А Valeriy - The effect of discharge chamber geometry on the characteristics of low-pressure rf capacitive discharges - страница 1
IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 35, NO. 2, APRIL 2007
The Effect of Discharge Chamber Geometry on the Characteristics of Low-Pressure
RF Capacitive Discharges
Valeriy A. Lisovskiy, Jean-Paul Booth, Karine Landry, David Douai, Valerick Cassagne, and Vladimir D. Yegorenkov
Abstract—We report the measured extinction curves and current-voltage characteristics (CVCs) in several gases of RF capacitive discharges excited at 13.56 MHz in chambers of three different geometries: 1) parallel plates surrounded by a dielectric cylinder ("symmetric parallel plate"); 2) parallel plates surrounded by a metallic cylinder ("asymmetric confined"); and 3) parallel plates inside a much larger metallic chamber ("asymmetric unconfined"), similar to the gaseous electronics conference reference cell. The extinction curves and the CVCs show differences between the symmetric, asymmetric confined, and asymmetric unconfined chamber configurations. In particular, the discharges exist over a much broader range of RF voltages and gas pressures for the asymmetric unconfined chamber. For symmetric and asymmetric confined discharges, the extinction curves are close to each other in the regions near the minima and at lower pressure, but at higher pressure, the extinction curve of the asymmetric confined discharge runs at a lower voltage than the one for the discharge in a symmetric chamber. In the particular cases of an "asymmetric unconfined chamber" discharge or "asymmetric confined" one, the RF discharge experiences the transition from a "weak-current" mode to a "strong-current" one at lower RF voltages than is the case for a "symmetric parallel-plate" discharge.
Index Terms—Chamber configuration, current-voltage characteristics (CVCs), extinction curve, RF capacitive discharge.
LOW-PRESSURE RF capacitive discharges are widely used for etching and surface modification of various materials, for depositing oxides, nitrides and other films, for cleaning
Manuscript received November 18, 2005; revised November 23, 2006. This work was supported by the Unaxis France—Displays Division, Palaiseau, France.
V. A. Lisovskiy was with the Laboratoire de Physique et Technologie des Plasmas, Ecole Polytechnique, 91128 Palaiseau, France. He is now with the Department of Physics and Technology, Kharkov National University, 61077 Kharkov, Ukraine.
J.-P. Booth is with the Laboratoire de Physique et Technologie des Plasmas, Ecole Polytechnique, 91128 Palaiseau, France, and also with the Lam Research
Corporation, Fremont, CA 94538 USA.
K. Landry is with the Unaxis, 38100 Grenoble, France.
D. Douai is with the Association Euratom-CEA, Departement de Recherches sur la Fusion Controlee, CEA Cadarache, 13108 Saint Paul lez Durance, France.
V. Cassagne is with the Riber, 95873 Bezons, France.
V. D. Yegorenkov is with the Department of Physics, Kharkov National University, 61077 Kharkov, Ukraine.
Digital Object Identifier 10.1109/TPS.2007.893261 deposition chambers, in plasma chemistry, and for medical tool sterilization.
A great variety of R&D problems under solution and a large number of research groups occupied with studying the properties of RF capacitive discharge in experiment resulted in a series of discharge vessels differing not only in the electrode diameter and interelectrode gap but also in the design (geometry) of the vessels. Four types of discharge vessels are mostly applied. The parallel-plate RF discharge, in which flat electrodes are surrounded by a dielectric cylinder (in the case of a cylindrical shape with flat electrodes), is the simplest design -. Here, the discharge can burn only inside the interelectrode gap. The asymmetric confined RF discharge, where in the plane-parallel design of the electrodes a metal cylinder connected to one of the electrodes (conventionally to a grounded one) is used instead of a dielectric one, may be attributed to the second type -. The plasma enhanced chemical vapor deposition devices (aimed for depositing semiconductor and dielectric films), in which the vessel has a rectangular design, are attributed to the same type -. We can attribute to the third type the design when the flat electrodes (of comparatively small size) are introduced inside the grounded vessel of large dimensions -. The fourth type—gaseous electronics conference (GEC) reference cell—is similar to the third type (flat electrodes in a large vessel) -. In the vessels of the third and fourth types, the RF discharge can usually burn not only inside the gap between the flat electrodes but also outside it in a large vessel with grounded walls. Obviously, with the fixed values of the gas pressure, the RF field frequency, RF voltage amplitude across the electrodes, and the interelectrode gap RF discharge characteristics (discharge current, delivered power, inner plasma characteristics as well as the extinction curve of the discharge) in different vessels will not be identical.
In our recent paper , we presented the breakdown curves of the RF discharge in chambers of different geometry. Breakdown curves, of course, are very important for estimating the RF voltage values at which a gas breakdown occurs and a self-sustained discharge ignites. However, they give no information on the range of RF voltage and gas pressure values within which a discharge can be sustained. It is just an extinction curve that shows under which conditions a discharge burns in a chamber of certain geometry and when the technological processes can be performed. RF discharge extinction in a symmetric chamber was studied in rather sufficient detail (see  and papers cited
0093-3813/$25.00 © 2007 IEEE
therein). However, the available references do not contain the data on extinction curves of the RF discharge registered for identical electrodes placed in discharge chambers of different geometries.
The effect of chamber geometry on current-voltage characteristics (CVCs) and charge burning regimes remains not to be studied in full. In the papers - listed above, the research was performed with a fixed chamber geometry. Boeuf and Merad  obtained numerically the distributions of potential and ionization rate for GEC cell-like chamber and for asymmetric confined discharge. Beneking  registered the CVCs for two electrodeless discharge configurations in which the field-supplying electrodes were separated from the discharge volume by a dielectric wall, as well as for a symmetric parallel-plate chamber with inner electrodes.
The objective of this paper, therefore, is to study how the discharge chamber geometry affects the shape of the extinction curves, CVCs, and burning modes for RF capacitive discharges. We have studied three different chamber geometries: 1) parallel plates surrounded by a dielectric cylinder ("symmetric chamber"), 2) parallel plates surrounded by a metallic cylinder ("asymmetric confined chamber"), and 3) parallel plates inside a much larger metallic chamber ("asymmetric unconfined chamber"), similar to the GEC reference cell.
We observed that the symmetric, asymmetric confined, and unconfined chambers have different extinction curves and RF CVCs. For the asymmetric unconfined chamber, the discharge exists over a much broader range of RF voltages and gas pressures. In the case of the asymmetric confined discharge, the extinction curve at higher pressure runs below one for the discharge in the symmetric chamber. However, near the minima for the symmetric and asymmetric confined discharges and to the left of them, the extinction curves are close to each other. For the discharge transition to a strong current mode in the "symmetric parallel-plate" case, there are required higher RF voltages than those for "asymmetric unconfined chamber" and "asymmetric confined" cases.
II. Experimental A. Experimental Setups
Experiments were performed for three different configurations, which we shall denominate as the following: 1) the symmetric chamber, 2) the asymmetric confined chamber, and 3) the asymmetric unconfined chamber.
The vacuum vessel consists of a 315-mm diameter and 231-mm high steel chamber, with a view port to observe breakdown. Two parallel 143-mm diameter aluminum electrodes are installed in this vessel. The upper (powered) electrode is 10-mm thick and is separated from a grounded shield (20-mm thick) by a layer of dielectric material. The lower (grounded) electrode is 30 mm thick. The gas was input through small holes in the upper (powered) electrode and evacuated via the external chamber. The gas pressure was monitored with capacitance manometers (10 and 1000 torr, MKS Instruments). A constant gas flow of 5 sccm was set with a mass flow controller. The gas pressure was set by a feedback-controlled valve on the pumping outlet.
Fig. 1. Schematic of the symmetric RF discharge.
Fig. 2. Schematic of the asymmetric unconfined chamber configuration.
In the "symmetric chamber" configuration, the two electrodes were surrounded by a 145-mm internal diameter fused silica tube, so that the effective discharge volume resembles Fig. 1. In the "asymmetric unconfined chamber" configuration (Fig. 2), the same electrodes were used, but the fused silica tube was removed so that the electrodes were located within a metallic chamber with grounded walls. Similar chamber configurations are widely used for studying the characteristics of the RF discharges (see, e.g., -). Experiments were performed in argon, nitrogen, and hydrogen over the pressure range p « 0.01-10 torr with an RF field frequency of f = 13.56 MHz. The extinction curves and the CVCs of the discharge were recorded for interelectrode gap values of L = 11.9 mm and L = 27 mm.
"Symmetric chamber" was also used to achieve the third "asymmetric confined chamber" configuration (Fig. 3). In this case, the inner surface of the discharge tube was covered with a grounded aluminum foil. A gap of 2 mm was left between the foil edge and the surface of the RF electrode. Similar chambers (from the viewpoint of the asymmetric distribution of the vacuum RF field) have also been widely used in other studies of RF discharges -.
B. Electrical Measurements
An RF voltage-current probe (Advanced Energy Z'SCAN) was used to measure the RF voltage Urf, the RF current Irf,
Fig. 3. Schematic of the asymmetric confined chamber, with foil on the radial walls.
Fig. 4. Equivalent circuit (a) for the symmetric RF discharge and (b) for the discharge with a large stray capacitor (asymmetric unconfined chamber).
the phase shift ф between the RF voltage and RF current and the delivered power. It was situated as close as possible to the RF electrode. The RF power (13.56 MHz) was supplied by an RF generator (RF Power Products Inc. RF5S) via an L-type matchbox (Huttinger Elektronik Gmbh PFM).
In an asymmetric unconfined chamber (Fig. 3), the RF voltage-current probe was located outside a large grounded chamber, and the electrodes were inside it. In this case, a large stray capacitance between the RF electrode and the grounded chamber walls affects the measurements of electric characteristics of the discharge considerably made with an external RF voltage-current probe (RF current amplitude Irf, phase shift angle ф between the RF voltage and current, and discharge impedance Z). In order to decrease the effect of the stray capacitance on electric measurements, Miller and co-workers ,  and Sobolewski ,  proposed to include an external shunt into the discharge circuit, thus removing a large displacement current and allowing to obtain more accurate values of the current Irf, phase shift angle ф, and discharge impedance Z.
Let us consider the way the stray capacitance affects the electric measurements. Fig. 4(a) shows the equivalent circuit of the RF discharge in a symmetric chamber, where R and L are the resistance and inductance of the discharge plasma and C is the capacitance of the near-electrode sheaths. For the given circuit, the impedance Z0, the current amplitude I0 ,the
phase shift angle ф0, and the active current I0 • cos ф0 for the symmetric chamber are equal to
I0 = Щ = Ur'/\l"+^£f
cos фо = RI ]j R2 + (^ujL - ^jc^j~
I0 • cos ф0 = Urf • R / R2 + ( ujL
(1) (2) (3) (4)
where Urf is the RF voltage amplitude and и = 2nf is the angular frequency of the RF electric field.
For the equivalent circuit of the discharge with the asymmetric unconfined chamber [Fig. 4(b)] possessing a large stray capacitance Cst, we have the following expressions:
cos ф = R/Q
(5) (6) (7)
I• cos ф = Urf • RJ R2 + [ujL-
= Iq • cos фо.
It follows from formulas (1)-(9) that the stray capacitance Cst affects the measurements of the impedance Z, RF current amplitude Irf, and phase shift angle ф, but does not affect the measurements of the active current Irf • cos ф,aswellasofthe delivered power
2 Urf Irf cos ф.
The same conclusion may be drawn considering the results by Sobolewski [25, Table II]. Without the external shunt, he obtained Irf = 1.816 A and ф = -88.3° ± 0.6°, whereas in its presence, he got Irf = 0.24 A and ф = -75.7° ± 0.6°. Then, without the shunt, we have Irf • cos ф = 53.9 mA (more precisely, taking into account the error in measuring the phase shift, the active current values were within the range of 3566.6 mA). With the shunt, we have Irf • cos ф = 59.3 mA (more precisely, within the range of 56.8-61.7 mA). Thus, the value of the active RF current with the shunt and without it happened to be constant to the accuracy of measuring the phase shift angle. The values of the delivered power  also almost coincided: without the shunt, Pd\v = 2.66 W, whereas
10 U-■ ■ ■ ■ ■ -■ ■ » . f..l-■ ■ ■ ■ .»■■!
0.01 0.1 1 10
p, Ton-Fig. 5. Extinction curves of the RF discharge in argon at L = 11.9 mm: Full circles depict the results for the symmetric discharge; empty circles depict the results for the asymmetric unconfined chamber, respectively.
with it, Pdlv = 2.88 W. In our paper studying the regimes of discharge burning in chambers of different geometry, we were primarily interested in registering the active RF current, but not the current amplitude, phase shift angle, or impedance separately; therefore, we did not apply an external shunt in our measurements.
Let us consider how the discharge chamber design affects the extinction curves and the CVCs of an RF discharge. The region of existence of a discharge is limited by its extinction curve. The extinction curves of an RF discharge in argon are shown in Fig. 5. For pressures greater than 0.4 torr, the extinction curves for the symmetric and asymmetric unconfined chambers coincide and possess a single minimum. However, at a lower pressure, the extinction curve for the symmetric chamber deviates abruptly to higher voltages and even possesses a multivalued section. Similar behavior of the extinction curves of the symmetric discharge was also observed in . The RF voltage for extinction in the asymmetric unconfined chamber increases, but only slowly, as the argon pressure is lowered. For a tenfold decrease of the argon pressure (from 0.2 to 0.02 torr), the extinction RF voltage increases by only 10 V.
For nitrogen (Fig. 6), the extinction curves of the symmetric and asymmetric unconfined chambers coincide for pressures above 1 torr. As the pressure is lowered, the extinction curve of the symmetric chamber, after passing through the minimum, deviates to higher RF voltages. At the same time, the extinction curve in the asymmetric unconfined chamber, after passing the minimum, drops to lower RF voltages and approaches a second minimum at p ~ 0.05 torr.
Figs. 5 and 6 show that, at low gas pressure, a broad range of RF voltages exists within which a discharge cannot be sustained in the symmetric chamber, but may exist in the asymmetric unconfined chamber. For example, in Fig. 6, we see that a discharge can be sustained in the asymmetric unconfined chamber with a nitrogen pressure of 0.05 torr and an RF voltage of 60 V, whereas a discharge cannot exist in the symmetric chamber at this RF voltage for any pressure.
0i-.........-■ ■ ■ ......-.........
0.01 0.1 1 10
Fig. 6. Extinction curves of the RF discharge in nitrogen at L =11.9 mm: Full circles depict the results for the symmetric discharge; empty circles depict the results for the asymmetric unconfined chamber, respectively. The a-j transition curves for the RF discharge (UaY) for the symmetric (full triangles) and asymmetric unconfined chambers (empty triangles).
Fig. 7. Active RF current against RF voltage for the symmetric RF discharge at different nitrogen pressures, L =11.9 mm.
Now, let us consider the CVCs of discharges in the two chambers. Specifically, we shall consider the dependence of the active RF current Irf • cos ф on the RF voltage. The experiments were performed for nitrogen at L = 11.9 mm. Fig. 7 shows the CVC for the symmetric chamber. For gas pressures less than 0.15 torr, a small increase in the RF voltage causes a rapid increase of the active RF current, and the discharge is always in a high-current mode. At higher gas pressure, the rate of increase of the RF current is somewhat lower, and a weak-current mode exists at low voltage (in agreement with the conclusions in  and , the weak-current mode cannot exist at too low gas pressure). Dogleg features are clearly observed, indicating the transition from the weak-current (a-) mode to the strong-current (7-) one. Usually, the discharge covers the entire surface of the electrodes, except at pressures above 0.8 torr and low RF voltages when the discharge was in the weak-current "normal" discharge regime.
Fig. 8 shows the CVC of the asymmetric unconfined discharge (with the electrodes in the chamber). At low pressure (p « 0.01 torr), the discharge occurred solely in the large chamber (outside of the electrode gap) over the whole range
О 100 200 300 400 500 600 700
Fig. 8. Active RF current against RF voltage for the asymmetric unconfined chamber at different nitrogen pressures, L =11.9 mm.
of the RF voltage studied. At higher gas pressure, as the RF voltage is increased, the discharge occurred first in the large chamber and, then, (at a sufficiently high RF voltage) also penetrated the gap between the planar electrodes. This was accompanied by an abrupt increase of the active RF current. Such a situation was observed, for example, at p « 0.023 torr and Urf « 600 V, and at p « 0.038 torr and Urf « 260 V(see Fig. 8). At a pressure of ~0.1 torr, the discharge was ignited both in the outer chamber and in the interelectrode gap. For pressures below 1 torr, the discharge in the gap is nonuniform across the electrode surface, having a toroidal shape located close to the radial boundaries of the electrodes. On increasing the RF voltage and gas pressure, the discharge glow expanded toward the electrode axis. In the strong-current mode, a bright glow was observed only near the boundary of the RF electrode sheath, whereas the glow was weak at the grounded electrode (as well as near the planar surface of the electrode and other grounded parts of the large chamber). However, at higher pressure (p> 1 torr), the glow intensity near the grounded surfaces increases. At p> 1 torr, the gas breakdown occurs in the gap between the electrodes. The discharge at the smallest voltages (near the extinction voltage) was in the normal regime having the same toroidal shape. With an increasing RF voltage, the discharge also expanded into the outer chamber and transformed into the anomalous mode. Nonuniform radial distribution of plasma parameters in GEC cell was predicted in theory in  and , whereas the paper of Overzet and Hopkins  demonstrated in experiment the presence of the plasma density maximum near the radial boundary of electrodes in agreement with the results of our paper.
Fig. 9 compares the CVCs for the symmetric and asymmetric unconfined chambers at p = 0.3 torr and p = 1.52 torr. The smallest RF voltage that could maintain a discharge at p = 0.3 torr was about 350 V, whereas the discharge in the asymmetric unconfined chamber was already in the strong-current mode. At a nitrogen pressure p = 1.52 torr, the discharges in both the symmetric and asymmetric unconfined chambers had approximately the same extinction voltage, and at the smallest voltages, they operated in the normal regime, with similar values of the discharge current. However, the active current of the discharge in the asymmetric unconfined chamber increased
Fig. 9. Active RF current against RF voltage at nitrogen pressure p = 0.3 torr (full and empty circles) and p =1.52 torr (full and empty triangles) for the symmetric RF discharge (full circles and triangles) and for the asymmetric unconfined chamber (empty circles and triangles), gap width L = 11.9 mm.