HK1079898B - Method of constructing a plasma reactor/generator - Google Patents

Description

Method of constructing a plasma reactor/generator

Technical Field

The present invention relates to gas ionization apparatus and methods, and more particularly to capacitively coupled gas plasma reactors and methods. The invention is based on divisional application of 'small-sized capacitive coupling plasma reactor/generator and method' filed on 3, 23 of 2001, international application number is PCT/US01/09497, national application number is 01811073.8.

Background

Capacitively coupled plasma reactors are generally constructed from a pair of parallel plate electrodes facing each other, positioned in parallel, and placed within a vacuum chamber. An external electric field (dc or ac) is applied to the electrodes of opposite polarity. Under low pressure conditions, and with the electrodes properly positioned, the initial ionization creates a stable plasma between the two electrodes, which then creates a glow discharge in the gas stream. Pairs of alternately poled parallel plates may be placed apart and (or) stacked together to form multiple zones where plasma discharge may occur. Such capacitively coupled plasma reactors have been used in a very wide range of industries, such as substrate etching, substrate cleaning, substrate thin film deposition, gas cleaning, ion beam sources, and various chemical reactions.

As indicated by the name "capacitively coupled plasma", the electrodes form a capacitor, typically of the parallel plate type. The most basic type is two plates of opposite electrical polarity, often referred to as "planar diodes". Its electrodes can be placed in a variety of geometries, including shapes with curved surfaces, such as concentric parallel cylinders or concentric spheres with parallel tangents. Typically, the electrode surfaces of alternating polarity are equally spaced throughout the structure, thereby maintaining the parallel plate relationship. In such a configuration, geometric regularity and symmetry between electrode representations is highly advantageous for producing a uniform electric field, and thus a more uniform ionosphere. Rugged flat-plate electrode pairs are often used to enhance or attenuate the plasma layer concentration in a particular area to suit particular applications, such as: focused sputtering, focused etching, or providing a focused ion source. Several capacitively coupled parallel plate electrode designs of different geometries are described in the prior art in U.S. patent number 4735633, entitled Method and System for Vapor Extraction FromGases, assigned to the assignee of the present invention and patented to the present inventors. The electrode structure taught in the' 633 patent provides a large surface area to volume ratio for a compact plasma generator. Reactors employing the electrode configuration described in the' 633 patent have been successfully employed in the industry where reaction efficiencies can be greater than 99%.

In addition to electrode distance, another key parameter for generating and maintaining a plasma in a capacitively coupled plasma reactor is operating pressure. At lower pressures, stable glow discharge plasma can be more efficiently and more easily maintained. This is because the generation and maintenance of the plasma is determined by the ionization of the gas molecules in the reactor to generate enough secondary electrons to participate in the cascade impact ionization process to compensate and balance the electrons (and ions) disappearing at the electrode surface. The mean free path, i.e., the mean path traveled in the reactor before a primary electron collides with a molecule to generate a secondary electron. The mean free path is determined by the operating pressure. Generally, the higher the pressure, the smaller the value of the mean free path. Under the action of the potential energy of the electrode electric field, the primary electrons are accelerated within a certain distance to obtain ionization potential energy required for promoting the ionization process. And the size of the mean free path limits this distance. Thus, the smaller the value of the mean free path, the smaller the ionization potential that an electron acquires before colliding with a gas molecule, and thus the less secondary ionization that occurs, given the operating potential energy.

The electrode distance determines the number of mean free path ionizing collisions an electron encounters before reaching and disappearing at the electrode surface for a given operating pressure. When the electron distance is short, glow discharge is not generated and maintained. This space is known as dark space. Once the plasma is ignited in the reactor, the plasma itself becomes a conductive sheet corresponding to the electrode. There is always an empty space between the plasma and the electrodes where no glow discharge ionization occurs. The ions and electrons are accelerated only in this gap without further glow ionization, and this space is known as the "dark space shield". The thickness of the dark space shield is also dependent on the pressure.

Therefore, the conditions under which the gas molecules decompose, generating and maintaining a stable glow discharge plasma are determined by the relationship between the applied electric field, the breakdown voltage, the electrode distance, and the operating pressure. Paschen found experimentally: the breakdown voltage (V) varies with the product of the pressure P (unit: Torr) and the electrode distance d (unit: cm). Paschen finds this relationship, which is known as the law of glow discharge, and is reflected by the "Paschen curve" shown in FIG. 1. FIG. 1 shows Paschen curves 10 for several different gases. The electrode design of a capacitively coupled, parallel plate plasma reactor must meet the requirements of the paschen curve.

The paschen curves 10 of fig. 1 show: when pd is approximately 1Torr-cm, i.e., at point 15, there is a minimum breakdown voltage (V) for each gas. Therefore, in practical conditions, if the distance between the parallel plate electrodes is fixed at 1cm, the applied voltage of the electrodes required to initiate ionization and decompose the gas in vacuum is minimized at a pressure of about 1 Torr. It can be seen from the Pascal curve 10 that for a given electrode distance d, the minimum applied voltage required to meet the 1Torr-cm breakdown parameter also increases slowly as the pressure P increases. However, as the pressure subsides, the minimum required voltage increases dramatically (linearly proportional to Pd). For example, in neon lamps, the maximum operating pressure for neon for a reactor with a fixed electrode spacing of about 1cm can be up to about 300Torr assuming a maximum voltage of 1000V that can be provided by a given power supply. However, at pressures below 0.1Torr, if the electrode spacing is not multiplied such that the Pd value at the breakdown voltage 15 of the Paschen curve 10 is below the maximum power limit of 1000V, the same 1000V power supply will not be sufficient to generate and sustain a plasma in neon.

Thus, in practice, the relationships shown by the Paschen curves 10 determine the minimum electrode spacing and, hence, the minimum size of the reactor for a given power supply rating and operating pressure range. In most applications, it is desirable to use a low voltage power supply, whether direct current or alternating current, rather than a high voltage power supply because of the inherent low power consumption characteristics of low voltage power supplies. It is also desirable that the distance between the electrodes is small, which allows the reactor to be smaller. However, when the operating voltage is below 0.5Torr, which is required in many specific applications such as semiconductor processing, the electrode distance must be increased to several centimeters or more, which makes the reactor size large, or else requires the use of a very expensive high voltage power supply. Although additional magnetic field sources can be used to confine the plasma in low pressure applications, this approach is very expensive, complicates interfering capacitive coupling, dissipates plasma energy, and introduces more edge effects.

The aforementioned' 633 patent teaches maximizing the surface area of the electrodes in a reactor of a given size in a particular manner to increase the efficiency of the reaction and thereby maximize the efficiency of the reactor. Although the' 633 patent is primarily directed to the decomposition and treatment of hazardous exhaust gases from semiconductor manufacturing, the plasma treatment process described therein also provides an effective method of treating materials, such as by sputtering, etching, deposition, surface treatment, and the like. It also provides an efficient gas chemical reaction process, such as chemical synthesis, polymer formation, chemical decomposition, etc., to produce desired by-products. Advantages of this plasma treatment process over other chemical processes include: actually reducing energy loss, improving reaction efficiency at lower temperature, and the like. The plasma reactor taught in the' 633 patent has entered commercial use under the trademark "PebaxAnd sold by the agents of the present invention. As taught in the' 633 patent,the ratio of electrode surface area to plasma volume of the reactor is large. It takes advantage of this and the longer gas flow path to enhance the chemical reaction at the electrode surface. This approach maximizes reaction speed and reaction efficiency compared to gas phase reactions within the gas stream itself.

Thus, as described in the' 633 patent, for a pair of parallel plate electrodes, the area of one surface of each electrode is A, and the total surface area of the opposing surfaces of the pair of electrodes is 2A. The volume between the two surfaces is 2Ad when the distance between the electrodes is fixed. In the case of low voltages, the electrode distance d must be increased for the reasons described above. The amount of plasma also increases with increasing electrode distance d, and thus the surface area to volume ratio also decreases inversely with increasing distance d. Thus, if the operating pressure is reduced and the surface area of the electrode cannot be managed to increase, some or all of the benefits of the surface reaction are lost. Of course, one way to increase the surface area of the electrodes is to increase the size of the reactor and thus the electrodes. However, it is often undesirable or simply not feasible to do so for various reasons, such as cost, application or design limitations. Therefore, for low voltage applications, it is necessary to find a new way to increase the surface area of the electrodes in the reactor without increasing the size of the reactor.

The present invention addresses this problem by providing a novel and unique electrode design. The main goal of the new electrode design is to increase the surface area of the electrode without increasing the reactor volume size. In a capacitively coupled parallel plate plasma reactor, the novel electrode is capable of providing efficient electrode surface reactions over a wide range of operating parameters without significantly increasing the size of the reactor, using the method described in the' 633 patent. Likewise, the new electrode design also greatly increases the range of applications for such reactors and processes.

Disclosure of Invention

As illustrated in the' 633 patent, the conventional idea is to extend the opposing faces of a parallel plate electrode pair laterally without extending into the open area between the opposing or adjacent faces of the electrodes. It is undesirable to have any surface protruding into the space between the electrodes, as that would reduce the distance between the electrodes at these points. There is a great concern that a short circuit path may be generated, resulting in arcing between the electrodes. Therefore, it is believed that the electrode design should have their opposing surfaces as flat and curved as smooth as possible to avoid this problem. Moreover, for the reasons discussed above for satisfying the Paschen curves, there is a concern that shortening the distance between the electrodes may affect the generation, maintenance, and quality of the glow discharge plasma. This has become a common wisdom for electrode design and construction.

The present invention is in contradiction to this common general knowledge of electrode design and construction. In the present invention, a pair of alternating polarity electrodes is formed of a plurality of "L" and "7" shaped tabs to form a so-called "L7" shaped electrode structure. The tabs extend to an open area intermediate adjacent opposing electrodes and are positioned in a staggered pattern. The electrodes with interleaved tabs form an approximately square "L7" channel with one or more gaps in one or more diagonal corners. Another more expansive embodiment employs a grid-like design in which there are many pairs of electrodes of opposite polarity, each having wings, which are stacked together so that the wings are interleaved between the electrodes. The "L7" shape with interleaved fins maintains the electrode distance d between the parallel opposing faces of the electrode pair, increasing the electrode surface area by a factor of four or more for a given volume. A variety of shapes may be used including a continuous curved surface or "W" shaped surface, which may provide more surface per unit volume.

A plasma reactor/generator apparatus and method employing the "L7" shaped design of the present invention includes a reactor body with an open interior space. The reactor body includes an inlet port and an outlet port. The electrode assembly is preferably provided as an integral unit which may be inserted into the interior of the reactor or removed therefrom as a unit. The electrode assembly is typically insulated from the reactor body. In an open system, the electrode assembly is housed inside the reactor body, defining a plurality of sub-paths between the gas inlet and the gas outlet. In a stationary or closed system, the electrode arrangement divides the gas volume into a number of cells, which may be in accordance with a certain desired pattern. A power supply coupled to the electrode assembly generates a voltage differential across adjacent pairs of oppositely polarized electrodes sufficient to ignite and sustain a plasma in a selected gas to be treated in the reactor. In another aspect of the invention, a gas stream is introduced at a selected pressure, flow rate and temperature into the gas inlet and passes through the sub-gas flow paths associated with the electrode surface to the gas outlet. A plasma can be generated having a plurality of separate regions, preferably partially interconnected. So that the gas on the surface of the electrode can carry out efficient and complete chemical reaction. The reactor utilizes or performs a selected process on such gas. In another aspect of the invention, the plasma generator is a closed system containing a gas. Plasma is formed in such a gas for light emission or the like. In this respect, the invention is a plasma generator.

To achieve the objects and effects of the present invention, the present invention provides a method of constructing a plasma reactor/generator, comprising:

providing a chamber defining an interior space;

providing a plurality of anode structures and cathode structures disposed within the chamber; each of the anode and cathode structures having a plurality of anode and cathode surfaces, respectively, wherein adjacent anode and cathode surfaces are opposed to each other and separated from each other to define a gas space therebetween;

providing a common electrical connection to the anode surfaces and a common electrical connection to the cathode surfaces;

providing a plurality of electrically conductive fin elements at the opposing surfaces of the anode structure and the cathode structure, the fin elements extending into the space between the opposing surfaces thereby dividing the space into a plurality of cells; and

the cells are arranged and formed to form a shape selected to control the characteristics of the plasma in the space.

Drawings

FIG. 1 is a graphical illustration of Paschen curves for some typical gases.

FIG. 2 illustrates a portion of a conventional capacitively coupled parallel plate electrode pair in a conventional capacitively coupled parallel plate plasma reactor.

FIG. 3 illustrates a portion of a capacitively coupled parallel plate electrode pair implementing a preferred "L7" structure according to the present invention.

FIG. 4 is a rear view of the "L7" electrode pair shown in FIG. 3.

FIG. 5 is a side cross-sectional view of a stacked grid of "L7" electrode pairs whose interleaved fins comprise a preferred embodiment of the present invention.

FIG. 6 is a partial plan view of a preferred embodiment of a first electrode forming an "L7" electrode pair with the second electrode of FIG. 7 for use in the stacked grid of "L7" electrode pairs shown in FIG. 5.

FIG. 7 is a partial plan view of a preferred embodiment of a second electrode forming an "L7" electrode pair with the first electrode of FIG. 6, for use in the stacked grid of "L7" electrode pairs shown in FIG. 5.

Fig. 8 is a cross-sectional plan view of a preferred "L7" electrode pair comprising the electrodes of fig. 6 and 7.

Fig. 9 is a side elevational view of a preferred electrode comprising the stacked grid of electrode pairs of "L7" of fig. 6-8.

Fig. 10 is another elevational view of the different sides of the preferred electrode of fig. 9.

FIG. 11 is a side elevational view of a preferred embodiment of a capacitively coupled parallel plate gas plasma reactor to which the present invention is applied.

Detailed Description

A preferred embodiment of the invention will now be described with the aid of the accompanying drawings.

Fig. 2 illustrates a conventional capacitively coupled parallel plate electrode pair 20 of the type employed in almost all conventional plasma reactor designs today. Electrode pair 20 includes a first plate electrode 22 and a second plate electrode 24. First and second plate electrodes 22 and 24 have first and second surfaces, respectively, each having an area A. The distance between the two opposing faces of the first and second electrodes 22 and 24 is a fixed value d. The plate electrodes comprise the two plates of a parallel plate capacitor, which are connected to opposite poles of a power supply 26, which may be either ac or dc. Thus, at any given time, the polarity of electrodes 22 and 24 is reversed. A voltage (V) is present between them, which is capable of igniting and maintaining a glow discharge plasma in the gas flow between the two electrodes. Simple calculations show that the total electrode surface area adjacent to the open space between the two electrodes is 2A and the volume of the space between the two electrodes is Ad. Thus, the ratio of electrode surface area to volume is 2/dcm-1, the electrode distance is typically about 1cm and the surface area per volume is about 2.

The basis of the surface reaction principle taught in the prior art' 633 patent is: stable glow discharge can be easily maintained in a low pressure environment. Maximizing the electrode surface area per volume of plasma in the reactor allows the electrode surface to react most strongly. The large surface area provides large reaction sites for the gases to react at the surface. The adsorbed gas molecules and adsorbed gas molecules can easily find their location and total effective range on the surface to ensure that when ions or electrons strike the surface, the probability of a chemical reaction is very high. As shown by the paschen curve in fig. 1, the initiation and maintenance of the glow discharge is dependent on the operating pressure and electrode spacing. As a general rule of thumb, for a parallel plate electrode, when the product of the distance d (cm) between the anode and cathode and the operating voltage P (Torr), i.e., Pd Torr-cm, is about 1Torr-cm, the minimum breakdown voltage or plasma ignition voltage occurs for most gases when the electrode voltage is between about 250 and 350V.

For example, when the operating pressure is between 500 and 1000mTorr, the optimal distance between parallel plate electrodes is about 1 cm. If the operating pressure is greater than this range, the range over which the optimum breakdown or plasma ignition voltage is maintained may be slightly narrowed. However, if the voltage is below this range, the distance d must be increased sharply to maintain the optimum breakdown voltage. In other words, if it is desired to have the breakdown or excitation voltage as close to a minimum as possible when the required operating pressure varies, the distance d must be varied to keep the product pd as close to unity 1 as possible. Otherwise, when the pressure is low, the power supply must provide a high voltage well above 1000V.

Therefore, increasing the electrode distance d is a typical method used industrially in recent years. For example, Reactive Ion Etching (RIE) processes for plasma etching substrates such as semiconductors employ capacitively coupled electrodes operating at pressures in the range of 10 to 100 mTorr. In commercial reactors, the electrode distance is between 5cm and 15 cm. This maintains a relatively low breakdown voltage and minimizes self-induced bias. This deviation can cause undesirable radiation damage to the substrate being etched by the high energy electron and ion bombardment.

The difficulty in initiating gas ionization, plasma ignition and maintenance is overcome by the use of a larger electrode spacing d, according to paschen relationships. But a larger electrode distance has the undesirable effect of increasing the plasma volume, which can require more reactor space and larger electrode surface area, which can greatly increase cost. In order to construct an electrode having the same surface area and operating at a low voltage, the electrode distance and volume must be multiplied. However, if the reactor is designed with large surface areas for its electrodes, the high pressure range for practical applications may be excluded because the product Pd will increase to the high pressure end of the Paschen curve.

Almost all practical parallel plate plasma reactor designs today are based on the aforementioned principles. For example, if the electrode distance d is about 1cm, it is relatively easy to generate and maintain a glow discharge plasma at a pressure of about 1Torr for almost all gases available. Similarly, if the distance is 2cm, the optimum operating pressure will be 0.5 Torr. It can also be seen from the paschen curve that for a fixed electrode distance d, as the operating pressure increases, it will slowly become difficult to generate and sustain a plasma, and the minimum breakdown voltage will slowly increase with increasing voltage. Conversely, when the operating pressure is reduced, the minimum breakdown voltage for striking the plasma will increase dramatically and it becomes very difficult to generate and sustain a plasma. The lowest breakdown voltage increases with increasing operating pressure, which is physically interpreted as: the higher the operating pressure, the shorter the gas molecules, atoms, ionized ions between the electrodes, and the mean free path between the electrodes. Thus, there are multiple collisions between excited molecules, atoms, ions and electrons before reaching the electrodes. In each collision, the excited particle loses energy and jumps from a high energy state to a low energy state. Thus, after a certain time, the particles with sufficient energy to excite secondary ionization become less and less, and the generation of secondary ions and electrons becomes localized and more difficult. In this case, a higher external operating voltage is required to generate and sustain the ionization breakdown process to sustain the plasma.

The longer mean free path reduces the number of collisions of excited particles between the electrodes at lower operating voltages. In this case, the primary electrons absorb energy from an externally applied voltage (direct current or alternating current), are accelerated between the electrodes, and have a greater chance of colliding with the electrodes before colliding with gas molecules to ionize neutral particles and generate more secondary electrons. The fast speed of the primary electron extinction coupled with the small number of secondary electrons generated requires the external power supply to provide a higher voltage to generate a stronger electric field and high energy electrons to ensure the ionization process is initiated and the plasma is maintained. But higher voltages will accelerate the electrons and thus shorten their time before they disappear to the electrode, thereby reducing the number of collisions that generate secondary electrons. As a result, the minimum breakdown voltage required to ignite and sustain a plasma increases dramatically as the operating pressure decreases.

In many plasma reactor applications today, such as semiconductor manufacturing applications, the operating pressure of the plasma reactor is required to be below 100 mTorr. To operate at such low pressures, the distance of the electrodes in the reactor must be increased in order to generate and maintain a plasma at reasonable voltages below about 1000V. Even so, the cost of a high voltage plasma generator is much higher than the cost of a low voltage generator. Furthermore, in order to consume the same power P as IV, a high operating voltage means that the current consumed in the plasma is low. Since the chemical reactions involved require electron exchange, low current means that the chemical reaction rate is reduced. Therefore, to increase the efficiency of the plasma reactor, it is better to use a low voltage, high current plasma.

At the same time, it is also desirable to retain the successful features of the reactor described in the' 633 patent, such as parallel plate design, optimal electrode spacing, high surface area to volume ratio, long flow paths, and short tracks, among others.

Fig. 3 and 4 illustrate the construction of a substantially "L7" parallel plate electrode. Which comprises the basic preferred embodiment of the invention. As will be explained below, this "L7" electrode arrangement overcomes the disadvantages of the prior art described above while retaining the successful features of the prior art design described in the' 633 patent.

The "L7" design includes a first parallel plate electrode 32 and a second parallel plate electrode 34, which are oppositely disposed in a conventional manner. Electrodes 32 and 34 are connected to opposite electrodes of a power supply 36. The power supply can be alternating current or direct current and is suitable for the capacitive coupling parallel plate plasma reactor. Thus, the polarity of electrodes 32 and 34 is reversed. Each electrode has a first surface 33a, 35a and a second surface 33b, 35 b. Although the configuration and angular relationship are foreseeable, in the preferred embodiment, the first and second surfaces are integral and at right angles. The distance between surfaces 33a and 35a is fixed at d with their opposing faces parallel to each other. Similarly, the distance between surfaces 33b and 35b is fixed at d, and their opposing surfaces are also parallel to each other. Surfaces 33b and 35b extend into the open space between parallel surfaces 33a and 35a such that the end of surface 33b reaches surface 35a and the end of surface 35b reaches surface 33 a. Thus, the "L7" electrode pair arrangement forms an approximately square shape and divides a channel into individual cells. Preferably, the channels or cells are not completely enclosed. A small gap is provided at one or more of the pair of corners (diagonalhorns) to separate the electrodes by a distance d' between the end of surface 33b and surface 35a and the end of surface 35b and surface 33 a. The surfaces 33a, 33b, 35a and 35b each have an area of about a.

In this manner, the preferred "L7" electrode arrangement greatly increases the ratio of surface area to plasma volume while maintaining the electrode distance d. For example, the total area of the four surfaces of a square channel or cell is 4A. The total volume is kept at Ad, the same as the conventional parallel plate electrode arrangement in fig. 2. Thus, the "L7" electrode pair is disposed at a surface area to volume ratio of about 4/dcm-1, which is twice that of a conventional flat plate electrode pair combination.

This "L7" electrode pair arrangement is contrary to the conventional concept of parallel plate electrode pair design. Conventional thinking has opposed the surfaces 33b and 35b extending between the parallel surfaces 33a and 35a, and certainly not so close to these surfaces. This conventional thought is because when high voltages are used in many applications, arcing occurs between adjacent electrodes. However, the inventor of the "L7" electrode design determined, by a renewed study and an in depth understanding of the meaning of the Paschen curves in FIG. 1, that a low pressure range may be applied for the "L7" electrode. The gap d 'between the opposite polarity electrodes 32 and 34 at the corners can be made small enough so that the product Pd' can only generate and sustain a plasma in the gap when the power supply is slightly above the breakdown voltage. Thus, under these conditions, no plasma is present in the voids. In addition, arcing and short circuits are avoided in these cases, since the gap distance d' is too small and the acceleration distance of the electrons is not sufficient for it to cause ionization. Only after a continuous ionization, creating a conductive path between the electrodes (like a lightning discharge path), an arc or short circuit phenomenon may occur. These conditions are not actually met in the low voltage operating range where the "L7" electrode is primarily used. For example, when the gap distance d 'is 0.5cm and the electrode distance d' is 2, and the power supply voltage is less than 1000V, the operating pressure of the electrode may be less than 0.1 Torr. When the operating pressure is raised to a higher value, such as 2Torr, a suitable plasma can still be generated and maintained in the gap region, so long as the gap d' remains sufficiently small that arcing or shorting does not occur in the gap.

Another reason for the inventors to be confident that arcing and shorting do not occur within the gap distance d' is: a key requirement of arc discharge is that the discharge should be focused at one point, so that a high current conducting path is created between the discharge point and the opposing electrode. In the "L7" electrode pair design, the electrodes form a conductive line, and the discharge between them will be distributed over the entire conductive line (rather than at one point). Thus, under expected operating conditions, a high voltage sufficient to excite field radiation is not established between any given point on the electrodes, such that the potential of the electrons is higher than the breakdown potential, sufficient to ionize the entire conductive path between the electrodes. In other words, because the plasma generated between "L7" type electrodes is well distributed over the surface of the electrodes, at any given point, including in the corners very close to the electrodes, there is not a strong enough potential generated so that a complete conductive path is not ionized between the electrodes, resulting in arcing. It is therefore not necessary to take into account arcing and short-circuiting problems as in conventional thinking.

Another advantage of the "L7" electrode design is that it allows the distance between the preferred vertical surfaces of the electrode pair to be varied. The effective distance may vary from the gap distance d' to the distance d between the opposing parallel surfaces of the electrodes (i.e., surfaces 33a and 35a, or 33b and 35 b). In fact, the distance between the electrodes may be greater on a diagonal line from the corner of the closed end of one electrode to the corner of the closed end of the other electrode, i.e., on a line from the intersection of surface 33a and surface 33b to the intersection of surface 35a and surface 35 b. This new design thus allows the distance between the electrodes to be varied so that optimum operation can be achieved at various pressures. This electrode design can simply and efficiently ignite and sustain a plasma over a wide range of operating conditions. Moreover, as will be seen in greater detail below, this design feature can be further extended. For example, with a flared electrode design having an open end (with a larger cross-section than the closed end), a greater range of variation in distance can be achieved, thus allowing the plasma to be selected for optimal distance under operating conditions, thereby allowing the plasma to be readily ignited.

The variable electrode distance of the "L7" design, as well as its expanded form, is a significant feature. Once the plasma is excited, it becomes a conductive layer, which itself acts as an electrode with many conduction electrons. The plasma itself is therefore an additional source of electrons that can be used to replenish the electrons lost to the electrode. Thus, an easily ignited plasma is easily sustained, which means that an "L7" type electrode can easily and efficiently ignite and sustain a plasma over a wide range of operating conditions.

Fig. 5 illustrates the extension of the basic "L7" electrode pair design to a stacked electrode pair configuration. Thus, fig. 5 shows that the four electrodes 52, 54, 56, and 58 are stacked in a vertical configuration. Each electrode has two opposing surfaces, i.e., electrode 52 has opposing surfaces 52a and 52b, electrode 54 has opposing surfaces 54a and 54b, electrode 56 has opposing surfaces 56a and 56b, and electrode 58 has opposing surfaces 58a and 58 b. The electrodes are stacked such that their surfaces are parallel to each other, i.e., surfaces 52a, 52b, 54a, 54b, 56a, 56b, 58a and 58b are all parallel to each other. Electrodes 52, 56 are commonly connected to one pole of a suitable ac or dc power source 60, and electrodes 54 and 58 are connected to the other pole of power source 60. The polarities of adjacent electrodes in the stacked configuration alternate such that each adjacent pair of stacked electrodes forms an electrode pair of opposite polarity. I.e., electrodes 52 and 54 form one opposite polarity pair, electrodes 54 and 56 form another pair, and electrodes 56 and 58 form another pair. Notably, this configuration employs both opposing faces (e.g., 54a and 54b) of an electrode (e.g., electrode 54) which greatly increases the surface area of the electrode in the reactor for chemical reactions to occur. Consistent with the basic "L7" design concept, each electrode has a plurality of tabs extending between adjacent electrodes at right angles to the opposing surfaces. Thus, the tabs 64 extend outwardly from the surface 52b of the electrode 52 and the tabs 64 extend outwardly from the opposing surface 54a of the adjacent electrode 54, both of which extend into the open space between the adjacent electrodes and are in close proximity to the adjacent electrodes. As shown, it may be better for these tabs to be disposed in a staggered fashion over adjacent electrodes of an electrode pair for various reasons. One reason for this is that it helps to distribute the plasma over the adjacent electrodes of each electrode pair. This in turn helps to ensure that there is no point ionization source that causes arcing or shorting between the electrodes. As will be explained further below, it also helps to separate the plasma, making the plasma of better quality and thus improving the efficiency of the reaction. Also, it ensures a long and continuous gas flow path in the reactor, which also contributes to the increase of reaction efficiency.

This extension of the basic "L7" electrode design still increases the surface area to plasma volume ratio relative to prior parallel plate electrode structures and doubles it by stacking electrode pairs within the reactor. Assume that the distance between adjacent fins 64 of the same electrode (e.g., electrode 52) is d. The distance between the opposing surfaces of adjacent electrodes of one electrode pair (e.g., surface 52b of electrode 52 and surface 54a of electrode 54) is also d. Also assume that the surface area of each electrode between adjacent fins is a, and that the surface area of each fin is also approximately a. The total electrode area is 4A in each "cell" bounded by adjacent fins of one electrode (e.g., electrode 52) and the opposing parallel surfaces (e.g., surfaces 52b and 54A) of the adjacent electrode of each electrode pair. As with conventional parallel plate electrode designs, the electrode distance is still d, so that the ratio of electrode surface area to plasma volume per cell or channel is about 4/d cm-1 for the plasma.

The "L7" design concept can be further expanded to provide higher surface area to volume ratios. The approach is to further divide the approximately square "L7" channel into three-dimensional cells with a spacing approximately equal to d by adding additional fin elements. The result is a surface area to volume ratio of approximately 6/d cm-1 in each approximate spatial separation.

Although the stacked "L7" electrode arrangement is shown in cross-section in FIG. 5, it will be understood by those skilled in the art that electrodes 52-58 may be of various shapes. For example, as described below, each electrode may be circular. Similarly, although the tab 64 is also shown in cross-section, the tab may be of various shapes including flat, curved, "U" -shaped, "V" -shaped, "W" -shaped, and flared. Also, the electrodes need not be continuous surfaces, but may include one or more openings to facilitate gas flow. Similarly, the fins 64 need not be a continuous surface, but may include a plurality of openings to facilitate gas flow and plasma communication.

Furthermore, the size and geometry of the chamber may be varied to tune the plasma. For example, the plasma intensity within the chamber may be adjusted according to a desired pattern. Areas of plasma focus and dispersion can be created. Plasma grids and picture elements can also be generated. This adjustment may be made periodically or according to other desired patterns.

Similar to the basic "L7" design, the tabs 64 of adjacent electrodes (e.g., electrodes 52 and 54) are in close proximity to the opposing surfaces (e.g., surfaces 52b and 54a) of the electrode pair, but maintain a gap distance d'. For the same reasons discussed above for the basic "L7" electrode design, if the gap distance d' is small enough, arcing and shorting between the electrodes is not a concern.

FIGS. 6-8 are examples of one preferred configuration of the airfoil assembly, and reference may be made to the figures for illustrating the electrodes and electrode arrangements. Fig. 6 is a plan view of a portion of an electrode in an opposite polarity electrode pair, such as electrode 52 in fig. 5. The surface of the electrode 52 shown in fig. 6 is a surface 52 b. The preferred form of the electrode 52 is circular as shown in figure 8. The tab 64 is a flat surface extending outwardly at right angles to the surface 52 b. FIG. 7 is a top view of another preferred fin shape with the open end being "trumpet" shaped. The flared fin shown in fig. 7 extends outwardly from surface 54a of electrode 54 in fig. 5. As shown in fig. 8, the surface 54a of electrode 54 has a plurality of such flared tabs 64, such that the tabs 64 extending outwardly from the surface 52b of electrode 52 are inset on either side of the flared tabs and surround the entire surface of the adjacent electrodes 52 and 54. Fig. 8 illustrates, in a cross-sectional view, how the straight fins 64 and the flared fins 64 are inserted when the circular electrodes (e.g., electrodes 52 and 54) are adjacent. Fig. 8 also illustrates that the preferred form of at least one electrode in each electrode pair, in this case electrode 52, has a central opening 80 for the passage of gas flow. Further shown in fig. 8 is that the diameter of the other electrode 54 is slightly smaller than the diameter of electrode 52, allowing gas flow past the edge of the electrode to the next stacked electrode pair.

Inserting additional fin surface area between the electrodes increases the ratio of electrode surface area to plasma volume. In addition, the fins greatly increase the flow path of the gas flow by converting the planar wider path between the electrodes into a plurality of narrower paths. As mentioned previously, the channel may be further divided into approximately solid chambers by inserting additional fin elements within the "L7" shaped channel, if desired. This division of the broad path greatly increases the electrode surface area encountered by the gas stream as it flows in the reactor without increasing the volume or size of the reactor.

This can be seen from different sides in fig. 9 and 10. Fig. 9 and 10 are two side views of the stacked "L7" electrode type illustrated in fig. 5-8. As can be seen in fig. 9 and 10, the stacked electrodes comprise a series of interleaved stacked electrodes 92 and 94. As shown in fig. 5, electrodes 92 are commonly connected to one pole of a suitable power supply (not shown) and electrode 94 is connected to the other pole of the power supply, so that the polarities of adjacent stacked electrodes 92 and 94 are opposite. Each electrode 92 has a disk shape with an opening in the center to allow gas to flow from one layer of the stack to another. Each electrode 94 is also disk-shaped, but has no central opening. The preferred diameter of electrode 95 is slightly smaller than the diameter of electrode 92, allowing gas to flow from one layer of the stack to the next through the outer edges of electrode 94. It will be apparent that electrode 92 is electrically insulated from electrode 94 by means such as an insulating spacer. The electrodes 92 are provided with fins 64 which extend orthogonally outward from each face of each electrode 92, proximate to, but not intersecting, the adjacent electrode 94 of each electrode 92. Similarly, electrodes 94 are provided with "horn" shaped fins 64 which extend orthogonally outward from each face of each electrode 94, proximate to, but not intersecting, electrodes 92 adjacent to each electrode 94. Moreover, the straight fins 64 and the flared fins should preferably be staggered so that they are staggered in the space between adjacent electrodes 92 and 94. With this arrangement, the gas stream entering the central opening 96 of the first electrode 92 must first follow a tortuous path in the interleaved fins between the first electrode 92 and the first electrode 94 before passing over the outer edge of the first electrode 94 to the next layer in the stack. At the next level, the gas flows into a plurality of tortuous paths in the interleaved fins between adjacent second electrodes 92 and 94 to the central opening of second electrode 92. From there, into the next layer of the stack, and then repeat the same meandering path within each layer of the stack until the last layer is traversed.

Electrodes 92 and 94 may be composed of suitable conductive materials familiar to those skilled in the art that have been used in the past in plasma reactors. The electrodes 92 and 94 illustrated in fig. 9 and 10 are made of stainless steel, which is relatively inexpensive. When the power source is a radio frequency power source, an insulating material may also be sandwiched between the conductive cores.

Another advantage of the preferred stacked "L7" electrodes of the present invention is that the spacing between the electrodes is segmented, increasing the quality of the plasma and thus the reaction efficiency of the reactor. A conceptual breakthrough is that a conventional parallel plate electrode pair can be viewed as two long parallel wires. Typically, the plasma forms a "layer" in the central region of the space between the electrodes and has a significant "dead zone" near the electrodes. The division of the space between the electrodes by the fin assembly breaks up the parallel plate blind area. To this end, the fin elements dividing the space between the electrodes may be two opposite plates having the same potential as the cathode or anode electrons, the two L-shaped and two 7-shaped plates facing each other like the "L7" opposite plates. This combination of "L7" type plates allows the plasma in the center of the channel to see the entire space to the surface of the electrode, even if the plasma is generated and sustained in the center of the channel. This is because ions and electrons generated in the center of the channel radiate into the closed channel formed by the electrodes and the segmented fin elements of each "cell" or partition, thereby causing a chemical reaction. The reaction efficiency is thus greatly improved.

Furthermore, in an extended parallel plate electrode reactor, the plasma generated and sustained corresponds to a lateral layer between the opposing surfaces of the electrodes. This allows the plasma to undergo significant additional chemical reactions between the two electrodes and in the space where the electrodes are equal. Because such chemical reactions proceed in the gas phase, they readily form molecular groups and condensed particles known as "plasma dust". They will certainly condense in the gas stream and be present in the reactor together with the gas stream. This can create serious problems for certain devices, such as downstream pumps, especially when these "plasmas" are corrosive. By providing plasma "layers" as individual cells or segments, the present invention greatly enhances the control of the additional chemical reactions that generate plasma dust. In fact, the "L7" design allows the designer to control the length of the flow path and the number of partitions with relative ease, which allows for better control of the balance between the desired surface reaction and the gas phase reaction.

However, it may sometimes be desirable to have some communication between the plasma between adjacent cells or sections. For example, variations in the size of adjacent sections or other factors may cause the plasma formed in one cell or section to be weaker than adjacent cells or sections due to manufacturing errors or other reasons. Allowing some communication of plasma between adjacent cells or partitions allows the plasma of the weaker cells and partitions to be enhanced by the stronger plasma in the adjacent cells or partitions, thereby achieving a balancing effect. This communication may be accomplished partially or completely through the central gap shown in fig. 4 or the gap between the adjacent surfaces between the fins 64 and the electrodes 52-58 shown in fig. 5. If more communication is desired, a plurality of communication holes may be provided in the surface of other fins or all of the fins. The size of these holes will of course depend on the application, the size of the tabs and the electrodes themselves, and the desired operating parameters.

Another advantage of the preferred "L7" electrode structure of the present invention is in material strength considerations. The formation and maintenance of the plasma generates heat. Further, the larger the surface area of the electrode, the higher the operating voltage, and the more heat is generated. Therefore, the electrodes must be stressed with thermal deformation. Conventional parallel plate electrodes, which comprise relatively large continuous metal sheets, accumulate much thermal stress and are therefore prone to structural deformation. This deformation can change the electrode distance and thus the capacitance, electrical characteristics, and plasma properties. In some severe cases, structural deformation can lead to short circuits. These factors must be taken into account when selecting the electrode material. In contrast, in "L7" electrode structures, particularly in stacked "L7" electrode arrangements, which include multiple small surfaces and angularly interconnected planes, they improve the structural support and stability relative to large planar layers. In addition, since the plasma is divided into small portions, thermal stress accumulation to the structure is reduced. This increase in structural strength results in: the sheet metal from which the electrodes are fabricated can be thinner than that used in conventional parallel plate reactor designs under the same operating parameters and conditions. This in turn allows the space inside the reactor to be increased, allowing more surface area to be placed in a given small volume, resulting in better performance.

Fig. 10 is a side view of a capacitively coupled parallel plate electrode plasma reactor incorporating a preferred "L7" electrode configuration of the present invention. The reactor 110 has a chamber enclosing an inner space (not shown) inside which the electrodes shown in fig. 9 and 10 are installed. The reactor chamber can be opened or closed by conventional means. The chamber may also be provided with a cooling surface for air cooling, if desired. A conventional gas inlet 115 is provided to receive the gas stream to be treated. A gas outlet 120 is also provided for the flow of treated gas out of the reactor. External electrodes (not shown) are also provided for connecting the electrodes of a suitable power supply to the internal electrodes of the reactor, as shown in figure 5.

Fig. 10 and 11 illustrate a gas plasma reactor assembled in accordance with the present invention. In assembled form, the reactor has a chamber defining a cylindrical interior space. The reactor has an external height of about 420mm and a diameter of about 290 mm. The internal height is about 305mm and the diameter is about 254 mm. The chamber is made of aluminum and defines an interior space of approximately 15-436 square centimeters. As shown in fig. 10, one electrode made of 316L stainless steel includes 6 disk-shaped anode-cathode pairs. The distance between adjacent anodes and cathodes was about 1 inch near the inlet of the reactor. Slightly below the dimensions of the reactor in the vicinity of the outlet opening, which improves the efficiency of the treatment of the gas stream passing between the inlet and the outlet. The outer dimension of the electrode is slightly less than 254mm and about 300mm high. The anode disk has a central aperture and the cathode disk has an outer dimension slightly smaller than the anode disk to provide a tortuous gas flow path between adjacent anode-cathode pairs. The 16 flared fin elements are evenly distributed around each face of each cathode and the 16 planar fin elements are evenly distributed around each face of each anode. The planar fins are interleaved between each flared fin and each leg of each flared fin (see fig. 8). The flared and planar fins are positioned and dimensioned such that the size of the gas flow section between adjacent positive and negative electrodes is about one cubic inch. The total electrode area within the reactor interior space was about 27700 square centimeters and the ratio of electrode surface area to volume was about 1.8.

The reactor was tested for exciting and sustaining plasma in air over a range of pressures and voltages. The test thus used a model 2500E advanced energy industry model power supply modified to operate at 100 Khz. After modification, the power supply has a load rating of about 1500W. In the test, if the voltage is about 1000V and the load impedance is about 100 ohms, the reactor will successfully ignite and sustain a plasma in air when the pressure reaches 500 Torr. If the voltage is about 1400V, the impedance to load is about 1000 ohms and the pressure rises to about 18 mTorr.

The foregoing description of the preferred embodiment of the invention is intended to be illustrative of the nature of the invention and is not to be taken in a limiting sense. Those skilled in the art will understand that: various changes and modifications can be made to the preferred embodiment without departing from the spirit of the invention. For example, the dimensions of the illustrations may be varied, and various suitable materials may be used. The electrodes, reactor chamber, fin assembly and the like may be chosen from different geometries, such as "U" -shaped, "V" -shaped, "W" -shaped, and even cylindrical, spherical or conical. Thus, the present invention is not limited to the aforementioned electrode geometries and designs. The operating parameters may also vary. The preferred embodiments do not limit the scope of the invention, which is defined by the claims.

Claims (5)

1. A method of constructing a plasma reactor/generator, comprising:

providing a chamber defining an interior space;

providing a plurality of anode structures and cathode structures disposed within the chamber; each of the anode and cathode structures having a plurality of anode and cathode surfaces, respectively, wherein adjacent anode and cathode surfaces are opposed to each other and separated from each other to define a gas space therebetween;

providing a common electrical connection to the anode surfaces and a common electrical connection to the cathode surfaces;

providing a plurality of electrically conductive fin elements at the opposing surfaces of the anode structure and the cathode structure, the fin elements extending into the space between the opposing surfaces thereby dividing the space into a plurality of cells; and

the cells are arranged and formed to form a shape selected to control the characteristics of the plasma in the space.

2. The method of claim 1, wherein controlling the characteristic of the plasma in the space comprises controlling a shape of the plasma in the space.

3. The method of claim 2, wherein controlling the characteristic of the plasma in the space further comprises controlling a focus of the plasma.

4. The method of claim 1, wherein controlling the characteristic of the plasma in the space comprises controlling an intensity of the plasma.

5. The method of claim 1, wherein controlling the characteristic of the plasma in the space comprises controlling the plasma to generate a selected geometric pattern.