing regions from mixing with each other, which may cause unwanted chemical reactions resulting in unwanted deposition, etching or residue generation.

First gap height H₁Providing clearance between the radial spacer 270 and the workpiece support 114. The first gap height may be determined based on providing sufficient clearance for workpiece 115 to pass through while reducing leakage of process gas into pumping zones 282 and 284. For example, the first gap height may be in the range of 2-4mm, such as 3 mm.

Referring back to fig. 2A, plasma processing region 285 is configured to process workpiece 115 with a plasma. For example, an electrode assembly 220a located within plasma processing region 285a may generate plasma for processing the surface of workpiece 115. The precursor-treated surface of workpiece 115 that has moved through gas isolation region 281 is treated with the plasma generated by electrode assembly 220 a. In some implementations, the plasma treatment completes a deposition cycle of a single atomic layer of the first ALD film.

In some implementations, the electrode assembly 220 is formed in a rectangular shape as shown. In some implementations, the electrode assembly 220 is formed in a wedge shape.

Referring back to fig. 2B, in some implementations, the process gas for the plasma processing region 285 is provided through a gas inlet 210 formed adjacent to the electrode assembly 220. Specifically, the gas inlet 210 can be disposed at the edge of the gas isolation region 281 adjacent to the plasma processing region 285 a. For example, a channel may be formed between one of the separators 270 and the outer wall 221 of the electrode assembly 220 a.

Second gap height H₂A gap is provided between the electrode assembly 220 and the workpiece support 114. May be based on providing sufficient clearanceThe second gap height is determined by passing the workpiece 115 and providing a process gas to the interior region of the electrode assembly 220 while reducing leakage of the process gas into the pumping zones 282 and 284. For example, the second gap height may be in the range of 1-3mm, such as 2 mm. In some implementations, the gas inlet is formed on the entry side of the workpiece 115. In some implementations, the gas inlet is formed toward a radially outer edge of the electrode assembly, which is near the chamber wall 202. In some implementations, the gas inlet is formed toward the center of the workpiece support 114, such as near the axis 150.

In some implementations, the top electrode 208 is formed as part of the electrode assembly 220a or is supported by the electrode assembly 220 a. For example, the top electrode 208 may be supported by the top plate 221 a.

Referring to FIG. 2C, a cross-sectional view of a portion of the processing tool 200 along section line C is illustrated. In some implementations, as shown, the support 206 is configured to provide mechanical support for the electrode assemblies 220a and 220 b.

In some implementations, the processing tool 200 includes a second precursor processing region 280b and a second plasma processing region 285 b. Regions 280b and 285b may be configured to deposit a second ALD film. In some implementations, the second ALD film is the same as the first ALD film deposited in regions 280a and 285 a. Such implementations may provide improved deposition rates for a single ALD film. In some implementations, the second ALD film is different from the first ALD film. In such implementations, two different ALD films may be deposited in an alternating manner. In general, the process tool 200 may be configured to deposit 2, 3, 4, or more types of ALD films.

In general, workpiece 115 may make a single pass or may make multiple passes through the processing region. For example, the direction of rotation may be alternated to pass through a particular processing region multiple times.

In general, the treatment zones may be arranged in any order. For example, the precursor processing region may be followed by 2 different plasma processing regions having the same or different plasma characteristics.

With respect to fig. 1 or fig. 2A-2C, the electrode assembly 120 or 220 includes one or more coplanar wires 300, the one or more coplanar wires 300 extending laterally in the chamber and over a support surface of the workpiece support. At least a portion of the coplanar filaments of the electrode assembly above the workpiece support extend parallel to the support surface. The wire 300 may be at a non-zero angle relative to the direction of motion, such as substantially perpendicular to the direction of motion.

The electrode assembly may include a sidewall 221 surrounding the electrode plasma chamber region. The sidewalls may be formed of a process compatible material, such as quartz. In some implementations, the filaments project laterally out of the sidewall. In some implementations, the ends of the wire 300 extend out of the top plate of the electrode assembly and turn to provide a portion that is parallel to the support surface of the workpiece (see fig. 2C).

Figures 3A-C are schematic illustrations of various examples of filaments of an electrode assembly within a chamber. Referring to fig. 3A, a wire 300 of the electrode assembly 120 within the chamber is shown. The filament 300 includes a conductor 310 and a cylindrical shell 320, the cylindrical shell 320 surrounding the conductor 310 and extending along the conductor 310. The channel 330 is formed by the gap between the conductor 310 and the cylindrical shell 320. Cylindrical shell 320 is formed of a non-metallic material that is compatible with the process. In some implementations, the cylindrical shell is semi-conductive. In some implementations, the cylindrical shell is insulating.

The conductor 310 may be formed of various materials. In some implementations, the conductor 310 is a solid wire, such as a single solid wire having a diameter of 0.063 ". Alternatively, the conductor 310 may be provided by a multi-stranded wire. In some implementations, the conductor contains 3 parallel 0.032 "strands. Stranded wires can reduce RF power loss through the skin effect. In some implementations, the conductors 310 are formed of Litz wire (Litz wire), which may further reduce the skin effect.

Using materials with high conductivity (e.g. higher than 10)⁷Siemens per meter (Siemen/m)), which may reduce resistive power losses. In some implementations, the conductor 310 is made of copper or a copper alloy. In some implementations, the conductor is made of aluminum.

Unnecessary sputtering or etching of materials may lead to process contamination or particle formation. Whether the electrode assembly 120 is used as a CCP source or as an ICP source in the chamber, unnecessary sputtering or etching may occur. Unnecessary sputtering or etching may be caused by excessive ion energy at the electrode surface. When operating as a CCP source, an oscillating electric field around a cylindrical shell is required to drive the plasma discharge. This oscillation causes sputtering or etching of the material because the sputtering energy threshold of all known materials is below the corresponding minimum operating voltage of the CCP source. When operating as an ICP source, the capacitive coupling of the filament 300 with the plasma generates an oscillating electric field at the nearby surface, which also causes sputtering of the material. By using process compatible materials for the outer surface of the wire 300 that is exposed to the interior space 104 (e.g., cylindrical shell 320), problems caused by unnecessary sputtering or etching of materials may be mitigated.

In some implementations, the cylindrical shell 320 is formed from a process compatible material, such as silicon (e.g., high resistivity silicon), an oxide material, a nitride material, a carbide material, a ceramic material, or a combination thereof. Examples of oxide materials include silicon dioxide (e.g., silica, quartz) and aluminum oxide (e.g., sapphire). Examples of carbide materials include silicon carbide. For certain chemical environments, including fluorine-containing environments or fluorocarbon-containing environments, ceramic materials or sapphire may be desirable. In a chemical environment containing ammonia, dichlorosilane, nitrogen and oxygen, silicon carbide or quartz may be required.

In some implementations, the cylindrical shell 320 has a thickness of 0.1mm to 3mm, such as 2 mm.

In some implementations, a fluid is provided in the channel 330. In some implementations, the fluid is a non-oxidizing gas to purify oxygen to mitigate oxidation of the conductor 310. Examples of non-oxidizing gases are nitrogen and argon. In some implementations, the non-oxidizing gas is continuously flowed through the channels 330, such as by the fluid supply 146, to remove residual oxygen or water vapor.

Heating the conductor 310 may make the conductor more susceptible to oxidation. The fluid may provide cooling to the conductor 310 that may be heated from the supplied RF power. In some implementations, a fluid is circulated through the channels 330, such as by the fluid supply 146, to provide forced convection temperature control, such as cooling or heating.

In some implementations, the fluid may be near or above atmospheric pressure to prevent fluid breakdown (breakthrough). For example, gas decomposition or unwanted plasma formation in the tube can be prevented by providing a fluid pressure above 100 Torr.

Referring to fig. 3B, in some implementations of the wire 300, the conductor 310 has a coating 320. In some implementations, the coating 320 is an oxide of the material forming the conductor (e.g., aluminum oxide on an aluminum conductor). In some implementations, the coating 320 is silicon dioxide. In some implementations, the coating 320 is formed in situ in a plasma reactor of the processing tool 100, such as by reaction of silane, hydrogen, and oxygen to form a silicon dioxide coating. In situ coating may be advantageous because it may be replenished during etching or sputtering. The in situ coating may have a thickness in the range of 100nm to 10 μm.

Referring to fig. 3C, in some implementations of the wire 300, the conductor 310 is hollow and a hollow channel 340 is formed inside the conductor 310. In some implementations, as depicted in fig. 3A, the hollow channel 340 can carry a fluid. A coating of a process compatible material may cover the conductor 310 to provide a cylindrical shell 320. In some implementations, the coating 320 is an oxide of the material forming the conductor (e.g., aluminum oxide on an aluminum conductor). In some implementations, the hollow conductor 310 has an outer diameter of 2mm, with a wall thickness of 0.5 mm.

FIG. 4A is a schematic view of a portion of an electrode assembly within a chamber. The electrode assembly 400 in the chamber includes a plurality of coplanar wires 300 attached at a support 402. The electrode array is formed from a plurality of coplanar wires 300. Electrode assembly 400 may provide electrode assembly 120. In some implementations, the filaments 300 extend parallel to each other, at least over an area corresponding to where the workpiece has been processed.

The filaments 300 are separated from each other by a filament spacing 410. The spacing 410 may affect plasma uniformity. If the spacing is too large, the filaments can produce shadowing and non-uniformity. On the other hand, if the spacing is too small, the plasma cannot move between the top gap 130 and the bottom gap 132, and the non-uniformity will increase or the radical density will decrease.

In general, the desired value of the filament spacing 410 depends on several factors. Examples of such factors include chamber pressure, RF power, distance between the filament 300 and the workpiece 115, and process gas composition. For example, the filament spacing 410 may be increased when operating at lower pressures (e.g., below 2Torr) and having a large distance (e.g., greater than 3mm) between the filament and the workpiece.

In some implementations, the wire spacing 410 is uniform throughout the assembly 400. The filament spacing 410 may range from 3mm to 20mm, such as 8 mm.

Fig. 4B-C are schematic cross-sectional views of electrode assemblies in a chamber having different plasma region states. Referring to fig. 4B, under some operating conditions, a plasma region 412 surrounds the filament 300. An example of such an operating condition may include driving all filaments with the same RF signal (i.e., "monopoles") and having a grounded top electrode. Plasma region 412 has an upper plasma region 414 and a lower plasma region 416. Upper plasma region 414 may be located at top gap 130 and lower plasma region 416 may be located at bottom gap 132. As shown in fig. 4B, upper plasma region 414 and lower plasma region 416 are connected by a gap between filaments 300 to form continuous plasma region 412. This continuity of plasma region 412 is expected because regions 414 and 416 "communicate" with each other through the exchange of plasma. The exchange of the plasma helps to maintain electrical balance of the two regions, which helps in the stability and repeatability of the plasma.

Referring to fig. 4C, in this state, the upper plasma region 414 and the lower plasma region 416 are not connected to each other. Such "necking" of the plasma region 412 is undesirable for plasma stability. The shape of the plasma region 412 can be altered by various factors to remove plasma region discontinuities or improve plasma uniformity.

In general, regions 412, 414, and 416 may have a wide range of plasma densities and are not necessarily uniform. Furthermore, the discontinuity between upper plasma region 414 and lower plasma region 416 shown in FIG. 4C represents a substantially low plasma density relative to the two regions, and is not necessarily completely free of plasma in the gap.

Under some operating conditions, such as when the top electrode is not present or floating, and the workpiece support electrode is grounded, the plasma region 414 may not be formed, or may have a low plasma density.

In some implementations, the electrode assembly 400 may include first and second sets of filaments 300. The first and second groups may be spatially arranged such that the filaments alternate between the first and second groups. For example, the first group may include filament 302 and the second group may include filaments 300 and 304. The first group may be driven by a first terminal 422a of the RF power supply 422 and the second group may be driven by a second terminal 422b of the RF power supply 422. The RF power supply 422 may be configured to provide a first RF signal at terminal 422a and a second RF signal at terminal 422 b. The first and second RF signals may have the same frequency and a stable phase relationship with each other. For example, the phase difference between the first and second RF signals may be 0 degrees or 180 degrees. In some implementations, the phase relationship between the first and second RF signals provided by the RF power supply 422 may be adjustable between 0 and 360. In some implementations, the RF supply 422 can include two separate RF power supplies that are phase locked to each other.

Under some operating conditions, for example, when the phase difference between the first and second RF signals is 180, the resulting plasma region may be concentrated between the filaments.

The top gap 130 is a factor that affects the shape of the plasma region. Reducing the top gap 130 generally results in a reduction in plasma density in the upper plasma region 414 when the top electrode 108 is grounded. The particular value of the top gap 130 may be determined based on computer simulations of the plasma chamber. For example, the top gap 130 may be 3mm to 8mm, such as 4.5 mm.

The bottom gap 132 is a factor that affects the shape of the plasma region. Reducing the bottom gap 132 generally results in a reduction in plasma density in the lower plasma region 416 when the workpiece support electrode 116 is grounded. The particular value of the bottom gap 132 may be determined based on computer simulations of the plasma chamber. For example, the bottom gap 132 may be 3mm to 9mm, such as 4.5 mm.

In general, chamber pressure is a factor that affects the shape of the plasma region.

Fig. 5A and 5B are schematic diagrams of various examples of electrode assembly configurations within a chamber. Referring to fig. 5A and 5B, in some implementations, the electrode assembly 106 can include a first set of conductors 120a and a second set of conductors 120B. The first and second sets of conductors 120a, 120b may be arranged in an alternating pattern, at least within the plasma chamber 104. The first group may be driven by a first terminal 122a of the RF power supply 122 and the second group may be driven by a second terminal 122b of the RF power supply 122. The RF power supply 122 may be configured to provide a first RF signal at terminal 122a and a second RF signal at terminal 122 b. The first and second RF signals may have the same frequency and a stable phase relationship with each other. For example, the phase difference between the first and second RF signals may be 180 degrees. By driving the conductors 120a, 120b with RF signals having a phase difference of 180 degrees, the resulting plasma distribution may have less sensitivity to imperfect RF grounding of the electrode 116. Without being bound by any particular theory, this may be because the RF current returns through adjacent electrodes due to the differential nature of the drive signal. In some implementations, the phase relationship between the first and second RF signals provided by the RF power supply 122 may be adjustable between 0 and 360.

To generate the signal, the unbalanced output signal from the oscillator of the RF power supply may be coupled to a balun (balance-unbalanced transformer) 124, which balun 124 outputs a balanced signal on terminals 122a, 122 b. Alternatively, the RF supply 122 may comprise two separate RF power supplies that are phase locked to each other.

Referring to fig. 5A, the electrode assembly 120 includes a first electrode subassembly 510 and a second electrode subassembly 520, the first electrode subassembly 510 including a first set of conductors 120a, and the second electrode subassembly 520 including a second set of conductors 120 b. The conductors 120a of the first electrode subassembly 510 and the conductors 120b of the second electrode subassembly 520 are interdigitated.

The subassemblies 510, 520 each have a plurality of parallel conductors 120a, 120b, the plurality of parallel conductors 120a, 120b extending across the chamber 104. Every other electrode 120, such as electrode 120a, is connected to a first bus 530 on one side of the chamber 104. The remaining (alternating) electrodes 120 (i.e., electrode 120b) are each connected to a second bus 540 on the other side of the chamber 104. The ends of the respective conductors 120 not connected to the RF power supply bus may remain unconnected, e.g., floating.

The first bus 530 may be connected to the first terminal 122a, and the second bus may be connected to the second terminal 122 b. The first electrode subassembly 510 and the second electrode subassembly 520 are oriented parallel to each other such that the conductors of the subassemblies 510 and 520 are parallel to each other.

In some implementations, the buses 530, 540 connecting the conductors 120a, 120b are located outside of the interior space 104. This is better for improving uniformity within the chamber 104. However, in some implementations, the busses 530, 540 connecting the conductors 120a, 120b are located in the interior space 104.

FIG. 5B depicts an electrode assembly 106 that is similar to the implementation shown in FIG. 5A, but the ends of the respective conductors 120 that are not connected to the RF power supply bus may be grounded, such as a bus connected to ground. For example, the electrode 120a may be connected to the chamber 104 as a third bus 550 on the side of the second bus 550, and the electrode 120b may be connected to a fourth bus 560 on the same side of the chamber 104 as the first bus 530. Each bus 550, 560 may be grounded through an adjustable impedance 580, such as an impedance matching network.

With respect to fig. 5A or 5B, optionally, a low frequency common mode bias may be applied between the electrode subassemblies 510, 520. This can controllably increase the plasma potential.

Fig. 5C depicts the in-chamber electrode assembly 106, the in-chamber electrode assembly 106 including a first electrode subassembly 522 and a second electrode subassembly 532, the first electrode subassembly 522 and the second electrode subassembly 532 configured such that the filaments of the subassemblies 522 and 532 extend at a non-zero angle (e.g., perpendicular to each other).

The electrode assembly 106 within the chamber may be driven with the RF signal in various ways. In some implementations, the subassembly 522 and the subassembly 532 are driven with the same RF signal relative to RF ground. In some implementations, subassembly 522 and subassembly 532 are driven with differential RF signals. In some implementations, the subassembly 522 is driven with an RF signal, and the subassembly 532 is connected to an RF ground.

Fig. 5D illustrates the in-chamber electrode assembly 106, the in-chamber electrode assembly 106 including a first electrode subassembly 524 and a second electrode subassembly 534, the first electrode subassembly 524 overlapping (overlapping) the second electrode subassembly 534. The first and second electrode subassemblies 524 and 534 each have a plurality of parallel wires 300, the plurality of parallel wires 300 being connected at respective ends of respective bus bars by bus bars 530, 540, 550 and 560. First electrode subassembly 524 and second electrode subassembly 534 are configured such that the filaments of subassemblies 524 and 534 are parallel to each other, the filaments of subassemblies 524, 534 being arranged in an alternating pattern.

The electrode assembly 106 within the chamber may be driven with the RF signal in various ways. In some implementations, the subassembly 524 and the subassembly 534 are driven with the same RF signal relative to RF ground. In some implementations, the subassembly 524 and the subassembly 534 are driven with differential RF signals. In some implementations, the subassembly 524 is driven with an RF signal, and the subassembly 534 is connected to an RF ground.

In some implementations, the chamber inner electrode assembly 106 is driven with an RF signal in a single-ended manner using the center feed 590. The center feed 590 is connected at the center to an X-shaped current splitter 592. The four corners of the subassemblies 524 and 534 are connected to X-shaped current splitters 592 using a vertical feed structure.

Generally, differential driving of the subassemblies 510, 522, 524 and the corresponding subassemblies 520, 532, 534 can improve plasma uniformity or process repeatability when sufficient RF grounding cannot be provided (e.g., RF grounding by rotating mercury couplers, brushes, or slip rings).

Fig. 6A is a schematic top view of an interior region of an example of a processing tool 650. In the processing tool 650, the workpiece support 114 rotates about the axis 150, and rotation of the workpiece support 114 causes the workpiece 115 to move beneath the electrode assembly 600, through the plasma region generated by the electrode assembly 600. Unless otherwise noted, the process tool 650 is similar to the process tool 200, and the electrode assembly 600 is similar to the electrode assembly 400.

As workpiece 115 rotates about axis 150 through the plasma region, the velocities experienced by different surface regions of the workpiece vary as a function of their radial distance from axis 150. For example, regions of the workpiece that are farther from the axis 150 move faster than regions closer to the axis 150. For a rectangular or linear plasma region, regions of the workpiece further from the axis 150 experience correspondingly shorter residence times in the plasma region. This radial non-uniformity in residence time results in non-uniformity in the amount of plasma dose received on the workpiece, causing unnecessary process non-uniformity.

One way to compensate for the aforementioned non-uniformity of residence time is to vary the local density of the plasma region in proportion to the local velocity of the wafer. For example, the local plasma density may be increased in proportion to the radial distance from the axis 150. By increasing the plasma density at the local higher velocity regions, these regions receive equal amounts of plasma for their respective shorter residence times. However, the spatial non-uniformity of the plasma density may result in non-uniform charging of the workpiece surface, thereby creating a potential difference across the workpiece surface. Depending on the grain size and device sensitivity, a sufficiently large potential difference across the surface (e.g., greater than 2 volts, 5 volts, 10 volts, 15 volts, 25 volts) may cause damage to devices fabricated on the workpiece, such as dielectric breakdown of a thin transistor gate dielectric layer.

Another way to compensate for residence time non-uniformities is by changing the geometry of the plasma region. The geometry of the plasma region may be varied such that regions of higher local velocity travel through a correspondingly longer portion of the plasma region to equalize (equalize) the residence times of different regions of the workpiece surface. For the configuration shown in fig. 6A, a wedge shaped plasma region may be used to achieve residence time equalization. In such a configuration, the radial increase in local velocity by moving away from the axis 150 can be counteracted by a proportional increase in the arc length of the wedge-shaped plasma region on each region.

The aforementioned wedge-shaped plasma region may be formed by configuring the coplanar filaments and openings 627 of the electrode assembly 600 in various ways. One way is to configure the electrode array formed by the filaments of the electrode assembly 600 in a wedge-like manner. For example, the respective lengths of the individual coplanar filaments of the electrode array may be varied such that the overall profile of the electrode array defines a wedge shape. In some implementations, the support 206 may provide support at respective ends of the coplanar wires of the electrode array.

Another way to form a wedge-shaped plasma region is to have a wedge shape by forming opening 627 and to form a plasma region larger than the size of opening 627 using an electrode array of electrode assemblies 600 larger than opening 627 (e.g., electrode assembly 400). A portion of the generated plasma region may then be blocked by the wedge-shaped opening to generate a wedge-shaped plasma region. For example, the support 206 may provide a wedge-shaped opening 627.

In general, various factors may affect the size of the wedge-shaped plasma region. In some applications, partial or incomplete plasma coverage over the workpiece surface can lead to adverse results. For example, workpiece 115 may contain devices that are sensitive to charging damage, such as transistors with thin gate dielectric layers. In this case, the potential generated between the region of workpiece 115 exposed to the plasma and the region not exposed to the plasma may cause dielectric breakdown of the gate dielectric layer, causing permanent damage to sensitive devices. By sizing the plasma region larger than the workpiece, this problem can be mitigated, thereby achieving complete plasma coverage over the entire workpiece surface. In some implementations, the plasma region is sized such that the workpiece can move through the plasma region while maintaining complete plasma coverage.

In some implementations, for example where the plasma area is larger than the workpiece, the timing of applying RF power to electrode assembly 600 can be coordinated with the movement of workpiece 115 to ensure that the entire surface experienced by the workpiece is uniformly exposed to the plasma. For example, a plasma can be generated (ignited) after the entire workpiece is moved under the opening 627 or electrode assembly 600, and turned off (extinguished) before the workpiece leaves the plasma region. In this case, the plasma region need not be wedge-shaped.

However, in some cases, it can be challenging to generate a large plasma region (e.g., greater than 300mm by 300mm) using the electrode assembly 600. The size of the plasma region can be adjusted to be smaller than the workpiece surface in one direction of the workpiece if the workpiece to be treated can withstand incomplete plasma coverage on the surface of the workpiece. For example, as shown in FIG. 6A, the wedge electrode assembly 600 (and thus the plasma region) is smaller than the workpiece diameter in the direction of travel of the workpiece 115, but larger in the radial direction relative to the axis 150 to achieve full coverage in the radial direction.

Other considerations for adjusting the size of the plasma region include workpiece movement speed, target processing rate, and target plasma exposure time to achieve a desired processing duration or throughput.

In some implementations, the plasma may be coordinated with the movement of the workpiece to ensure that a stable plasma is established before the workpiece enters the plasma region. For example, in processes requiring relatively short plasma exposure times, the time it takes to strike the plasma can be a significant fraction of the total plasma exposure time. Because the plasma is relatively unstable during the strike phase, the resulting process repeatability may be compromised. By establishing a stable plasma prior to introduction into the workpiece, plasma exposure time and dose can be precisely controlled by controlling the speed of the workpiece as it moves through the plasma region. For such implementations, whether the plasma region is larger or smaller than the workpiece, it is advantageous that the plasma region be tapered to compensate for the difference in exposure time. In some implementations, the generated plasma is maintained over the processing of multiple workpieces.

Given a fixed plasma region size of the processing tool 650, various process parameters can be controlled to achieve desired plasma processing characteristics. Examples of process parameters that may be controlled include process rate, exposure time, workpiece travel speed profile, number of plasma exposure passes, and total plasma exposure dose. For example, the workpiece may pass through the plasma region multiple times or may oscillate at a location within the plasma region.

Fig. 6B is a schematic top view of an example of a wedge-shaped electrode assembly for generating a wedge-shaped plasma region. Wedge-shaped electrode assembly 600 has a plurality of coplanar wires 610 and a frame 620. Electrode assembly 600 is similar to electrode assemblies 120, 220, and 400, unless otherwise noted. Frame 620 has a first end 602, a second end 604, a center angle θ ca, and an inner radius R₁Outer radius R₂And a bisector line 605. The first end 602 is the short end of the electrode assembly 600, sometimes referred to as the apex. The second end 604 is the longer end, sometimes referred to as the base, of the electrode assembly 600. The plurality of coplanar filaments 610 is similar to the filament 300, unless otherwise noted. Each coplanar filament 610 has a respective length L and a respective angle θ (theta) with respect to the bisector 605. The length L is defined as the linear portion of the coplanar filaments 610 in a region parallel and adjacent to the workpiece support surface (e.g., 114 a). Each pair of adjacent coplanar filaments 610 is separated by a respective spacing S, defined as the center-to-center distance between adjacent filaments. For non-parallel filaments, spacing S is defined as the minimum center-to-center spacing along the length of the pair of adjacent filaments.

There are various considerations for determining the angle theta of the filament 610. One consideration in determining angle theta is the trajectory of workpiece 115 as workpiece 115 moves under electrode assembly 600. In some cases, the plasma generated by the electrode assembly 600 may have non-uniformities in the plasma that extend along the direction of the filament 610. For example, under certain operating conditions, there may be an elongated region of reduced plasma density between a pair of filaments 610. If a point on the workpiece surface travels along such a region of reduced plasma density, the point will receive a reduced amount of plasma exposure, resulting in process non-uniformity. By arranging the filaments to have an appropriate value of theta (e.g., less than or greater than 90 deg., but not including 90 deg.), such tangential travel along the region of reduced plasma density can be reduced, thereby improving process uniformity. For example, by setting theta to 60 °, a point on the workpiece surface passes under the plurality of filaments, exposed to a localized plasma region having reduced density and nominal density along the way, such that there is a time-averaging of the plasma exposure dose. In some implementations, the respective theta of the plurality of coplanar filaments 610 is equal, i.e., the filaments are parallel.

In some implementations, the respective θ of the filaments 610 differ based on their respective positions within the electrode assembly 600. For example, for filaments near apex 602 to filaments near base 604 of assembly 600, the respective theta increases monotonically to maintain equal lengths of filaments 610 across electrode assembly 600. Filaments having equal lengths may improve uniformity when the assembly 600 is operated as an ICP source.

Generally, the number of coplanar filaments 610 is determined by the size of the plasma region, theta, and spacing S to achieve desired plasma region characteristics, such as plasma density, uniformity.

In general, the spacing S may be determined based on considerations discussed in fig. 4 with respect to the filament spacing 410.

The frame 620 defines the shape of the electrode assembly 600 and the shape of the plasma region formed by the electrode assembly 600. The inner radius, outer radius and central angle determine the size of the wedge electrode, which in turn defines the size of the plasma region. The dimensions of the frame may be determined based on the previous discussion of adjusting the dimensions of the plasma region with respect to fig. 6B.

The frame 620 may be formed of different process compatible materials. Suitable process compatible materials include those described with respect to cylindrical shell 320, such as quartz. Other examples of process compatible materials include ceramics (e.g., alumina, aluminum nitride) and various silicon nitrides (e.g., SiN, Si₃N₄)。

Although the frame 620 has been described with respect to the wedge-shaped electrode assembly 600, the wire 610 may be formed and arranged to have the described wedge-shape without the frame 620 to achieve a similar result.

Examples of wedge-shaped electrode assemblies have the following design characteristics: r₁＝91mm、R₂427mm, 31 ° center angle, 60 ° theta, 15mm filament center-to-center spacing, 20 filaments, and quartz.

Referring to fig. 6C, in some implementations, the frame 620 has a cutout 622. The cut 622 may be shaped to fit the wedge-shaped top electrode 624. The wedge shaped top electrode 624 may be grounded or biased to a bias voltage. The tapered top electrode 624 may be formed from various process compatible materials, such as silicon. In some implementations, the wedge electrode is shaped to be inserted into the cut 622 to fill (fill) the cut 622.

Referring to fig. 6D, a cross-sectional view of a portion of the frame 620 along section line a is shown. In some implementations, the frame has an upper portion 625, an inner sidewall 626, and an opening 627.

In general, the respective lengths L of the plurality of coplanar filaments 610 are set to produce a desired shape of the plasma region. Frame 620 may be shaped to provide support for coplanar wires 610. In some implementations, the ends of the coplanar wires 610 are supported by the inner side walls 626 of the frame 620, similar to the configuration shown in fig. 6B. In some implementations, the ends of coplanar wires 610 are bent (e.g., 90 °) to be supported by an upper portion 625 of frame 620, as shown in electrode assembly 220a of fig. 2B. In some implementations, the opening 627 of the frame 620 may determine the shape of the plasma region.

In some implementations, theta is close to 0, e.g., <20 °. Referring to fig. 6E, assembly 601 has two filaments and the filaments are arranged at θ ═ 0 °, i.e., the filaments are parallel to bisector 605. The frame 620 of the assembly 601 has a cutout 622 and a wedge electrode 624. Wedge electrode 624 may be grounded. In such a configuration, the shape of the plasma region created by the electrode assembly 601 is affected by the interaction between the filament 610 and the wedge electrode 624, thereby creating a wedge-shaped plasma region. In configurations where theta approaches 0 deg., the effect of plasma non-uniformity parallel to filament 610 may be reduced as the direction of travel of workpiece 115 approaches substantially 90 deg. relative to the orientation of filament 610.

Fig. 7A-7D are conceptual illustrations of various electrical configurations of a wedge-shaped electrode assembly. The filaments of the electrode assembly may be electrically connected in a variety of different configurations. Referring to fig. 7A, an electrode assembly 700 is similar to electrode assembly 600 and has a first bus bar 730 and a second bus bar 740. The first bus 730 and the second bus 740 may be located on opposite sides of the chamber body 102, such as outside the chamber.

First bus 730 has a first end 750 and a second end 751 opposite first end 750. The first bus 730 and the second bus 740 are electrically connected to respective opposite ends of the respective filaments 710 of the electrode assembly 700. Unless otherwise noted, filament 710 is similar to filament 300. The electrode assembly 700 can be driven in various ways using one or more RF power sources.

In some implementations, a first RF power source drives the first bus 730, and the second bus 740 is connected to RF ground. In such a configuration, RF current flows through the filament 710, and the electrode assembly may be used primarily as an ICP plasma source.

In some implementations, the first RF power source drives the first bus 730, and the second bus 740 is electrically floating. In such a configuration, the electrode assembly may be used primarily as a CCP plasma source. The RF current return path may be provided by the chamber body 102, the upper electrode 108, the wedge-shaped top electrode 624, or the workpiece support electrode 116.

In some implementations, a first RF power source drives first bus 730 at first end 750, a second RF power source drives first bus 730 at second end 751, and second bus 740 is connected to RF ground. In such a configuration, the electrode assembly may be used primarily as an ICP plasma source.

In some implementations, a first RF power source drives the first bus 730 and a second RF power source drives the second bus 740.

Generally, the RF drive point at which the RF power supply is connected to the bus is selected to optimize the uniformity of the resulting plasma. For example, the drive point location may be selected based on minimizing non-uniformity in the RF signal amplitude experienced by the individual filaments 710.

In some implementations, the intraluminal electrode assembly can include first and second sets of coplanar filaments. The filaments of the first and second groups may be arranged in an alternating pattern in a direction perpendicular to the longitudinal axis of the filaments. As such, the coplanar filaments alternate between the first and second sets.

Referring to fig. 7B, electrode assembly 702, which is similar to electrode assembly 600, has a first group and a second group, the first group may include co-planar filaments 710 and 714, and the second group includes co-planar filaments 712. The first set is electrically connected to a first bus 732 and the second set is electrically connected to a second bus 742. The end of each wire remote from the bus to which it is connected may be "floating" or grounded. If the ends of the filaments are floating, then the two sets of filaments can be considered to form an interdigitated array.

The first bus 732 may have a first end 752 and a second end 753 opposite the first end 752. In some implementations, the first RF power supply drives the first bus 732 with a first RF signal and the second RF power supply drives the second bus 742 with a second RF signal. The first and second RF signals may have the same frequency and stable phase relationship with each other. For example, the phase difference between the first RF signal and the second RF signal may be 0 degrees or 180 degrees. In some implementations, the phase relationship between the first and second RF signals provided by the RF power supply 422 may be adjustable between 0 and 360. In some implementations, the RF supply 422 can include two separate RF power supplies 422a and 422b that are phase locked to each other.

In some implementations, a first RF power source drives the first bus 732 and the second bus 742 is connected to RF ground. In this case, the second bus 742 and the even array of wires connected to the second bus 742 can be used as RF current return paths.

In some implementations, a first RF power source drives the first bus 732 at a first end 752 and a second RF power source drives the first bus 732 at a second end 753, and the second bus 742 is connected to RF ground.

In some implementations, a first RF power source drives the first bus 732 and a second RF power source drives the second bus 742. In this case, the electrode assembly 702 may be used primarily as a CCP plasma source. The RF current return path may be provided by the chamber body 102, the top electrode 108, the wedge-shaped top electrode 624, or the workpiece support electrode 116.

Referring to FIG. 7C, electrode assembly 704, similar to electrode assembly 600, has a single bus 734. Bus 734 is electrically connected to both ends of filament 710.

In some implementations, a first RF power source drives the first bus 734. The first bus 734 may have a first end 754 and a second end 755, and in some implementations a first RF power supply drives the first bus 734 at the first end 754 and a second RF power supply drives the first bus 734 at the second end 755. In such a configuration, the electrode assembly may be used primarily as a CCP plasma source. The RF current return path may be provided by the chamber body 102, the top electrode 108, the wedge-shaped top electrode 624, or the workpiece support electrode 116.

Referring to FIG. 7D, an electrode assembly 706, similar to electrode assembly 600, has a first bus 736 and a second bus 746. The first bus 736 and the second bus 746 are electrically connected to respective opposite ends of the wire 710 of the electrode assembly 706. The first RF power source drives the first bus 736 at a drive point 756. The second bus 746 may be connected to RF ground.

The first RF signal generated by the first RF power source may be attenuated by various RF loss sources. The RF transmission lines forming the bus 736 are lossy, for example, due to the finite conductivity of the conductors or due to the loss tangent (loss tangent) caused by the dielectric material forming the transmission lines. As another example, the plasma load of the RF transmission line affects the RF losses. Thus, wires 710 connected at different locations along the direction of propagation of the RF signal may experience different RF signal amplitudes. For example, referring to fig. 7A, the RF signal transmitted at first end 750 will attenuate as the RF signal propagates down the length of first bus 730. As such, the RF signal amplitude at the filament 710 near the second end 751 will be less than the RF signal amplitude at the filament 710 near the first end 750 where the RF signal is being transmitted.

Standing waves generated by reflection of the RF signal due to imperfect RF impedance matching/termination may also produce non-uniformities in the RF signal amplitude along the length of the first bus 730. For example, an RF signal emitted at the first end 750 upon reaching the second end 751 may be reflected back to the first end 750 due to the lack of an impedance matching termination, thereby creating a standing wave along the length of the first bus 730.

Such non-uniformity in RF signal amplitude across the length of the first bus 730 may cause plasma non-uniformity.

By using a recursive RF feed structure, non-uniformities in the amplitude of the RF signal across the first bus 730 may be reduced. Referring back to fig. 7D, the first bus 736 is configured to form a recursive RF feed structure to pass the first RF signal generated by the first RF power supply to the filaments 710 such that the signal path length from the drive point 756 to each filament 710 and the loss experienced by the RF signal are approximately equal for all filaments 710. Such approximately equal path lengths may enable approximately equal RF signal amplitudes at the driven end (i.e., the end connected to first bus 736) of filament 710. In some implementations, non-uniformity in RF signal amplitude is further mitigated by configuring the recursive RF feed structure such that each branch (brachnch) of the structure is connected to a filament of approximately equal total length. For example, 7, 6, 5, 4 filaments are connected from left to right to the respective branches of the recursive RF feed structure. Such an approximately equal overall length of each branch may help improve uniformity when the electrode assembly 706 is operated as an ICP source. In some implementations, each level (level) of the feed structure is recursively shielded by a respective ground plane, and vertical vias that penetrate the ground plane connect the respective levels of the structure.

In the case where the electrode assembly is driven by two RF signal sources, various factors affect the shape of the plasma region produced. Examples of factors include the frequency and phase relationship of the two RF signals. Referring to fig. 7B, for example, when the frequencies of the first and second RF signals driving the first and second buses 732 and 742 are the same and the phase difference is set to 0 degrees ("unipolar" or "single-ended"), the plasma region is pushed out of the gap between the coplanar filaments 710, resulting in discontinuities or non-uniformities, e.g., in some cases where the spacing between the cylindrical shells is small. When the phase difference of the RF signals driving the adjacent coplanar filaments 710 is set to 180 degrees ("differential"), the plasma region is more strongly confined between the coplanar filaments 710. Any phase difference between 0 and 360 degrees may be used to affect the shape of the plasma region.

Generally, grounding of the workpiece support electrode 116 is a factor that affects the shape of the plasma region. The imperfect RF grounding of electrode 116, combined with the 0 degree phase difference between the RF signals driving the adjacent coplanar filaments, pushes the plasma region toward the top gap. However, if adjacent coplanar filaments (e.g., coplanar filaments) are driven with RF signals having a phase difference of 180 degrees, the resulting plasma distribution is much less sensitive to imperfect RF grounding of the electrode 116. Without being bound by any particular theory, this may be because the RF current returns through adjacent electrodes due to the differential nature of the drive signal.

The electrical configuration and characteristics of the aforementioned electrode assemblies (e.g., 400, 500, 502, 504, 600, 601, 700, 702, and 704) can be dynamically changed using RF switches coupled to various locations of the electrode assemblies in various configurations.

Referring to fig. 8A, an electrode assembly 800 includes a wire 810, a first bus 820, and a second bus 824. As shown, buses 820 and 824 may have respective third ends 821 and respective fourth ends 822. Wire 810 is similar to wires 610 and 300, unless otherwise noted. Each wire 810 has a respective first end 811 and a respective second end 812. The first bus 820 and the second bus 824 may be located inside the chamber body 102, in a ceiling of the chamber, or outside the chamber, and may form electrical connections between respective ends of the wires 810 to various locations along the buses 820 and 824 (e.g., along the length of the buses 820 and 824).

The filaments 810 may be divided into a first multiplicity 816 of filaments and a second multiplicity 817 of filaments. In some implementations, the filaments 810 of the first multiplicity 816 and the second multiplicity 817 can be arranged in an alternating pattern along a direction perpendicular to their longitudinal axes such that the coplanar filaments alternate between the first and second sets, as shown.

The first ends 811 of the filaments of the first multiplicity 816 can be coupled to a first bus 820. The first ends 811 of the wires of the second multiplicity 817 can be coupled to a second bus 822. The coupling between the wire 810 and the bus can be accomplished using a simple wire or metal ribbon (if the length is short relative to a fraction of the wavelength of the RF frequency), or by using an RF transmission line (such as a coaxial cable).

In some implementations, the electrode assembly 800 additionally includes a third bus bar 826 and a fourth bus bar 828. In such an implementation, the second ends 812 of the filaments of the first multiplicity 816 can be coupled to a third bus 824. The second ends 812 of the filaments of the second multiplicity 817 may be coupled to a fourth bus 826.

Buses 820, 824, 826, and 828 are configured to electrically couple to respective filaments 810 coupled to buses 820, 824, 826, and 828. The RF transmission lines forming the bus may have a length (e.g., >1/10 wavelengths) comparable to or greater than a significant portion of the wavelength of the RF frequency, as well as having losses due to intentional plasma loading of the filament array, i.e., absorption of RF power. Accordingly, wires 810 connected at different locations along the direction of propagation of the RF signal may experience different RF signal amplitudes. For example, an RF signal transmitted at the third end 821 of the first bus 820 will attenuate as the RF signal propagates down the length of the first bus 820. As such, the RF signal amplitude at the filament 810 near the second end 822 will be less than the RF signal amplitude at the filament 810 near the first end 821 where the RF signal is being transmitted. Such non-uniformity in the RF signal amplitude across the length of the first bus 820 or 824 may cause plasma non-uniformity.

In general, the plasma region generated by the electrode assembly 800 over a substantial area may contain significant (substitional) non-uniformities in plasma density. For example, for a plasma region 40cm long by 40cm wide, a significant difference in plasma uniformity can be observed between RF signal frequencies of 13.56MHz and 60 MHz. When driven at a lower frequency (e.g., 13.56MHz), the plasma density may decrease away from ends 811 and 812 toward the center portion of filament 810. However, the time-average plasma density remains substantially spatially uniform in a direction perpendicular to the longitudinal axis of the filament. When driven at higher frequencies (e.g., 60MHz), the plasma density becomes more non-uniform both along the filament and perpendicular to the longitudinal axis of the filament. For example, a periodic distribution of local maxima and minima may be formed in two directions. Without wishing to be bound by theory, this pattern of non-uniformity may be caused, at least in part, by the presence of standing waves.

By dynamically changing the electrical characteristics of the electrode assembly 800 using the RF switch, it may be possible to mitigate such non-uniformities. It may also be possible to intentionally introduce non-uniformities in the voltage signal to compensate for other sources of non-uniformity in the workpiece, such as non-uniform layer thicknesses, or plasma density (e.g., non-uniform gas distribution).

Referring to fig. 8B, switched electrode system 802 includes a first RF switch 830, a second RF switch 834, a third RF switch 836, a fourth RF switch 838, a first tap 840, and a second tap 842. Generally, the first and second taps 840 and 842 may be connected to various signals and potentials to generate a plasma, such as to first and second RF signals, RF ground.

Each RF switch includes a first terminal 831 and a second terminal 832. Generally, the RF switch 830 operates bi-directionally, and the first and second terminals 831 and 832 are not dependent on a particular physical termination of the RF switch, but are used to represent two different terminations of the RF switch. Various RF switch components may be used to provide the RF switches 830, 834, 836, and 838. Examples of RF switching components include mechanical relays or switches, PIN diodes, saturable inductors/reactors, MOSFETs, electronic circuits including these components, and frequency dependent impedance circuits when combined with RF power generators having tunable RF signal frequencies.

Generally, the first and second tap 840 and 842 may be positioned along the respective lengths of the buses 820, 824, 826, and 828, e.g., at the middle of the buses. In some implementations, the first tap 840 is located in the middle of the first bus 820 and the second tap 842 is located in the middle of the fourth bus 828.

In some implementations, the first and second tap 840 and 842 are differentially driven by two RF signals having the same frequency (e.g., 60MHz) and having a relative phase difference of 180 degrees.

Generally, the first and second terminals 831 and 832 of the RF switch can be coupled to the bus in various ways to achieve various effects. For example, respective first terminals of RF switches 830, 834, 836, and 838 are connected to ends of buses 820, 824, 826, and 828 as shown. In such a configuration, the closing of any one of the RF switches 830, 834, 836, and 838 electrically connects or "shorts" the respective ends ("corners") of the bus. The shorting of the corners may cause the RF reflection coefficient at the location to vary such that the RF signal amplitude and power coupling at the local region of the filament 810 near the shorted corners is reduced, thereby reducing the local plasma generation. The short-circuiting of the corners may also shift and/or change the spatial distribution of maxima and minima in the plasma density.

Generally, the electrical connection and coupling may be provided by wires, coaxial cables, waveguides or by physical contact (e.g., swaging, welding, one-piece fabrication).

Generally, process uniformity of a workpiece can be improved by time averaging of plasma exposures. One way to achieve time averaging of plasma exposure is by moving the spatial distribution of the non-uniformities in the plasma region. For example, the plasma density distribution (non-uniformity) can be shifted by turning on and off ("modulating") RF switches coupled to the four corners of the electrode assembly.

The RF switches 830, 834, 836, and 838 can be modulated in various ways to achieve a desired time-averaged plasma density. An example of a procedure for modulating the RF switch is to cyclically connect pairs of different buses. For example, the system may operate as follows: (1) turn off RF switch 830 for a first duration and then on, (2) turn off RF switch 834 for a second duration and then on, (3) turn off RF switch 836 for a third duration and then on, (4) turn off RF switch 838 for a fourth duration. The first to fourth durations may be determined based on a desired program repetition rate. For example, the repetition rate may be set much faster than the time scale of certain actions (e.g., device charging). For example, in a program with 4 states, the duration of each state including the dead time may be set to 50 μ s to achieve a repetition rate of 5 kHz.

In some implementations, slack time is inserted between steps of the procedure. The dead time may provide a "break before make" contact to prevent two or more generators from shorting in some configurations. In some implementations, the closing of the switches may overlap in time. For example, two switches, such as paired diagonal switches (830-. As another example, all four switches may be turned on and off simultaneously.

Referring to FIG. 8C, an example of a switched electrode system 804 is shown. Switched electrode system 804 is similar to system 802, unless otherwise noted. Switched electrode system 804 includes a first RF switch bank 850, a second RF switch bank 854, a third RF switch bank 856, and a fourth RF switch bank 858. The first RF switch set 850 includes sub-switches 860a and 860b, the second RF switch set 854 includes sub-switches 860c and 860d, the third RF switch set 836 includes sub-switches 860e and 860f, and the fourth RF switch set 838 includes sub-switches 860g and 860 h. The sub-switches are similar to the RF switch 830.

A first terminal 831 of the sub-switch is connected to the ends of buses 820,824, 826 and 828. In some implementations, the second terminal 832 of the sub-switch is connected to RF ground. In such a configuration, the closing of any one of the sub-switches electrically couples the respective end of the bus line to the RF ground or to the ground of the bus line. Grounding of the ends of the bus may cause a reduction in the amplitude of the RF signal in a local region of the filament 810 near the RF ground end of the bus, and result in a reduction in the amplitude of the electric field or lower power coupling in that region. A reduction in the magnitude of the electric field may result in a reduction in the plasma generation in said region.

The RF switch bank and individual subswitches may be modulated in various ways to provide modulation of the plasma density profile. For example, each RF switch bank may operate as a single unit, with the sub-switches of the RF switch bank turned on and off as a single unit. As another example, the sub-switches of each RF switch group may be independently turned on and off.

The switches may be modulated in a variety of different processes in a manner similar to the various processes described with respect to fig. 8B. For example, the switched electrode system may be operated in the following manner: one switch set may be closed cyclically at a time (optionally with a time delay), the switch sets may be closed cyclically during the time that the different sets are closed, the switch sets may be alternated, or all switches may be opened and closed synchronously.

As another example, the system may operate as follows: (1) turn off first and third RF switch sets 850 and 856 for a first duration and then on, (2) turn on all switches, (3) turn off second and fourth RF switch sets 854 and 858 for a second duration and then on.

As yet another example, the system may operate as follows: (1) turning off the first switch set 850 for a first duration and then on, (2) turning off the second switch set 854 for a second duration and then on, (3) turning off the third switch set 856 for a third duration and then on, (4) turning off the fourth switch set 858 for a fourth duration and then on, (5) turning on all switch sets, (6) turning off all switch sets.

In some implementations, the feeding of RF signals to various locations of the bus may be dynamically reconfigured using RF switches. Referring to FIG. 8D, an example of a switched electrode system 806 is shown. Unless otherwise noted, switched electrode system 806 is similar to system 804 and can operate in a similar manner.

The first multiplicity 816 is driven with an RF signal at taps 844 and 846. The RF signals driving taps 844 and 846 may be the same frequency or different frequencies. For the same frequency, the phase relationship of the two signals may be 0, 180, or any value between 0 and 360. For some implementations, the phase relationship may be modulated over time. As shown, the second terminals 832 of the sub-switches 860a, 860c, 860f and 860h are connected to the respective taps 844 and 846.

In such a configuration, the ground characteristics of the second plurality 817 may be modulated using the respective subswitches, and RF signals may be transmitted to the buses 820 and 826 from different locations (e.g., from the terminals 821 and 822). A combination of the grounding characteristics and modulation of the RF signal distribution can be used to modulate the plasma density to improve process uniformity through time averaging.

In such a configuration, it may be advantageous to maintain at least one of the sub-switches 860 in a closed state to provide a continuous RF signal supply to the assembly 800.

Referring to FIG. 8E, an example of a switched electrode system 808 is shown. Unless otherwise noted, switched electrode system 808 is similar to system 804 and can operate in a similar manner. The second terminal 832 of the sub-switch is connected to a single tap 848. A symmetric distribution network as shown can be used to improve the uniformity of the RF signal delivered to the four corners of the system 808. The sub-switches can be tuned in various ways as previously described to change the plasma distribution and improve process uniformity.

In some implementations, switches may be distributed across the bus to allow for finer control of the instantaneous plasma uniformity, thereby improving the time-averaged plasma uniformity. Referring to fig. 8F, an example of a switched electrode system 801 is shown. Switched electrode system 801 is similar to system 808 and may operate in a similar manner unless otherwise stated. The first bus 820 is coupled to a first set of RF switches 870, such as three or more sub-switches. Each RF switch group includes a plurality of sub-switches 860. The first terminals of the sub-switches 860 of the first RF switch set 870 are electrically coupled to the first bus at various locations across the length of the first bus 820. In some implementations, the coupling points are approximately equally spaced, as shown. The second terminal of the sub-switch 860 of the first set of RF switches 870 is electrically coupled to the tap 848 to receive the RF signal.

Second, third and fourth buses 824, 826 and 828 are connected to second, third and fourth RF switch banks 874, 876 and 878, respectively, each connected in a manner similar to first bus 820 and first RF switch bank 870.

In such a configuration, additional hierarchical control of the emission location of the RF signal along the length of the bus may result in improved time-averaged plasma uniformity.

In general, the number of subswitches included in the RF switch bank may be determined based on, for example, the length of the bus, the size of the plasma region, the RF signal frequency and power, and the chamber pressure.

In some implementations, the RF signal feed and ground positions can be dynamically reconfigured using RF switches to provide a mode-selectable plasma source that can be switched between a primary CCP mode and a primary ICP mode. Referring to fig. 9A, an example of a switched electrode system 900 is shown. Unless otherwise noted, switched electrode system 900 is similar to system 802 and may operate in a similar manner. First terminals 831 of the RF switches 830 and 834 are connected to respective third and fourth terminals 821 and 822 of the second bus 824, and first terminals 831 of the RF switches 836 and 838 are connected to respective third and fourth terminals 821 and 822 of the third bus 826, as shown. The second terminal 832 is connected to RF ground.

The RF switches 830, 834, 836, and 838 can be controlled in various ways to vary the primary mode of plasma generation by the switching electrode assembly 900. For example, by turning off all four RF switches, RF current flows along the length of the filament 810, generating a magnetic field and producing a predominantly inductively coupled plasma. By turning on all four switches, the RF current is reduced and the assembly 900 generates a predominantly capacitively coupled plasma.

In some implementations, the first and second RF signals driving the respective taps 840 and 842 have a phase difference of 180 degrees, i.e., are driven differentially. In this case, the alternating filaments 810 belonging to the first and second multiples 816 and 817 are fed from opposite ends of the RF signal with a phase difference of about 180 degrees, resulting in the generation of an auxiliary RF magnetic field. In some implementations, the first and second RF signals driving the respective taps 840 and 842 have a phase difference of about 0 degrees. In this case, the alternating filaments 810 belonging to the first and second multiples 816 and 817 are fed from opposite ends of the RF signal with a phase difference of about 0 degrees, resulting in the generation of opposing RF magnetic fields.

In some implementations, switches may be distributed across the bus to allow for finer control of the instantaneous plasma uniformity, thereby improving the time-averaged plasma uniformity. Referring to FIG. 9B, an example of a switching electrode assembly 902 is shown. Unless otherwise noted, the switched electrode assembly 902 is similar to the system 801. The first bus 820 is coupled to a first RF switch set 870 that includes a plurality of sub-switches 860.

The first terminals of the sub-switches 860 of the first RF switch set 870 are electrically coupled to the first bus at various locations across the length of the first bus 820. In some implementations, the coupling points are approximately equally spaced, as shown. The second terminal of the sub-switch 860 of the first RF switch set 870 is electrically coupled to the tap 940 to receive the first RF signal.

The second bus is connected to the second RF switch bank 874 at the first terminal of the sub-switch and the second terminal of the sub-switch is connected to RF ground.

The third bus is connected to the third RF switch group 876 at the first terminal of the sub-switch 860 and the second terminal 832 of the sub-switch 860 of the third RF switch group 876 is connected to RF ground.

The fourth bus is connected to a fourth set 878 of RF switches at the first terminal of the sub-switch and the second terminal of the sub-switch is electrically coupled to tap 942 to receive the second RF signal.

The first and second RF signals driving taps 940 and 942 may be at the same frequency or different frequencies. For the same frequency, the phase relationship of the two signals may be 0, 180, or any value between 0 and 360. For some implementations, the phase relationship may be modulated over time.

The RF switch banks 870, 874, 876, and 878 can be controlled in various ways to change the primary mode of plasma generation by the switching electrode assembly 902. For example, by turning off at least one sub-switch from each of the first set 870 and the fourth set 878, and turning on the second and third RF switch sets 874 and 876, the assembly 902 generates a primarily capacitively coupled plasma.

As another example, assembly 902 generates a primarily inductively coupled plasma by turning off at least one sub-switch from each of first set 870 and fourth set 878, and turning off all sub-switches of second and third RF switch sets 874 and 876. In some implementations, the first and second RF signals driving respective taps 940 and 942 have a phase difference of 180 degrees, i.e., are driven differentially. In this case, the alternating filaments 810 belonging to the first and second multiples 816 and 817 are fed from opposite ends of the RF signal with a phase difference of about 180 degrees, resulting in the generation of an auxiliary RF magnetic field. In some implementations, the first and second RF signals driving respective taps 940 and 942 have a phase difference of about 0 degrees. In this case, the alternating filaments 810 belonging to the first and second multiples 816 and 817 are fed from opposite ends of the RF signal with a phase difference of about 0 degrees, resulting in the generation of opposing RF magnetic fields.

In some processing applications, ICP, using opposing RF magnetic fields that can deposit RF power in the plasma in a manner generally parallel to the ribbon of filaments, can provide a more uniform plasma, particularly when the workpiece is close to the plasma source (e.g., electrode assembly), i.e., a small bottom gap 132. Therefore, it may be beneficial to have the ability to change the phase relationship of the first and second RF signals.

In general, the individual subswitches of the first and fourth sets 870 and 878 may be modulated to change the plasma density profile. Additionally, where the switched electrode assembly 902 is configured to generate a primarily inductively coupled plasma, the sub-switches of the second and third groups 874 and 876 may be individually modulated to further alter the plasma density profile.

Generally, although the figures show the bus driven near the center and the ends floating or with grounded terminations, it may be advantageous to be driven or terminated at other locations (such as driven end, terminated end or center) depending on the application, RF configuration, frequency and operating area (plasma load).

Generally, where the second terminal of the RF switch is connected to RF ground, a variable impedance may be placed in series to the RF ground to provide a variable RF termination impedance to further control changes in plasma density.

In general, although the illustrations depict taps connected to the center of the respective busses, the taps used to apply RF power to the electrode assembly may be located at one or more ends of the busses, at the center, or at other locations.

The switch may be used to improve the time-averaged plasma uniformity of the wedge electrode assembly. Referring to fig. 10, an example of a switching electrode assembly 1000 is shown. The switching electrode assembly 1000 includes a wedge-shaped electrode assembly 1010. The wedge electrode assembly 1010 is similar to the wedge electrode assembly 704, unless otherwise noted. The assembly 1010 includes a wedge-shaped top electrode 624, and the wedge-shaped top electrode 624 may be grounded. The switched electrode assembly 1000 includes a first RF switch 1030, a second RF switch 1034, a third RF switch 1036, a fourth RF switch 1038, and a tap 1040. The RF switch is similar to RF switch 830. First terminals of RF switches 1030 and 1034 are connected to a first end 754 of component 1010 and first terminals of RF switches 1036 and 1038 are connected to a second end 755 of component 1010. Second terminals of the first and fourth RF switches 1030 and 1038 are connected to each other and to the tap 1040, and second terminals of the second and third RF switches 1034 and 1036 are connected to the RF ground.

The first and fourth RF switches 1030 and 1038 may be turned on and off to selectively feed RF signals to the first end 754, the second end 755, or both ends of the assembly 1010. The second and third RF switches 1034 and 1036 can be turned on and off to selectively ground either the first end 754 or the second end 755 of the component 1010.

The RF switch may be modulated in various ways to improve time averaged plasma uniformity. The following are examples of procedures: (1) closing the RF switch 1030 for a first duration and opening the switches 1034, 1036, and 1038 (e.g., for 30 microseconds), (2) closing 1030, 1036, opening 1034, 1038 (e.g., for 40 microseconds), and then (3) closing 1036, opening 1030, 1034, and 1036 (e.g., for 30 microseconds). Alternatively, the unpowered terminal may be grounded after a short delay after the RF signal is applied to the other terminal, and the ground terminal may not be grounded before the RF signal is applied to the terminal.

The following is another example of a procedure: (1) turn ON (ON) at 1030, turn OFF (OFF) at 1038, 1034, 1036 for 30 microseconds, (2) turn ON at 1030, 1038, turn OFF at 1034, 1036 for 40 microseconds, and (3) turn ON at 1038, turn OFF at 1034, 1030, 1036 for 30 microseconds, and then repeat the cycle multiple times until the process steps are completed or the cycle is alternately inverted. Alternatively, the non-power terminal may be grounded after a short delay after applying power to the other terminal, and the ground terminal may not be grounded before applying power to the terminal.

In general, the wedge-shaped electrode assembly 1010 may be similar to an electrode. In general, the switch may be applied to other electrode assemblies, such as 600, 601, 700, 702, 704.

Various circuit implementations may be used to provide an RF switch suitable for switching an RF signal for plasma generation. There are various considerations for implementing the RF switches (e.g., RF switch 830, sub-switch 860) to be used in a switched electrode system. Examples of such considerations include RF power handling capability, switching speed, ON-state impedance, OFF-state impedance, and bidirectionality.

Generally, a switch is considered to be in an "ON" or closed (closed) state when the impedance presented between its two terminals is low, and in an "OFF" or open (open) state when the impedance is high.

A PIN diode switch may be used to provide a suitable RF switch. Referring to fig. 11A, a PIN diode switch 1100 includes a PIN diode 1110, a first capacitor 1120 having a capacitance C1, a second capacitor 1122 having a capacitance C2, and an inductor 1140 having an inductance L1. Switch 1100 has a first terminal 1131, a second terminal 1132, and a control terminal 1134. The first terminal 1131 may provide a first terminal 831, and the second terminal 1132 may provide a second terminal 832 of the RF switch 830.

The first capacitor 1120 and the inductor 1150 may be connected in parallel between the first terminal 1131 and the second capacitor 1122. PIN diode 1110 may then be connected in parallel with first capacitor 1120, inductor 1150, and second capacitor 1122 between first terminal 1131 and second terminal 1132. The control terminal 1134 may be connected between the second capacitor 1122 and the first capacitor 1120.

The PIN diode 1110 is a diode with a wide undoped intrinsic semiconductor region between the p-type semiconductor and the n-type semiconductor region, and may be well suited for fast switching of high power RF signals. The PIN diode has an anode (+) and a cathode (-) and can provide a low impedance conduction path, such as <1 ohm, for RF signals when a forward bias is established between the anode and cathode (e.g., >0.7V and/or diode current >100 mA).

The PIN diode switch 1100 operates based on the following operating principle. The impedance of PIN diode 1110 may be controlled by providing a control signal to control terminal 1134. The control signal is a quasi (quasi) static voltage that is switched between a first level (e.g., -0.7V) and a second level (e.g., -2 kV). Due to the quasi-static nature of the control signal, the control voltage and any resulting diode current may be conducted through inductor 1140. In addition, the second capacitor 1122 prevents the control voltage from reaching the cathode. PIN diode 1110 may be set to an "OFF state" by providing a sufficiently large negative control voltage (e.g., -2kV) to the anode relative to the cathode, presenting a high impedance between the cathode and anode of PIN diode 1110. PIN diode 1110 can be set to an "ON" state when a sufficiently large positive control voltage (e.g., 0.7V) is applied, presenting a low impedance path (e.g., <1 ohm) for RF signals between terminals 1131 and 1132.

The first capacitor 1120 and inductor 1140 connected in parallel as shown form a parallel LC resonator 1150. The resonator 1150 has the equation

The determined resonance frequency. At the resonance frequency f₀The resonator 1150, in turn, exhibits a near open circuit (e.g., open circuit, such as>1000 ohms) depending on the quality factor of the resonator. By selecting the values of C1 and L1 such that the resonant frequency is aligned (align) with the frequency of the RF signal present at terminal 1131 or 1132, the RF signal may be prevented from passing through resonator 1150.

In general, the capacitance C2 of the second capacitor 1122 may be set to provide a low impedance path at the frequency of the RF signal.

In some implementations, the first capacitor 1120 is a variable capacitor ("capacitor") having an adjustable capacitance C1, and the adjustable capacitance C1 can be varied to optimize the resonance of the parallel LC circuit formed by the first capacitor 1120 and the inductor 1140 to align with the frequency of the RF signal.

In some implementations, a control signal buffer amplifier 1136 may be provided to buffer and/or amplify the control signal applied at control terminal 1134 to the anode of PIN diode 1110.

Generally, multiple PIN diode switches may be used in combination to achieve a range of impedance values between first and second terminals 1131 and 1132. The control signal may also be set between the first and second levels to provide a variable impedance.

In some implementations, first terminal 1131 is connected to a bus (e.g., bus 820), and second terminal 1132 is connected to an RF ground, forming a path to the RF ground. In some implementations, the first terminal is connected to a first bus (e.g., bus 820) and the second terminal 1132 is connected to a second bus (e.g., bus 824), in which case the switch may be considered "floating," and the potential of the second terminal 1132 is defined by external factors.

As another example, saturable inductor switches may be used to provide suitable RF switching. Referring to fig. 11B, the saturable inductor switch 1102 includes a saturable inductor 1160, a first capacitor 1124 having a capacitance C1, and a second capacitor 1126 having a capacitance C2. Switch 1102 has a first terminal 1131, a second terminal 1132, and a control terminal 1135. First terminal 1131 may provide first terminal 831 and second terminal 1132 may provide second terminal 832.

Saturable inductor 1160 has a primary winding (primary winding)1162 having an inductance of L1 and a control winding 1164 having an inductance of L2. In some documents, saturable inductors may also be referred to as saturable reactors or magnetic amplifiers. A saturable inductor is an inductor having a magnetic core that is intentionally saturated by passing current through control winding 1164. Once saturated, the inductance L1 of the primary winding 1162 drops significantly. The reduction of the inductance of the primary winding results in a reduction of the impedance of the RF signal, which can be used to implement the switch.

The primary winding 1162 of the inductor 1160 may be connected in series with the second capacitor 1126, and the first capacitor 1124 may be connected in parallel with the primary winding 1162 and the second capacitor 1126 between the first and second terminals 1131, 1131. Control terminal 1135 is connected to control winding 1164, which control winding 1164 may in turn be connected to ground.

Saturable inductor switch 1102 operates based on the following operating principle. First capacitor 1124 in parallel with the series combination of primary winding 1162 and second capacitor 1126 forms a parallel LC resonator that operates similarly to LC resonator 1150. For example, the values of C1, C2, and L1 may be set such that when the control signal is set to an "OFF" or low state, where no current flows through the control winding 1164, resonance of the switch 1102 occurs at the RF signal frequency (e.g., 60 MHz). In this state, switch 1102 is in an "open" state, presenting a high impedance between first and second terminals 1131 and 1132. When the control signal applied to control terminal 1135 is set to an "ON" or high state, the magnetic field generated by the current flowing through secondary winding 1164 saturates the magnetic core of saturable inductor 1160, thereby reducing the inductance L1 of primary winding 1162. The reduction in inductance L1 changes the resonant frequency of switch 1102, presenting a low impedance between first and second terminals 1131 and 1132 at the same RF signal frequency. This low impedance state can be used as a closed state for switch 1102.

In some implementations, a control signal buffer amplifier 1137 may be provided to amplify and/or buffer the control signal applied at control terminal 1135 so that a current sufficient to saturate the saturable inductor 1160 may be applied to control winding 1164.

In some implementations, a low pass filter 1138 may be provided between the control signal terminal 1135 and the control winding 1164 to mitigate noise coupling from the control signal and/or RF signal propagating to the control signal terminal.

In general, the impedance of the switch present between first terminal 1131 and second terminal 1132 may be controlled between the "ON" state and the "OFF" state by adjusting the control signal to provide a range of currents to control winding 1164.

In some implementations, first terminal 1131 is connected to a bus (e.g., bus 820), and second terminal 1132 is connected to RF ground. In some implementations, the first terminal is connected to a first bus (e.g., bus 820) and the second terminal 1132 is connected to a second bus (e.g., bus 824).

The impedance presented by the aforementioned switches 1100 and 1102, and their switching states, are controlled by the application of control signals. However, in some implementations, the characteristics of the switches may remain static, but rather the frequency of the RF signal may be modulated such that the switches present an "open" or "closed" state to RF signals having different frequencies. For example, the frequency dependent impedance of the circuit may be used to provide such frequency based switching.

Referring to fig. 12A, a frequency-based switch 1200 includes a first capacitor 1220 having a capacitance C1, a second capacitor 1222 having a capacitance C2, a first inductor 1240 having an inductance L1, and a second inductor 1242 having an inductance L2. Switch 1200 has a first terminal 1231 and a second terminal 1232.

The first capacitor 1220 and the first inductor 1240 may be connected in series, and the second capacitor 1222 and the second inductor 1242 may be connected in series. The pair of circuits may be connected in parallel between the first terminal 1231 and the second terminal 1232.

The combination of L1, C1, L2, and C2 may be arranged such that at a first frequency (e.g., 58MHz), a low impedance (e.g., <0.1 ohm) is present between first and second terminals 1231 and 1232, and at a second frequency (e.g., 62MHz), a high impedance (e.g., >100 ohm) is present. For example, the values of L1-L2-0.1 μ H, C1-75.3 pF, and C2-58.6 pF may provide low impedance resonance at 58MHz and high impedance resonance at 62 MHz.

Without wishing to be bound by theory, a low impedance resonance may be provided by a series LC resonance, and a high impedance resonance may be provided by a parallel LC resonance.

The capacitance and inductance may be arranged to form a frequency-based switch having an approximately complementary response to the examples provided above. For example, the following values L1-L2-0.1 μ H, C1-65.9 pF, and C2-87.8 pF may provide a low impedance resonance at 62MHz and a high impedance resonance at 58MHz, exhibiting approximately complementary or opposite responses with respect to the first example. This complementary behavior can be used to form various frequency-switched electrode systems.

In some implementations, discrete capacitors and inductors may be implemented with distributed circuit elements (e.g., transmission line segments, stubs).

Referring to fig. 12B, the frequency-switched electrode system 1202 includes an electrode assembly 800, a first frequency-based switch 1200a, a second frequency-based switch 1200B, and a tap 1260. RF signals of different frequencies may be provided to tap 1260, for example, using a variable frequency RF generator with a matching network and series-connected isolators or circulators.

In this configuration, the frequency of the RF signal supplied through tap 1260 may alternate from a first frequency to a second frequency such that more RF signal is coupled to the left side of electrode assembly 800 through switch 1200a or to the right side of electrode assembly 800 through switch 1200 b. Alternatively, the frequency of the RF signal provided through tap 1260 may be driven, such as with a ramp function, to vary between a first frequency and a second frequency.

For example, by setting the component values to L1 a-L2 a-0.1 μ H, C1 a-75.3 pF, and C2 a-58.6 pF, the first switch 1200a can provide low impedance resonance at 58MHz and high impedance resonance at 62 MHz. The component values of the second switch 1200b may be set to L1-L2-0.1 μ H, C1-65.9 pF, and C2-87.8 pF to provide low impedance resonance at 62MHz and high impedance resonance at 58 MHz. In such a configuration, a majority of the RF signal may be coupled to the left side of the electrode assembly 800 through the first switch 1200a by switching the frequency of the RF signal to a first frequency (e.g., 58MHz), and to the right side of the assembly 800 through the second switch 1200b by switching the frequency of the RF signal to a second frequency (e.g., 62 MHz). When the frequency is midway between the two frequencies, approximately at about 60MHz, then the power is approximately similarly coupled across and may cause high center non-uniformity.

In some implementations, the transmission line segment may be used to change the frequency dependent impedance of the switch 1200. For example, a transmission line segment of a length of a quarter wavelength may be used to connect the corners of the electrode assembly 800 to the terminals of the switches 1200a and 1200b, taking into account the speed factor of the transmission line. By using a quarter-wavelength transmission line, the impedances presented at the first and second frequencies can be swapped. For example, a low impedance of the series resonance may translate into a high impedance of about 1000 ohms, and a high impedance of the parallel resonance may translate into a low impedance of about 1 ohm.

In some implementations, the frequency-based switch 1200 can be used as a frequency selective termination to provide impedance matching termination at different frequencies to control coupling of RF signals into the electrode assembly. Referring to fig. 12C, the frequency-switched electrode system 1204 includes an electrode assembly 800, a first frequency-selective termination 1250a, a second frequency-selective termination 1250b, and a tap 1260. Unless otherwise noted, frequency selection terminals 1250a and 1250b may be provided by frequency-based switch 1200 and operate in a similar manner.

In some implementations, the component values of the frequency selective terminations 1250a and 1250b may be set such that at a first frequency, the termination 1250a presents the characteristic impedance of the RF generator and transmission line, while the termination 1250b presents a high impedance. In such a configuration, the terminal 1250a provides an impedance matched terminal to RF ground, minimizing RF signal reflection and coupling to the RF signal on the left side of the electrode assembly 800. At the same time, the high impedance presented by terminal 1200b allows the RF signal to be coupled to the right side of electrode assembly 800.

In some implementations, the component values of the frequency selective terminals 1250a and 1250b can be set such that at a first frequency, terminal 1250a presents a low impedance path to RF ground while terminal 1250b presents a high impedance. In such a configuration, the low impedance path provided by terminal 1250a to RF ground minimizes RF signals coupled to the left side of electrode assembly 800. At the same time, the high impedance presented by terminal 1200b allows the RF signal to be coupled to the right side of electrode assembly 800.

In general, a frequency-based switch and a frequency-selective terminal may be coupled to various locations along a bus. For example, an additional pair of coupling points to the taps may be provided at the approximate center of the bus, and additional switches or terminals may be provided at those coupling points.

In general, the frequency switching is not limited to 2 states corresponding to a high impedance state and a low impedance state, but may advantageously operate continuously between or outside the first and second switching frequencies.

In general, various combinations of frequency-based switches with various resonant frequencies can be used to extend frequency-based switching to 3, 4, or more frequencies.

In some plasma chambers, a workpiece is moved through a plasma processing region on, for example, a linear or rotary workpiece support. In such a chamber, the moving workpiece support may be dc grounded by, for example, a rotating mercury coupler, a brush, or a slip ring. However, the moving workpiece support may not be sufficiently grounded at radio frequencies. The RF ground path should have a much lower impedance than the plasma so that the RF ground path becomes a sufficient RF ground. The lack of a sufficient RF grounding path can make it difficult to control the ion energy at the workpiece and reduce the repeatability of the process.

Therefore, a plasma source having the following characteristics is required: the plasma source can efficiently produce a uniform plasma with desired characteristics (plasma density, electron temperature, ion energy, dissociation, etc.) over the workpiece size; the plasma source may adjust the uniformity (e.g., pressure, power, gas composition) over the operating window; it has stable and repeatable electrical performance even if the workpiece is moving; and it does not produce excessive metal contamination or particles.

Fig. 13 is a schematic side view of another example of a plasma reactor. The plasma reactor 2100 has a chamber main body 2102, and the chamber main body 2102 encloses an inner space serving as a plasma chamber. The chamber body 2102 can have one or more side walls 2102a, a top panel 2102b, and a bottom panel 2102 c. The interior space 2104 can be cylindrical, as used for processing round semiconductor wafers. The plasma reactor includes a top electrode array assembly 2106 located at the top of the plasma reactor 2100. The top electrode array assembly 2106 can abut the top plate (as shown in figure 13), or be suspended within and spaced apart from the inner space 2104, or form a portion of the top plate. Portions of the side walls and bottom of the chamber body 2102 may be separately grounded.

The gas distributor is located near the ceiling of the plasma reactor 2100. The gas dispenser may include one or more ports 2110 in the side wall 2102, the one or more ports 2110 connected to a process gas supply 2112. Alternatively or additionally, the gas distributor may be integrated with the top electrode assembly 2106 as a single component. For example, a channel connected to the process gas supply 2112 may be formed through a dielectric plate in the assembly 2112 to provide an opening in the ceiling of the plasma chamber. The gas supply 2112 delivers one or more process gases to the gas distributor 2110, the composition of which may depend on the process to be performed, such as deposition or etching.

A vacuum pump 2113 is connected to the interior space 2104 to evacuate the plasma reactor. For some processes, the chamber is operated in the Torr range, and the gas distributor supplies argon, nitrogen, oxygen, and/or other gases.

Depending on the chamber configuration and the process gases supplied, the plasma reactor 100 may provide an ALD apparatus, an etching apparatus, a plasma processing apparatus, a plasma enhanced chemical vapor deposition apparatus, a plasma doping apparatus, or a plasma surface cleaning apparatus.

The plasma reactor 2100 includes a workpiece support 2114 (e.g., pedestal) for supporting a workpiece, a top surface of which is exposed to the plasma formed in the chamber 2104. The workpiece support 2114 has a workpiece support surface 2114a facing the top electrode 2108. In some implementations, the workpiece support 2114 includes a workpiece support electrode 2116 located inside the support 2114, and a workpiece bias voltage supply 2118 is connected to the workpiece support electrode 2116. The voltage supply 2118 may apply a voltage to clamp the workpiece 2115 to the support 2114 and/or supply a bias voltage to control characteristics of the generated plasma, including ion energy. In some implementations, the RF bias power generator 2142 is AC coupled to the workpiece support electrode 2116 of the workpiece support 2114 through an impedance match 2144.

Additionally, the support 2114 may have internal channels 2119 for heating or cooling the workpiece 2115, and/or an embedded resistive heater (2119).

An electrode assembly 2106 is positioned at the top plate of chamber 2104. The electrode assembly 2106 includes a plurality of conductors 2120, the plurality of conductors 2120 extending laterally over the workpiece support 2114. The conductors 2120 are coplanar, at least in the area of the workpiece over the expected location on the support 2114. For example, in this region, the conductor may extend parallel to the support surface 2114 a. The plurality of conductors 2120 may be arranged as a parallel line array. In some implementations, the conductors may have a "U-shape," with both ends connected to respective busses on the same side of the chamber 2104. Alternatively, the conductors may be arranged in staggered spirals (staggered circular spirals or staggered rectangular spirals). The longitudinal axis of the conductor 2120 may be disposed at a non-zero angle (e.g., an angle greater than 20 degrees) with the direction of motion of the workpiece 10 below the electrode assembly 2106. For example, the longitudinal axis of the conductor 2120 may be substantially perpendicular to the direction of motion of the workpiece 10.

A gap 2132 is formed between the workpiece support 2114 and the electrode assembly 2106. For high pressures (e.g., 1-10Torr), the gap 2132 can be 2-25 mm. A larger minimum gap, for example about 5mm, may be required to hold the workpiece, depending on the electrode-to-electrode spacing on the source and the thickness of the dielectric cover. At lower pressures (e.g., less than 100mTorr), the gap 2132 can be 1cm to 50 cm.

In some implementations, fluid supply 2146 circulates fluid through electrode assembly 2106. In some implementations, a heat exchanger 2148 is coupled to the fluid supply 2146 to remove heat or supply heat to the fluid.

The electrode assembly 2106 is driven by an RF power supply 2122. The RF power supply 2122 can apply power to the conductor 2120 of the electrode assembly 2106 at a frequency of, for example, 1 to 300 MHz. For some processes, the RF power supply 2122 provides a total RF power of greater than 2kW at a frequency of 60 MHz.

In some implementations, a heat spreader 2150 (e.g., aluminum plate) is attached to the top plate 2102b of the chamber body 2102. Channels 2152 may be formed through heat spreader 2150 and coolant may be circulated through channels 2152. A heat exchanger 2154 may be connected to the channel 152 to remove heat or supply heat to the coolant.

Fig. 14A-14C are schematic diagrams of another example of a plasma reactor. In this example, which operates the same as fig. 13, the multi-chamber processing tool 200 includes a plasma reactor 100, unless otherwise noted.

The processing tool 2200 has a body 2202 that surrounds an interior space 2204. The main body 2102 can have one or more sidewalls 2202a, a top plate 2202b, and a bottom plate 2202 c. The inner space 2204 may be cylindrical.

The processing tool 2200 includes a workpiece support 2214 (e.g., a pedestal) for supporting one or more workpieces 10 (e.g., a plurality of workpieces). The work piece support 2214 has a work piece support surface 2214 a. The workpiece support 2214 can include a workpiece support electrode 2116, and a workpiece bias voltage source 2118 can be connected to the workpiece support electrode 2116.

The space between the top of workpiece support 2214 and top plate 2202b may be divided into multiple chambers 2204a-2204d by barrier 2270. Barrier 2270 may extend radially from the center of workpiece support 2214. Although four chambers are shown, there may be two, three or more than four chambers.

The workpiece may be rotated about an axis 2260 by a motor 2262. As such, any workpieces 10 on the workpiece support 2214 will be sequentially carried through the chambers 2204a-2204 d.

The chambers 2204a-2204d may be at least partially isolated from each other by a pump-purge system 2280. Pump-purge system 2280 may include a plurality of channels formed through barrier 2210 that flow a purge gas (e.g., an inert gas, such as argon) into and/or pump gas out of the space between adjacent chambers. For example, the pump-purge system 2280 may include a first passage 2282, such as by a pump, forcing purge gas through the first passage 2282 and into the space 2202 between the barrier 2272 and the workpiece support 2214. Either side of the first channel 2282 (relative to the direction of movement of the workpiece support 2214) may flank the second and third channels 2284, 2286, which are connected to a pump to pump gases (including purge gas and any gases from adjacent chambers, such as chamber 2204 a). Each channel may be an elongate slot, the elongate slot extending generally radially.

At least one of the chambers 2204a-2204 provides a plasma chamber of the plasma reactor 2100. The plasma reactor includes a top electrode array assembly 2106 and an RF power supply 2122, and may also include a fluid supply 2146 and/or a heat exchanger. The process gas may be supplied through a port 2210 positioned along one or both barriers 2270 to the chamber 2104. In some implementations, the port 2210 is positioned only on the front side of the chamber 2104 (relative to the direction of motion of the workpiece support 2214). Alternatively, or in addition, the process gas may be supplied through a port in the sidewall 2202a of the tool body 2202.

Fig. 15A shows an example of an electrode assembly 2106. Electrode assembly 2106 includes a top dielectric plate 2130, a plurality of conductors 2120, and a bottom dielectric plate 2132. As described above, the conductors 2120 may be arranged as parallel linear strips that extend laterally over the workpiece support 2114. The dielectric top plate 2130 may be a ceramic material.

Dielectric base plate 2132 provides a window for RF power, i.e., is substantially transparent to RF radiation at the frequencies used to generate the plasma. The base plate 2132 may be, for example, quartz or silicon nitride. The base plate can protect the plasma process and workpiece environment from metal contamination or particle formation that might otherwise occur if the conductor or ceramic were exposed to the plasma. The bottom plate 2132 may be a consumable part that is periodically replaced. The base plate may be relatively thin, such as 0.25mm to 2mm, for example 0.5 mm.

The conductors may have a width of 1-5mm and the spacing W between the conductors 120 may be 0.5 to 3 mm. The conductor may be wider than the space, such as about twice as wide.

The thickness T of the lower dielectric plate 2132 should be less than twice the spacing W between the conductors 2120, e.g., less than the spacing W between the conductors. At higher pressures, the gap between lower dielectric plate 2132 and upper dielectric plate 2130 should be "small," such as less than 0.5mm, for example less than 0.25mm, to avoid plasma from occurring behind the plates.

Conductor 2120 may be formed directly on the lower surface of dielectric top plate surface 2130. For example, conductor 2120 can be formed by depositing (e.g., plating, sputtering, or CVD) a thin layer across the bottom surface, and then patterning by etching to form a stripline structure. The conductors may then be covered by a dielectric bottom dielectric plate 2132.

Conductor 2120 may also be embedded (i.e., buried) below the surface of the dielectric top plate. For example, the top plate 2130 may be a ceramic structure that is similar in structure to an electrostatic wafer chuck. For buried conductors, the dielectric bottom plate becomes optional, but can still be used as a dielectric cap (e.g., made of quartz) to protect the bottom surface of the top plate.

In an exemplary implementation, 45 pairs (90 total) of parallel conductors 2120 are deposited on square structural ceramic top plate 2130. The conductors 2120 each have a line width of 3mm with a spacing of 1.5mm (so the conductors are arranged at a pitch of 4.5 mm). The conductors may be 400mm long with vertical feed-throughs through the ceramic top plate 2130 and electrical connections made on the back side at atmospheric pressure. Every other electrode is connected to a bus on one side, and the remaining (alternating) electrodes are each connected to a bus on the other side, thereby forming two arrays. RF power at 60MHz and 180 degrees out of phase is connected across the two arrays.

Referring to fig. 15B, a plurality of grooves 2136 may be formed in the bottom surface 2130a of the dielectric top plate 2130, and the conductors 2120 may be cut into the grooves. The slots 2136 may be arranged in parallel linear stripes (strips).

In some implementations, each conductor 2120 is a portion of a filament 2150. The filament 2150 may be cut into a corresponding groove 2136 of the filament 2150. Filament 2150 may include a shell that surrounds and protects conductor 2120. The filament 2150 may be provided by the various filaments 300 described with reference to fig. 3A-C.

Referring to fig. 15C, in some implementations, conductor 2120 may be provided by a conductive coating on top plate 2130. For example, conductor 2120 may be a stripline electrode plated on ceramic top plate 2130. Each conductor 2120 may be a coating on one or more interior surfaces of the respective groove 2136. The space between the conductor 2120 and the bottom plate 2132 may provide a conduit 2450. The conduit 2450 can carry a fluid as described in fig. 3A.

Plasma simulations were performed using 2-D models to study the dependence of plasma parameters on gas pressure. The computational domain exceeds two half pairs (two half-pairs) of electrodes. Assume process conditions of 1450sccm argon +50sccm N per source₂6Torr, 200W per pair of half electrodes (per pair of half electrodes). Simulations indicate that plasma density will generally be higher in the region below the electrode. Ar + density and electron density were similar (N)₂Much lower density) mainly due to argon versus N₂The proportion of gas supply is high.

Specific embodiments of the present invention have been described. Although this description contains many specific implementation details, many other variations are possible. For example:

the workpiece may be moved linearly through a series of chambers, for example on a belt or linear actuated stage, rather than a rotating stage. Additionally, the workpiece may be stationary, e.g., the workpiece support does not move relative to the wire.

The RF power is connected to the conductor bus at the center, end, or other location or combination of locations of the bus.

Grounding of the electrode bus may be performed at the center, end, or other location or combination of locations of the bus.

The RF power supply may apply signals in the RF, VHF, UHF or microwave range.

Other implementations are within the scope of the following claims.

Claims (30)

1. A processing tool for plasma processing, the processing tool comprising:

a chamber body having an interior volume providing a plasma chamber, the chamber body having a ceiling and an opening on a side opposite the ceiling,

a workpiece support that holds a workpiece such that at least a portion of a front surface of the workpiece faces the opening;

an actuator that produces relative motion between the chamber body and the workpiece support such that the opening moves laterally across the workpiece;

a gas distributor to deliver a process gas to the plasma chamber;

an electrode assembly comprising a plurality of coplanar filaments extending laterally through the plasma chamber between the workpiece support and the top plate, each filament of the plurality of filaments comprising a conductor; and

a first RF power source that provides a first RF power to the conductor of the electrode assembly to form a plasma.

2. The processing tool of claim 1, wherein the workpiece support is rotatable about an axis of rotation and the actuator rotates the workpiece support such that rotation of the support carries the workpiece across the opening.

3. The processing tool of claim 2, wherein the plurality of coplanar filaments extend across a wedge-shaped region.

4. The processing tool of claim 3, wherein the workpiece is completely engaged within the wedge region such that, in operation, an entire front surface of the workpiece is exposed to plasma.

5. The processing tool of claim 3, wherein the workpiece is larger than the wedge region such that, in operation, a wedge portion of the front surface of the workpiece is exposed to plasma.

6. The processing tool of claim 3, wherein the opening is wedge-shaped.

7. The processing tool of claim 3, wherein the plurality of coplanar filaments comprises linear filaments, and different filaments have different lengths so as to define the wedge-shaped region.

8. The processing tool of claim 7, wherein the plurality of coplanar filaments extend in parallel.

9. The processing tool of claim 3, wherein the plurality of coplanar filaments are oriented to have a longitudinal axis at a non-zero angle relative to a direction of motion of the portion of the substrate below the opening.

10. The processing tool of claim 1, wherein a bottom of the chamber is open.

11. The processing tool of claim 1, wherein ends of the plurality of coplanar wire conductors are connected to the first RF power source through a recursive RF feed structure.

12. A plasma reactor, comprising:

a chamber body having an interior volume providing a plasma chamber;

a gas distributor to deliver a process gas to the plasma chamber;

a pump coupled to the plasma chamber to evacuate the chamber;

a workpiece support that holds a workpiece;

an in-chamber electrode assembly comprising a plurality of filaments extending laterally through the plasma chamber between a ceiling of the plasma chamber and the workpiece support, each filament comprising a conductor surrounded by a cylindrical insulating shell, wherein the plurality of filaments comprises a first plurality of filaments and a second plurality of filaments, the second plurality of filaments being arranged in an alternating pattern with the first plurality of filaments,

a first bus and a second bus, the first bus coupled to the first plurality of filaments, the second bus coupled to the second plurality of filaments;

an RF power source that applies an RF signal to the electrode assembly within the chamber; and

at least one RF switch configured to controllably electrically couple and decouple the first bus to one of: i) ground, ii) an RF power source, or iii) the second bus.

13. The plasma reactor of claim 12 wherein said at least one RF switch is configured to controllably electrically couple and decouple said first bus to said second bus, and wherein said at least one RF switch comprises a plurality of switches connected in parallel between different pairs of locations of said first bus and said second bus to controllably electrically couple and decouple said first bus to said second bus.

14. The plasma reactor of claim 12 wherein said at least one RF switch comprises a first switch configured to controllably couple and decouple said first bus to ground and at least one second RF switch configured to controllably couple and decouple said second bus to ground.

15. The plasma reactor of claim 14 wherein said at least one RF switch comprises a first plurality of switches connected in parallel between different locations of said first bus and ground, and said at least one second switch comprises a second plurality of switches connected in parallel between different locations of said second bus and ground.

16. The plasma reactor of claim 12 wherein said at least one RF switch comprises a first switch configured to controllably electrically couple and decouple said first bus to said RF power source and at least one second RF switch configured to controllably electrically couple and decouple said second bus to said RF power source.

17. The plasma reactor of claim 16 wherein said at least one RF switch comprises a first plurality of switches connected in parallel between different locations of said first bus and said RF power source, and said at least one second switch comprises a second plurality of switches connected in parallel between different locations of said second bus and said RF power source.

18. The plasma reactor of claim 12, comprising:

a third bus coupled to the first plurality of filaments and a fourth bus coupled to the second plurality of filaments,

wherein the plurality of filaments have a plurality of first ends and a plurality of second ends, and the first end of each respective filament is closer to the first sidewall of the plasma chamber than the second end of the respective filament, an

Wherein the first bus is coupled to a first end of the first multifilament, the second bus is coupled to a first end of the second multifilament, the third bus is coupled to a second end of the first multifilament, and the fourth bus is coupled to a second end of the second multifilament.

19. The plasma reactor of claim 18 wherein said at least one RF switch is configured to controllably electrically couple and decouple said first bus to said second bus and comprises at least one second RF switch configured to controllably electrically couple and decouple said third bus to said fourth bus.

20. The plasma reactor of claim 18 wherein said at least one RF switch comprises a first switch configured to controllably couple and decouple said first bus to ground and at least one second RF switch configured to controllably couple and decouple said third bus to ground.

21. The plasma reactor of claim 18 wherein said at least one RF switch comprises a first switch configured to controllably couple and decouple said first bus to ground, and at least one second RF switch configured to controllably couple and decouple said second bus to said RF source, at least one third RF switch configured to controllably couple and decouple said third bus to ground, and at least one fourth RF switch configured to controllably couple and decouple said fourth bus to said RF source.

22. The plasma reactor of claim 18 wherein said at least one RF switch comprises a first switch configured to controllably electrically couple and decouple said first bus to said RF source, and at least one second RF switch configured to controllably electrically couple and decouple said second bus to said RF source, at least one third RF switch configured to controllably electrically couple and decouple said third bus to said RF source, and at least one fourth RF switch configured to controllably electrically couple and decouple said fourth bus to said RF source.

23. A plasma reactor, comprising:

a chamber body having an interior volume providing a plasma chamber;

a gas distributor to deliver a process gas to the plasma chamber;

a pump coupled to the plasma chamber to evacuate the chamber;

a workpiece support that holds a workpiece;

an in-chamber electrode assembly comprising a plurality of filaments extending laterally through the plasma chamber between a ceiling of the plasma chamber and the workpiece support, each filament comprising a conductor surrounded by a cylindrical insulating shell,

a bus outside the chamber and coupled to opposite ends of a plurality of filaments;

an RF power source that applies an RF signal to the electrode assembly within the chamber; and

a plurality of RF switches configured to controllably electrically couple and decouple a plurality of different locations of a bus to one of: i) ground or ii) the RF power supply.

24. A plasma reactor, comprising:

a chamber body having an interior volume providing a plasma chamber;

a gas distributor to deliver a process gas to the plasma chamber;

a workpiece support that holds a workpiece;

an electrode assembly including a plurality of conductors spaced apart from and extending laterally across the workpiece support in a parallel coplanar array;

a first RF power source providing a first RF power to the electrode assembly; and

a dielectric baseplate between the electrode assembly and the workpiece support, the dielectric baseplate providing an RF window between the electrode assembly and the plasma chamber.

25. The plasma reactor of claim 24 comprising a dielectric ceiling, wherein said plurality of conductors are positioned between said dielectric ceiling and said dielectric window.

26. The plasma reactor of claim 25 wherein said bottom plate has a lower surface with a plurality of parallel slots, and wherein said plurality of parallel coplanar conductors are positioned in said plurality of parallel slots.

27. The plasma reactor of claim 26 wherein said shell forms a conduit and said conductor is suspended in and extends through said conduit, or wherein said conductor comprises a hollow conduit, or a plurality of wires in said plurality of slots, each wire comprising a conductor and a non-metallic shell surrounding said conductor.

28. The plasma reactor of claim 25 wherein said plurality of conductors are coated on said dielectric ceiling.

29. The plasma reactor of claim 25 wherein said plurality of conductors are embedded in said dielectric ceiling.

30. The plasma reactor of claim 24 wherein said plurality of conductors comprises a first plurality of conductors and a second plurality of conductors, said second plurality of conductors being arranged in an alternating pattern with said first plurality of conductors, and said RF power supply is configured to apply a first RF input signal to said first plurality of conductors and a second RF input signal to said second plurality of conductors.

Images