Download Operating a radio-frequency plasma source on water vapor

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Plasma display wikipedia , lookup

Transcript
REVIEW OF SCIENTIFIC INSTRUMENTS 80, 083503 共2009兲
Operating a radio-frequency plasma source on water vapor
Sonca V. T. Nguyen,1 John E. Foster,2 and Alec D. Gallimore1
1
Plasmadynamics and Electric Propulsion Laboratory, University of Michigan, Ann Arbor,
Michigan 48108, USA
2
Plasma Science and Technology Laboratory, University of Michigan, Ann Arbor, Michigan 48108, USA
共Received 18 May 2009; accepted 17 July 2009; published online 24 August 2009兲
A magnetically enhanced radio-frequency 共rf兲 plasma source operating on water vapor has an
extensive list of potential applications. In this work, the use of a rf plasma source to dissociate water
vapor for hydrogen production is investigated. This paper describes a rf plasma source operated on
water vapor and characterizes its plasma properties using a Langmuir probe, a residual gas analyzer,
and a spectrometer. The plasma source operated first on argon and then on water vapor at operating
pressures just over 300 mtorr. Argon and water vapor plasma number densities differ significantly.
In the electropositive argon plasma, quasineutrality requires ni ⯝ ne, where ni is the positive ion
density. But in the electronegative water plasma, quasineutrality requires ni+ = ni− + ne. The positive
ion density and electron density of the water vapor plasma are approximately one and two orders of
magnitude lower, respectively, than those of argon plasma. These results suggest that attachment and
dissociative attachment are present in electronegative water vapor plasma. The electron temperature
for this water vapor plasma source is between 1.5 and 4 eV. Without an applied axial magnetic field,
hydrogen production increases linearly with rf power. With an axial magnetic field, hydrogen
production jumps to a maximum value at 500 W and then saturates with rf power. The presence of
the applied axial magnetic field is therefore shown to enhance hydrogen production. © 2009
American Institute of Physics. 关DOI: 10.1063/1.3202250兴
I. INTRODUCTION
Plasma discharges operated on water vapor, in either liquid or vapor form, can potentially be used in a wide range of
applications.1–10 For example, the decomposition of organic
contaminants in waste water, such as phenol, into more benign byproducts has been studied using a pulsed streamer
corona discharge.1 In the intermediate pressure regime between tens of millitorr to one torr, water vapor plasmas are
used in plasma-assisted chemical vapor deposition for diamond film formation. Currently, CH4 and H2 are injected into
a plasma chamber to fabricate diamond films. When O2 is
introduced, the presence of O2 and OH radicals is stipulated
to increase the rate of diamond growth. Instead of O2, other
researchers have shown that addition of H2O vapor in the
main gas feed favorably contributes to diamond growth.6,7
Ultraviolet 共UV兲 light source plasma is another example of
an application of water vapor plasma. Oh et al.10 investigated
the potential of using water vapor plasmas excited by microwaves as a UV light source.
In the work presented here, the potential use of a radiofrequency 共rf兲 plasma discharge to directly dissociate water
molecules for hydrogen production is investigated to address
the need to reduce greenhouse gas emission and humanity’s
dependence on nonrenewable natural resources. Hydrogen as
an energy carrier has the potential to address many aspects of
the energy problem, particularly in the transportation sector.
Currently, 96% of hydrogen is produced from steam methane
reformation, gasification, or oxidation. In an effort to address
the energy and climate challenge, the production of hydrogen
via renewable methods is researched. Hydrogen produced
0034-6748/2009/80共8兲/083503/8/$25.00
from electrolysis is still more expensive than hydrogen produced from reformation methods. Presently, only 4% of the
total hydrogen produced worldwide is from electrolysis. It is
argued that the widespread adoption of conventional electrolysis systems has a physical limitation. In these systems,
hydrogen must diffuse in liquids, but their diffusion rate in
liquids is slow compared to their diffusion rate in a gas.11
Chaffin et al.11 investigated hydrogen production via
plasma electrolysis by placing metal electrodes above a liquid water surface. Through this proposed method, the cost of
electrolytes and catalysts that are required in conventional
electrolysis are removed and there is even the potential to
increase hydrogen production rates. Yan et al.12 have produced hydrogen by using a plasma discharge operating on
methanol solutions. Methanol is chosen instead of water because methanol is easier to vaporize. Both Chaffin and Yan
attribute the presence of high-energy electrons as the key in
electrolyzing water or methanol.
In plasma electrolysis, plasma processing of the liquid is
localized to the region within the gas bubbles around the
electrodes. In contrast, the present work proposes to investigate a method of hydrogen production by dissociating water
vapor 共as opposed to liquid water兲 in a high-density inductive plasma source. Givotov et al.13 investigated dissociating
water vapor in a plasma using a microwave discharge in the
1980s. In their work, the plasma source was a quartz tube, 20
mm in diameter and 30 mm in length. The operating pressure
was 50 torr and the power input was 2 kW at 2400 MHz.
Similarly, this current work characterizes the behavior of a rf
plasma discharge operating on water vapor, and evaluates its
potential application as a renewable hydrogen production
80, 083503-1
© 2009 American Institute of Physics
Downloaded 25 Jun 2010 to 141.212.191.3. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/rsi/copyright.jsp
083503-2
Rev. Sci. Instrum. 80, 083503 共2009兲
Nguyen, Foster, and Gallimore
FIG. 1. A photograph of the quartz tube and three solenoid magnets with
one pulled away to show the antenna.
method. The plasma source used in this experiment has a
diameter of 15 cm, a length of 50 cm, and operates at 13.56
MHz. This source also has an electromagnet to allow the
investigation of an applied magnetic field on hydrogen production. With the applied magnetic field, dissociative attachment is speculated to be enhanced by greater electron confinement. Although the current system is not optimized for
energy efficiency, it has a potential to achieve higher efficiency than the method employed by Givotov using a microwave discharge because of this axial applied magnetic field.
This rf source in principle is simpler than a microwave system, as it only requires a 13.56 MHz power supply. These
supplies have efficiencies greater than 90% compared to
70%–85% of the microwave supplies. Further, the rf system
does not require any waveguide, reducing the complexity of
the overall system.
II. EXPERIMENTAL SETUP
Testing of the rf water vapor plasma source is performed
at the plasmadynamics and electric propulsion laboratory
共PEPL兲 in the cathode test facility 共CTF兲. The vacuum facility utilized is a cylindrical aluminum chamber that is 2.44 m
in length and 0.61 m in diameter. An Edwards XDS 35i dry
pump is used to evacuate the chamber. With a maximum
pumping speed of 35 m3 / h on N2, the base pressure is below 3 mtorr. The plasma source is attached to a 15-cmdiameter side port on the CTF.
A. Plasma discharge
A photograph of the plasma source is shown in Fig. 1.
The source consists of a quartz tube that is 15 cm in diameter
and 50 cm in length. Three magnetic coils connected in series to a 60 A dc power supply produces a peak axial magnetic field on axis of the plasma source of approximately 200
and 400 G at 30 and 60 A magnet current, respectively. Figure 2 shows the mapping of the magnetic field for the 60 A
magnet current setting in the horizontal plane of the quartz
tube at the centerline. The outline of the quartz tube and the
flange that is used to connect the quartz tube to the CTF
vacuum chamber are also shown in this figure. The magnetic
field is measured by a three-axis Hall probe that is connected
to a three-channel Gaussmeter from Lakeshore 共model 460兲.
The mapping is achieved by placing the Hall probe on three
motion tables. The magnetic field mapping for a magnet cur-
FIG. 2. 共Color online兲 Mapping of magnetic field generated from three
solenoid magnets at 60 A magnet current with an outline of the quartz tube
and an adapter flange.
rent of 30 A is exactly identical to that at 60 A, with the
exception that the magnetic field strength is reduced in half.
A double helical antenna circumscribes the quartz tube, and
is sandwiched between the quartz tube and the magnetic
coils, as shown in Fig. 3. A 13.56 MHz rf power supply is
used to excite the antenna up to 1.5 kW. A ␲-matching network is employed to match the impedance of the rf power
supply output with the impedance of the antenna, reducing
the reflected power to less than 5% of input power for all
conditions investigated.
B. Water feed system
An Eldex water pump 共a positive displacement, reciprocating piston兲 with a flow rate range of 0.01–20 ml/min 关12.5
SCCM 共SCCM denotes cubic centimeter per minute at STP兲
to 25 slm H2O兴 is used to meter liquid water into the plasma
source. A back pressure regulator from Upchurch Scientific
is placed in the gas line between the water pump and the
vacuum chamber to maintain a pressure of 250 psi 共1723
kPa兲. At room temperature, water is in the vapor phase at a
pressure of hundreds of millitorr. Nonetheless, the line between the water pump and the quartz tube is heated with
heating tape up to 400 K to ensure that water molecules are
FIG. 3. Diagram of plasma source showing the rf power supply and matching network connecting to the double-helix antenna, three solenoid magnets
connecting in series to a dc power supply, and the water pump connecting to
the quartz tube.
Downloaded 25 Jun 2010 to 141.212.191.3. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/rsi/copyright.jsp
083503-3
Rev. Sci. Instrum. 80, 083503 共2009兲
Nguyen, Foster, and Gallimore
FIG. 4. 共Color online兲 共〫兲 A typical Langmuir probe curve for water vapor
plasma operating with 75 SCCM H2O and 10 SCCM Ar at 500 W rf power
and 30 A magnet current. 共䊊兲 Probe current squared vs voltage for the same
plasma, indicating the slopes in the ion and electron saturation regions taken
to calculate ion and electron densities, respectively.
not absorbed onto the metal surface of the feed line and that
water remains in vapor state upon entry into the plasma
source.
C. Langmuir probe
A commercial rf-compensated single Langmuir probe
共LP兲 system from Hiden Corporation is used to obtain
plasma density, floating potential, plasma potential, and electron temperature. This commercial system includes a data
acquisition software package, rf compensation circuitry, a
driver power supply, and an ammeter. The LP collector is
0.12 mm in diameter and 3 mm in length.
The probe sheath thickness is determined to be approximately over an order of magnitude smaller than the local
electron and ion mean free paths. Therefore, a collisionless
probe model is used. Further, magnetic effects are neglected
because rLE ⬎ r P and rLE Ⰷ ␭D, where rLE is the Larmor radius of the electron, r P is the probe radius, and ␭D is the
Debye length. LP data are analyzed using standard methods
reported in a number of references.14–18 Figure 4 shows a
typical LP trace. As shown, the floating potential is taken as
the biased voltage where the net current to the probe is zero.
The ion and electron number densities are calculated using the orbital motion limited analysis model because the
ratio of the probe radius to Debye length, r P / ␭D, is less than
three for all operating conditions. This technique of calculating the number density assumes that the electron distribution
is isotropic, Maxwellian, and that the sheath is
collisionless.17,18
冑 冉 冊
冑
ni =
␲
Ap
ne =
␲
Ap
M i dI2i
,
2e3 dV
共1兲
me dI2e
.
2e3 dV
共2兲
In Eqs. 共1兲 and 共2兲, ni is the ion number density, A p is the
probe surface area, M i is the ion mass, ne is the electron
number density, me is the electron mass, and e is the electron
FIG. 5. 共Color online兲 Semilog plot of Langmuir probe trace showing the
linear region where electron temperature is calculated and the point where
the plasma potential is derived from.
charge. Finally, dI2i / dV and dI2e / dV are the slopes of current
squared versus bias voltage for ion and electron, respectively,
as shown in Fig. 4.
In this work, an effective ion mass, M i, is calculated
from a weighted average of the product of mass 共mi兲 and
partial pressure 共pi兲 obtained from the residual gas analyzer
共RGA兲, which is to be discussed in Sec. IID,
Mi =
兺 im i p i
.
兺i pi
共3兲
Finally, the electron temperature is derived from the
slope of the semilog plot, and the plasma potential is the
biased voltage where the two lines intersect in the semilog
plot, as shown in Fig. 5.
D. Differential pump system and RGA
A commercial RGA 共Stanford Research Systems
RGA100兲 is used to identify gas species inside the chamber.
The maximum operating pressure of the RGA is 10−5 torr,
significantly lower than water vapor plasma pressure of a
few hundred millitorr. Therefore, a differential pump system
is utilized to reduce the pressure in the RGA chamber to
allow for RGA operation. Figure 6 depicts the layout of this
FIG. 6. Differential pump system including a variable leak valve, 70 l/s
turbo pump, RGA, and pressure transducer.
Downloaded 25 Jun 2010 to 141.212.191.3. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/rsi/copyright.jsp
083503-4
Rev. Sci. Instrum. 80, 083503 共2009兲
Nguyen, Foster, and Gallimore
FIG. 7. 共Color online兲 Calibration curves for hydrogen, oxygen, and water
relating the RGA partial pressure ratio to the flow rate ratio.
system. A variable leak valve isolates the RGA chamber
from the plasma chamber.
Without a proper calibration, the RGA only yields qualitative information, i.e., the presence of certain gas species in
the plasma. To estimate the amount of hydrogen along with
other gases produced in the plasma quantitatively, the RGA
is calibrated by injecting a known amount 共10 SCCM兲 of
argon. A calibration factor 共CF兲 relating the ratio of the partial pressure of the gas of interest to the partial pressure of
argon could therefore be obtained from the ratio of the flow
rates of the two gases. Figure 7 shows the calibration curves
of hydrogen, oxygen, and water. The pressure ratio varied
linearly with mass flow rate ratio as expected. With CF
known for each gas, the mass flow rate of a certain gas
produced in the water vapor plasma source is calculated using Eq. 共4兲.
ṁi =
ṁAr Pi
CF PAr
共4兲
In this equation, PAr is the partial pressure of argon, Pi is
the partial pressure of the gas of interest, ṁAr is the mass
flow rate of Ar, and ṁi is the rate of production of the gas of
interest. Effects of the addition of argon to the plasma properties and hydrogen production have been studied. As can be
seen in Fig. 8, the rates of hydrogen production for the addition of 5 and 10 SCCM Ar are almost identical at all rf
power levels except for 750 W. Even at this power level,
however, the two results are within the uncertainty of the
measurement technique. Furthermore, Figs. 9 and 10 illustrate the invariance of electron density, ion density, plasma
potential, and floating potential to argon flow rate. Therefore,
the addition of 10 SCCM Ar to calibrate the RGA is shown
to have a negligible effect on the plasma.
FIG. 8. 共Color online兲 Hydrogen production for 75 SCCM H2O with 5 and
10 SCCM Ar operating with 30 A magnet current.
light into a fiber optic cable that is connected to the spectrometer via a SubMiniature version A⫽SMA connector.
III. RESULTS
In this section, results from a plasma source operating on
argon and on water vapor are presented. The operating pressure for the two cases is ⬃300 mtorr. Results given in this
paper are the mean values of a sample of measurements.
Error bars represent one standard deviation among the spread
in the measurements sample. For the Langmuir probe data,
the sample size includes four sets of 5–10 Langmuir probe
traces. The measurement is taken in a random order of rf
power. While the absolute error in the Langmuir probe diagnostics is 20%–30% for electron temperature and 50%–60%
for plasma densities,19 the relative error from point to point
is small, which allows for a study of trends in plasma properties.
For the RGA results, the sample size is four sets of three
spectra, and the error bars also represent one standard deviation of spread in each set of measurement. Based on the
calibration of the RGA, the error for determining hydrogen
and oxygen production is 5%–10%, and for determining water vapor flow rate is 10%–20%.
E. Spectrometer
In addition to the RGA, two spectrometers from Ocean
Optics are used to obtain optical emission spectra from the
plasma: USB2000 and HR4000. The USB2000 model spectrometer has a 25 ␮m entrance slit and a detector wavelength range of 200–850 nm. The HR4000 model spectrometer has a 5 ␮m entrance slit and a detector wavelength
range of 400–850 nm. A collimating lens is used to focus
FIG. 9. 共Color online兲 Ion and electron densities for 75 SCCM H2O with 5
and 10 SCCM Ar operating with 30 A magnet current.
Downloaded 25 Jun 2010 to 141.212.191.3. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/rsi/copyright.jsp
083503-5
Rev. Sci. Instrum. 80, 083503 共2009兲
Nguyen, Foster, and Gallimore
FIG. 10. 共Color online兲 Floating potential and plasma potential for 75
SCCM H2O with 5 and 10 SCCM Ar operating with 30 A magnet current.
FIG. 12. 共Color online兲 Plasma and floating potentials of argon plasma
operating with 400 SCCM Ar. The plasma potential is consistently 5 V
higher than the floating potential.
A. Argon plasma
The ion densities for the argon plasma as a function of rf
power for 0, 30, and 60 A magnet current setting are shown
in Fig. 11. The magnetic field strength does not influence the
ion density within the certainty of the measurement. The
magnetic field also does not appreciably influence the plasma
or floating potential, as shown in Fig. 12. These potentials
increase linearly as a function of rf power. The plasma potential increases from 21 V at 250 W to 27 V at 1000 W.
Similarly, the floating potential increases from 17 V at 250
W to 24 V at 1000 W. The plasma potential is consistently 5
V higher than the floating potential. Finally, the electron temperature ranges between 1.2–1.6 eV. Experimental results of
electron temperature, plasma potential, and floating potential
are consistent with the following relationship:
冉冑 冊
V p − V f = ln
Mi
Te ,
me
共5兲
where Te is the electron temperature, V p is the plasma potential, and V f is the floating potential.
B. Water plasma
Figure 13 shows two photographs of the water plasma
operated at 500 W rf power and at 0 and 60 A magnet current
FIG. 11. 共Color online兲 Plasma density of argon plasma operating with 400
SCCM Ar.
setting. In the case of 0 A magnet current, the plasma is
diffused throughout the quartz tube, but in the case of 60 A
magnet current, the plasma is more concentrated near the
edge of the quartz tube. The outline of the antenna is clearly
observed in the presence of the applied magnetic field.
Argon and water vapor plasma number densities differ
significantly. Figures 14 and 15 show the electron and ion
densities for the water vapor plasma. Here, both electron
density and ion density are presented because attachment and
dissociative attachment are expected to be nontrivial in electronegative water plasma. Under a similar operating pressure, the ion density for the water vapor plasma is approximately one order of magnitude less than that of the argon
plasma. For the electron density, the difference is approximately two orders of magnitude. The reduced density in the
water vapor plasma is attributed to the increase in number of
modes of energy absorption relative to atomic argon including rotation, vibration, and dissociation. It is speculated that
electron density of the water vapor plasma is lower than the
calculated ion density because of attachment driven by high
electron affinity species such as atomic oxygen and hydroxyl
radicals.
For the water vapor plasma, both plasma and floating
potentials are higher than those of the argon plasma in the
presence of a magnetic field as shown in Fig. 16. Without a
magnetic field, there is a negligible effect of rf power on both
the plasma and floating potentials. However, at 30 and 60 A
magnet current, there is an increase in both plasma and float-
FIG. 13. 共Color online兲 Photographs of water plasma operating with 75
SCCM H2O and 10 SCCM Ar at 500 W rf power and 共a兲 0 A magnet current
and 共b兲 60 A magnet current.
Downloaded 25 Jun 2010 to 141.212.191.3. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/rsi/copyright.jsp
083503-6
Rev. Sci. Instrum. 80, 083503 共2009兲
Nguyen, Foster, and Gallimore
FIG. 14. 共Color online兲 Ion number densities of water vapor plasma operating with 75 SCCM H2O and 10 SCCM Ar.
FIG. 16. 共Color online兲 Plasma and floating potentials of water vapor
plasma operating with 75 SCCM H2O and 10 SCCM Ar.
ing potentials. Finally, electron temperature is not affected by
rf power, but is a strong function of water vapor input mass
flow rate, hence chamber pressure as well. The electron temperature first increases, then reaches a peak value, and finally
decreases with water vapor input flow rate. The electron temperature ranges between 1.5 and 4 eV.
trum confirms the dissociation of water molecules and substantiates the presence of electron affinity species such as
OH.
2. Residual gas analyzer
One way to identify the species inside a mixed gas, either neutral or charged particles, is by using a RGA. Another
way to identify the plasma species is by using an optical
spectrometer to capture the emission released when an excited atom or molecule is relaxed to a lower state. Both
methods are employed for species identification in this work.
Figure 17 shows a spectrum of a water vapor plasma with 75
SCCM water vapor and 10 SCCM argon flow rates, operating at 500 W rf power and 30 A magnet current. In this
spectrum, the OH band is clearly identified with the intense
peaks in the range 302.1–308.9 nm. H␣ at 656.3 nm and H␤
at 486.1 nm are also clearly identified. The emission spec-
The RGA allows for both qualitative identification of
species and quantitative estimate of the amount of each gas
species in the water plasma source. Figure 18 shows the
RGA spectrum for a water vapor plasma operating in the
same condition as that shown in Fig. 17. Quantitatively, the
amount of each gas can be estimated using Eq. 共4兲 if the
input argon flow rate and the partial pressures of argon and
hydrogen are known.
In Fig. 19, the production of hydrogen is characterized at
different magnet current settings 共0, 30, and 60 A兲 and rf
power levels 共250, 500, 750, and 1000 W兲 for again 75
SCCM water vapor and 10 SCCM argon gas flow rates. At
maximum rf power and magnetic field strength, the water
vapor plasma produced approximately 23 SCCM of hydrogen. With magnetic field, hydrogen production saturates at
high rf power. Without an applied magnetic field, hydrogen
production varies linearly with rf power in the range inves-
FIG. 15. 共Color online兲 Electron number densities of water vapor plasma
operating with 75 SCCM H2O and 10 SCCM Ar.
FIG. 17. An optical emission spectrum identifying the presence of hydrogen
and hydroxyl radicals in water vapor plasma operating with 75 SCCM H2O
and 10 SCCM Ar at 500 W rf power and 30 A magnet current.
C. Hydrogen production
1. Water plasma spectrum
Downloaded 25 Jun 2010 to 141.212.191.3. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/rsi/copyright.jsp
083503-7
Rev. Sci. Instrum. 80, 083503 共2009兲
Nguyen, Foster, and Gallimore
FIG. 18. A RGA spectrum indicating the presence of hydrogen, hydroxyl,
oxygen, and argon in a water vapor plasma operating with 75 SCCM H2O
and 10 SCCM Ar at 500 W rf power, and 30 A magnet current.
tigated, but the values are lower than those in the case with a
magnetic field. When a magnetic field is applied, there is a
40% jump in hydrogen production, from 12 SCCM at 250 W
to 17 and 18 SCCM for 30 and 60 A magnet current settings,
respectively. This jump is observed in all rf power levels.
Figure 20 shows a breakdown of gases for the 75 SCCM
water vapor and 10 SCCM argon flow rates operating with a
magnet current of 30 A. Most of the gases in the plasma
source is in fact still water vapor. There is a small concentration of oxygen and hydroxyl radicals. However, when the
input water flow rate is reduced to 25 SCCM, Fig. 21 shows
that the dominant species is hydrogen.
The optical emission spectrum shows a strong presence
of atomic hydrogen, especially at 656 nm. However, in the
RGA spectrum, only molecular hydrogen is detected. This is
mostly due to atomic hydrogen recombining to become molecular hydrogen before entering the RGA’s detector. OH is
detected by both the RGA and the spectrometer. Diatomic
oxygen is detected by the RGA but the spectrometer only
shows a faint signal of atomic oxygen 共777 nm兲. This is due
to both the small relative intensity of oxygen lines within the
spectrometer’s wavelength range 共200–850 nm兲 and the
spectrometer’s low sensitivity at 777 nm.
FIG. 19. 共Color online兲 Effect of rf power and magnet current on hydrogen
production for water vapor plasma with 75 SCCM H2O and 10 SCCM Ar.
FIG. 20. 共Color online兲 Production of all gases in water vapor plasma operating with 75 SCCM H2O and 10 SCCM Ar. Note that the dominant
species is water.
Although careful consideration has gone into the design
of the differential pump system, it is inevitable that recombination occurs as the gas particles travel to the RGA chamber. Therefore, the results shown here are for the worst case
scenario. It is expected that there is in fact more dissociation
inside the plasma source. Nevertheless, the results presented
here give general trends to show the effect of rf power and
magnetic field strength on hydrogen production. The presence of the axial magnetic field enhances hydrogen production in the vapor plasma. However, with the magnetic field,
hydrogen production saturates at high rf power due to the
loss of electrons to attachment.
ACKNOWLEDGMENTS
Sonca Nguyen is grateful for the financial support of the
National Science Foundation, Zonta International Foundation through the Amelia Earhart Fellowship and the Gates
Millennium Scholars Program. We acknowledge the support
of the technical staff in the Aerospace Engineering Department, especially Thomas Griffin, and from our colleagues at
PEPL.
FIG. 21. 共Color online兲 Production of all gases in water vapor plasma operating with a lower rate, 25 SCCM H2O and 10 SCCM Ar. Note that the
dominant species is hydrogen.
Downloaded 25 Jun 2010 to 141.212.191.3. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/rsi/copyright.jsp
083503-8
1
J. Clements, M. Sato, and R. Davis, IEEE Trans. Ind. Appl. IA-23, 224
共1987兲.
A. T. Sugiarto and M. Sato, Thin Solid Films 386, 295 共2001兲.
3
T. Maehara, H. Toyota, M. Kuramoto, A. Iwamae, A. Tadokoro, S.
Mukasa, H. Yamashita, A. Kawashima, and S. Nomura, Jpn. J. Appl.
Phys., Part. 1 45, 8864 共2006兲.
4
A. Ivannikov, V. Lelevkin, A. Tokarev, and V. Yudanov, High Energy
Chem. 37, 115 共2003兲.
5
A. Samokhin, N. Alekseev, T. Korovkina, E. Troitskaya, and Y. Tsvetkov,
Theor. Found. Chem. Eng. 41, 613 共2007兲.
6
R. Rudder, G. Hudson, J. Posthill, R. Thomas, R. Hendry, and D. Malta,
Appl. Phys. Lett. 60, 329 共1992兲.
7
R. Singh, D. Gilbert, R. Tellshow, P. Holloway, R. Ochoa, J. Simmons,
and R. Koba, Appl. Phys. Lett. 61, 2863 共1992兲.
8
B. Sun, M. Sato, and J. Clements, Environ. Sci. Technol. 34, 509 共2000兲.
9
B. Sun, M. Sato, and J. Clements, J. Phys. D: Appl. Phys. 32, 1908 共1999兲.
10
J. Oh, K. Kawamura, B. Pramanik, and A. Hatta, IEEE Trans. Plasma Sci.
2
Rev. Sci. Instrum. 80, 083503 共2009兲
Nguyen, Foster, and Gallimore
37, 107 共2009兲.
J. Chaffin, S. Bobbio, H. Inyang, and L. Kaanagbara, J. Energy Eng. 132,
104 共2006兲.
12
Z. Yan, C. Li, and W. Lin, Int. J. Hydrogen Energy 34, 48 共2008兲.
13
V. Givotov, A. Fridman, M. Krotov, E. Krasheninnikov, and B. Patrushev,
Int. J. Hydrogen Energy 6, 441 共1981兲.
14
M. Lieberman and A. Lichtenberg, Principles of Plasma Discharges and
Materials Processing 共Wiley, New York, 2005兲.
15
S. R. V. I. Demidov and K. Rypdal, Rev. Sci. Instrum. 73, 3409 共2002兲.
16
L. Schott, in Plasma Diagnostics, edited by W. Lochte-Holtgreven 共NorthHolland, Amsterdam, 1968兲, pp. 668–731.
17
I. Sudit and R. Woods, J. Appl. Phys. 76, 4487 共1994兲.
18
F. Chen, in Plasma Diagnostic Techniques, edited by R. H. Huddlestone
and S. L. Leonard 共Academic, New York, 1965兲, pp. 113–200.
19
I. Hutchinson, Principles of Plasma Diagnostics 共Cambridge University
Press, Cambridge, England, 2002兲.
11
Downloaded 25 Jun 2010 to 141.212.191.3. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/rsi/copyright.jsp