Decomposition of silicon tetrachloride by microwave plasma jet at atmospheric pressure

Article abstract of DOI:10.1134/S0020168509120188

A new method for destroying silicon tetrachloride has been developed, which based on a microwave plasma jet that operates at atmospheric pressure using hydrogen as work gas. Experiments showed that the silicon tetrachloride was dissociated into silicon an

Full text of DOI:10.1134/S0020168509120188

ISSN 0020-1685, Inorganic Materials, 2009, Vol. 45, No. 12, pp. 1403–1407. © Pleiades Publishing, Ltd., 2009.
Decomposition of Silicon Tetrachloride by Microwave Plasma
Jet at Atmospheric Pressure
Lifeng Wu , Zhibin Ma, Aihua He, and Jianhua Wang
Province Key Lab of Plasma Chemistry and Advanced Materials, School of Material Science and Engineering,
Wuhan Institute of Technology, Wuhan, China
e-mail: mazb@mail.wit.edu.cn, linwl28@yahoo.com.cn
Received November 2, 2008
Abstract—A new method for destroying silicon tetrachloride has been developed, which based on a microwave
plasma jet that operates at atmospheric pressure using hydrogen as work gas. Experiments showed that the sil-
icon tetrachloride was dissociated into silicon and hydrogen chloride under the effect of the plasma jet. The sil-
icon was deposited on the molybdenum substrate of the plasma reactor, which was yellow and typical nano-
sized particles. These solid samples have been analyzed by SEM, EDX, XRD, FT-IR and the active particle in
the plasma is detected by atomic emission spectroscopy. The results show that the silicon tetrachloride is mainly
transformed into nano-silicon with size of 54 nm. The dissociation efficiency reaches 50%.
DOI: 10.1134/S0020168509120188
1
INTRODUCTION
enhanced destruction and removal, plasma-based
chemical processes provide some advantages such as
low implementation, operating and capital costs, and
modest facility size requirements, in relation to conven-
tional volatile compounds destruction techniques [1].
In this letter, we report a system for destroying SiCl4
based on a microwave plasma jet (MPJ) operating at
atmospheric pressure. MPJ at atmospheric pressure has
gained huge potential from industry in recent years as a
clean, high-temperature intense energy source for vari-
ous applications of material processing [6].
The semiconductor industry in the large-scale use of
silane produces a large number of halogenated silane in
flue gases. Such undesirable waste is usually disposed
of by dumping in depots or discharge into the aquatic
medium or the atmosphere. This has resulted in increas-
ing contamination of ground water and degradation of
tropospheric ozone [1]. Silicon tetrachloride, one of
volatile compounds in flue gases, easily brings health
hazards in the air. Recently, a number of technologies
for destroying such environmental hazards ranging
from air pollutants to solid wastes have so far been
tested including thermal decomposition [2], catalytic
oxidation [3], biofiltration, carbon adsorption, mem-
brane separation, UV oxidation [4], condensation and
plasma-based procedures [5].
EXPERIMENTAL
The experimental set-up for SiCl4
destruction and
byproduct analysis consists of two distinct parts,
namely: (a) the destruction zone, which includes a
microwave plasma jet operating at atmospheric pres-
sure; (b) the analytical systems of atomic emission
spectroscopy.
As we known, non-thermal plasmas provide an
alternative method for generating highly reactive spe-
cies; in which electrons are accelerated by the applied
electrical fields and transfer their energy via elastic and
non-elastic collisions with neutral molecules. The reac-
tions occurring under these conditions are usually far
from thermodynamic equilibrium and result in destruc-
tion associated with higher electron temperatures of
10000–20000 K, by contrast, the neutral gas remains at
much lower temperatures. This mechanism is useful for
the applied energy to generate radicals and excited
atomic and molecular species, thereby facilitating the
decomposition of contaminant molecules. Although the
temperature of a non-thermal plasma is not so highly
influential on the destruction efficiency, an increased
temperature facilitates the process. In addition to
Figure 1 depicts the proposed SiCl4 destroying sys-
tem. The former comprises a magnetron connected to a
rectangular waveguide (WR340). They are operated at
a frequency of 2.45 GHz and their power is approxi-
mately 800 W. The power supply, consisting of a full-
wave voltage double circuit, provides the electrical
power to the magnetron, which generates the micro-
wave radiation and is cooled by a water-cooled
matched load. The generated microwave radiation from
the magnetron is guided through the waveguide, passes
through the three-stub tuner, and enters the discharge
part [7]. The resonant microwave is adjusted by sliding
short circuits and is induced into a copper nozzle
located at the centre of the reaction section. The center
1
The article is published in the original.
1403

1404
Cool–water
LI-FENG et al.
Cooper nozzle
Exhaust gas
Plasma
3–Stub tuner
Circulator
20 mm
2.45 GHz
magnetron
Molybdenum
substrate
Waveguide
Photomultiplier
Flow meters
SiCl
4
H
Spectrometer
2
Carrier
Gas
Fig. 1. Schematic diagram of the propmosed SiCl destruction set-up.
4
axis of nozzle is located one-quarter wavelength from
the shorted end and is perpendicular to the wide
waveguide walls.
H + SiCl2
H + SiCl
SiCl + HCl.
Si + HCl.
(2)
H2 is to deoxidize the silicon in SiCl4 to Si. The powder
deposited on the substrate was yellow in color.
For the 2.45 GHz cavity, the modeling results gave
the following optimum condition. The nozzle position
from the waveguide cavity short circuit wall is 45 mm,
nozzle tip diameter is 2 mm, nozzle hole is 0.5 mm and
cavity hole is 20 mm in diameter. This diameter has
been optimized to prevent the electrical field from
breaking down inside the cavity and also to stop the jet
The SEM images of as-produced Si nanoparticles
by the microwave plasma jet are shown in Fig. 2. By
comparing samples extracted at different regions, also
the difference between the different agglomeration
degrees can be found. An agglomerate is equal to just
reflected back into the cavity during the material pro- one primary particle at the centre of substrate, whereas
it is built up out of many primary particles at the edge.
When the plasma has a high degree of particle loading
(a high precursor feed rate and higher temperatures),
the sample containing particles with a small degree of
agglomeration will show a more open and fluffy struc-
ture (Fig. 2a), while the sample containing a high
degree of agglomeration shows a dense structure
(Fig. 2b). Quenching is the more effective at higher
temperatures for centre, and much smaller particles in
the agglomerates can be obtained and have a more open
structure. In order to obtain the information of element
species in the powder. EDX survey spectra were used to
determine which elements were present in the powder.
Si and oxygen species were detected with atomic frac-
tion of 61.42% oxygen and 31.72% silicon consistently.
The EDX spectra showed an additional line character-
istic of the existence of Cl (due to the chlorine atoms of
SiCl4). In addition, the Cu and Al signal is significantly
detected due to a little erosion of copper nozzle and alu-
minum fixture around the molybdenum substrate.
cessing [8]. Hydrogen was used as a plasma gas and
carrier gas with 100 kPa pressure and a 0.4–0.6 l/min
flow rate. The powder was collected by quenching from
molybdenum substrate.
Surface morphology of the powder was studied by a
scanning electron microscopy (SEM) with JSM
5510LV microscope equipped with an EDX analyzer.
Crystal structure was detected with aYB-XD-5A X-ray
diffraction. Further information about the chemical
structure was obtained by Impact 420 FT-IR. Atomic
emission spectroscopy was used for monitoring excited
species in the discharge with WDS-8A grating spec-
trometer.
RESULTS AND DISCUSSION
We primarily considered H2 as a reaction gas, which
was injected as the carrier gas of the microwave plasma
jet. The desired reaction pathway is the following pro-
cess.
Representative FT-IR spectra that show the trans-
mittance against wavenumber for specimen is given in
Fig. 3. Figure 3 are the FT-IR spectra corresponding to
SEM images in Fig. 2b, respectively. Strong absor-
bance peaks occur at 584, 1110 and 3440 cm–1. These
H2
H + SiCl4
H + SiCl3
2H,
(1)
SiCl3 + HCl,
SiCl2 + HCl,
INORGANIC MATERIALS Vol. 45 No. 12 2009

DECOMPOSITION OF SILICON TETRACHLORIDE BY MICROWAVE PLASMA
1405
Si
1.9
1.5
1.2
0.8
0.4
0
O
Cl
Al
Cu
8
2
4
6
10 12 14
Energy, keV
1 μm
1 μm
(b)
(a)
Fig. 2. SEM pictures of Si particles extracted at different regions, (a): Si nanopowder as collected at the centre of substrate and at
the edge for (b) with EDX spectra of the powder deposited on the molybdenum substrate. Plasma conditions: input power 800 W,
100 kPa, flow rate 500 ml/min.
11
features are associated with Si–H2 bonds that have a
characteristic bend mode (584 cm–1) .The system was
10
9
8
7
6
5
4
3
open to the atmosphere in the process of SiCl4 destruc-
tion, as shown in Fig. 1. Therefore, the constituents of
the atmosphere may be involved in the destroying pro-
cess, resulting in possible byproducts during the pro-
duction of Si nanoparticles, and Si–OH bonds that have
a bend mode (1620 cm–1) and a stretch mode
(3440 cm−1) could be detected in FT-IR spectra. We also
find a non-symmetric stretching vibration absorbance
peak (1110 cm–1) corresponding to asymmetric stretching
Si–O–Si bond. Moreover, for the sample the Si–O–Si
bending bond peak situated at 826 cm–1 does not appear.
So the as-collected powder is of a lower level of oxida-
tion.
1620
1110
584
3440
500 1000 1500 2000 2500 3000 3500 4000
–1
Wavenumber, cm
Fig. 3. FT-IR spectra illustrate the microwave plasma torch
abatement of SiCl .
4
XRD is a powerful tool for determining the structure
of as-produced material clearly. Therefore, an XRD
pattern of the collected powder was taken, and is shown
in Fig. 4. As shown in Fig. 4, black circles indicate the
peaks corresponding to Si [space group: Fd3m(227)]
with lattice constants a = 5.430 Å, which are in good
agreement with the reported data (JCPDS file
no. 65-1060) [9] and white circles correspond to
CuCl2 · H2O (JCPDS file no. 33-0451) [10]. And in the
process of discharge, detected CuCl2 · H2O in the pow-
der was caused by a little erosion of copper nozzle. In
the role of the plasma jet, the high activity copper atoms
through, high-temperature evaporation move up with
the chlorine atoms decomposed from silicon tetrachlo-
ride, and precipitate as CuCl2 · H2O crystal when touch
the cold-Mo substrate. The Willamson–Hall method
[11] can be used to determine the crystallite size and
lattice strain when there are three more reflections
available for measurement. Contributions of the crys-
1800
1600
1400
1200
1000
800
600
400
200
0
Si
CuCl × 2H O
2
2
10
20
30
40
50 2θ, deg
Fig. 4. XRD patterns for the powder deposited on substrate.
INORGANIC MATERIALS Vol. 45 No. 12 2009

1406
LI-FENG et al.
300
250
200
(a)
252.8
250.45
251.55
252.35
150
100
50
0
250.0
250.5
325.1
251.0
251.5
252.0
252.5
253.0
(b)
3500
3000
2500
2000
1500
510.4
327.5
521.7
515.1
515
1000
500
0
320
325
330
505
510
520
525
Wavelength , nm
Fig. 5. A part of a typical atomic emission spectrum as measured from a discharge in gases at atmospheric pressure (a) atomic emis-
sion spectrum of Si during destroying SiCl (b) detected atomic lines of copper in the process of discharge.
4
tallite size and lattice strain are given by the following can detect the emission spectra of silicon, as shown in
equation:
Fig. 5a. Experimentally, it is found that metal such as
copper in plasma is more sensitive to changing condi-
tions than silicon. Atomic lines in different wavelength
intervals with comparable weak intensities might easily
be lost in the analysis process [12]. Therefore, In the
process of silicon atomic lines detection, it should try to
avoid the interference of copper. To improve the inten-
sity of silicon atomic lines signal, we operate under the
optimum condition mentioned above. The range for sta-
ble operation in dependence on forward microwave
power and H2 gas flow was determined experimentally.
Stable plasma operation thereby means that the
reflected power is less than a few percent of the input
power. Moreover, in this region the plasma should
exhibit no filaments and a fully symmetrical shape. As
shown in table in the different state of discharge (arc,
spark and MPJ), the copper atoms and the silicon atoms
AES sensitive line show different intensity. At the same
time, we can calculate destruction and removal effi-
ciency (DRE) of SiCl4 from the expression:
kλ
βcosθ = ----- + ηsinθ,
(3)
L
where λ is the wavelength of the X-rays (λ = 1.54 Å in
this experiment), θ is the diffraction angle, η is the lat-
tice strain, L is the crystallite size, k is a constant (0.94
for Gaussian line profiles and small cubic crystals of
uniform size), and β is the full width at half maximum
(FWHM). In Eq. (3), the Si crystallite size was calcu-
lated to be approximately 54 nm in the powder.
A part of a typical atomic emission spectrum mea-
sured from a discharge in gases is given in Fig. 5. The
spectrum measurement system can be used to follow
relative fluctuations in the gas composition in time. In
the case of the MPJ, the carrier gas and SiCl4 are mixed
before entering the plasma, and the nozzle material has
a strong impact on the plasma properties. The atomic
line of copper can be easily distinguished in the shown
spectral range between 320 and 525 nm. Several strong
atomic line (325.1 and 510.4 nm) are dominantly
present in the measured spectra, as for instance can be
seen in Fig. 5b. The number of different atomics which
can be observed by emission spectroscopy in the micro-
wave-induced plasmas studied at atmospheric pressure
is very limited. In the case of not exceeding the detec-
tion limit of grating spectrometer, the wavelength of
Win – Wout
⎛
⎝
⎞
⎠
-------------------------
DRE =
100%,
(4)
Win
where Win is the mass of waste inserted and Wout tnat
remaining after the plasma treatment, and DRE can
detection was transferred to a more modest scale. We reach 50% under the optimum condition.
INORGANIC MATERIALS Vol. 45 No. 12 2009

DECOMPOSITION OF SILICON TETRACHLORIDE BY MICROWAVE PLASMA
1407
*
The copper atoms and the silicon atoms AES sensitive line in the different state of discharge (arc, spark and MPJ)
Intensity
Intensity
Wavelength, nm
* Wavelength, nm
arc
spark (discharge tube) MPJ
Si
arc spark (discharge tube) MPJ
Cu
390.5528
288.158
20
500
400
400
500
300
300
15
400
500
400
500
200
200
–
–
–
521.8202
515.3235
510.5541
327.3962
324.7540
224.6995
219.2260
700
600
–
–
2400 U3
2000 U4
2500 U5
2500R U2
2500R U1
U1
252.851
200 U2
200
200 U3
200
210 U4
500
–
252.4112
251.6111
251.4320
250.6899
–
3000R
5000R
30
1500R
2000R
500
500
–
–
–
V3
V2
25
* U1 – the most sensitive lines of neutral atoms of the element, U2 – the subaltern-sensitive line of neutral atoms of the element , so is the
rest; V1 – the most sensitive lines of ions of the element, V2 – the subaltern-sensitive line of ions of the element, so is the rest; R – the
width reflexive line.
4. Mohseni, M., Gas Phase Trichloroethylene (TCE) Pho-
tooxidation and Byproduct Formation: Photolysis vs.
Titania/Silica Based Photocatalysis, Chemosphere,
2005, vol. 59, pp. 335–342.
CONCLUSIONS
The proposed system for destroying volatile con-
taminants is an effective tool for the removal of silicon
tetrachloride. By using the microwave power (800 W)
and a flow rate of gas within the optimum range, one
can reach an efficiency decomposition levels. Based on
the analysis of byproducts, the SiCl4 is mainly trans-
formed into nano-silicon and nanometer-sized amor-
phous silica particles with size of 54 nm under optimum
conditions and such nano-materials have very great
prospects for industrial application. In addition, the
main pathway for the loss of chlorine is the formation
of a copper chloride deposit on the molybdenum sub-
strate. This is consistent with the spectroscopic mea-
surements and XRD, based on the intensity of the
atomic line of Cu and diffraction peaks of CuCl2 · H2O.
And CuCl2 · H2O is less toxic and easier to handle than
SiCl4. Also, it can be used for various purposes. There-
fore, the proposed method suppresses the most serious
hazards of the initial compound by fixing chloride in a
solid, non-volatile form.
5. Oda, T., Takahashi, T., and Tada, K., Decomposition of
Dilute Trichloroethylene by Nonthermal Plasma, IEEE
Trans. Ind. Appl., 1999, vol. 35, pp. 373–379.
6. Al-Shamma’a, A.I., Wylie, S.R., Lucas, J., and Van, J.D.,
Atmospheric Microwave Plasma Jet for Material Pro-
cessing, IEEE Trans. Plasma Sci., 2002, vol. 30,
pp. 1863–1871.
7. Hong, Y.C., Kim, H.S., and Uhm, H.S., Reduction of
Perfluorocompound Emissions by Microwave Plasma-
Torch, Thin Solid Films, 2003, pp. 329–334.
8. Al-Shamma’a, A.I., Wylie, S.R., Lucas, J., and Pau, C.F.,
Design and Construction of a 2.45 GHz Waveguide-
Based Microwave Plasma Jet at Atmospheric Pressure
for Material Processing, J. Phys., Ser. D, 2001, vol. 34,
pp. 2734–2741.
9. Kohno, A., Aomine, N., Soejima, Y., and Akasaki, A.,
Anomalous Behaviour of Silicon Single-Crystals
Observed by X-Ray Diffraction, Jpn. J. Appl. Phys.,
part 1, 1994, vol. 33, p. 5073.
REFERENCES
10. Natl. Bur. Stand, Monogr, Plane-Wave Scattering-
Matrix Theory of Antennas and Antenna-Antenna Inter-
actions, J. Chem. Phys., 1981, pp. 18–33.
1. Rubio, S.J., Quintero, M.C., Rodero, A., and Fernandez,
J.M. Rodriguez, Assessment of a New Carbon Tetrachlo-
ride Destruction System Based on a Microwave Plasma
Torch Operating at Atmospheric Pressure, J. Hazard.
Mater., 2007, pp. 419–427.
11. Williamson, G.K., Hall, W.H., X-Ray Line Broadening
from Filed Aluminium and Wolfram L’elargissement des
raies de rayons x obtenues des limailles d’aluminium et
de tungsteneDie verbreiterung der roentgeninterferen-
zlinien von aluminium- und woiframspaenen, Acta Met-
all., 1953, vol. 1, p. 22.
2. Moretti, E.C. and Mukhopadhyay, N., VOC Control-
Current Practices and Future-Trends, Chem. Eng. Prog.,
1993, vol. 7, p. 89.
12. Timmermans, E.A.H., de Groote, F.P.J., Jonkers, J., et al.,
Atomic Emission Spectroscopy for the On-Line Moni-
toring of Incineration Processes, J. Phys. D, part B,
2003, vol. 58, pp. 823–836.
3. Spivey, J.J., Complete Catalytic-Oxidation of Volatile
Organics, Ind. Eng. Chem. Res., 1987, vol. 26, pp. 2165–
2180.
INORGANIC MATERIALS Vol. 45 No. 12 2009