PLASMA GENERATION

BACKGROUND

[0001] As is known in the art, the use of plasma in the processing of materials has become prevalent in a wide range of industries, including electronics, aerospace, automotive, steel, biomedical, and toxic waste management. For example, various processes in semiconductor manufacturing facilities, such as reactive ion etching, utilize plasmas that activate a chain of chemical reactions at the surface of a substrate. A commonly used means of generating these plasmas is to inductively or capacitively couple energy from a radio-frequency (RF) power amplifier into the chamber containing the gas to be ionized, (e.g., by driving RF current through a coil wound around the chamber).

[0002] Some challenges in RF plasma generation include efficiently generating and controlling the RF power delivered into the plasma, while maintaining acceptable loading of the associated RF power amplifier under the highly-variable conditions in a plasma system. For example, Inductively Coupled Plasma (ICP) loads represent a dynamically- variable load impedance that depends on gas type and pressure, operating mode, power level and other features. In processes such as the Bosch process in reactive ion etching, there is a frequent change of the gas being ionized, which can contribute to the variation in the impedance presented by the coil. The effective load impedance can vary substantially in both its real and reactive component, making matching challenging.

SUMMARY

[0003] In accordance with the concepts, circuits and techniques described herein a dynamic matching system for plasma generation (e.g., inductively coupled plasma (ICP) generation) that losslessly maintains a near-constant driving point impedance across an entire plasma operating range is provided. The matching system includes a Resistance Compression Network (RCN), an impedance transformation and reactance compensation stage and a plurality of loads (e.g., plasma drive coils or electrodes) and can be configured to dynamically adjust the system operating frequency to maintain the match across plasma load variations during operation.

[0004] The matching system can include three stages, an RCN stage, an impedance transformation and reactance compensation stage (e.g., an impedance transformation and tuning stage), and a load stage and can be used for radio-frequency (RF) plasma generation to provide efficient power delivery into a plurality of plasma loads, while eliminating the need for mechanical or switched components.

[0005] The RCN stage can include a single-input multi-output matching network that utilizes resonant impedance transformation to losslessly transfer energy from the single input port to each of the multiple output ports loaded with variable but identical loads (ideally resistive), such that the impedance looking into the input port is (ideally) resistive and varies much less than the resistances at the plurality of loads. Thus, the variation of the input resistance at the RCN appears "compressed" as compared to the resistance variation at the plurality of loads. In an embodiment, the plurality of loads may see equal portions of the input power, though the relative phases of the load voltages may vary as the impedance values of the plurality of loads vary.

[0006] The impedance transformation and reactance compensation stage can include a plurality of impedance transformation and reactance compensation networks ((also sometimes referred to herein as a "load network" or a "network" since it provides reactance compensation and impedance transformation of a load), each having an input coupled to at least one of the outputs of the RCN. The networks can be configured to perform reactance compensation and/or impedance scaling (transformation) of the load impedances such that a desire impedance/reactance is presented to the RCN. For example, the networks can be configured to provide dynamic reactance cancellation and thus tune the impedances of a respective one of the plurality of loads they are coupled to, to appear resistive, by utilizing narrowband (e.g., less than 5%) adjustments to the system operating frequency. The reactive impedance at the plurality of loads can vary somewhat across operating conditions (e.g., owing to the screening out of the RF fields from the interior of the plasma chamber as the plasma density increases). In embodiments, a load network may be provided from a first functional portion which performs a reactance compensation function (and hence is sometimes referred to as a reactance compensation stage or component of the load network) and a second functional portion which performs an impedance compensation function (and hence is sometimes referred to as an impedance compensation stage or component of the load network). Thus, reactance compensation components included in an impedance transformation and reactance compensation stage can utilize variations in frequency to adaptively tune to achieve matched resistive loading with the plurality of loads.

[0007] The networks can also perform impedance scaling by transforming the reactance compensated plasma impedance at each of the plurality of loads to a desired range over which the RCN can be designed to operate.

[0008] The load stage (i.e. the loads coupled to the impedance transformation and reactance compensation stage) includes a plurality of loads that can be coupled to a plasma container in a variety of different arrangements to provide independent loads having equal impedances. For example, the loads can be disposed having an orthogonal relationship with respect to each other or a coaxial relationship with respect to each other, such that the loads are magnetically uncoupled. In an embodiment, by arranging the plurality of loads in particular geometries such that they are magnetically uncoupled, the RCN can be used to provide passive, high bandwidth matching (e.g., matching can be achieved in, for example, less than 100 RF cycles. In some embodiments, matching can be achieved in less than 10 RF cycles) with no mechanically-controlled or moving parts. Thus, the matching system can provide good matching across the entire range of plasma operation.

[0009] The plurality of loads can include two (or more) plasma drive coils (or electrodes) that act as independent and matched loads that vary in impedance together across the operating conditions (e.g., as the state of the plasma changes). This is accomplished through appropriate design of the plurality of loads that are independent (magnetically uncoupled) and present driving point impedances that are as close to identical as possible across all operating conditions. For example, in a plasma system, the total impedance seen at the terminals of a load can be a function of the power absorbed by the plasma as well as plasma composition, pressure, temperature and other conditions. To realize the multiple matched loads for inductively-coupled plasmas, the matching systems described here utilize multiple loads that meet two conditions: (a) the loads should have similar self-inductances and provide low magnetic coupling to each other (i.e., to the other loads that are ideally uncoupled); and (b) each load can drive plasma having substantially the same conditions, such that the loads have similar driving-point impedances. In an embodiment, this can be accomplished by disposing the loads such that they drive the same physical region of plasma. Together, these conditions can provide a set of loads having identically (or nearly identical) varying load impedances. It should be appreciated, that the loads can be disposed in a variety of different arrangements (e.g., geometries) to provide one or both of the aforementioned conditions, beyond those arrangements described herein.

[0010] The systems described herein may include one or more of the following features independently or in combination with another feature.

[0011] In a first aspect, a matching system for generating plasma using multiple loads having matched impedance values includes a plurality of loads with each of the loads disposed so as to be substantially independent from other ones of said plurality of loads and such that each of said plurality of loads present substantially equal impedance characteristics in response to signals provided thereto. The matching system further includes one or more impedance transformation and reactance compensation networks coupled to the plurality of loads. The one or more impedance transformation and reactance compensation networks can be configured to modify the reactive components of the impedances of the plurality of loads. In some embodiments, the one or more impedance transformation and reactance compensation networks can be configured to scale the range of the reactance-compensated plasma impedance at each of the plurality of loads to a desired range over which a resistance compression network (RCN) can be designed to operate. The RCN is coupled through the one or more impedance transformation and reactance compensation networks to the plurality of loads. The RCN can be configured to compress the resistive components of the impedance range for the plurality of loads such that the range of the resistive component of the impedance at the network's input is much narrower than that of the plurality of loads. In some

embodiments, the matching system includes a controller configured to adjust the operating frequency of the system to modify the reactive portion of at least one of the input impedance of the matching system and the individual loads.

[0012] The impedance characteristics of each of the plurality of loads can vary at the same rate over an operating range of the matching system. The plurality of loads can be provided from at least one of: a coil or an electrode. The coils can be disposed having a geometric relationship with respect to each other such that the coils are substantially magnetically uncoupled from each other and provide approximately equal impedance values. The geometric relationship can include at least one of the coils being disposed orthogonal with respect to each other, or the coils being disposed having a coaxial relationship with respect to each other, or the coils being disposed as two phases in a Halbach cylinder arrangement.

[0013] The plurality of loads can be independently coupled to at least one of the one or more impedance transformation and reactance compensation networks to generate independently driven loads. The one or more impedance transformation and reactance compensation networks can be configured to reduce the reactive components of the loads to zero or substantially zero. The one or more impedance transformation and reactance compensation networks can be coupled to each of the plurality of loads or to each of the reactance-compensated polarity of loads. In an embodiment, the one or more impedance transformation and reactance compensation networks may include a transformation stage and a reactance compensation stage. Each of the transformation stage and a reactance compensation stage may include a capacitive element, an inductive element, a

combination of both, or a transmission line transformer. [0014] The resistance compression network may include an input and multiple outputs. The number of outputs may correspond to the number of the plurality of loads, and the resistance compression network can be configured to transfer energy from the input to each of the outputs having identical load values. In some embodiments, the resistance compression network can be configured to present a near-constant resistive impedance at its input.

[0015] The matching system may include an impedance transformer coupled to the input of the resistance compression network. The impedance transformer can be configured to scale the near-constant impedance value.

[0016] In some embodiments, the controller can be communicatively coupled to a matching sensor at the input of the matching system. The controller can be configured to adjust the operating frequency of a power amplifier based on the matching measure obtained from the matching sensor.

[0017] In another aspect, a method for driving plasma using a matching system is provided. The method includes generating a radio-frequency drive signal and providing it to the input of a matching system, and utilizing a resistance compression network as part of the matching system to split that signal for provision to a plurality of loads, the plurality of loads each comprised of a plasma drive coil/electrode and monitoring the state of the impedance match of the matching system. The method further includes varying an operating frequency of the generated RF signal to maintain a matched input impedance characteristic.

[0018] The plurality of loads may be disposed with respect to each such that the loads are magnetically uncoupled from each other and provide the approximately equal impedance values. For example, the plurality of loads may be disposed orthogonally with respect to each other or the plurality of loads may be disposed having a coaxial relationship with respect to each other, or the plurality of loads may be disposed in a Halbach cylinder arrangement.

[0019] In some embodiments, each of the plurality of loads may be driven independent of each of other using a plurality of impedance transformation and reactance compensation networks coupled to each of the plurality of loads.

[0020] The method may include reducing the reactance components of the loads to zero or substantially zero. A reduction of the reactance components may correspond to a change in the operating frequency of the RF signal driving the input of the matching system.

[0021] In some embodiments, the impedance at the input of the resistance

compression network can be maintained to be resistive and nearly constant.

[0022] In another aspect, a matching system for generating plasma using multiple loads includes a plurality of loads, each of the loads disposed so as to be substantially independent from other ones of said plurality of loads and one or more reactance compensation networks coupled to the plurality of loads. The one or more reactance compensation networks can be configured to modify reactive components of impedances of the plurality of loads. The matching system further includes a resistance compression network coupled through the one or more reactance compensation networks to the plurality of loads. The resistance compression network can be configured to compress the resistive components of the impedance range for the plurality of loads such that the range of the resistive component of the impedance at the network's input is much narrower than that of the plurality of loads.

[0023] In a still first aspect, a matching system for generating plasma using multiple loads having matched impedance values includes one or more impedance transformation and reactance compensation networks configured to be coupled to the plurality of loads. The one or more impedance transformation and reactance compensation networks can be configured such that in response varying impedances of the plurality of loads, the impedance transformation and reactance compensation networks modifies at least a reactive component of the load impedances and transforms the load impedance for desired interaction with a resistance compression network (RCN) coupled to the one or more impedance transformation and reactance compensation networks. In some embodiments, the one or more impedance transformation and reactance compensation networks can be configured to transform (or scale) a range of the reactance-compensated plasma impedance at each of the plurality of loads to a desired operational range of the resistance compression network (RCN). The RCN is configured such that is it coupled through the one or more impedance transformation and reactance compensation networks to the plurality of loads. The RCN can be configured to compress the resistive components of the impedance range for the plurality of loads such that the range of the resistive component of the impedance at the network's input is much narrower than that of the plurality of loads. In some embodiments, the matching system includes a controller configured to adjust the operating frequency of the system to modify the reactive portion of at least one of the input impedance of the matching system and the individual loads. In some embodiments, the multiple loads are provided as part of the matching system with ones of the loads disposed so as to be substantially independent from other ones of the loads and such that each of the loads present substantially equal impedance characteristics in response to signals provided thereto.

[0024] It should be appreciated that elements of different embodiments described herein may be combined to form other embodiments not specifically set forth above. Various elements, which are described in the context of a single embodiment, may also be provided separately or in suitable combination. Other embodiments, not specifically described herein are also within the scope of the following claims.

BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The foregoing features may be more fully understood from the following description of the drawings in which:

[0026] FIG. 1 shows a plasma drive and matching system;

[0027] FIG. 1A shows one embodiment of the structure of the matching system of FIG. 1 and its respective stages;

[0028] FIGs. IB-ID show conceptual plots illustrating the effective impedances seen at inputs of different components of the plasma drive and matching system of FIG. 1;

[0029] FIGs. 2-2A show circuit diagrams of two embodiments of the Resistance Compression Network of the matching system of FIG. 1;

[0030] FIG. 2B is a plot illustrating the variation of an input resistance of a resistance compression network as a function of load resistances;

[0031] FIG. 3 shows a circuit diagram of one embodiment of the reactance compensation network of the matching system of FIG. 1 ; [0032] FIG. 3A shows a plot illustrating the effect of small frequency adjustments from a nominal operating frequency of the matching system of FIG. 1 on the magnitude of the impedance value of one or more loads;

[0033] FIG. 4 shows one embodiment of an arrangement of coils having an orthogonal relationship;

[0034] FIG. 4A shows one embodiment of an arrangement of coils having a coaxial relationship;

[0035] FIG. 4B shows one embodiment of an arrangement of one of two coils having a magnetically orthogonal relationship;

[0036] FIG. 4C shows a top view of the arrangement of one of the two coils having a magnetically orthogonal relationship of FIG. 4B;

[0037] FIG. 4D shows a top view of the arrangement of the two coils having a magnetically orthogonal relationship of FIG. 4B;

[0038] FIG. 5 is one embodiment of a conceptual circuit diagram of the plasma drive and matching system of FIG. 1;

[0039] FIG. 5A is a simplified circuit diagram of an embodiment of the plasma drive and matching system of FIG. 1; and

[0040] FIG. 6 is a flow diagram of a method for driving plasma using the matching system of FIG. 1. DETAILED DESCRIPTION

[0041] Referring now to FIG. 1 , a matching system 102 includes a resistance compression network (RCN) 106 having at least one input and a plurality of outputs, each of the outputs coupled to an input of at least one of a plurality of impedance

transformation and reactance compensations networks 108a-108n (sometimes referred to herein simply as "load networks" or even more simply as "networks" 108a-108n, or "network" 108). In some embodiments, the impedance transformation and reactance compensations networks 108a-108n may be referred to herein simply as "load networks." Each of the plurality of networks 108a-108n have outputs coupled to an input of at least one of a plurality of loads 1 12a-112n. In some embodiments, matching system 102 may include an optional impedance transformer 104 having an input coupled to an RF power amplifier 116 (which generates an RF signal) and an output coupled to the input of RCN 106. In other embodiments, impedance transformer 104 may be omitted and an input of RCN 106 may be coupled to the output of power amplifier (PA) 1 16.

[0042] A controller 114 is coupled to a first matching sensor 1 10a, which monitors a state of an impedance match (e.g. by measuring reflected power or by otherwise determining an impedance match between an output of PA 116 and impedance transformer 104 or between an output of PA 1 16 and RCN 106 in the event that impedance transformer 104 is omitted). In response to information measured or otherwise provided thereto, matching sensor 1 10a can feed back a command signal to RF amplifier 116 via a frequency control circuit 1 14.

In embodiments, matching system 100 may include a plurality of matching sensors 110a- 1 10η disposed at its input and/or outputs to monitor the state of the impedance match. For example, and as illustrated in FIG. 1, controller 1 14 may be coupled to matching sensor 110a to monitor an impedance match at or near an input to matching system 100. In some embodiments, controller 114 may in addition be optionally coupled to a plurality of other matching sensors 1 lOb-110η that monitor the state of the impedance match at other points in the circuit (e.g., at the inputs to the plurality of loads). Thus, in some embodiments it may be desirable or advantageous to use all of sensors 1 lOa-1 lOd while in other embodiments it may be desirable or advantageous to only sensor 110a, while in still other embodiments it may be desirable or advantageous to use only sensors 1 lOb-1 lOd, while in still other embodiments it may be desirable or advantageous to use some combination of sensors H Oa-l l Od.

[0043] Controller 1 14 can be configured to generate a command signal and provide the command signal to amplifier 116. In some embodiments, the command signal can be based at least in part on measurements provided by one or more of the matching sensors 110a- 1 10η, and the command signal can be used to modify the frequency of the RF signal generated by amplifier 1 16. For example, controller 114 can monitor the measurements from one or more of the matching sensors 110a- 11 On and actively send response signals to amplifier 116 to adjust the frequency of the RF signal generated by it. This adjustment of the operating frequency enables dynamic cancellation of the reactive components of the impedances of the plurality of loads 112a-112n across plasma operating conditions, and allows the match to be maintained.

[0044] In operation, loads 1 12a-l 12n (e.g., plasma drive coils, plasma-drive electrodes) provide independent and variable but matched input impedances for driving plasma contained within a plasma container that the loads 112a-l 12n are coupled to (e.g., on which the drive coils/electrodes are mounted). Loads 1 12a-l 12n are coupled to networks 108a-108n, which can be configured to perform impedance scaling and reactance compensation to cancel the reactive impedance components of the impedance values experienced at loads 1 12a-l 12n. In an embodiment, networks 108a-108n can each include multiple stages, such as but not limited to an impedance transformation stage and a reactance compensation stage, to perform the impedance scaling and reactance cancellation, respectively.

[0045] RCN 106 is coupled to each of the networks 108a-108n and can be configured to transform the multiple matched but variable transformed and reactance-compensated resistive loads at the RCN's outputs into a single, nearly constant input resistance. In some embodiments, impedance transformer 104 can be disposed between power amplifier 116 and the input of RCN 106 to provide further impedance scaling if required.

[0046] It should be appreciated that some of the components and/or stages of the matching system can be combined or omitted, depending on received specific

implementation and system requirements. Each of the different stages of matching system 100 will be described in greater detail below.

[0047] Referring now to FIG. 1 A, illustrative embodiments of circuit representations for RCN 106, one impedance transformation and reactance compensation network 108 of the plurality of networks 108a-108n (also referred to herein as a load network) and one load 1 12 of the plurality of loads 1 12a-112n of FIG. 1 are shown. FIG. 1 A, is organized in three stages, a resistance compression stage, a tuning and impedance transformation stage, and a matched plasma loads stage.

[0048] RCN 106 includes first and second reactive elements 130, 132 (e.g., conjugate reactance elements), having impedances Z=-jX and +jX, respectively. In the illustrative embodiment of FIG. 1A, reactive elements 130, 132 are coupled to input 128 as shown. In operation, RCN 106 can losslessly transform the impedance presented by multiple matched but variable resistive loads (e.g., the resulting impedances of the loads 112a-112n of FIG. 1 after reactance compensation and impedance scaling) into a single, constant or nearly-constant input resistance. The range of impedances seen at the input 128 of RCN 106 is illustrated in FIG. IB. In the plot 170 provided in FIG. IB, RCN 106 transforms the varying resistances of a set of two or more loads (the impedance of which is shown in FIG. 1C) into a single narrow-range driving point resistance as illustrated, here between Rd-Δ and Rd+Δ, where A=Rd(b-l) ² /(b+l) ² .

[0049] Referring back to FIG. 1A, an embodiment of each network of the plurality of networks 108 includes a first reactive element 134 (having impedance Ζ=-)Χτ), a second reactive element 136 (having impedance Ζ=+)Χτ), a third reactive element 138 (having impedance Ζ=-)Χτ), and a fourth reactive element 140 (having impedance Z=-)XL).

[0050] Network 108 can utilize frequency tuning to null the reactive components of the plasma load input impedances at load 1 12. Network 108 can also be configured to transform the reactance-compensated impedance values to a desired range over which the RCN stage 106 is configured to operate. For example, impedance network 108 can include an inductive element, a capacitive element or a combination of inductive elements and capacitive elements, to perform reactance tuning. Now referring to FIG. 1 C, plot 180 illustrates a range of effective impedances observed at the input of network 108, here maintained within a range of impedance values from Rofb to bRo, where b represents the extent over which the matched loads vary in resistance and Ro represents a nominal resistance of RCN 106. In an embodiment, network 108 can transform the reactance- compensated impedance values to the range from Ro/b to bRo, such that the resulting impedance observed at the input 128 of RCN 106 can be compressed and remains purely resistive.

[0051] Referring back to FIG. 1A, load 1 12 is modeled as an inductive element 142 (having impedance Z=+)XL), and a resistive element 144 coupled in series. Referring now to FIG. ID, a plot 190 illustrates the input impedances seen at the input of load 1 12. For example, and as illustrated in plot 190, the input impedances can vary over a wide range in effective resistance (e.g., 1 : 10), here from RL, min to RL, max, and over a narrower range in reactance (e.g., 1 : 1.1), here from XL, min to XL, max.

[0052] In some embodiments, the structure and/or components of a matching system (e.g., matching system 100 of FIG. 1) are selected based at least in part on the properties and/or attributes of RCN 106. For example, RCN 106 utilizes the multiple matched resistive loads to achieve the narrow-range input resistance illustrated in FIG. IB, thus the matching system includes a plurality of loads 1 12 (e.g., two or more) to provide identical or nearly identical but dynamically-varying driving-point impedances.

[0053] Referring now to FIG. 2, an illustrative embodiment of a first RCN 200 topology includes a first reactive component 204 coupled in series with a first resistive component 208, and a second reactive component 206 coupled in series with a second resistive component 210. First and second resistive components 208, 210 correspond to two matched resistances, and may be coupled to a ground reference 212. Briefly referring to FIG. 2A, an illustrative embodiment of a second RCN 250 topology includes a first reactive component 254 coupled in parallel with a first resistive component 258, and a second reactive component 256 coupled in parallel with a second resistive component 260. It should be appreciated that each of matched resistances 208, 210, 258, 260 may be the same as or substantially similar to at least one of the matched resistances resulting from the transformed and reactance-compensated plurality of loads 112a-l 12n of FIG. 1.

[0054] As illustrated in FIGs. 2-2A, an RCN can include a single-input multi-output matching network. In operation, the RCN can utilize resonant impedance transformation to losslessly transfer energy from a single input port, here represented by inputs 202, 252 in RCN 200 and 250 respectively, to a pair of output ports (here represented by first and second matched loads 208, 210, 258, 260) loaded with equal or substantially equal resistances, such that the resistance looking into input port 202, 252 varies much less than the matched-load resistances (i.e., resistances of the transformed and reactance- compensated plurality of loads 112a- 112n of FIG. 1). Thus, the variation of the input resistance of the RCN appears "compressed" as compared to the resistance variation of the matched loads. In an embodiment, first and second matched loads 208, 210, 258, 260 may ideally see equal portions of the input power, and the relative phases of the load voltages vary as the matched load impedances vary.

[0055] For example, and referring to RCN 200 of FIG. 2, with its characteristic impedance value X chosen to be at a geometric center of the resistive load range, Ro. In an embodiment, if the first and second resistive matched loads 208, 210 vary together about a nominal resistance Ro, from Ri = Ro/b to R2 = bRo, the resulting impedance observed at the input of RCN 200 can be purely resistive across the entire RCN load range, and can vary from Ro to kRo, where k = (1 + b ² )/2b.

[0056] This is illustrated in FIG. 2B, which shows a plot 270 illustrating the variation in input impedance Zin = Rin of RCN 200 as a function of its matched load impedance R. In the illustrative embodiment of FIG. 2B, b = V(R2/Ri), which represents the extent over which the matched loads vary. Further, b determines k, which represents the extent of input resistance variation. In an embodiment, for a requested degree (e.g., highest degree) of compression in the input impedance to be achieved, the geometric center of the matched load resistance range Ro can be made equal to the characteristic impedance X of RCN 200.

[0057] In an impedance matching system (e.g., matching system 100 of FIG. 1), the input impedance can be configured to be as close as possible to a desired resistance value Rd (e.g., 50 Ω). For example, for a given RCN load range [Ri, R2], we can compute b = V(R2/Ri), and thus k can be determined. The RCN characteristic impedance X can be chosen to produce a compressed input resistance range between Rd-Δ and Rd+Δ. In some embodiments, this can be achieved by selecting X=2Rd/(k+l), resulting in A=(k- l)Rd/(k+l). Thus, the nominal RCN load resistance can be computed as the geometric mean of the minimum and maximum resistance values in the given RCN load range, (e.g., Ro = V(RiR2)). Achieving this is one of the functions of the tuning and impedance transformation stage described below. Alternatively, to provide greater flexibility in the impedance levels selected for the RCN stage, an additional impedance transformation stage (i.e., impedance transformer 104 of FIG. 1) can be disposed between the RCN and an input of the respective matching system or to select other RCN structures that can simultaneously or near simultaneously provide both transformation and compression.

[0058] Referring now to FIG. 3, a reactance compensation network 300 includes a resonant network 304 and a reactance-cancellation stage 310 coupled in series to a load 312. In an embodiment, the resonant network 304 may include an inductive element 306 and a capacitive element 308 and reactance-cancellation stage 310 may include a capacitive element 320. Reactance compensation network 300 may be the same as or substantially similar to part of one of the plurality of networks 108a-108n of FIG. 1, and load 312 may be the same as or substantially similar to one of the plurality of loads 1 12a- 112n of FIG. 1.

[0059] In operation, network 300 can be configured to provide reactive tuning. For example, the impedance of load 312 includes a resistive component and a reactive component, and the impedance of load 312 can vary in both resistance and reactance, as operating conditions change (e.g., due to the screening out of radio frequency (RF) fields from an interior of a plasma chamber on which the load 312 (e.g., drive coils/electrodes) can be mounted, as properties of the plasma changes). Therefore, network 300 can enable narrow-band dynamic frequency tuning, whereby small variations in a matching system operating frequency are used to adaptively tune the matching system such that the input impedances at network 300 appear substantially resistive. In an embodiment, network 300 includes reactance-cancellation stage 310 to perform reactive tuning to cancel (null) or otherwise reduce a reactive component of the impedance of load 312, and network 300 also includes resonant network 304 to make the reactance very frequency selective, such that narrow-band frequency tuning can be used to achieve resistive loading of the RCN stage.

[0060] As indicated above, in some embodiments, the loads for an RCN stage can be purely resistive or near-purely resistive. Thus, reactance stage 310 performs reactive tuning to cancel (null) or otherwise reduce the reactive component of the impedance of load 312. In the illustrative embodiment of FIG. 3, reactance stage 310 includes capacitive element 320 (e.g., capacitor) to perform the reactive tuning. The properties of capacitive element 320 can be selected based at least in part on the impedance values of load 312. It should be appreciated that FIG. 3 illustrates one embodiment of a reactance stage and that other circuit components can be used to form reactance stage 310 and be configured to cancel (null) or otherwise reduce the reactive component of the impedance of load 312.

[0061] The resonant network 304 is configured to increase the frequency selectivity of the reactance. For example, network 300 coupled to load 312 can be frequency selective (e.g., network having capacitive element 320 in series with inductive element 314).

However, in some applications, given a limited allowable range of frequency adjustment, that frequency selectivity may not provide sufficient reactance compensation. Thus, network 304 having inductive element 306 and capacitive element 308 in series can increase the overall quality factor ("Q") of the network formed by coupling network 300 in series with load 312, thereby increasing the frequency selectivity of the network.

[0062] Now referring to FIG. 3A, a plot 350 illustrates the effects of frequency adjustments, Αΐ, from the nominal operating frequency, fo, on the maximum achievable reactance. In an embodiment, depending on a given range of frequency adjustment, 0≡fo/Af (typically »1) can be set, and a characteristic impedance Zo for a reactance compensation network coupled to a load can then be selected to provide a desired maximum amount of reactance cancellation. Thus, impedance network 300 can be configured to dynamically tune a net impedance of load 312 to be resistive.

[0063] As indicated above, referring briefly to FIG. 1, a second function of each of the plurality of networks 108a-108n is to transform/scale the reactance-compensated impedance values to a desired range over with the RCN stage is configured to operate. In the illustrative embodiment of FIG. 3, the reactance cancellation functionality was described, and the same or substantially similar embodiment of network 300 can be used to perform the reactive tuning to each of the plurality of networks 108a-108n. For example, and referring to FIG 1 A, this can correspond to conceptual element 140. The impedance transformation functionality can be accomplished using a variety of circuit implementations (e.g., utilizing transformers, transmission-line transformers, immittance converters or other transformation networks that function well with variations in loading impedance). This can correspond to the conceptual elements 134, 136, 138 in FIG 1 A. Thus, for a predetermined compression amount, the RCN matched loads should vary from Ro/b to bRo, as illustrated in FIG. 1C. The result of this is a matched pair of resistive load impedances that can be compressed into a desired range by the RCN.

[0064] Now referring to FIG. 4, a first plasma drive coil 402 and a second plasma drive coil 410 are coupled to a plasma container 414. In the illustrative embodiment of FIG. 4, first and second drive coils 402, 410 are disposed around an outer surface 414a of plasma container 414 having a geometrically orthogonal relationship with respect to each other. In an embodiment, plasma drive coils 402, 410 may be the same as or substantially similar to loads 112a-l 12n of FIG. 1.

[0065] First coil 402 includes a first and second winding 406a, 406b, with an input terminal 404a coupled to first winding 406a and an output terminal 404b coupled to second winding 406b. Input terminals 404a, 410a can be coupled to an output of at least one reactance compensation network (e.g., one of networks 108a-108n of FIG. 1) to receive RF signals.

[0066] Second coil 410 includes a first and second windings- 412a, 412b, with an input terminal 410a coupled to first winding 412a and an output terminal 410b coupled to second winding 412b. Input terminals 404a, 410a can be coupled to an output of at least one reactance compensation network (e.g., one of networks 108a-108n of FIG. 1) to receive RF signals.

[0067] Each of input terminals 404a, 410a can be coupled to an output of at least one reactance compensation network (e.g., one of networks 108a-108n of FIG. 1) to receive RF signals. In an embodiment, each of input terminals 404a, 410a can be coupled to outputs from different reactance compensation networks. In an embodiment, in response to the received RF signals, the orthogonal relationship between first and second plasma drive coils 402, 410 can generate two independent loads that are magnetically -uncoupled.

[0068] For example, the orthogonal relationship between first and second drive coils 402, 410 provides loads that are magnetically uncoupled (magnetically "orthogonal") but spatially close (such that they drive similar or identical portions of the plasma). For example, and as illustrated in FIG. 4, first and second drive coils 402, 410 physically overlap the same space or surface area of plasma container 414 and drive similar (or the same) plasma volume, but are wound such that net flux from each load does not link the other respective drive coil (or coils). In one embodiment, first and second coils are disposed geometrically orthogonal having a similar center point (i.e., the respective center portions cross each other), but rotated by 90 degrees.

[0069] As first and second plasma drive coils 402, 410 are wound around the same plasma container 414 and are spatially orthogonal, they can excite the same or substantially similar physical plasma region(s) within plasma container 414. This ideally results in two magnetically uncoupled loads (coils) with matched driving-point impedances. [0070] First and second drive coils 402, 410 are magnetically orthogonal in the sense that a current in one does not induce voltage on the other, resulting in no (or very small) mutual inductance between the two drive coils. In the illustrative embodiment of FIG. 4, both first and second drive coils 402, 410 surround the same region of plasma, and therefore for a given plasma power, the impedance seen at each load terminal should ideally be the same or substantially close.

[0071] It should be appreciated that although FIG. 4 provides two drive coils, first and second drive coils 402, 410, disposed around plasma container 414, in other embodiments, more than two drive coils may be used. Further, in the illustrative embodiment of FIG. 4, first and second plasma drivers 402, 410 are represented by coils. However, it should be appreciated that in other embodiments, first and second plasma drivers 402, 410 may include electrodes.

[0072] Now referring to FIG. 4A, a first plasma drive coil 432 and a second plasma drive coil 436 are disposed around an outer surface 440a of a plasma container 440. First plasma drive coil 432 includes a plurality of windings 434a-434c with an input 432a coupled to a first winding 434a and an output terminal 432b coupled to a third winding 434c. Second drive coil 436 includes a plurality of windings 438a-438c with an input 436a coupled to a first winding 438a and an output terminal 436b coupled to a third winding 438c. Each of input terminals 432a, 436a can be coupled to an output of at least one reactance compensation network (e.g., one of networks 108a-108n of FIG. 1) to receive RF signals. In an embodiment, each of input terminals 432a, 436a can be coupled to outputs from different reactance compensation networks. [0073] The plurality of windings of each of the first and second drive coils 432, 436 can be disposed such that they are wound coaxially around outer surface 440a of plasma container 440. Thus, first and second drive coils 432, 436 can share generally the same plasma volume (i.e., plasma container 440) and provide mutual inductance cancellation.

[0074] For example, and as illustrated in FIG. 4A, the windings of first and second coils 432, 436 are disposed having forward turns and backward turns to cancel mutual inductance between first and second coils 432, 436. In an embodiment, first and second windings 434a-434b of first coil 432 can be disposed as forward-turns and third winding 434c of first coil 432 can be disposed having a backward-turn (or opposite turn of first and second windings 434a-434b). Likewise, first and second windings 438a-438b of second coil 436 can be disposed as forward-turns and third winding 438c of second coil 436 can be disposed as a backward-turn (or opposite turn of first and second windings 438a-438b).

[0075] If we denote the indices for the forward-turns of first and second coils 432, 436 by i and j, respectively, and denote back- ward-turns by b, then the net mutual inductance between the two coils can be written as follows:

[0076] Where Nf represents windings having forward-turns and Nb represents windings having backward-turns, and M represents mutual inductance.

[0077] In some embodiments, the order or arrangement of the plurality of windings of first and/or second coil 432, 436 can be modified to create other coil structures and provide mutual inductance cancellation. For example, in one embodiment, the forward windings of first and second coils 432, 436 can be disposed around outer surface 440a such that they are alternating with respect to each other.

[0078] It should be appreciated that although FIG. 4A illustrates two plasma drive coils, the techniques here can be extended to higher-order systems having more than two magnetically uncoupled drive coils.

[0079] Now referring to FIGs. 4B-4D, in which like reference numerals indicate like elements, a first drive coil 452 is disposed around an outer surface 460a of a plasma container 460. First drive coil 452 includes an inner layer of windings 454, an outer layer of windings 456, an input terminal 452a and an output terminal 452b. Input terminal 452a can be coupled to an output of at least one network (e.g., one of networks 108a-108n of FIG. 1) to receive RF signals. In an embodiment, inner and outer windings 454, 456 of first coil 452 can be disposed around outer surface 460a such that they form a "Halbach array."

[0080] For example, and as illustrated in FIGs. 4B and 4C, first coil 452 is arranged having two layers, inner and outer windings 454, 456, that are constructed to focus the fields inside plasma container 460 and cancel the (unused) external fields, thus benefiting the potential electromagnetic interference that could otherwise be generated. Inner and outer windings 454, 456 are disposed such that they cover the same or substantially similar surface area of plasma container 460.

[0081] One or a combination of the wire arrangement, number of wires, wire spacing, and/or the layer spacing can be selected to minimize the energy stored in the stray external fields (which ideally should be zero) as compared to the fields within plasma container 460, while yielding a pair of separate Halbach coil phases that are ideally magnetically uncoupled but drive similar regions within plasma container 460.

[0082] In some embodiments, a second drive coil or more than two coils can be disposed around outer surface 460. For example, and as illustrated in FIG. 4D, a top view shows first drive coil 452 and a second coil 462 disposed around outer surface 460a of plasma container 460. First drive coil 452 includes inner and outer windings 454, 456 and second drive coil 462 includes an inner layer of windings 464 and an outer layer of windings 466, thus forming a two layer coil similar to first coil 452. In an embodiment, the windings of each of first and second coil 452, 462 form two "Halbach arrays" around the outer surface 460a.

[0083] First and second drive coils 452, 462 are disposed such that they are magnetically orthogonal but cover substantially the same surface area of plasma container 460. The windings of first and second coils 452, 462 can be arranged such that their respective drive areas share the same plasma volume with respect to outer surface 460a, and can be spatially-rotated from each other such that (ideally) no net flux links from first coil 452 to second coil 462 and vice versa.

[0084] For example, in some embodiments, the windings of first and second coils 452, 462 can be rotationally offset such that any fields from first coil 452 that link second coil 462 are cancelled by equal and opposite fields from second coil 462 that link first coil 452, as illustrated in FIG. 4D. In such an embodiment, no magnetic coupling is provided between first and second coils 452, 462, and allows both coils to drive the same region of plasma within plasma container 460. [0085] Now referring to FIG. 5, a matching system 500 includes an RCN 506, a first impedance transformation and tuning network 512, a second impedance transformation and tuning network 530 and a plurality of loads 544. Matching system 500 may be the same as or substantially similar to matching system 100 of FIG. 1. For example, matching system 500 may correspond to one illustrative conceptual embodiment of matching system 100 of FIG. 1.

[0086] In the illustrative embodiment of FIG. 5, an RF signal generator 502 is coupled to an input of an amplifier 504, and an output of amplifier 504 is coupled to an input 506a of RCN 506. RCN 506 includes a capacitive element 508 and an inductive element 510. RCN 502 includes two outputs, with a first output 506b coupled to a first impedance transformation and tuning network 512 and a second output 506c coupled to a second impedance transformation and tuning network 530.

[0087] First network 512 includes a transformation stage 514 and a tuning stage 524. Transformation stage 514 includes a first inductive element 518, second inductive element 520 and a capacitive element 522. In the illustrative embodiment of FIG. 5, first and second inductive elements 518, 520 are coupled as shown and a first terminal of capacitive element 522 is coupled to a common node between first and second inductive elements 518, 520 and a second terminal of capacitive element 522 coupled to a ground potential 570. An output of transformation stage 514 is coupled to an input of tuning (or reactance compensation) stage 524. Tuning stage 524 includes an inductive element 526 and a tunable capacitive element 528 coupled in series to illustrate dynamic reactance compensation (in practice, this dynamic reactance compensation may be achieved using a fixed capacitor 528 with frequency adjustments). An output of tuning stage 524 is coupled to an input of a first load 546. First load 546 includes a capacitive element 548, an inductive element 552 and a resistive element 554 coupled in series (which models the effective plasma drive coil impedance), with a second terminal of resistive element 554 coupled to ground potential 570. In an embodiment, inductive element 552 and resistive element 554 can correspond to an impedance value 550 for first load 546.

[0088] In the illustrative embodiment of FIG. 5, capacitive element 548 of first load 546 can be used to tune a reactance component of first load 546. In an embodiment, the impedance transformation can be performed using a T-network. This network can be designed to transform the range of plasma resistance into a range suitable for use by RCN 506. Inductive element 526 in series with capacitor 528 can be used for frequency-based tuning cancellation of the reactive component of first load 546 and/or to account for any mismatch in the implementation and/or geometry of first and second loads 546, 556.

[0089] Second network 530 includes a transformation stage 532 and a tuning stage 540. Transformation stage 532 includes a first inductive element 534, second inductive element 536 and a capacitive element 538. First and second inductive elements 534, 536 are coupled as shown and a first terminal of capacitive element 538 is coupled to a common node between first and second inductive elements 534, 536 and a second terminal of capacitive element 538 coupled to a ground potential 570. An output of transformation stage 532 is coupled to an input of tuning stage 540. Tuning stage 540 includes an inductive element 542 and a capacitive element 574 coupled in series. An output of tuning stage 540 is coupled to an input of a second load 556.

[0090] Second load 556 includes a capacitive element 558, an inductive element 562 and a resistive element 564 coupled in series, with a second terminal of resistive element 564 coupled to ground potential 570. In an embodiment, inductive element 562 and resistive element 564 may correspond to an impedance value 560 for second load 556.

[0091] In the illustrative embodiment of FIG. 5, capacitive element 558 of second load 556 can be used to tune a reactance component of second load 556. Inductive element 542 in series with capacitor 574 can be used for frequency-based tuning cancellation of the reactive component of second load 546 and/or to account for any mismatch in the implementation and/or geometry of first and second loads 546, 556.

[0092] In an embodiment, as the load reactance varies across an operating range due to the plasma conductivity of plasma within a plasma container that first and second coils 546, 556 are coupled to (e.g., mounted on), a sensitivity of one or more of the components (e.g., inductive elements 526, 542, capacitive elements 528, 574) of matching system 100 to the operating frequency can be exploited to adjust the tuning of first and second loads 546, 556 across the plasma regions of operation by making small variations to the operating frequency of power amplifier 504. In some embodiments, a closed-loop control strategy that dynamically selects the optimal power amplifier operating frequency for which the reflection is minimal can be used to tune the operating frequency of matching system 500.

[0093] It should be appreciated that FIG. 5 illustrates one embodiment of a matching system. For example, in other embodiments, a matching system may include more than two impedance transformation and tuning networks 512, 530, and/or more than two loads 544.

[0094] Further, one of ordinary skill in the art will appreciate that other circuit components and/or circuit topologies may be used to form matching system 500 and/or each of the stages of matching system 500. For example, different combinations of the circuit elements (e.g., inductive elements, capacitive elements, resistive elements, etc.) and/or different circuit components may be used to form one or a combination of RCN 506, first and second impedance transformation and tuning networks 512, 530, transformations stages, 514, 532, tuning stages 524, 540, and first and second loads 546, 556.

[0095] Now referring to FIG. 5 A, a matching system 551 includes an RCN 507. Matching system 551 may be the same as or substantially similar to matching system 100 of FIG. 1 and matching system 500 of FIG. 5. For example, matching system 551 may correspond to one illustrative embodiment of the conceptual matching system 100 of FIG. 1 and matching system 500 of FIG. 5.

[0096] In the illustrative embodiment of FIG. 5 A, an RF signal generator 503 is coupled to an input of an amplifier 505 and an output of amplifier 505 is coupled to an input 507a of RCN 507. RCN 507 includes a capacitive element 509 and an inductive element 511. RCN 507 includes two outputs, with a first output 507b coupled to capacitive element 517, capacitive element 523 and capacitive element 529. A second terminal of capacitive element 517 is coupled to a ground potential 571. Second output 507c is coupled to capacitive element 519, capacitive element 539 and capacitive element 575 and a second terminal of capacitive element 519 is coupled to ground potential 571.

[0097] An output 521 of capacitive element 523 and capacitive element 529 is coupled to a first load 547, and an output 525 of capacitive element 539 and capacitive element 575 is coupled to a second load 557. First load 547 includes an inductive element 553 and a resistive element 555 coupled in series (which models the effective plasma drive coil impedance), with a second terminal of resistive element 555 coupled to ground potential 571. In an embodiment, inductive element 553 and resistive element 555 can correspond to an impedance value for first load 547.

[0098] Second load 557 includes an inductive element 561 and a resistive element 565 coupled in series (which models the effective plasma drive coil impedance), with a second terminal of resistive element 565 coupled to ground potential 571. In an embodiment, inductive element 561 and resistive element 565 can correspond to an impedance value for second load 557.

[0099] It should be appreciated that FIG. 5A illustrates one embodiment of a matching system. Further, one of ordinary skill in the art will appreciate that other circuit components and/or circuit topologies may be used to form matching system 551 and/or each of the stages of matching system 551.

[00100] Now referring to FIG. 6, a method 600 for driving plasma using a matching system begins at block 602 by generating an RF drive signal and providing it to the input of the matching system. The signal (e.g. RF signal) can be split by the matching system to drive multiple loads that are each comprised of a plasma drive coil/electrode mounted on a plasma container for the purposes of driving or otherwise modifying the properties of plasma contained within the plasma container. For example, the matching system can include an RCN having an input coupled to the output of an amplifier. The RCN includes multiple outputs, with each output coupled to at least one of a plurality of impedance transformation and reactance compensation networks (also referred to herein as networks). Each of the networks can be coupled to at least one of a plurality of loads (e.g., coils, electrodes) that are mounted on an outer surface of a plasma container. In an embodiment, the multiple loads formed by the effective input impedances of the plasma coils/electrodes can act as independent and variable but matched impedances for driving plasma in the plasma container. The networks can each include one or more stages, such as an impedance transformation stage (or impedance transformation network) and a reactance compensation stage (or reactance compensation network), to perform impedance scaling and/or cancel a reactive impedance component of the multiple loads. The RCN can be configured to losslessly transform the impedance presented by the multiple loads into a single, nearly -constant input resistance.

[00101] At block 604, the RCN can be utilized as part of the matching system to split the RF signal for provision to the plurality of loads. In an embodiment, the plurality of loads can have approximately equal impedance values and be independent of each other. For example, the loads (e.g., drive coils/electrodes) can be disposed around the outer surface of the plasma container such that they are magnetically uncoupled from each other and provide the approximately equal impedance values. To provide independent loads, each of the loads can be coupled to a different network and thus driven by the respective network. To provide the magnetically uncoupled relationship, the loads can be disposed in different arrangements with respect to each other. In some embodiments, the loads can be disposed such that they are orthogonal with respect to each other. In other embodiments, the loads can be disposed such that they have a coaxial relationship with respect to each other. It should be appreciated that other arrangements that provide independent and magnetically uncoupled loads are within the scope is this disclosure.

[00102] At block 606, the state of the impedance match of the matching system can be monitored. The matching system can include and/or be coupled (e.g., communicatively) to a controller that receives measurements from one or more matching sensors which monitor the state of the impedance match of the system. The controller can feed a command signal back to the RF amplifier and can thus control the operating frequency of the matching system. The matching sensor can be placed at the input of the matching network to monitor the state of the impedance match. Other embodiments may optionally have additional matching sensors placed at other locations in the circuit, such as the inputs of the plurality of loads, or the input of the RCN. Based on the current and previous state of impedance match, the controller determines the system operating frequency that achieves the best impedance match, and provides a command signal to the RF amplifier to produce the required RF signal at the desired frequency.

[00103] At block 608, an operating frequency of the matching system can be varied to maintain a matched input impedance characteristic. In an embodiment, the controller can generate a command signal to provide to the RF amplifier to adjust the operating frequency of the matching system. The networks (i.e., impedance transformation and reactance compensation networks) can be configured to dynamically null the reactance of the loads. For example, the impedance at each of the loads includes a resistive component and a reactive component and the impedance can vary in both, resistance and reactance, as operating conditions change (e.g., as properties of plasma within the plasma container change). The reactance compensation networks of the impedance transformation and reactance compensation networks can perform dynamic frequency tuning, for example using small variations in the matching system operating frequency to adaptively tune the matching system such that the input impedances at the reactance compensation networks appear substantially resistive.

[00104] In an embodiment, the matching system can provide efficient power delivery into plasma loads and good loading of the power amplifier, while achieving small size, low cost and high bandwidth, and eliminating mechanical or switched components. For example, the matching system utilizes loads having a particular arrangement with respect to each other, along with the RCN to compress the variable plasma impedance to an almost fixed value that results in minimal reflected power to the power amplifier. By arranging the loads in certain geometries such that they are magnetically uncoupled, the RCN can be used to provide passive, high bandwidth matching with no mechanically- controlled or moving parts. This can permit the design of efficient, compact, fast- response, and cost-effective plasma generation systems that provide good matching across an entire range of plasma operation.

[00105] Having described preferred embodiments, which serve to illustrate various concepts, structures and techniques, which are the subj ect of this patent, it will now become apparent that other embodiments incorporating these concepts, structures and techniques may be used. Accordingly, it is submitted that the scope of the patent should not be limited to the described embodiments but rather should be limited only by the spirit and scope of the following claims.