JPH1197430A - High-density plasma processing chamber - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a processing chamber which provides a high-density, highly directive plasma. SOLUTION: A processing chamber 55, which processes a semiconductor substrate 60 in a plasma, has a process-gas distributor 100 for distributing the process gas into the plasma zone of the chamber. By using an inductor antenna 135, an inductive plasma is formed from the process gas in the plasma zone. A primary bias electrode 145 located at a ceiling 140 of the chamber has a conducting surface 150 exposed in the plasma zone 65. A dielectric member 155 having a power electrode embedded inside has the receiving surface for receiving the substrate. A secondary bias electrode 170 under the dielectric member 155 has a conducting surface 175 exposed in the plasma zone. An electrode voltage source 180 maintains the power electrode 165, the primary bias electrode 145 and the secondary bias electrode 170 at the different potentials and provides a high-density, highly directive plasma to the plasma zone of the chamber.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to a process chamber for processing a semiconductor substrate.

[0002]

2. Description of the Related Art Process chambers include plasma enhanced chemical vapor deposition (CVD), reactive ion etching (RIE), and ion implantation.
Used in the semiconductor fabrication process. FIG.
1 shows a conventional process chamber 20 having a gas distributor 22 that provides process gas to the chamber. A coil power supply 24 powers an inductor coil 26 adjacent the chamber that inductively couples RF energy to the process gas to form a plasma. The process electrodes used to couple the RF power to the plasma are:
It typically includes a cathode 28 below the substrate and an anode 32 surrounding the cathode. Cathode 28 is electrically insulated from anode 32 by one or more quartz or silicon dioxide insulator shields 34 extending below and around it. Power supply 36 applies impedance matched RF bias power to the cathode, the anode being formed by the electrically grounded side walls and top wall of chamber 20. Cathode 28 includes an electrostatic chuck 38 placed thereon, substrate 3
0 and capacitively coupled to the anode 32 via a plasma sheath formed at the cathode boundary. The capacitive coupling electric field energizes the plasma ions to accelerate in the direction of the substrate 30.

[0003] Conventional chambers 20 often have a low plasma ion density and a plasma ion energy distribution with a large number of energy level peaks that diffuse to a wide range of energy levels. Plasma ion density and energy level depend on the plasma sheath electron density or energy distribution, the power and frequency of capacitively coupled RF power applied to electrodes 28, 32, process gas composition and pressure, and components inside the chamber. The plasma sheath is an electron deficient region caused by the difference in mobility between electrons and positive ions in the plasma and has an associated voltage waveform caused by resonance, impedance loading, and other electrical properties of the chamber. This voltage waveform is normally (n) due to the harmonic content of the plasma sheath.
ormal). RF bias and impedance matching circuit, power transmission line, electrodes 28, 32
And the components inside the chamber 20 all result in the generation of harmonics of the plasma sheath, which in turn disturbs the sheath voltage waveform over time, producing perturbed waveforms. The perturbation waveform may generate a “double dominant” multi-mode plasma energy distribution where the plasma ion energy distribution curve has more than one peak. Such a multiple peak energy distribution is undesirable. This is because the large spread of plasma ion energy results in poor plasma performance, for example, low average plasma ion energy levels that result in etch stops in the process of etching high aspect ratio trenches.

[0004] The low plasma density may be due to a number of components within the chamber 20, such as a dielectric or insulator shield 34, an electrostatic chuck 38, a focus ring, and a gas seal. These components provide a large chamber impedance load due to the number of capacitive couplings formed across the components. For example, a capacitive coupling 40 extending from the side wall of the cathode 28, through the adjacent dielectric component, to the surrounding anode wall 32, forms an individual "leaky" capacitor that adds an impedance load to the chamber. As a result of these parasitic or stray capacitances, the impedance loading of stray capacitances is usually much greater than the impedance loading of the plasma, making it difficult to adjust the RF bias supply to match the plasma impedance. is there. Also, stray capacitance may alter the plasma sheath waveform as a function of the magnitude of the capacitive effect. Thus, it is desirable to provide a more stable and controllable plasma characteristic by having a chamber that minimizes stray capacitance and impedance loading between the cathode, dielectric structure and other components within the chamber.

In some cases, components that are capacitively coupled to each other in the chamber greatly reduce the electric field in the chamber, thereby reducing the vector directional energy of the plasma ions. The low directional energy of the plasma ions is due, at least in part, to capacitively coupled surfaces that make angles other than perpendicular to the plane of the substrate. In particular, the cathode 28
The vertical outer edge coupling deviates from normal the electric field components perpendicular to the plane of the substrate 30,
The cathode 28 may be curved toward the outer edge.
Also, the electrostatic chuck 38 used to securely hold the substrate in the chamber provides an additional interface that is capacitively coupled to each other. Plasma species are also attracted in the direction of capacitive coupling formed across the insulator shield 34 surrounding the substrate 30. These capacitive couplings result in an asymmetric distribution of plasma ion energy levels or plasma ion densities across the surface of the substrate, such that the periphery and center of the substrate are processed at different rates.

Another problem with conventional chambers arises from the use of a quartz insulator shield 34 surrounding the cathode 28. The insulator shield 34 reduces the ratio of the surface area of the cathode 28 to the surface area of the anode 32 by occupying space in the chamber and preventing the cathode 28 from extending across the width of the chamber. A low anode to cathode surface area ratio reduces plasma ion density and energy in the chamber. Further, when the diameter of the cathode is reduced, the plasma ion density and energy at the peripheral portion of the substrate 30 may be reduced as compared with the central portion of the substrate. Further, the quartz insulator shield 34 is complex in shape and expensive to manufacture, often requiring replacement by erosion in a corrosive plasma environment. The insulator shield 34 reduces the flow of heat into and out of the substrate,
It also acts as a thermal insulator surrounding the cathode 28. Thus,
It is desirable to have a plasma processing chamber in which there is no insulator shield 34 around the cathode 28 and the cathode 28 extends substantially across the width of the chamber. However, this is not possible with current chamber designs. 3 because the cathode 28 typically comprises a metal base plate that is exposed to the plasma and that will short circuit in the plasma without the presence of an insulator shield 34 around the cathode.

[0007]

Conventional chambers also suffer from problems arising from the placement of inductor coil 26 adjacent chamber 20. The inductor coil 26 parallel to the side wall of the chamber 20 has a strong
Provides a non-uniform electric field across the substrate surface with a weak induced electric field at the periphery. On the other hand, an inductor coil that inductively couples energy through a flat dielectric ceiling (not shown) that penetrates the RF induced electric field, because the ceiling is made of a non-conductive dielectric material, Do not allow capacitive coupling. It is desirable that both the capacitive and inductive electric field components within the chamber have a highly directional vector electric field component that is substantially perpendicular to the surface of the substrate and extends uniformly over the entire substrate surface.

Another problem with conventional designs for inductor coil 26 arises from the relatively large volume of space required in chamber 20 to form an inductively coupled plasma from conventional inductor coils. The inductor coil 26 typically surrounds the chamber 20, and the chamber must have a large volume to provide a large skin depth for the RF induction field from the coil. Without a large penetration thickness, the magnitude of the RF induced electric field created in the chamber 20 would not be high enough to generate a plasma. However, a plasma chamber 2 having a large internal volume
0 must use a relatively large power level RF bias voltage to energize the plasma ions,
This is undesirable because it causes the substrate to overheat. Also, it is more difficult to stably maintain a uniform plasma ion density and energy distribution over a large plasma volume. In addition, it is more difficult to precisely adjust high-density plasma occupying a large volume of space due to an increase in energy perturbation generated in a large-sized plasma volume. To provide a high density plasma in a small volume chamber having a diameter slightly larger than the diameter of the substrate 30, a new inductor coil design is required.

Thus, there is a need for a plasma process chamber that provides a high density plasma having a uniform energy distribution and less perturbation of the ion energy distribution. There is also a need for an apparatus that provides a uniform distribution of plasma ions having a highly directional energy vector over the entire surface of a substrate. There is also a need for designing process electrodes and inductor coils that can generate a stable and controllable plasma in the chamber. Further, there is a need for a plasma processing chamber that eliminates the use of an insulator shield around the cathode while providing a high anode to cathode surface area ratio within the chamber.

[0010]

SUMMARY OF THE INVENTION The present invention is directed to a process chamber for providing a high density, highly directional plasma for processing semiconductor substrates. The process chamber includes a process gas distributor for distributing a process gas capable of forming a plasma to a plasma zone of the chamber. The primary bias electrode on the ceiling of the chamber has a conductive surface exposed to the plasma zone. A single monolithic dielectric member is located below the primary bias electrode. The dielectric member has a power electrode embedded therein,
It has a receiving surface for receiving a substrate. The secondary bias electrode is below the dielectric member. The electrode voltage source keeps the power, primary and secondary bias electrodes at different potentials to provide a high density, highly directional plasma to the chamber.

In a preferred embodiment, a unitary monolithic dielectric member including a power electrode includes a bonding coating having a receiving surface for receiving a substrate to face a primary bias electrode. The tie layer consists of RF and D applied to the electrodes.
The electric fields from the C voltage each have capacitive absorption through the coupling layer to energize the plasma in the chamber and have low enough electric field absorption to hold the substrate electrostatically. The decoupling layer surrounds the other surfaces of the power electrode and has a high electroabsorption to substantially eliminate capacitive coupling of the RF voltage from the power electrode to the surrounding chamber walls. A single monolithic dielectric member provides a much lower chamber impedance of less than 3000 picofarads because it eliminates the need for a separate insulator shield around the power electrode. This substantially improves chamber plasma performance by reducing the chamber impedance load added to the plasma impedance load.

It is desirable to generate an inductive electric field in the chamber using an inductor antenna to form an inductive plasma from the process gas in the plasma zone. More preferably, the inductor antenna is adjacent to a semiconductor ceiling having (i) electrical susceptibility low enough to allow penetration of an RF induced electric field, and (ii) a conductive surface exposed to the plasma zone. A method of processing a substrate using a process chamber includes the steps of (1) placing the substrate on a surface of a single dielectric member, (2) generating an induced electric field in the chamber by passing a current through an inductor antenna, and (3) A) forming a high-density, highly directional plasma in the chamber while maintaining the power electrode and the primary and secondary bias electrodes at different potentials.

[0013] Instead of a single unitary dielectric member having electrodes embedded therein, the process chamber comprises a plurality of dielectric members positioned below a semiconductor ceiling with electrodes embedded therein. A dielectric member may be provided. The first dielectric member has a receiving surface for receiving a substrate thereon. The plasma in the chamber is energized by maintaining the semiconductor ceiling and the electrodes at different potentials using an electrode voltage source.

In another form of the process chamber, the primary bias electrode is made from a semiconductor material that forms a wall on the ceiling of the chamber. The semiconductor wall has a conductive surface and R
It has low electrical sensitivity to allow penetration of the F-induced electric field. One or more inductor antennas are positioned adjacent to the semiconductor wall to generate an RF induced electric field that is transmitted through the semiconductor ceiling and forms a plasma in the chamber. A single monolithic dielectric member with a power electrode embedded therein is within the chamber. The dielectric member has (i) a substrate receiving surface for receiving the substrate, and the RF and DC voltages applied to the power electrode are capacitively coupled to energize the plasma in the chamber, respectively, and A coupling layer having a low electroabsorption that is only electrostatically held;
i) a non-coupling layer surrounding the other surface of the electrode, the non-coupling layer having high electric absorption so as to reduce capacitive coupling therethrough. An electrode voltage source is used to maintain the primary bias electrode and the power electrode at different potentials, thereby providing a high density, highly directional plasma to the chamber.

In yet another form, the process chamber comprises:
A gas distribution system for distributing process gas to the chamber is provided. The primary bias electrode is on the ceiling of the chamber and has a conductive surface exposed to the plasma zone. The monolithic monolithic dielectric member with a power electrode embedded therein has (i) a substrate receiving surface for receiving a substrate, and RF and DC voltages applied to the electrodes are each plasma-enhanced in the chamber. And (ii) a non-bonding layer surrounding the other surface of the electrode, which is capacitively coupled through the bonding layer to bias And a non-coupling layer having high electric absorption only to reduce capacitive coupling therethrough.
The primary bias electrode and the power electrode are maintained at different potentials using an electrode voltage source. A multi-directional magnetic field generator adjacent to the chamber generates a multi-directional magnetic field in the plasma zone having a time-varying angular orientation and magnitude. The magnetic field generator may include a plurality of electromagnets positioned adjacent to the chamber, and an electromagnet power supply that varies a current applied to the electromagnets. Alternatively, the magnetic field generator comprises a plurality of movable permanent magnets arranged adjacent to the chamber and means for moving the permanent magnets.

These and other features, aspects, and advantages of the present invention will be better understood with regard to the following description, appended claims, and accompanying drawings which illustrate preferred embodiments of the invention.

[0017]

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a process chamber 55.
Substrate 30 in a high-density, highly directional plasma
To a plasma processing apparatus 50 used to process the plasma. A high-density plasma is a plasma having an ion energy density exceeding 10 ¹¹ ions / cm ³ compared to a conventional plasma having a low ion density on the order of 10 ¹⁰ ions / cm ³ . High directivity means that charged plasma ions and nuclides are energized in the plasma zone 65 by the electric field vector component 60 and accelerated in a direction substantially perpendicular to the plane of the substrate 30. The high-density, highly directional plasma provides a large amount of reactive plasma nuclides that actively collide with the substrate 30 and efficiently react with the substrate or transfer energy to the substrate. Using a highly directional plasma, etching, implantation,
Material can be deposited thereon.

FIG. 2 schematically illustrates a typical plasma processing apparatus 50 of the present invention. The figures are provided solely to illustrate examples of the present invention and should not be used to limit the scope of the invention. Apparatus 50 generally comprises a closed chamber 55 having a side wall 70 and a bottom wall 75 made from any one of a variety of materials, including metals, ceramics, glass, polymers, and composites. Process chamber 5
The metals commonly used to fabricate No. 5 are aluminum, anodized aluminum, "HAYNES24
2 "," Al-6061 "," SS304 "," SS3 "
16 ", and" INCONEL ", of which anodized aluminum is preferred. Process gas is introduced into the chamber 55 via a gas distributor system 80 that includes a process gas source 85 and a gas flow control system 90 that operates a gas flow meter 95 to distribute the gas to the chamber. Gas distributors
It is desirable to have a silicon ring 100 surrounding the substrate 30. As shown in FIG.
0 is an annular passage 105 for holding the process gas
And a series of gas injection holes 110 for distributing the process gas around the substrate 30. Silicon ring 10
0 uses an optical pyrometer for temperature feedback, a series of lamps 115 under the ring to radiantly heat the ring, and cooling by conduction to the bottom wall 75, about 25 minutes.
It is desirable that the temperature be controlled with an accuracy of ± 5 ° C. in a temperature range of 0 to about 600 ° C. The temperature controlled silicon ring 100 also serves as a controlled silicon source for scavenging free fluorine ions and radicals from the process environment. One or more pumps 125 (typically 100
An exhaust system 120 including a roughing pump (0 liter / sec.) And a throttle valve 130 is used to exhaust the used process gas and to control the pressure of the process gas in the chamber 55. It is desirable to exhaust gas from chamber 55 using an asymmetric pumping channel 132 to provide a uniform distribution of gaseous nuclides around the surface of substrate 30.

In the embodiment shown in FIG.
Comprises an inductor antenna 135 adjacent to the chamber for generating an inductively coupled electric field within the chamber to form a high density inductive plasma therein. Inductor antenna 135 has R
It is desirable to have multiple coils located adjacent to the chamber ceiling 140 for inductively coupling F power to the chamber 55. The plurality of process electrodes are electrically biased with respect to each other and capacitively coupled via a plasma sheath and a dielectric layer separating the electrodes to energize the plasma ions. The primary bias electrode 145 has a first conductive surface 150 exposed to the plasma zone 65. A single monolithic dielectric member 155 located directly below the primary bias electrode 145 has a receiving surface 140 for receiving the substrate 30 thereon. The power electrode 165 is provided so that the dielectric material completely surrounds the power electrode.
Embedded inside 55. The chamber 55 further includes a second conductive surface 175 that is disposed directly below the dielectric member 155 and that is preferably exposed to the plasma zone 65.
A bias electrode 170 is provided. An electrode voltage source 180 is provided to keep the power electrode 165 and the first and second bias electrodes 145, 170 at different potentials. A combination of three or more electrodes is vertically aligned to form a stacked cylindrical electrode arrangement to provide an intense plasma with highly directional plasma ions in chamber 55.

The primary bias electrode 145 comprises a conductive element disposed directly above the substrate 30 with a large area to cover substantially the entire area of the substrate. Primary bias electrode 145 is capacitively coupled to power electrode 165 within dielectric member 155 to provide a highly directional electric field component 60 substantially perpendicular to the plane of substrate 30. In a preferred embodiment, the primary bias electrode 145 includes a semiconductor ceiling 140 that acts as a conductor that can be biased or grounded to create an electric field in the chamber 55. Semiconductor ceiling 140
Is the R transmitted by the ceiling inductor antenna 135
It also acts as an induction field transmitting window that provides a low impedance to the F induction field. The semiconductor ceiling 140 is sufficiently conductive to act as a primary bias electrode 145 and has a low field sensitivity to transmit the induced electric field generated by the inductor antenna 135 through the ceiling with minimal power loss. .
The ceiling 140 is preferably made of silicon, which is less likely to be a source of contamination for processing the silicon substrate 30 than other materials. However, other well-known semiconductor materials such as silicon carbide, germanium, or III-V compound semiconductors, such as gallium arsenide or indium phosphide, or II-III-V compound semiconductors, such as mercury cadmium telluride, can also be used. . In a preferred embodiment, the semiconductor ceiling 140 is less than about 500 Ω-cm (room temperature), more preferably from about 10 Ω-cm to about 300 Ω-cm.
cm, most preferably from about 20 ohm-cm to about 200 ohm
It consists of a slab of semiconductor silicon with a resistivity of -cm.

The temperature of the silicon ceiling 140 is maintained in a temperature range such that the semiconductor material has semiconducting properties and the carrier electron concentration is fairly constant with temperature. For silicon, the preferred temperature range is from about 100K (below which silicon begins to have dielectric properties) to about 600K (above which silicon begins to have metal conductor properties). At temperatures above this range, silicon gives conductance similar to metals,
Above this temperature range, it behaves as a very low conductivity dielectric material. Active control of the temperature of the semiconductor ceiling 140 is desirable to act as both an induction field window and an electrode. Active temperature control of windows ensures consistent and stable plasma and good "cold start" for plasma
(cold start) condition is also provided. The temperature of the ceiling 140 is controlled using a plurality of radiant heaters, such as a tungsten halogen lamp 115, and a heat transfer plate 185 made of aluminum or copper having a passage (not shown) through which the heat transfer fluid flows. The heat transfer fluid source supplies the heat transfer fluid to the passage,
The chamber 55 is maintained at a constant temperature by heating and cooling the heat transfer plate 185 as needed. The semiconductor ceiling 140 has a plurality of high thermal conductivity rings 190 seated on the ceiling 140 on the bottom surface and supporting the plate on the top surface.
Is in thermal contact with the plate 185. Disposed around the lower portion of the heat transfer ring 190 is an inductor antenna 135. The height of the heat transfer ring 190 is
The plate 185 is selected to be supported on the inductor antenna 135 at least one half of the total antenna height. This would otherwise result from access to the conductive surface of plate 185, antenna 135
To reduce or eliminate the reduction in inductive coupling between the plasma and the plasma.

The power electrode 165 embedded in the single monolithic dielectric member 155 is located apart from and directly below the primary bias electrode 145. The power electrode 165 has two functions, from the power electrode to the primary bias electrode 1.
45 as well as serving as a plasma electrode for forming a capacitively coupled electric field extending to the dielectric member 15.
It also serves as an electrostatic member that generates an electrostatic charge for electrostatically holding the sheet. It is desirable to use the power electrode 165 to carry both the DC chucking voltage and the RF bias voltage, both voltages being applied by an electrical connector 195 such as a banana jack inserted through a dielectric material 155 that connects to the power electrode 165. Is done. The voltage source 180
An AC voltage source for providing an RF voltage for plasma generation to the power electrode 165 and a DC voltage source for providing a chucking voltage to the electrode 165 are included. AC voltage source is 1
An RF generating voltage for forming a capacitively coupled plasma in a chamber having one or more frequencies from 3.56 MHz to 400 KHz is provided. The power level of the RF bias current applied to electrode 165 is typically from about 50 to about 3000W. Independent DC voltage applied to electrode 165
To form a static charge that holds the substrate on the chuck to form a DC blocking capacitor circuit electrically connected to the DC chuck power supply. RF power to bridge circuit and D
Connects to a C converter to provide DC chuck power to the electrodes. The voltage source 180 is a system controller for controlling the operation of the electrodes by directing DC and / or RF currents to the electrodes for chucking and dechucking of the substrate 30 and generation of plasma in the plasma chamber 55. May be included. The DC chuck power supply typically provides a voltage for the DC chuck between 250 and 2000 V to the electrode 165.

The active area of the power electrode 165 is selected to couple to the primary bias electrode 145 to maximize the area of the electric field therebetween. Since there is no insulator shield in the chamber, the active area of the power electrode can be increased to cover an area having a diameter that extends across the entire bottom of the chamber. This provides a much larger power electrode active area than the conventional cathode active area. Desirably, the ratio of the surface area of the primary bias electrode 145 to the surface area of the power electrode 165 is at least about 0.9: 1. More preferably, the power electrode is at least about 9% of the area of the primary bias electrode 145 in the chamber.
8%. For a circular substrate 30 having a diameter of about 200 mm (8 inches), a suitable diameter for the power electrode 165 is about 180 to about 220 mm.

The plasma ion density of the plasma zone 65 can be set such that the power electrode 165 is located within a short distance from the primary bias electrode 145 which is smaller than the substrate diameter, and is only a fraction of the substrate diameter (for example, within a few centimeters for a 20 cm diameter wafer). ) By placing them closer to each other. The vertical confinement of the plasma near the substrate surface will result in a parasitic capacitance from the power electrode 165 to the chamber 70 sidewall 70 by making the power electrode much closer to the overhead primary bias electrode 145 than to the chamber 55 sidewall. Reduce energy loss due to Preferably, the distance between the two electrodes is between about 1 and 20 cm, more preferably between about 2 and about 10 cm.

The unitary monolithic dielectric member 155 with the embedded power electrode 165 offers several advantages over prior art electrode designs. First, the combined plasma energizing and chucking electrode 165 is "hot" with the other electrodes in the chamber 70 maintained at a different potential, including electrical ground or floating potential, with respect to the power electrode 165. The only conductor to which a (hot) RF voltage is applied. Because the power electrode 165 is embedded in the unitary dielectric member 155, the power electrode need not be electrically isolated from electrical ground by an insulator shield as in prior art chambers. The absence of any insulator shield indicates that the power and secondary bias electrodes 16
5, 170 and the number of parasitic capacitances between the chamber walls 70, 75 are reduced. This eliminates the parasitic chamber impedance load that would otherwise occur between the power electrode 165, the insulator shield, and the ground chamber wall. Since the parasitic RF loss to ground is minimized,
The circuit of the power supply 180 that matches the impedance of the RF load to the impedance of the power supply is simplified, thereby allowing more efficient transfer of power. Also, the complex shape of the quartz insulator shield 34 is expensive to manufacture and can quickly erode in many process environments. Finally, by combining the functions of the electrostatic chuck 38 and the electrode of the power electrode 165, the use of a single electrostatic chuck is no longer required.

Referring to FIGS. 3 (a) and 3 (b), the electrical properties of the dielectric member 155 are selected to obtain a low conductivity of about 10 ⁸ to 10 ¹⁴ Ω-cm.
The dielectric member 155 includes a coating layer 200 for electrically insulating the electrode 165 to prevent a short circuit of plasma in the chamber 55 and for insulating the substrate 30 from the electrode. The dielectric covering layer 200 has a low electric absorption enough to capacitively couple the RF bias voltage applied to the electrode 165 to the primary bias electrode 145 through the plasma sheath formed on the dielectric member 155 via the covering layer. Formed from body material. The dielectric constant, field sensitivity, and thickness of the cladding layer 200 may affect the primary bias electrode 145 through the cladding layer.
It is selected to enhance the capacitive coupling of the RF voltage applied to the power electrode 165. Further, the cover layer 200 allows a DC voltage applied to the power electrode 165 to electrostatically hold the substrate 30 by Coulomb or Johnsen-Rahbek electrostatic attraction. Desirably, coating layer 200 has a dielectric constant of at least about 2. Power electrode 1
The dielectric support layer 205 below 65 is sufficiently thick and has high electroabsorption to reduce or limit the electrical coupling between the power electrode and the underlying secondary bias electrode 170. The peripheral edge 208 of the power electrode 165 facing the electrical ground side wall 70 of the chamber 55 is preferably surrounded by a side decoupling layer 210 having a high electric field absorption to prevent coupling of the electrode to the chamber side wall 70. Support layer 205
Desirably, the dielectric material of the layers surrounding the power electrode 165, such as the side layer 210, has a dielectric constant of at least about 1.

The RF reactance of each layer of the unitary dielectric member 155 is adjusted to achieve a desired coupling or decoupling effect through each portion of the dielectric member. Dielectric coating layer 200 has a thickness of about 1 to about 500 ohms, more preferably
Having an RF reactance of 50 to 50 Ω,
Is about 100 to about 10,000 Ω, more preferably 10
It is desirable to have an RF reactance between 0 and 1000Ω. On or dielectric layer minimum thickness of the lower powered electrode 165 can be determined using the formula _{C = (E 0 E r A} ) / L. Here, L is the minimum thickness of the dielectric layer, A is the area, and E ₀ and _Er are the relative dielectric constants of the dielectric material and air. A suitable thickness for the dielectric cover layer 200 is about 1 to 1000 microns, and a suitable thickness for the dielectric support layer 205 is about 0.1 to about 15 mm. For a typical chamber design with an impedance load of about 8300 picofarads, the minimum thickness of the dielectric coating layer 200 is about 3.6 mm. About 400
For the present process chamber having a low impedance load of ~ 600 picofarads, the minimum thickness of the dielectric support layer 205 is preferably about 8.4 mm. This value of capacitance provides a current load of about 6.4 A, which is about 1/6 of the power level of the current supplied to the plasma of a conventional chamber.
To provide a more power-efficient, conditioned plasma.

The dielectric member 155 is fabricated as a monolithic structure from hot melt ceramic or polymer.
A single discrete structure including the embedded power electrode 165 is provided. A monolithic ceramic structure is preferred
This is because these materials typically have low porosity, highly reliable electrical properties, and completely surround the power electrode 165, thereby eliminating the need for an insulator shield in the chamber 55. The high breakdown strength of the dense ceramic structure also results in a high RF power level for the power electrode 165 compared to prior art dielectric members having a thin coating of Al ₂ O ₃ which typically has a relatively low breakdown strength. Enables the application of Preferably, the dielectric member 155 is made from a low porosity ceramic having a porosity of about 10% or less. Suitable ceramic materials include one or more of aluminum oxide, aluminum nitride, boron carbide, diamond, silicon oxide, silicon carbide, silicon nitride, titanium oxide, titanium carbide, yttrium oxide, and zirconium oxide. The thermal conductivity of the dielectric material ranges from about 80 to about 24 provided by, for example, diamond or aluminum nitride.
A high conductivity of 0 W / mK is desirable. The power electrode 165 embedded in the dielectric medium is made of a conductive metal such as aluminum, copper, gold, molybdenum, tantalum, titanium, tungsten, and alloys thereof, and more preferably, heats the dielectric member with the embedded electrode. Manufactured from refractory metals such as tungsten, tantalum, or molybdenum that allow sintering. The dielectric member 155 having the embedded power electrode 165 is formed by isostatic pressing from a mixture of ceramic powder and a low-concentration organic binder material.
ressing), hot press molding, mold casting,
Or it is produced by tape casting.

The unitary dielectric member 155 may also include a gas feedthrough hole 215 extending therethrough for providing a helium heat transfer gas to the receiving surface 160 below the substrate 30. Helium gas feed-through hole 21
5 has a diameter small enough to prevent plasma glow discharge of helium in the hole, and the diameter is about 1 to 10 m.
Il (mill, 25-250 microns) is desirable, otherwise helium gas will form a plasma glow discharge or electrical arc at the end of the hole. Typically, gas feedthrough holes 215 are provided around the outer periphery of the dielectric member to provide an even distribution of the heat transfer gas to the area under the substrate 30.

The secondary bias electrode 170 is connected to the power electrode 1
It serves as a bias or reference electrode located below 65. Secondary bias electrode 170 has a diameter or width substantially equal to or greater than the diameter or width of power electrode 165. When the secondary bias electrode 170 is maintained at a slightly negative or positive potential with respect to the power electrode 165, the secondary bias electrode 170
5 serves as a secondary biasing means for controlling the bias electric field between the power electrode 165 and the power electrode 165. The secondary bias electrode also maintains a potential difference between the power electrode and the secondary electrode that redirects the above-described capacitive coupling effect to the secondary bias electrode by a controllable electric field strength; It also serves to reduce stray capacitance that may occur between power electrodes 165. The electric field strength between the two electrodes is controlled by adjusting the relative potential difference of the voltage applied to the two electrodes.

The secondary bias electrode 170 is connected to the power electrode 1
And a conductive element made of a conductive material such as aluminum disposed directly under the dielectric member 155 including the first and second dielectric members. A thermally conductive adhesive bonds the dielectric member to the secondary bias electrode 170. Preferably, the adhesive comprises a pressure-sensitive adhesive loaded with a thermally conductive inert powder on an aluminum carrier having a slit through which it flows. Adhesive bonding takes place in an autoclave to obtain a uniform pressure and to avoid air pockets in the bonding layer. Secondary bias electrode 1
70 is embedded in or covered by a dielectric material, as shown in FIG. Alternatively, when the secondary bias electrode is maintained at a low voltage equal to the voltage applied to the primary bias electrode or electrically grounded,
As shown in FIG. 5A, the secondary bias electrode 170
It has a conductive surface 175 exposed in the plasma zone 65, for example a side wall 225 of the conductive element.

In another function, as shown in FIG.
The secondary bias electrode 170 includes a dielectric member 155 that circulates a heat transfer fluid and is heated by bombardment with strong plasma ions, and a channel 230 for controlling the temperature of the embedded power electrode 165. In this configuration, the electrode power supply 180 desirably maintains the secondary bias electrode 170 at a floating potential, or is otherwise grounded to avoid discharge to the normally water circulating fluid circulating in the channel 230. Since the dielectric member 155 is completely electrically insulated from the secondary bias electrode 170,
No additional electrical insulation layer is required between the secondary bias electrode and the power electrode 215.

The electrode voltage source 180 includes a power electrode 165, a primary bias electrode 145, and a secondary bias electrode 170.
Are kept at different potentials from each other. The electrode voltage source 180 is shown in FIG.
One or more conventional voltage supply systems having output terminals for directing different voltages to different electrodes, as shown in the block diagram of FIG. Electrode voltage source 180 maintains primary bias electrode 145 and power electrode 165 at different potentials by at least about 1000 volts,
It is desirable to provide a strong capacitively coupled electric field between the two electrodes to energize the plasma in the chamber 55. Power supply 18
0 is also the power electrode 165 and the secondary bias electrode 170
And at different potentials by at least 100 V, to provide a weak capacitive coupling between the electrodes and
5 to reduce stray capacitance. Chamber 55
The parasitic capacitance in the secondary bias electrode 170
Is maintained at a different potential than the overlying power electrode 165. For example, when the power electrode is “hot” with a high voltage applied, the secondary bias electrode 170 may be electrically insulated, floating, or grounded. The use of a ground bias electrode below the power electrode 165 provides vector directivity to the electric field component extending from the power electrode to the secondary bias electrode 170. Since the secondary bias electrode 170 is directly below the power electrode 165, any stray field components resulting from capacitive coupling between the power electrode and the chamber walls 70, 75 will turn in the direction of the secondary bias electrode. As a result, the electric field component 60 between the primary bias electrode 145 and the power electrode 164 is directed substantially vertically between the two opposing surfaces of both electrodes with little dilution of the electric field strength or little stray capacitance effect. In this manner, the power supply 180 maintains the primary and secondary bias electrodes 145, 170 at the first low potential and the power electrode 165 at the second high potential, thereby creating a highly directional electric field within the chamber. provide.

The arrangement of the vertically stacked cylindrical multi-electrode system is substantially parallel to the axis of the cylinder of the process chamber 55 and substantially perpendicular to the plane of the processing surface of the substrate 30.
It provides a strong vertical electric field component 60. By controlling the magnitude of the electrical bias voltage applied to the primary, secondary and power electrodes, the magnitude and direction of the strong plasma nuclide vector in the chamber 55 can also be controlled more accurately. This is achieved by three electrodes 145,
This is because the relative potential applied to each of 165 and 170 controls the magnitude and directivity of the electric field component 75 between the electrodes. For example, a high potential difference between the primary bias electrode 145 and the power electrode 165 provides a very strong electric field between the two electrodes that accelerates the plasma nuclide in the direction of the substrate 30. However, the increase in potential difference between the power electrode 165 and the secondary bias electrode 170 can be reduced by redirecting some electric field components radiating downward from the power electrode in the direction of the underlying secondary bias electrode. Will work to reduce the strength of the electric field component directed to the By controlling the relative potential of each electrode, the energy and directivity of the plasma nuclide can be precisely adjusted to make incident plasma ions and neutral particles
The desired energy level distribution of (neutrals) can be achieved.

In the process chamber 55, the plasma ion density and the plasma ion energy are determined by the power supply power level of the current applied to the inductor antenna 135 and the voltage applied to each of the three stacked electrodes 145, 165, and 170. And the bias power levels are controlled independently. Conventional chambers generally do not provide independent control of ion energy levels through the use of more than two plasma electrodes. Typically, the inductively coupled power determines the overall plasma characteristic or plasma ion density, and the capacitively coupled power determines the ion energy of ions incident on substrate 30. Dual control of the power and bias of the process chamber 55 results in ions in excess of 10 ¹¹ ions / cm ³ compared to conventional RIE (reactive ion etching) which results in much lower ion densities on the order of 10 ¹⁰ ions / cm ^3. Providing high-density plasma with energy and 3
Allows better control over the bias voltage applied to the two electrodes. The dominant ion-to-neutral nuclide density ratio and energy level distribution are also small for etched features, such as 0.1.
By providing superior etch performance at sub-5 micron feature sizes, it provides better etch anisotropy, etch profile, and etch selectivity.

Further, the power electrode 165 and the chamber wall 7
The indeterminate parasitic capacitance between 0 and 75 causes the potential difference between the power electrode 165 and the secondary bias electrode 170 to be changed so that the auxiliary electric field component extending from the power electrode 165 is preferentially directed toward the secondary bias electrode 170. By keeping it high, it can be reduced or completely eliminated. The effect of these stray capacitance losses on the impedance load of the chamber 55 can be significant. In fact, the parasitic capacitance load is much greater than the capacitive load of the plasma itself, making it difficult to accurately match the impedance of the power supply 180 to the impedance load of a much smaller plasma. To provide more efficient and stable power transfer to the plasma, the ability to match the impedance of the plasma to the impedance of the power supply 180 is also needed. Further, the magnitude of the parasitic capacitance loss described above depends on the type and arrangement of hardware mounted on the process chamber 55, such as insulator shields, focus rings, and gas distribution nozzles. As process hardware can change from process to process, the impedance load is also changed, making it difficult to match the impedance characteristics of each different chamber configuration. However, the single dielectric member 1
The use of 55 power electrodes 165 provides a chamber impedance load of less than about 3000 picofarads, more typically less than about 1000 picofarads, and most typically less than 500 picofarads, whereas conventional chambers have 7000 to 5000 picofarads. Provides a chamber impedance load of 10,000 picofarads. In this way, the present invention provides more efficient and controllable power transfer to the plasma and more consistent plasma load characteristics than conventional chambers by reducing stray capacitance.

A description will now be given of preferred characteristics and aspects of the process chamber 55 of the present invention, as well as alternative chamber designs.

"Inductor Antenna and Silicon Ceiling" FIGS. 4 and 5 (a) show a variation of the present process chamber having an inductor antenna 135 and an RF field transparent ceiling. The inductor antenna 135 includes a circularly symmetric inductor coil having a center axis that coincides with the longitudinal axis of the process chamber 55 and is perpendicular to the plane of the substrate 30. However, the circular symmetry of the inductor coil provides a spatial distribution of induced electric field vector components having zero or a minimum along the center axis of symmetry, reducing the number of electrons generated on the center of substrate 30. . Thus, the inductor antennas 135 desirably comprise non-planar solenoid coils stacked within each other, each having a central axis substantially coincident with the longitudinal axis of the chamber 55. As described in U.S. patent application Ser. No. 08 / 648,254, the process chamber 55 is a cylindrical chamber, near the ceiling 140 by stacking the coil windings of the inductor antenna 135 vertically in the form of two solenoids 135a, 135b. To increase the product of the current and the number of turns of the antenna (d / dt) (NI) to produce a strong inductive flux linkage tightly coupled to the plasma
And thus provide a higher plasma ion density in the plasma zone 65 adjacent to the substrate 30.

FIG. 5B shows a preferred coil configuration,
The windings of the antenna 135 are wound so that each individual winding is parallel to the horizontal plane of the substrate 30 except for the transition from the horizontal winding to the next horizontal winding. Substantially parallel to the plane of the substrate. Preferably, the inductor antenna 135 comprises one or more solenoid coils, each coil having 1 to 10 turns, more typically 2 to 6 turns. In a preferred configuration,
The inductor coil consists of two 4-turn solenoid coils, the inner coil 135a is about 9 cm, the outer coil 135a.
b has a diameter of about 25 cm. Each coil is liquid cooled to reduce heat transfer to the semiconductor inductor window 145. The coils 135a and 135b are more reliable
Powered by a three-channel RF oscillator that provides more accurate RF power control and eliminates active RF matching of impedance. The control system uses frequency tuning in mutually exclusive frequency ranges for each power supply coil and bias power supply in combination with true supply control.

Power Supply System The preferred power supply system shown in FIG.
a / 135b and a combined RF / DC power supply 180 for supplying power to both the process electrodes 145, 165 and 170. The combined RF / DC power supply 180 has five RF channels and a fixed impedance matching network 235 for each channel to control the true supply power.
RF power is provided to the inner and outer inductor antennas by two independent outputs 240,245, and RF bias power is provided to the process electrodes 145,165,170 by independent electrode outputs 250,255,260. In a preferred embodiment, the electrodes 145, 165, 17
Zero and each inductor antenna 135a, 135b operate at a distinct and mutually exclusive frequency to provide more precise RF power control and minimize RF interactions in the plasma. The combined RF / DC and coil power supply 180 includes a separate fixed impedance matching network 235 for each RF output. Elimination of active RF matching provides higher reliability and process repeatability. However, the present invention is not limited to the use of RF oscillators with fixed impedance matching circuits, but uses frequency tuning and related parameter servo methods. In fact, the present invention is generally applicable where any type of impedance matching circuit is used. For example, the impedance matching component may be variable rather than "fixed," in which case the conventional impedance matching controller of the RF oscillator will adjust one component of the impedance matching network (e.g., a? It operates by changing the reactance of one of the two capacitors.

The RF voltage oscillator module 265 provides an RF energy to the power electrode 165 to generate an electric field extending laterally through the substrate 30 to energize the plasma on the substrate. The RF frequency of the voltage applied to the power electrode 165 is mutually exclusive from the frequency applied to the inductor antenna 135, and the harmonic separation (harmonic separation)
isolation) applied through a fixed matching network 235 with a filter 270. The matching network components are selected to approximately match the impedance of the RF voltage module 256 to the load impedance of the power electrode 165 over a selected frequency range. To tune the RF voltage module 265 to the chamber 55, the RF frequency is varied within a selected range of frequencies to find the minimum reflected power. In a fixed network (power supply or bias) this is near-resonant, but does not necessarily result in zero reflected power. The set point power is compared to the difference between the forward power and the reflected power, and the forward power is automatically adjusted to provide a constant supply power at the set point level to compensate for the impedance magnitude mismatch. Both control loops are repeated to maintain a constant supply power level.

The DC chuck voltage module 275 of the power supply 180 provides a DC output "chuck voltage" of about 12 to 36 V DC to the power electrode 165 to allow the coulomb or Johnsen-Rahbek force to cause the substrate 30 to be a monolithic dielectric. An electrostatic charge that can be electrostatically held by the member 155 is generated in the electrode. For a unipolar electrode, the substrate 30
Is charged by the plasma ions in the chamber to form a residual charge having the opposite polarity that attracts and holds the substrate to the dielectric member 155. In a bipolar electrode, two parts of the electrode that are electrically insulated from each other are electrically biased relative to each other to generate an electrostatic force that holds the substrate 30 to the dielectric member 155.

The distortion of the sheath voltage waveform provides a specific impedance to the harmonic components of the sheath voltage and current (eg, the second and third harmonics) and at the same time the desired impedance matching at the fundamental frequency in cooperation with the impedance matching network. , A harmonic filter 270 that provides the RF output 250, 255 of the combined RF / DC power supply 180 (or impedance matching network).
By providing between 260, it is prevented. Choosing the harmonic filter 270 allows the phase and amplitude of the harmonics generated in the plasma sheath to be "normal" without significantly affecting the impedance matching at the fundamental frequency provided by the impedance matching network 235. Adjust for harmonics required to generate half-wave rectified sinusoidal sheath voltage waveform. In one embodiment of the invention, this is
This is achieved by selecting a harmonic filter to provide a much higher impedance to the desired harmonic (eg, the second, or second and third harmonics) than provided by the matching network 235.

In another embodiment, a plurality of harmonic filters (not shown) arranged in series may each be used to adjust the magnitude and phase of a particular impedance to, for example, 20 harmonic filters for accurate performance optimization. Up to a specific harmonic of the RF signal. This embodiment may be particularly advantageous because the design change requirements increase the aspect ratio of the openings etched into the substrate. Another advantage of using harmonic filter 270 is the ability to optimize impedance matching circuit 235 to match a wide range of plasma load impedances at the fundamental frequency, regardless of behavior at harmonic frequencies. This occurs because at higher order second and / or third harmonics, the impedance of the harmonic filter 270 is greater than the impedance of the matching circuit. Thus, the harmonic filter 270 is connected to the impedance matching circuit and the chamber 55.
In a preferred embodiment in-between, the matching circuit is isolated from harmonic frequency components generated in the plasma sheath.

In a preferred embodiment of the invention, harmonic filter 270 provides the required phase response at the second harmonic (and the third harmonic as well) to provide a normal or sharp half-wave rectified sinusoidal sheath voltage. Primarily inductive for generating waveforms. Further, according to a preferred embodiment, the magnitude of the impedance at the second and third harmonics is relatively high, typically about 50Ω. Although the preferred embodiment uses inductive impedance at the second harmonic (and / or third harmonic), it should be noted that
In any case, the impedance of the RF outputs 250, 255, 260 has a large real component, while the chamber load impedance has a small real component, and therefore is almost purely reactive and has any phase angle of ± 90 degrees. It is. Thus, in each case, the harmonic filter 2
The 70 elements are either purely capacitive or inductive. The latter is preferred because the load impedance of the chamber / plasma / sheath is generally primarily capacitive. A suitable circuit is disclosed in U.S. patent application Ser.
No. 43, entitled "Improvement of Plasma Process Performance by Filtering Harmonics Generated by Plasma Sheath".

In designing a system that includes a matching circuit 235 and a harmonic filter 270, the matching circuit component values are selected while taking into account the impedance matching of the harmonic filter to provide impedance matching at the fundamental RF frequency. This matching is precise enough to limit the reflected RF power to no more than 20%, and preferably no more than 10%, over the entire impedance range including the expected variation in the load impedance of the plasma or plasma chamber 55 being processed.
The component values of the harmonic filter 270 are selected to provide a high inductive impedance over the entire range of the second harmonic frequency. This range is determined by the matching circuit 235.
This is important when using an RF tuning method that changes the fundamental RF frequency to follow variations in the plasma chamber load impedance.

"Magnetic Confinement and Stirring of Plasma Ions" Referring to FIG. 2, the plasma ion density and energy distribution produce a magnetic field 280 that is substantially perpendicular to the plane of the Zone 65
It is further enhanced by confining the plasma to The magnetic field 280 is typically generated by a permanent magnet 285 or an electromagnetic antenna adjacent the chamber 55,
This is the case, for example, with U.S. Pat.
No. 4,842,683, which is incorporated herein by reference. The permanent magnet 285
5 is sealed in a removable chamber liner 290 inside the plasma, semiconductor window 140, silicon ring 1
00 and preferably comprises a double circular ring confined in the region between the substrates 30 and defining a plasma zone 65 in the internal chamber. Magnet 285 has a radially opposed magnetic field direction that produces a strong local magnetic field 280 near the magnet that decays rapidly with radial distance from the magnet. Typically, the actual component of the magnetic field is up to 10 Gauss above the substrate surface. Magnetic confinement reduces the plasma ion current,
Usually reduced by at least two orders of magnitude,
Helps block ion energy greater than 25 eV outside the inner chamber area. Magnetic confinement of the plasma to the processing surface of substrate 30 increases the directivity of plasma ions while minimizing contamination of substrate 30. Magnetic confinement also reduces the amount of polymer deposits that form on the chamber walls in the outer chamber area, thereby reducing the chamber processing steps required after etching.

FIGS. 6 and 7 schematically show a multi-directional magnetic field system for generating a highly directional magnetic field in the plasma zone 65 that enhances the uniformity of the distribution of plasma ions. The chamber 55 has a metal ceiling that acts as a primary bias electrode on a dielectric member 155 with a power electrode 165 incorporated therein. Electromagnetic coils 295, 300, 305, 3
A magnetic field generator 292 adjacent to or within chamber 55, such as an assembly of ten or permanent magnets (not shown), generates a magnetic field in the plasma zone that is the vector sum of the magnetic fields generated by each magnetic field generator. Magnetic field is plasma zone 6
5 depending on the position of the magnetic field generator 292 and the operating mode. It is desirable that the magnetic field generator 292 generates a multidirectional magnetic field having an angular azimuth and magnitude that change over time. In one embodiment, multidirectional magnetic field generator 292 includes a plurality of electromagnets 29 disposed adjacent sidewall 70 of chamber 55.
5, 300, 305, and 310. Electromagnet 295,
300, 305, 310 are powered by an electromagnet power supply that changes the current applied to the electromagnet coil to generate a multi-directional magnetic field in the plasma zone. A suitable coil configuration comprises a plurality of pairs of conductive coils, each pair of coils generating a magnetic field substantially parallel to the plane of the substrate 30. The power supply electrically generates a DC magnetic field and a plurality of paired coils in a selected order to selectively move the magnetic field by independently changing the angular orientation of the magnetic field generated by each coil pair. Turn on electricity. Instead of being located adjacent to the side wall 70 of the chamber 55, the electromagnetic field generator 292 can be located above the ceiling of the chamber and / or below the unitary dielectric member 155 having the power electrode 165 inside. This is disclosed, for example, in US Pat. No. 5,255,024, which is incorporated herein by reference. In addition, the magnetic field generator 292 can include a plurality of movable permanent magnets located adjacent to the side wall 70 of the chamber 55 (not shown). The magnet may be mounted on an armature that rotates and / or moves linearly in a circular or elliptical orbit to generate a multidirectional magnetic field in the plasma zone 65. Suitable permanent magnets comprise a ferromagnetic material such as nickel ferrite, cobalt ferrite, or barium ferrite having a magnetic field strength of about 35 to about 100 Gauss.

The electromagnetic field generator 292 generates a magnetic field that is the vector sum of the magnetic fields generated by each coil or magnet and that is dependent on their position with respect to the chamber 55 and the mode of operation. The electromagnetic coils or magnets can be configured to provide a magnetic field having a component substantially parallel to the surface of the substrate 30 and symmetric about an axis orthogonal to the substrate surface. In this embodiment, the EX given to the electrons in the plasma
The B drift velocity is azimuthal and energizes the electrons in the plasma sheath to move along a circular path in a plane parallel to and just above the processing surface of substrate 30. In another embodiment, the coil or magnet is configured to provide a magnetic field having a principal component that is substantially orthogonal to the plane of substrate 30. In yet another embodiment, the coils or magnets are at an angle to the plasma zone 65 or provide a magnetic field having a component above the plane of the substrate 30 that is curved across its space or volume. Be composed. The particular orientation of the magnetic field of the magnetic field generator 292 is selected to increase the uniformity and speed of processing of the substrate 30.

In the preferred embodiment, the magnetic field generator 29
5,300,305,310 respectively form magnetic vectors By, Bx which are substantially parallel and mutually perpendicular to the pedestal / cathode and the substrate, as generally described in commonly owned US Pat. No. 5,215,619. Disclosed and incorporated herein by reference. In the exemplary illustration of FIG. 6, computer 312 has lines 315, 32
0, 325, 330 via a conventional power system 33
By applying control signals to 5, 340, 345, 350, the magnitude and direction of the current supplied to electromagnets 295, 300, 305, 310 through conductors 355, 360, 365, 370, respectively, are controlled. The associated current determines the direction and magnitude of the magnetic field generated by each coil pair. In addition, using computer 312,
It is also possible to control the reciprocation of a magnetic field generator 292 with a set of ferromagnetic permanent magnets arranged in an armature that can rotate in an oval or reciprocate in a linear direction.
The vertical magnetic field vectors By and Bx generated by the electromagnetic field generator 292 are represented by the functions Bx = Bcos θ and By =
B sin θ. Magnetic field B and its angular orientation θ
When a desired value or a required value is given, the computer 3
12 independently solves the equations to obtain the associated magnetic field vectors By and Bx that provide the desired magnetic field strength and orientation, and then control the application of the required current in the coil or the movement of the permanent magnet to control the desired magnetic field. The magnetic fields By and Bx are provided.

In addition, the angular orientation and magnitude of the DC magnetic field can be independently changed as quickly or slowly as desired by changing the current in the electromagnetic coils 295, 300, 305, 310, or by rotating the magnets. Can be changed.
The control of the computer 312 is used to change the time the magnetic field is at each angular position, the direction of the angular stepping function, and the strength of the magnetic field. Thus, the magnetic field can be stepped around substrate 30 using the selected orientation and increment of time. If desired, the magnitude of the resulting magnetic field Bθ may be changed if the process conditions or chamber configuration requires a constant magnetic field strength. It is desirable to rotate the magnetic field at a slow rate of 2 to 5 seconds / revolution by sequentially changing the current to the electromagnetic coil or rotating the magnet. This causes the magnetic field to step around the substrate 30 and increase the etch uniformity at 360 degrees around the substrate rather than in one direction across the substrate. As a result, the chamber 5 over a wide range of pressures with directivity, selectivity, and uniformity over even low pressure etching systems.
5 can be used.

In an etching process, generally, as the strength of the magnetic field increases, the protective sidewall deposit formed in the plasma etching process becomes thicker (if an oxygen source is present) and the profile of the etched trench becomes tapered. And the curvature becomes smaller. The ability to change the magnetic field allows for control of the etched feature profile at large trench depths. For example, in very narrow and deep trenches, it may be desirable to have a wider trench opening to facilitate filling of the trench with dielectric material. The taper control provided by the adjustment of the magnetic field provides a funnel-shaped, wide-mouth, narrow-trench trench.

The computer control system 312 for the process chamber 55 may be, for example, Synnerg, Calif.
Operated using a computer program running on a conventional computer system with a central processing unit (CPU) interconnected with a storage system with peripheral control components, such as the 68040 microprocessor available from y Microsystems. The computer program code can be written in any conventional computer readable programming language, for example, 68000 assembly language, C, C ++, Pascal, and the like. Suitable program code is entered into a single file or multiple files using a conventional text editor and stored or incorporated in a computer usable medium, such as a computer storage system. If the entered code text is in a high-level language,
The code is compiled and the resulting compiler code is then compiled into the precompiled Windows Library
Linked to the routine's purpose code. To execute the linked compiled objective code, a system user calls the objective code, which causes the computer system 312 to load the code into memory and perform the tasks identified in the program.

"Curved Dome Ceiling" FIGS. 8 and 9 show a process chamber 55 having a cylindrical shape with a domed ceiling 140. FIG.
3 shows another embodiment of the present invention. The dome ceiling 140 can be constructed of a dielectric material that is transparent to the RF electric field, or can be constructed of an integral piece of silicon or silicon tiles bonded together using an adhesive to provide a curved shape. Ceiling 140 may be any suitable shape, including a flat surface, a dome, a cone, a truncated cone, a cylinder, or a combination of these shapes. The inductor antenna 135 has a configuration in which the upper surface of the chamber 55 is not conformable to the ceiling defining a three-dimensional shape, that is, the vertical pitch (the vertical height divided by the horizontal width) of the antenna is low. It is desirable that the pitch be larger than the pitch of the upper surface of the chamber 55. This configuration concentrates the induced electric fields of antennas 135 near their respective symmetry central axes. Also, the antenna 135 is not located around the side of the chamber, but
Being located above zero, the volume of the chamber required to provide the required induced penetration thickness is reduced, and the plasma zone 65 can be confined to a smaller space volume.

"Plasma ion density and energy distribution"
Various experiments were performed on the various chamber configurations described herein to determine the plasma ion density and energy density function on the substrate surface in the chamber. For example, the ion energy spectrum of the surface of the substrate 30 was measured in a simulated environment of the process chamber 55 of the present invention as compared to a simulated environment of a conventional process chamber. Whereas the conventional process chamber 20 had a chamber load impedance of 8300 picofarads, the chamber 70 of this design had a much lower chamber load impedance of about 450 picofarads (about 1/18 of the impedance of the conventional chamber). Both chambers had a constant plasma load impedance of about 358 picofarads. Thus, in the present invention, the load impedance due to the plasma is much larger than the impedance due to the chamber itself, so that the impedance of the RF voltage module 265 is matched to the impedance of the plasma by the fixed matching network 235.
It is possible to design In both chambers, the plasma ion energy spectrum was measured using a small battery powered electrostatic energy analyzer built into the dielectric member 155 and connected to the external environment with fiber optics. Since the dielectric member 155 was covered with a blank silicon wafer, the small holes machined through the wafer allowed access of the plasma ions to the analyzer.
This configuration makes it possible to measure the plasma ion energy distribution in the actual process state in the chamber 55 as it is.

FIG. 10 shows the energy distribution of a conventional process chamber 20 having a "hot" cathode electrode under the substrate and an insulator shield surrounding the cathode. Chamber 20 has a higher capacitive load resulting from insulator shield structures and other components within the chamber. This higher capacitive load significantly affects the plasma ion energy distribution, especially at higher RF bias power levels, resulting in a multi-mode or multi-peak ion energy distribution. Multiple peaks centered at different energy levels provide a large distribution and spread of ions with substantially different energy levels, and result in a relatively large number of plasma ions having low energy values. For example, the first at 600 W RF bias power
Has a peak at 230 eV, which is shown in FIG.
Approximately 100e compared to the ion energy distribution curve obtained for the same RF power level with this chamber design shown in
V lower. Thus, at the same RF bias power level,
The total number of ions at higher energy levels is lower in the conventional chamber design than in the present chamber 55. At higher RF bias power levels, a bimodal distribution with both large and small peaks is obtained. Such a bimodal energy distribution does not efficiently etch narrow or high aspect ratio holes in substrate 30 and often results in an "etch stop" process (where the etching process etches the entire required depth). Without stopping in the trench).

In contrast, the process chamber 55 of the present invention having a relatively low capacitive load and multiple electrodes provides substantially single mode ion energy regardless of the RF bias voltage applied to the chamber power electrode 165. Provided distribution. FIG. 11 shows the primary and secondary bias electrodes 14.
5, 170 are electrically grounded and the single inductor coil above the chamber 55 is operated at a constant power level of 2800 W, with the energy distribution as a function of the RF bias voltage applied to the power electrode 165 being increased. Show. In the figure,
The x-axis shows the measured plasma ion energy level, and the y-axis shows the number of plasma ions that reached the sensor as a function of increasing RF bias voltage on power electrode 165. The peak of the measured plasma ion energy level corresponds to the maximum negative voltage of the RF waveform. About 600
The peak of the first curve at an RF bias of W is 330
Construct a single high intensity peak that exhibits a monomode ion energy distribution with a peak centered at eV. RF
As the bias voltage increases, the average value of the peak of the single mode ion energy distribution shifts to the right, and the 800 W R
It has a peak at 400 eV at the F bias power level.
RF bias power levels of 1000, 1200, 140
As it increases to 0, 1600 W, the ion energy level moves to the right and peaks at about 600 eV. x
A single-mode ion energy distribution traveling transverse to the axis creates a highly directional ion where all power applied to the plasma etches high aspect ratio trenches, such as vertically long SAC etches with narrow openings and contact etches. To be used efficiently. It is clear from these experiments that the monomode ion energy distribution has a single peak with a relatively small spread of ion energy levels, and that the bimodal and spread-out ion energy increases with increasing RF voltage applied to the electrodes. It is to move uniformly to higher energy levels without changing its shape.

FIG. 12 shows that the RF induction electric field is applied to the semiconductor ceiling 14.
Experimental results are presented that represent low induced power loss or absorption when transmitted from inductor antenna 135 via zero.
In this experiment, two inductor antennas were placed on either side of a 13.5 inch diameter, 1 inch thick silicon slab. The inductor antenna 135 on one side of the slab was connected to a variable frequency RF power supply, and the antenna on the other side was connected to a multi-channel analyzer. RF power frequency is 1kHz
Swept from z to 10 MHz. The output of the resulting multi-channel analyzer is shown in FIG. In the curve labeled "magnitude", the vertical axis is the ratio of the magnitude of the received signal to the transmitted signal, falling from the value of 1 at the top of the scale at a scale of 0.1, while the horizontal axis is Shows RF frequencies from 1 kHz to 10 MHz. The axis labeled "Phase" represents the difference between the phase angle of the received signal and the transmitted signal, decreasing from the 20 ° value at the top of the scale on a 20 ° scale. FIG.
As is evident, there is virtually no loss of inductively coupled RF power through the silicon slab up to a frequency of about 2 MHz, and there is a relatively low loss of inductively coupled RF power at a frequency of about 2 MHz.

FIG. 13 shows a wide range of plasma ion saturation currents and plasma ion current uniformity as a function of radial position across the surface of substrate 30. In this experiment, the process chamber 55 of FIG. 2 was supplied with argon and oxygen gas at a ratio of 3: 1 at a flow rate of 600 sccm. The temperature of the silicon ceiling 140 was maintained at 210 ° C. The bias power supplied to the process electrode is maintained at 0 W,
The DC bias applied to the chucking electrode was -100 V DC. Total power supply was maintained at 1800W. The first bottom curve 400 has an external coil power of 1050 W and 75
For the internal coil power of 0 W, the second curve 405 is 1200/600 W, the third curve 41
0 is 1275 / 525W, 4th curve 415 is 135
0 / 450W, fifth curve 420 is 1500/300
W and the sixth curve 425 were obtained for 1800/0 W, respectively, of power to the external / internal coils. As can be seen, by varying the internal to external power coil power ratio, the ion saturation current is constant over a wide range of power and bias power levels. Also, the ion saturation current is linearly dependent on the coil power level and is essentially independent of the RF bias power level applied to the different electrodes. Further, the voltage and the DCRF negative peak voltage (a rough index of the average and peak ion energy in the substrate 30) at the substrate DC bias are RF
It depends linearly on the bias power level. A close correlation between ionic current uniformity and patterned oxide ion etch rate uniformity is also shown for a typical etch process.

In the next experiment, the ratio of the power applied to the external and internal antennas was kept at 2: 1 and the chamber pressure was
The process gas was maintained at 5 mTorr using a process gas consisting of Ar and O ₂ at a flow rate of 100 sccm at a ratio of 3: 1. It is clear from FIG. 14 that the ion saturation current varies linearly with increasing power supply power applied to the inductor coil, making it virtually independent of the RF power applied to the process electrode. RF bias power level is 12
Even if the power is increased from 00 W to 2400 W, the ion saturation current increases only slightly. In contrast, ion saturation current increases rapidly with increasing power supply power levels. However, as can be seen from FIG. 15, the RF bias power applied to the process
This results in a linear control of the bias voltage. Thus, independent control of the power supply power level for the inductor coil and the power level applied to the process electrode provides optimal control of the plasma ion density and energy distribution of the plasma zone.

FIGS. 16A and 16B show the chamber 5
5 clearly shows the magnetic plasma confinement of the plasma in the plasma zone 65 of FIG. FIG. 16A shows the effect of magnetic plasma confinement and bias power on O ₂ + plasma ion density and plasma ion energy inside and outside the plasma zone 65. Permanent magnet 2
The presence of 85 reduces the ion density outside the plasma zone 65 by two orders of magnitude and blocks ion energies above 25 eV. FIG. 16 (b) shows the magnetic plasma confinement measured with a Langmuir probe moving radially across the surface of the chamber 55. In these experiments, a fluorocarbon process gas was applied to chamber 5
5 and the chamber pressure was maintained at 90 mTorr. An RF voltage of 1600 W and a bias potential of −100 V DC was applied to the power electrode 165. 3100W
Is applied to the external inductor coil,
A current at a power level of 800 W was applied to the inner inductor coil. The ion saturation current and the plasma floating potential
It can be seen that it was uniform over the entire radius of zero. Further, the ion potential decreases when the radius exceeds the radius of the substrate 60,
It can be seen that beyond the plasma zone 65, it drops rapidly due to magnetic confinement, and the floating potential is slower than the ionic current, but drops with increasing radius.

[0062]

Although the present invention has been described in considerable detail with respect to its preferred form, other forms are possible. For example, the semiconductor ceiling 140 can be comprised of a thin metal layer of semiconductor material applied over the inductive field transmission window. Also, additional process electrodes operating at different bias levels can be used in the chamber 55 without departing from the scope of the present invention. In addition, top, bottom, center, ceiling, base,
The above terms of spatial orientation of floors and other structures can be changed to equivalent or opposite orientations without affecting the scope of the invention. Therefore, the appended claims should not be limited to the description of the preferred forms, materials, or spatial arrangements described herein for the purpose of describing the invention.

[Brief description of the drawings]

FIG. 1 is a schematic partial cross-sectional view of a prior art process chamber, showing an insulator shield around a cathode and a capacitance coupling formed in the chamber.

FIG. 2 is a schematic cross-sectional view of a process chamber according to the present invention, including an inductor coil, a primary bias electrode, a power electrode,
And a typical arrangement of secondary bias electrodes.

FIG. 3 (a) is a partial cross-sectional perspective view of the lower portion of the process chamber, showing a single monolithic dielectric member having a power electrode embedded therein and a lower secondary bias electrode having a cooling channel; (B) is a schematic cross-sectional view of the single-piece monolithic dielectric member of (a), showing an electrode embedded therein.

FIG. 4 is a schematic partial cross-sectional view of a process chamber showing a block diagram of a power supply for supplying power to a power electrode, primary and secondary bias electrodes, and an inductor coil.

FIG. 5 (a) is a schematic partial cross-sectional view of another embodiment of the process chamber, showing a preferred coil configuration;
(B) is a perspective view of the inductor antenna coil used in the process chamber of (a).

FIG. 6 is a schematic diagram of a magnetic field generation system for generating a multi-directional magnetic field in a process chamber.

FIG. 7 is a schematic partial cross-sectional view of the process chamber of FIG. 6, illustrating an arrangement of a multi-directional magnetic field generator.

FIG. 8 is a schematic cross-sectional view according to another embodiment of a process chamber, wherein the ceiling comprises a dome-shaped wall of dielectric material having an inductor coil adjacent thereto.

FIG. 9 is a schematic cross-sectional view according to another embodiment of a process chamber, wherein the ceiling comprises a dome-shaped wall of dielectric material having an inductor coil adjacent thereto.

FIG. 10 is a diagram illustrating the bi-mode plasma ion energy distribution obtained for different RF bias power levels in a conventional process chamber.

FIG. 11 is a diagram illustrating single mode plasma ion energy distributions obtained for different RF bias power levels in the process chamber of the present invention.

FIG. 12 shows a normalized forward voltage transmission coefficient from a transmitting RF inductor antenna to a receiving RF inductor antenna with a silicon semiconductor slab placed between the transmitting and receiving antennas. FIG.

FIG. 13 is a diagram showing ion saturation current as a function of position on the substrate surface and the power level of the current applied to the inductor antenna.

FIG. 14 illustrates ion saturation current as a function of inductor antenna power level and RF bias power level.

FIG. 15 is a diagram showing the influence of the inductor antenna power level and the RF bias power level on the DC bias potential applied to the substrate.

16A is a diagram showing the effect of magnetic plasma confinement and RF bias power on O ₂ + plasma ion density and ion energy in the plasma zone, and FIG. When measuring by moving the Langmuir probe in the radial direction of the chamber,
FIG. 3 is a diagram illustrating the effect of magnetic plasma confinement on ion saturation current and plasma floating potential.

[Explanation of symbols]

Reference Signs List 30 ... substrate, 34 ... quartz insulator shield, 50 ... plasma processing apparatus, 55 ... process chamber, 60 ... high directivity electric field component, 65 ... plasma zone, 70 ... side wall, 75 ...
Bottom wall, 85: process gas source, 90: gas flow control system, 100: silicon ring, 105: annular passage,
110: gas injection hole, 115: tungsten halogen lamp, 120: exhaust system, 125: exhaust pump, 1
30 ... Throttle valve, 132 ... Pump discharge channel, 135 ... Inductor antenna, 135a ... Internal coil,
135b: external coil, 140: receiving surface (semiconductor ceiling), 145: primary bias electrode, 155: dielectric member, 165: power electrode, 170: second bias electrode
175: conductive surface, 180: voltage source, 185: heat transfer plate, 200: coating layer, 205: dielectric support layer, 208:
Perimeter, 210 ... side layer, 215 ... helium gas feedthrough hole, 235 ... fixed impedance matching network, 250, 255,260 ... electrode output, 256 ... R
F voltage module, 265 RF voltage oscillator module, 270 harmonic filter, 275 DC chuck voltage module, 280 local magnetic field, 285 permanent magnet, 292 magnetic field generator, 295, 300, 305,
310 ... electromagnetic coil, 312 ... computer, 315,
320, 325, 330 ... line, 335, 340, 3
45, 350 ... power supply system, 355, 360, 36
5,370 ... conductor.

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Ananda H. Kumar United States, California, Milpitas, Norview Drive 1296 (72) Inventor Arnold Korodenko United States of America, California, San Francisco, Eucalyptus Drive 1747 (72) Inventor Dennis S.S. Grimado United States, Michigan, Ann Arbor, Liberty Point 511 (72) Inventor Jonathan Dee. Morne United States of America, California, Saratoga, Paseo Presada 13179 (72) Inventor Michael G. Chaffin United States, California, San Jose, Manzanita Drive 4120 (72) Inventor Kenneth S.S. Collins United States, California, San Jose, Knights Haven Way 165

Claims (20)

[Claims] 1. A process chamber for processing a semiconductor substrate in a plasma, comprising: (a) a process gas distributor for distributing a process gas capable of forming a plasma to a plasma zone of the chamber;
(B) having a conductive surface exposed to the plasma zone,
A primary bias electrode on the ceiling of the chamber; (c)
A unitary monolithic dielectric member with a power electrode embedded therein, the dielectric member positioned below the primary bias electrode and having a receiving surface for receiving a substrate;
(D) a secondary bias electrode below the dielectric member;
(E) an electrode voltage source for providing a highly directional plasma to the plasma zone of the chamber while maintaining the power electrode, the primary bias electrode, and the secondary bias electrode at different potentials. Process chamber. 2. The method of claim 1, wherein the electrode voltage source causes the power electrode to be at least about one potential below the potential of the primary and secondary bias electrodes.
The process chamber according to claim 1, wherein the potential is maintained at 000 volts higher. 3. The single monolithic dielectric member only energizes the plasma in the chamber by permitting capacitive coupling of the RF voltage applied to the power electrode to the primary bias electrode via a coating layer. RF applied to the power electrode, to the second bias electrode through a coating layer on the power electrode, and a support layer, having low electric field absorption
The process chamber of claim 1, comprising a support layer under the power electrode having a high electrical absorption to substantially reduce capacitive coupling of voltage. 4. The process chamber of claim 3, wherein said coating layer has an RF reactance of about 1Ω to about 500Ω, and said support layer has an RF reactance of about 100Ω to about 10,000Ω. 5. The power electrode includes a peripheral portion facing an electrical ground side wall of the chamber, and the dielectric member includes a non-coupling side layer covering the peripheral portion of the power electrode. 4. The process chamber of claim 3, wherein the uncoupled side layer has a high electrical absorption to reduce capacitive coupling of the power electrode to a sidewall of the process chamber. 6. The process chamber of claim 1, wherein said secondary bias electrode comprises a conductive element having a surface directly exposed to said plasma and having no surrounding insulator shield. 7. The simple monolithic dielectric member includes aluminum oxide, aluminum nitride, boron carbide, boron nitride, diamond, silicon oxide, silicon carbide, silicon nitride, titanium oxide, titanium carbide, yttrium oxide,
The process chamber of claim 1, comprising a monolithic ceramic made from one or more of zirconium oxide. 8. An inductor, wherein an inductor antenna abuts the ceiling of the chamber, and wherein the ceiling passes the inductor to the plasma zone of the chamber through the ceiling to generate plasma in the plasma zone. The process chamber of claim 1, comprising a semiconductor material having a low electric field sensitivity to permit inductive coupling of an induced electric field generated by the antenna. 9. A method for processing a substrate using the process chamber of claim 8, wherein: (a) placing the substrate on the receiving surface of the single dielectric member; and (b) applying a current to the inductive antenna. Generating an induced electric field in the chamber through (c) maintaining the power electrode, the primary bias electrode, and the secondary bias electrode at different potentials to provide a high density, highly directional plasma in the chamber. A method comprising: 10. The method according to claim 1, further comprising: applying RF and DC voltages to the power electrode embedded in the single monolithic dielectric member; and electrically grounding the primary and secondary bias electrodes. Claim 9
The method described in. 11. A process chamber for processing a semiconductor substrate in a plasma, comprising: (a) a process gas distributor for distributing a process gas to a plasma zone of the chamber; and (b) an RF induced electric field. Adjacent to a semiconductor ceiling having an electrical sensitivity low enough to transmit and a conductive surface exposed to said plasma zone;
An inductor antenna for forming a plasma from the process gas in the plasma zone, and (c) a first dielectric member having a receiving surface for receiving a substrate on an upper surface, each having an electrode therein. A plurality of dielectric members below the semiconductor ceiling; and (d) an electrode voltage source for energizing the plasma in the chamber while maintaining the semiconductor ceiling and the electrodes at different potentials. A chamber. 12. The electrode voltage source capacitively couples the electrodes by maintaining the ceiling and one or more of the electrodes at different potentials by at least about 1000 volts to form the plasma in the chamber. The process chamber according to claim 11, wherein the pressure is applied. 13. A process chamber for processing a substrate in a plasma, comprising: (a) a gas distribution system for distributing a process gas to a plasma zone of the chamber; and (b) a conductive film exposed to the plasma. A primary bias electrode on the ceiling of the chamber having a surface;
(C) A monolithic monolithic dielectric member with a power electrode embedded therein, comprising a bonding layer forming a receiving surface for receiving a substrate, and a non-bonding layer surrounding the other surface of the electrode. The coupling layer is configured such that the RF and DC voltages applied to the electrodes are capacitively coupled via the coupling layer to energize the plasma, respectively, and merely hold the substrate electrostatically. A dielectric member having low electric absorption, wherein the non-coupling layer has high electric absorption so as to reduce capacitive coupling therethrough, and (d) the primary bias electrode and the power electrode are interconnected. An electrode voltage source for maintaining different potentials, and (e) a multi-directivity adjacent to the chamber for generating a multi-directional magnetic field having a time-varying angular orientation and magnitude in the plasma zone. Magnetic field generator Process chamber, characterized in that it comprises a. 14. The process chamber according to claim 13, wherein the multidirectional magnetic field has a magnetic field moving in a circular direction. 15. The multi-directional magnetic field generator comprises: (1) a plurality of electromagnets positioned adjacent to a side wall of the chamber; and (2) changing a current applied to the electromagnets to change the plasma zone. 14. The process chamber according to claim 13, further comprising an electromagnet power supply for generating the multidirectional magnetic field. 16. The electromagnet includes a plurality of pairs of conductive coils wherein each pair of coils generates a magnetic field substantially parallel to a plane of the substrate, and wherein the power supply is configured to electrically generate a magnetic field. In order to selectively move the magnetic field by independently changing the angular orientation and magnitude of the magnetic field generated by each coil pair, the pair coils are energized in a selected order. The process chamber according to claim 15. 17. The multidirectional magnetic field generator includes: (1) a plurality of movable permanent magnets positioned adjacent to the chamber; and (2) moving the permanent magnets to provide the multidirectional magnetic field to the plasma zone. 14. The process chamber according to claim 13, further comprising: means for generating a magnetic field. 18. A process chamber for processing a semiconductor substrate in a plasma, comprising: (a) a process gas distributor for distributing a process gas capable of forming a plasma to a plasma zone of the chamber;
(B) a monolithic monolithic having a primary bias electrode having a conductive surface exposed to the plasma zone; and (c) having a power electrode embedded therein and having a receiving surface for receiving a substrate below the primary bias electrode. A dielectric member; and (d) an electrode voltage source for providing the highly directional plasma to the plasma zone of the chamber while maintaining the power electrode and the primary bias electrode at mutually different potentials, A process chamber, wherein there is no insulator shield surrounding the single monolithic dielectric member. 19. The electrode voltage source may be configured to cause the power electrode to be at least about 100 potentials higher than the potential of the primary bias electrode.
19. The process chamber of claim 18, wherein the potential is maintained at 0 volts. 20. The power electrode includes a peripheral portion facing an electrical ground side wall of the chamber, and the dielectric member includes a non-coupling side layer that covers the peripheral portion of the power electrode. 19. The process chamber of claim 18, wherein the uncoupling side layer has a high electrical absorption to reduce capacitive coupling of the power electrode to a sidewall of the process chamber.