Journal of Natural Gas Chemistry 15(2006)–

Article

Methane Conversion Using Dielectric Barrier Discharge:

Comparison with Thermal Process and Catalyst Effects

Antonius Indarto∗ , Jae-Wook Choi, Hwaung Lee, Hyung Keun Song

Korea Institute of Science and Technology, Clean Technology Research Center,
P.O. Box 131, Cheongryang, Seoul 130-650, Korea
[ Manuscript received January 17, 2006; revised March 16, 2006 ]

Abstract: The direct conversion of methane using a dielectric barrier discharge has been experimentally
studied. Experiments with different values of flow rates and discharge voltages have been performed to
investigate the effects on the conversion reaction products both qualitatively and quantitatively. Experi-
mental results indicate that the maximum conversion of methane has been 80% at an input flow rate of 5
ml/min and a discharge voltage of 3.5 kV. Experimental results also show that the optimum condition has
occurred at a high discharge voltage and higher input flow rate. In terms of product distribution, a higher
flow rate or shorter residence time can increase the selectivity for higher hydrocarbons. No hydrocarbon
product was detected using the thermal method, except hydrogen and carbon. Increasing selectivity for
ethane was found when Pt and Ru catalysts presented in the plasma reaction. Hydrogenation of acetylene
in the catalyst surface could have been the reason for this phenomenon as the selectivity for acetylene in
the products was decreasing.
Key words: plasma; dielectric barrier discharge; thermal process; methane conversion; catalyst

1. Introduction shipping by ocean-going vessels is expensive. Approx-

imately 11% of this gas is re-injected, and unfortu-
Methane, which is a primary constituent of most nately, another 4% is flared or vented [2], which is a
natural gas reserves, is currently being used for home waste of a hydrocarbon resource. Both methane and
and industrial heating as well as for the generation carbon dioxide derived from methane combustion are
of electrical power. In many respects, methane is an greenhouse gases.
ideal fuel for these purposes because of its availabil- Natural gas can be converted to fuels and chem-
ity in most populated centers, its ease of purification icals in two ways, either via synthesis gas or directly
to remove sulfur compounds, and the fact that among into higher hydrocarbons or methanol. Today, most
the hydrocarbons it has the largest heat of combustion commercial processes for natural gas conversion in-
relative to the amount of CO2 formed. On the other volve synthesis gas as an intermediate. However,
hand, methane is a greatly underutilized resource for extensive research is presently being carried out on
chemicals and liquid fuels. different possible routes for the direct conversion of
Moreover, the reserves are increasing more rapidly natural gas to improve the selectivity and yield of
than those of liquid petroleum. Much of the methane higher hydrocarbons to meet the industrial and eco-
is found in regions that are far removed from indus- nomical demands.
trial complexes and often it is produced off shore [1]. The investigations of methane conversion to
Pipelines may not be available for transporting this higher hydrocarbons in non-thermal plasmas have
remote gas to potential markets and liquefaction for been made worldwide by corona discharges, spark dis-
∗ Corresponding author. E-mail: indarto antonius@yahoo.com.
2 Antonius Indarto et al./ Journal of Natural Gas Chemistry Vol. 15 No. 2 2006

charges, gliding arc or dielectric barrier discharges was used in this experiment. In order to maintain
(DBD) at atmospheric pressure and ambient temper- a similar configuration for every experiment, for ex-
ature. Some researchers have tried to add auxiliary ample, the gap distance, the reactor capacitance was
gases in the process, such as hydrogen [3], air [3–5], checked and measured before the experiments. The
oxygen [4–7] and noble gas [5,8]. Among them, mix- reactor capacitance was in the range of 8.2–8.8 pF in
ing gas between methane and CO2 was found to be air-filled gap conditions. The same reactor size and
the most promising technique to produce more valu- configuration were used for the thermal (non-plasma)
able products, such as synthesis gas [9,10]. Some process.
others used different types of plasma discharges that
could possibly produce different product distribution 2.2. Power supply and heater
[10,11].
The maximum voltage and frequency of the AC
In the present research, a comprehensive study
power supply (Auto electric, model A1831) were
on the performance of plasma and thermal treatment
10 kV and 20 kHz. To measure voltage and
for methane conversion has been undertaken. Plasma
current waveforms, a digital oscilloscope (Agilent,
process was done on a dielectric barrier discharge
model 54641A), a voltage divider (Tektronix, model
(DBD). A series of metal catalysts (Ni, Ru, Pt), sup-
P6015A), and a current probe (Fluke, model i400 s)
ported by Al2 O3 , had been prepared to investigate
were used. The external input power was measured
the effect of the products distributions.
by a digital power meter (Metex, model M-3860 M)
inserted at the AC power input line. The amount
2. Experimental setup of actual power supplied were calculated by following
equation:
Figure 1 shows a schematic diagram of the exper- Z
imental setup. Methane, as the source gas, was in- Actual power = (V (t) × I(t))dt × frequency
troduced into a cylindrical reactor under atmospheric
pressure. Gases were analyzed by gas chromatogra-
The typical waveform of voltage and current, used
phy. Details of each part of the system have been
in this experiment, is shown in Figure 2. For the ther-
described in the next sections.
mal process, a tubular furnace equipped with a tem-
perature controller (Daepoong Industry, Korea) was
used.

Figure 1. Experimental Setup

2.1. Reactor Figure 2. Typical waveform of voltage and current

The reactor was a cylindrical pyrex tube (ID of 7.5 2.3. Materials preparation
mm) with two parallel-straight wires (0.2 mm diame-
ter, stainless steel) as the inner metal electrode and an All experiments were carried out using pure
outer electrode of silver film coated around the tube. methane (CH4 ) with purity higher than 99.99%. Flow
The effective gas volume and length of the reactor rate of the source gas was controlled by a cali-
were 8.8 ml and 200 mm, respectively. A high voltage brated mass flow controller (Milipore, model FC-
and frequency alternating current (AC) power supply 280SAV). Analysis of the gas sample was carried out
 Journal of Natural Gas Chemistry Vol. 15 No. 2 2006 3

using gas chromatography (YoungLin, model M600D) higher hydrocarbons.

with a thermal conductivity detector (TCD, Column: The evaluation of system performance was done
Hayesep D 80/100) for measuring H2 and CH4 , and a based on product selectivity and methane conversion
flame ionized detector (FID) for measuring CH4 and which are formulated as:
moles of H2 produced
Selectivity for H2 = × 100% (1)
2 × moles of CH4 converted
x × moles of Cx Hy produced
Selectivity for Cx Hy = × 100% (2)
moles of CH4 converted
moles of CH4 converted
Conversion of CH4 = × 100% (3)
moles of initial CH4
moles of CH4 converted
Energy efficiency = × 100% (4)
total power
A series of metal catalysts (Ni, Ru, Pt) were pre- more effective although the percentage of conversion
pared by the incipient wetness impregnation method is lower, as seen in Figure 4.
with Al2 O3 as the support. The aqueous solutions of
Ni, Ru, and Pt were made by dissolving NiCl2 ·6H2 O
(Kanto Chemical Co, Japan), RuCl3 ·3H2 O (Enger-
hard, England), and H2 PtCl6 ·6H2 O (Next Chimica,
South Africa) precursors in aqua regia. All catalysts
were dried at 388 K for 2 h, followed by calcination
at 873 K for 3 h in O2 -rich gas condition. Reduction
of catalysts was done by flowing hydrogen at 673 K
for 2 h. The amounts of metals loaded were all set-
tled to be 3% of the catalyst. In the plasma reaction
experiment, 0.5 grams of the catalyst was packed at
the end of plasma zone to prevent the decomposition
of products. Figure 3. Ef fect of methane f low rates and dis-
charge voltages on methane conversion
3. Result and discussion (1) 3 kV (2) 3.2 kV, (3) 4 kV

3.1. Plasma and thermal process

Methane conversion under plasma and thermal

process was conducted under atmospheric pressure.
In plasma experiments, the total flow rate was varied
from 5 ml/min to 40 ml/min and the voltage was set
at 3, 3.2, and 3.5 kV. Figure 3 shows the effects of
input methane flow rate and voltage on the plasma
process. Higher conversion of methane is produced at
higher voltages and lower flow rates. The maximum
methane conversion reaches 80% at initial methane
flow rate of 5 ml/min and voltage of 3.5 kV. This
shows that discharge voltage has a straight correlation Figure 4. Ef fect of methane f low rates and dis-
with methane conversion. The increase of the ampli- charge voltages on process ef f iciency
tude of voltage leads to the increase of plasma den- (1) 3 kV (2) 3.2 kV, (3) 4 kV
sity [12]. It can also be seen from the curve trend that
the conversion rate decreases when the flow rate is in- More numbers of methane molecules are con-
creased. Lower flow rates will give a longer chance for verted at higher flow rates than at lower ones. Instead
the molecules to collide with energetic species, such of electrons, energetic species from methane decompo-
as electrons. At a higher flow rate, the collisions are sition, e.g. methyl radical (CH3 ·) and H·, can initiate
4 Antonius Indarto et al./ Journal of Natural Gas Chemistry Vol. 15 No. 2 2006

another effective collision with CH4 . At lower flow reaction:

rates, the number of those species is less as the amount CH4 + e −→ CH4 · + e (5)
of methane that flows to the reactor is also small. CH4 · −→ CH3 · + H· (6)
However, by increasing discharge voltage, the energy
efficiency is slightly increased. Following the previ- Methyl radical (CH3 ·) will initiate the recombi-
ous statement, the plasma reaction is more effective nant reaction of higher hydrocarbons and hydrogen
to decompose methane at a higher plasma density. radical (H·) has a tendency to be a reducer or decom-
This leads to the conclusion that the optimum process poser reactant for higher hydrocarbons.
occurrs at higher flow rate with higher discharge volt-
CH3 · + CH3 · −→ C2 H6 (7)
ages.
CH3 · + CH3 · −→ C2 H4 + H2 (8)
Figure 5 shows the product distribution of
methane conversion by the plasma process. High se- Cx Hy + H· −→ Cx Hy−1 + H2 (9)
lectivity for H2 is produced at a low input flow rate. CH3 · + C2 H3 −→ C3 H6 (10)
Fragmentation of methane into smaller molecules, e.g.
Instate of H·, decomposition of higher hydrocar-
H2 and C, is more favorable that recombinant reac-
bon was also caused by electrons. Because electron
tion to produce higher hydrocarbons. At lower flow
collision with molecules is faster than that with ions or
rates or at longer residence time, higher hydrocarbons
radicals [14], the longer resident time of molecules in
that are also produced during the plasma reaction will
the plasma reaction will give a negative effect on the
be easily decomposed by electron and other energetic
formation of higher hydrocarbon and produce more
species.
H2 and C molecules than that in shorter resident time.
In case of coke (C solid), it was produced mostly at
the wall of the reactor and the surface wire of inner
electrode. The research on coke formation has been
experimentally done by Le et al. [15] and numeri-
cally simulated by Rhallabi and Chaterine [16]. Coke
will be formed at the plasma regions where the tem-
perature is relatively high, for example, at the sur-
face of the electrode. At the surface of the electrode,
methane will be converted into C and 4 H or H2 .
CH4 −→ C + 4H (11)
CH4 −→ C + 2H2 (12)

Figure 5. Ef fect of methane f low rates on products This phenomenon also occurred in our previous
distribution at discharge voltage of 3.5 kV research of methane conversion reaction by arc plasma
(1) C2 (2) H2 , (3) C3 (4) C4 [17] where the temperature was much higher than
room temperature.
Electron which has a key factor to initiate the However, numerous amounts of H2 and C produc-
radical reaction in DBD process [13] can lead to the tion were also found in the thermal process. Table 1

Table 1. Methane conversion and product distribution comparison between plasma and thermal processes
CH4 Selectivity (%)
conversion (%) H2 C2 H 2 C2 H 6 C3 H 6 C3 H 8 n-C4 H10 i-C4 H10
Plasma 54.97 35.29 4.85 20.89 2.01 12.42 5.90 4.64
51.11 29.22 7.36 23.19 3.36 13.33 6.11 7.37
25.01 34.77 5.06 25.48 1.73 13.89 4.42 4.95
22.37 38.46 9.30 42.75 2.98 19.79 5.85 7.67
Thermal 54.95 65.73
50.42 68.58
25.60 77.57
13.35 84.04
7.05 91.60
 Journal of Natural Gas Chemistry Vol. 15 No. 2 2006 5

shows the products comparison between plasma and the most driving factor in dissociating methane mole-
thermal process. Most of products of thermal process cules into H2 and C.
were dominated by H2 and C. Solid carbon was not
CH4 −→ C + 2H2 (13)
measured, but much of it was found in the output
line of the reactor. At a flow rate of 20 ml/min, the The production of higher hydrocarbon was rela-
conversion of methane was begun at temperature of tively small and un-detectable on using our analysis
700 ℃. To achieve similar conversion as in the plasma instrument. Methane can be converted directly into
experiment, it required the temperature to be more higher hydrocarbons, such as C2 hydrocarbons and
than 900 ℃. In the thermal process, supplied heat is hydrogen via thermal coupling reactions:

2CH4 −→ C2 H2 + 3H2 , ∆H298 =3765 kJ/mol (14)

2CH4 −→ C2 H4 + 2H2 , ∆H298 =2023 kJ/mol (15)
2CH4 −→ C2 H6 + H2 , ∆H298 =651 kJ/mol (16)

All these reactions are highly endothermic, and 3.2. Catalyst ef fect
high temperature operation is required to obtain fa-
vorable thermodynamics. In order to avoid the de- In order to improve the selectivity of products,
composition of products, very short time reaction, 0.5 grams of three different catalysts have been put
less than 0.01 s, should be performed [18] and con- at the end of plasma zone of the reactor. The pur-
sequently, this makes it difficult to achieve high con- pose of this is to prevent the production of higher
version of methane. hydrocarbons from decomposition reactions. Table 2
By comparing the experiment results, it can shows the CH4 conversion and product distribution
be seen that the plasma process has an advantage with different material packs. The experiment was
on higher hydrocarbons production. The existence conducted at a methane flow rate of 20 ml/min. It
of electrons and methyl radicals, produced during shows that the conversion of CH4 to other hydrocar-
plasma process, could be the primary role in initiat- bons was enhanced with the use of catalyst. Although
ing the decomposition of methane and recombination the difference was relatively small, the existence of
reactions of higher hydrocarbons. catalyst helped the cracking process of methane mole-

Table 2. Ef fects of dif ferent material packs loading on methane conversion and products distribution
CH4 Selectivity (%)
conversion (%) H2 C2 H 2 C2 H 6 C3 H 6 C3 H 8 n-C4 H10 i-C4 H10
Empty 43.42 34.52 5.13 25.73 0.00 11.73 5.25 4.81
γ-Al2 O3 42.50 39.69 7.62 26.31 2.69 15.85 5.74 5.56
Ni/γ-Al2 O3 48.06 38.89 5.23 27.97 2.31 14.02 4.92 5.31
Ru/γ-Al2 O3 48.13 33.55 2.65 30.22 0.00 13.47 5.08 5.36
Pt/γ-Al2 O3 47.00 32.92 0.00 34.87 0.00 16.36 5.51 6.00

cules. Interesting phenomena was found when Pt/γ- C2 H2(g) −→ C2 H2(surf) (17)
Al2 O3 was used as the catalyst. Compared to non-
catalyst treatment, the selectivity for ethane (C2 H6 ) The existance of numerous H atoms, also attached
was higher by a factor of 1.35. On the other hand, on the surface of catalyst, can initiate the series of hy-
no acetylene (C2 H2 ) was detected by our analysis in- drogenation reactions:
strument. According to this result, the existence of C2 H2(surf) + H(surf) −→ C2 H3(surf) (18)
Pt catalyst in the plasma reaction would convert the C2 H3(surf) + H(surf) −→ C2 H4(surf) (19)
produced acetylene into ethane by the hydrogenation C2 H4(surf) + H(surf) −→ C2 H5(surf) (20)
reaction on the surface of catalyst.
C2 H5(surf) + H(surf) −→ C2 H6(surf) (21)
Low activation energy (∼0 kJ/mol) of acetylene
adsorption on Pt will make acetylene attach easily to In our previous experiment, under the non-
the surface of the catalyst [19]. catalytic reaction, the above reactions rarely oc-
6 Antonius Indarto et al./ Journal of Natural Gas Chemistry Vol. 15 No. 2 2006

curred. This was shown by small yields of C2 H4 and References

C2 H6 from C2 H2 conversion [20]. The idea of this
reaction mechanism, for example, via intermediate [1] Stern J, Royal Institute of International Affairs, Re-
C2 H5 , was supported by Dent et al’s investigation port, July 2002
[21]. [2] Lercher J A, Bitter J H, Steghuis A G, Van Ommen
J G, Seshan K. Catalytic Science Series 1 (Environ-
However, a similar phenomenon also occurred
mental Catalysis.), 1999, 103
when Ru/γ-Al2 O3 catalyst was used. Although the [3] Liu C J, Mallinson R, Lobban L. J Catal, 1998,
condition was less clear than Pt, by adding Ru cata- 179(1): 326
lyst, the selectivity for ethane increased and followed [4] Zhou L M, Xue B, Kogelschatz U, Eliasson B. Plasma
by a decreased selectivity for acetylene. In term of Ni Chem Plasma Process, 1998, 18(3): 375
catalyst, no significant effects on product distribu- [5] Indarto A, Choi J W, Lee H, Song H K. Energy, in
tion of methane conversion were observed, which was press
in agreement with other experiment evidence that Ni [6] Cormier J M, Rusu I. J Phys D: Appl Phys, 2001,
is a less active catalyst for hydrogenation than Pt [19]. 34(18): 2798
[7] Nozaki T, Hattori A, Okazaki K. Catal Today, 2004,
98(4): 607
4. Conclusions [8] Thanyachotpaiboon K, Chavadej S, Caldwell T A,
Lobban L L, Mallinson R G. AIChE J, 1998, 44(10):
The conversion of methane into hydrogen and 2252
higher hydrocarbons was experimentally investigated [9] Hwang B B, Yeo Y K, Na B K. Korean J Chem Eng,
in a dielectric barrier discharge and a thermal process. 2003, 20(4): 631
[10] Song H K, Lee H, Choi J W, Na B K. Plasma Chem
In plasma process, the residence time of molecules
Plasma Process, 2004, 24(1): 57
had influenced the reaction mechanism of the prod-
[11] Lia X S, Zhu A M, Wang K J, Xu Y, Song Zh M.
ucts. Lower flow rate or longer residence time would Catal Today, 2004, 98(4): 617
make the reaction favor the dissociation of methane [12] Choi Y H, Kim J H, Hwang Y S. Thin Solid Films,
into smaller molecules than recombinant reaction of 2005, in press
higher hydrocarbons. By thermal process, the reac- [13] Eliasson B, Kogelschatz U. IEEE Trans Plasma Sci,
tion will produce H2 and C. The existence of electrons 1991, 19(6): 1063
in plasma process gave an advantage on the higher hy- [14] Lieberman M A, Lichtenberg A J, Principles of
drocarbon production. Plasma Discharges and Material Processing, 1st ed,
New York: John Wiley and Sons, 1994, 224
In order to increase the selectivity of products, the
[15] Le H, Lobban L L, Mallinson R G. Catal Today, 2004,
plasma-catalyst process has been performed. Pt and
89: 15
also Ru catalyst show a significant effect on increasing [16] Rhallabi A, Catherine Y. IEEE Trans Plasma Sci,
production of ethane by transforming the produced 1991, 19(2): 270
acetylene by hydrogenation reactions. [17] Indarto A, Choi J W, Lee H, Song H K. J Natur Gas
Chem, 2005, 14(1): 13
Acknowledgements [18] Holmer A, Olsvik O, Rokstad O A. Fuel Process Tech,
1995, 42: 249
[19] Shustorovich E, Bell A T. Surface Sci, 1988, 205: 492
This research was supported by the National Re- [20] Kim S S, Lee H, Na B K, Song H K. Korean J Chem
search Laboratory program of the Korea Minister of Eng, 2003, 20(5): 869
Science and Technology. [21] Dent A L, Kokes R J. J Physic Chem, 1970, 20: 3653