Manuscript

Partial oxidation of methane with Cu-Zn-Al catalyst in a dielectric barrier discharge

Antonius Indarto1,2,†, Jae-Wook Choi1, Hwaung Lee1, Hyung Keun Song1 1

Clean Technology Research Center, Korea Institute of Science and Technology 2

Department of Chemical and Biological Engineering, Korea University

†

Correspondence address: e-mail: [email protected], tel:+82-10-2296-3748

1

Abstract- Series of methanol synthesis catalyst containing Cu-Zn-Al (CZA) were prepared by co-precipitation method and applied for partial oxidation of methane into methanol using dielectric barrier discharge (DBD). The methanol synthesis process was occurred at ambient temperature and atmospheric pressure. In our experiment, CZA showed a high catalytic activity to increase the production of methanol. The methanol selectivity of CZA-assisted plasma process was twice higher than that of non-catalytic plasma process. The addition of other metals on CZA catalyst also produced a significant effect on the methanol production and it was found that yttrium could the best addition metal compared to Pt, Fe, and Ni. Instead of methanol, the reaction products of plasma reactions were dominated by H2, CO, CO2, C2 and water. The optimum methanol selectivity reached 31% when 3% yttrium metal was doped over CZA.

Keywords: Methane oxidation; methanol synthesis; dielectric barrier discharge; Cu-ZnAl; heterogeneous catalyst.

2

1. Introduction

The catalytic conversion of methane into methanol is one of the major challenges for chemists. Methane, as the major part of natural gas, is the cheapest and promising source for direct conversion chemical process to produce methanol. Nowadays, the need for methanol increases due to its importance as an intermediate material for industries. Currently, the production of methanol is done via synthesis gas (CO, CO2, and H2) over Cu-Zn-Al2O3 (CZA) catalyst. However, this process route can be changed drastically when the effective method to oxidize methane to methanol is found. Catalytic oxidation of methane at low temperatures is economically interesting, but, also very difficult to achieve as a result of the high stability of C-H bonds. The investigations of methane oxidation to methanol in non-thermal plasmas have been made worldwide by corona discharges, spark discharges, gliding arc, or dielectric barrier discharges (DBD) at atmospheric pressure and ambient temperature. Some researchers have tried to add auxiliary gases in the process, such as hydrogen [1], air [13], oxygen [2-5], noble gas [3, 6]. Some others used the different types of plasma discharges that possibly produce different products distribution [7-8]. However, the production of methanol was low and not economically meets the market requirement. In the present research, comprehensive study on the performance of plasma process and Cu-Zn-Al (CZA)-based catalyst for methane partial oxidation has been undertaken. Dielectric barrier discharge was chosen as the plasma media due to mild characteristic which probably offers better process for methanol synthesis. In methanol synthesis via synthesis gas, CZA has been claimed as the active component for methanol production

3

[9-10]. Although the catalytic reaction mechanism is still confusing, the existence of CZA in the process was proved to be better material catalyst than other metals. The synergy among Cu-Zn-Al was able to absorb the molecules and conducted a series of surface reactions on the catalyst surface. In the plasma environment where the radical and ion are abundant, the above mechanism would be able to boost the methanol production. In accordance with the above discussion, this study investigates the influence of CZA on direct partial oxidation of methane to produce methanol. CZA and metal-supported CZA catalysts were prepared and characterized, and the efficiencies were demonstrated by methanol selectivity.

2. Experimental setup Figure 1

Figure 1 shows a schematic diagram of the experimental setup. Methane and oxygen were introduced to the reactor at room temperature and atmospheric pressure. The products were analyzed by a gas chromatography (GC). Details of each parts of the system are described in the next sections.

2.1. Reactor The reactor is a cylindrical pyrex tube (ID of 7.5 mm) with 2 parallel-straight wires (0.2-mm diameter, stainless steel) as the inner metal electrode and silver film coated at the outer side of tube as the outer electrode. The effective volume and length of the

4

reactor were 8.8 mL and 200 mm, respectively. In order to maintain the similarity of the reactor configuration, e.g. electrodes gap distance, the reactor capacitance was checked first by an RCL meter (Fluke PM6304) before and after experiments.

2.2. Power supply Plasma was generated by an alternating current (AC) power supply (Auto electric, model A1831) which has a maximum voltage and current of 10 kV and 100mA, respectively. A digital power meter (Metex, model M-3860M) was inserted to the electricity line of the AC power supply in order to measure the total supplied power to the reactor.

2.3. Materials All experiments were carried out by introducing methane (CH4, purity >99.99%) and oxygen (O2, purity >99.9%) at fixed methane to oxygen ratio of 4:1 (volume basis). The input gases were controlled by calibrated mass flow controllers (Milipore, model FC280SAV). Analysis of the products was done by a gas chromatography (YoungLin, model M600D, Column: 10ft Hayesep D 80/100 + 10ft Carbowax 10% 80/100) with a thermal conductivity detector (TCD) for measuring H2 and CH4 and a flame ionized detector (FID) for measuring CH4, methanol, and higher hydrocarbons. The performance evaluation of the process was calculated on the basis of products selectivity and methane conversion, formulated as:

5

Selectivity of H 2 =

moles of H 2 produced × 100% 2 × moles of CH 4 converted

Selectivity of C x H y =

x × moles of C x H y produced moles of CH 4 converted

Selectivity of CH 3 OH =

Selectivity of CO x =

× 100%

moles of CH 3 OH produced × 100% moles of CH 4 converted

moles of CO x produced × 100% moles of CH 4 converted

Conversion of CH 4 =

moles of CH 4 converted × 100% moles of initial CH 4

(1)

(2)

(3)

(4)

(5)

2.4. Preparation and characterization of catalyst The precursors of methanol synthesis catalysts, denoted as CZA and CZA-[Pt Ni Fe Y], with the % mol ratios of Cu:Zn:Al = 6:2.5:1.5 were prepared by simultaneous coprecipitation method from their nitrates solution at a pH about 7.0. A buffer solution with a pH 8.50 containing a mixture of NaHCO3 and Na2CO3 was used as the precipitation agents throughout the works. In a typical catalyst preparation, metal nitrates of Cu(NO3)2⋅3H2O, Zn(NO3)2⋅6H2O, and Al(NO3)⋅9H2O were dissolved in H2O to get 250 mL of metal nitrate solution with total molar concentration of 0.3 M. In a 1000 mL beaker glass initially filled with 200 mL of distilled water, a solution of metal nitrates and a solution of precipitation agent were dropt simultaneously for 1.5 hours under vigorous agitation at room temperature while the pH of the slurry was maintained in a range from 6.90 to 7.10. Following the completion of co-precipitation process, the

6

slurry was stirred further, which is so-called aging, for 2 hours at room temperature. The formed precipitate was filtered and washed with distilled water under stirring for 12 hours. This procedure was repeated 6 times which were found to be sufficient for the slurry reached pH about 7.0. The cake was dried at room temperature for 24 hours and then oven dried at 110 C for 24 hours. The catalyst precursor was calcined under air condition at 350oC for 2 hours. The catalyst for methanol synthesis was grinded to about 0.18 mm and physically mixed with a weight ratio of 1 with the same particle size of γAl2O3 as the methanol dehydration catalyst. In order to compare the performance of prepared catalyst with commercial CZA catalysts, the CZA produced by ICI (ICI Katalco 51-8; noted as CZAi) and BASF (BASF K3-110; noted as CZAb) was ordered. Moreover, using the same method, some other metals, e.g. Pt, Ni Fe, and Y, were also doped to CZA to increase the activity of pure CZA.

2.4.1. XRD patterns Crystal structures were identified by powder X-ray diffraction spectroscopy on a Shimadzu 6000 X-ray diffractometer system (Cu Kα radiation, 40 kV, 30 mA). Figure 2a presents the XRD pattern changes of the prepared CZA at various conditions. The CZA structure was greatly changed before and after calcinations. After oven drying, the catalyst shows clearly a hydorotalcite-like (HTlc) and malachite phases. This result is consistent with the previous report which suggested that HTlc phase coexists with a malachite phase when the atomic ration of Cu/(Cu+Zn) is in the range from 0.5 to 0.9. CuO also appears due to the high Cu/Zn ratio. The presence of CuO along with HTlc occurs due to the Jahn-Taller effect. The existence of those phases on the fresh

7

uncalcined CZA catalyst will guaranty its activity of the after calcinations treatment. Those forms were also appeared at all series CZA catalyst with addition of Pt, Ni, Fe, and Y as shown in figure 2b. The catalyst structure was greatly changed after calcinations treatment which mostly produced CuO and ZnO as the most dominant phases. However, in figure 2a, the structure of catalyst before and after 6 hours plasma process was not so much different. It concludes that the deactivation of catalyst was not occurred during 6 hours operation due to plasma reactions.

2.4.2. SEM, EDS, and Cu surface area Dispersed copper, incorporated with ZnO structure, is believed as the active side on the catalyst [11]. In this experiment, the copper surface areas were determined by N2O titration using pulse flow experiment following to the previous method [11]. In a typical experiment about 50 mg of sample was loaded and reduced using 5 vol % H 2 in Ar at 250 °C for 1 h and purged the reactor in a He stream at 250 °C for 30 min and cooled down to 90 °C at which the N2O titrations were carried out. A surface copper density of 1.47.1019 atoms/m2 assuming Cu:O = 2:1 was used in this investigation. Surface analysis was done using Hitachi FE-SEM S-400 microscope at an accelerating voltage 0.5~30 kV. Figure 3 shows the surface images of the CZA catalyst. Figure 3 also present that there is no significant different of the catalyst structure before plasma reaction (figure 3b) and after plasma reaction (figure 3c). The elemental compositions of the samples were determined by an energy dispersive spectrometer (EDS) analysis using the same instrument. The typical result of EDS analysis of the catalyst is shown in figure 3d.

8

3. Results and discussion

3.1. Plasma stability and products distribution Figure 4, Table 1

The direct methane oxidation using dielectric barrier discharge (DBD) was conducted at room temperature and atmospheric pressure. The total flow rate and the volumetric ratio of methane to oxygen were maintained constant at 30 ml/min and 4:1, respectively. The supplied power to the reactor was also fixed at 50 watt. Our previous study showed that in this region, the selectivity of methanol, the most profitable products, reached the highest value for non-catalytic partial oxidation of methane by dielectric barrier discharge [12]. The products were analyzed at three different sampling times to ensure the reproducibility data and each point of the experimental variables was repeated twice. Figure 4 shows the comparison of the catalyst performance between CZA made by coprecipitation method with the commercialized one. By addition of CZA catalyst, the selectivity of methanol was sharply increased, almost 2 times higher compared to noncatalytic plasma reaction. In non catalyst reaction, the most products were dominated by CO, CO2, and H2. The selectivity of methanol was less than 11%. From the result of non catalytic process, it can be said that the non-catalytic plasma will convert methane and O2 into synthesis gas (CO, CO2, and H2). C2H6 was found as the single major product of higher hydrocarbon. Based on our previous kinetics calculation [12], the methanol from methane oxygen

9

plasma reaction can be possible produced from two reaction mechanisms: CH3 + OH  CH3OH

(6)

CH2 + H2O  CH3OH

(7)

Insertion reaction of single atom oxygen to methane was rarely occurred because methane is easily dissociated to CH3 and H or C and H2 in the bulk plasma. CH4  CH3 + H

(8)

CH4  C + 2 H2

(9)

Between above two reactions, the second reaction of methanol synthesis was higher chance to be occurred. By employing CZA catalyst, which is well-known catalyst for converting synthesis gas to methanol, the selectivity of methanol was higher. We can conclude here that the plasma-catalytic process of methane and oxygen is a 2-stages process which consists of methane and oxygen conversion to synthesis gas then with catalyst, it will be converted to methanol. Figure 4 shows that the value difference was relatively small means that the performance of CZA catalyst made by co-precipitation method was satisfying enough to be used further as the selectivity to produce methanol is similar to the commercialized catalysts. Instead of methanol selectivity, the stability of the catalyst was also good. The deactivation of catalyst, e.g. by carbon decomposition, was not occurred. Our previous simulation shows at the ratio of CH4:O2 = 4:1, the produced solid carbon from CH4 fragmentation will react easily with oxygen radical to form CO. C + O  CO

(10)

10

Figure 3c also proves that there is no significant amount of carbon attached on the surface of the catalyst after plasma reaction. Physical observation of the catalyst color shows that there is no significant different of the color change.

3.2. Effect of catalyst loading Figure 5, Table 2

The effect of catalyst loading on the methanol selectivity at 50W was shown in Figure 5. The experimental results represented that the methanol selectivity was increased with the increase of catalyst loading. The methanol selectivity was increase from 24% to 27.4% when the catalyst loading increased from 0.5 gram to 1.5 gram. However, the catalyst loading did not affect the conversion of methane. The methane conversion was between 28% and 33%. Increasing amount of catalyst in the reactor will increase the methanol selectivity means that more catalytic reactions of methanol production will be occurred on the surface of the catalyst. Although the selectivity increased, the efficiency of the catalyst, defined as a ratio of methanol production per weight of catalyst loading, was decreasing. The selectivity of other products is listed in Table 2. The main products were dominated by CO, CO2, and H2. The selectivity of C2H6 was between 3% and 9% or around 80% of the total C2 hydrocarbons. Similar with previous results, the increment of methanol production was followed by reducing concentration of CO, CO2 and H2 in the products stream. Compared to non-catalytic process, the CO and CO2 decreased around 20-30% while H2 around 30%.

11

3.3. Effect of power increment Figure 6, Table 3

In order to increase the methane conversion and probably the yield of methanol, the supplied power to the reactor was increased from 80 to 120W. The effects of applied power to the products are shown in figure 6 and table 3. The conversion rate of methane increased from 27% at 50W to 45% at 140W. It can be deduced trivially that by supplying more energy to the reactor, the conversion of reactant will increased. As conversion of methane and supplied power increased, the methanol selectivity of noncatalytic process tends to decrease. In other words, fragmentation reactions of molecules were preferable than re-combination or higher molecule synthesis reactions [13-15]. This idea was supported by the result presented in Table 3 which the concentration of CO, CO2, and H2 also increased. In non catalytic plasma reaction, this condition resulted lower production of methanol. The existence of CZA catalyst will help maintaining the production of methanol from synthesis gas produced by fragmentation reactions of methane and oxygen. As shown in figure, although the selectivity of methanol decreases as power increases, the decrement rate of plasma with CZA catalyst is lower than that of non-catalytic plasma process.

3.1. Effect of metal addition on CZA Figure 7, Table 4

In order to increase the activity of the catalyst, other transition metals, e.g. Pt, Ni, Fe,

12

and Y, have been to the CZA catalyst. The percentage of additional metal loading was only 3%, except for Pt which is only 1%. Meanwhile, the total catalyst weight loaded in the reactor was only 0.5 gram. As shown in Figure 5, except Ni, the addition of metal catalysts on CZA gives a positive effect by increasing the selectivity of methanol around 15 to 20% compared to the methanol selectivity of CZA catalyst. Among those, yttrium produced the highest value of methanol selectivity that able to convert methane to methanol around 31%. The addition of Y, also Pt and Fe, gave a significant effect on the direction of oxidation reaction. From table 4, the addition of Pt decreases the selectivity of CO2 and CO. The probability of methane conversion to methanol would be higher as there is no significant different found on the remaining products, e.g. C2 and higher hydrocarbons. Instead of above reaction mechanism of methanol production, another mechanism of methanol by CH3 and OH reaction could be occurred. Pt and Fe are well-know metal catalysts which were able to activate CH4 [16-17] and adsorb OH [18]. Methanol can be formed by surface reaction between methyl methane with OH: CHx  CHx (surf)

(11)

CHx (surf) + OH(surf) + (3-x) H(surf) CH3OH (surf)

(12)

For the case of yttrium addition, the above mechanism is not proved yet. However, we calculate of total produced CO and CO2 and the ratio of CO2 to CO is lower among other catalyst. Instead of lower selectivity of methane conversion to CO and CO 2, the oxidation of CO to CO2 in CZA catalyst is also the lowest value. The existence of yttrium inhibits the secondary oxidation of carbon from CO to CO2. This will create a chance for CO to be converted into methanol. The detail phenomena of partial oxidation

13

of methane with yttrium stabilized zirconia were observed by our previous research [19]. Different phenomena could be found in case of Ni-CZA catalyst. The selectivity of methanol production decreased compared to CZA catalyst. In this case, we can think about different actions of Ni on the reaction. Ni catalyst is also able to decompose methanol by oxidation reaction [20-21]

4. Conclusions

The well-known methanol catalyst – copper-zinc-alumna – was tested to produce methanol by direct methane conversion in a dielectric barrier discharge. The CZAs catalysts displayed higher activity and better performance to produce methanol than non-catalytic plasma process. The existence of CZA increased the selectivity and yields of methanol. Yttrium, also Pt and Fe, was found to be the best addition for the conventional CZA system in the plasma process for methanol synthesis.

Acknowledgements This study was supported by the National Research Laboratory Program of the Korea Ministry of Science and Technology. The authors thank to the Korea Institute of Science and Technology for supports.

14

References

[1]

D.J. Liu, R. Mallinson and L. Lobban, J. Catal. 179 (1998) 326.

[2]

L.M. Zhou, B. Xue, U. Kogelschatz and B. Eliasson, Plasma Chem. Plasma Process 18 (1998) 375.

[3]

A. Indarto, J.W. Choi, H. Lee and H.K. Song, Energy 31 (2006) 2986.

[4]

J.M. Cormier and I. Rusu, J. Phys. D: Appl. Phys. 34 (2001) 2798.

[5]

T. Nozaki, A. Hattori and K. Okazaki, Catal. Today 98 (2004) 607.

[6]

K. Thanyachotpaiboon, S. Chavadej, T.A. Caldwell and R. G. Mallinson, AIChE J. 44 (2004) 2252.

[7]

H.K. Song, H. Lee, J.W. Choi and B.K. Na, Plasma Chem. Plasma Process 24 (2004) 57.

[8]

X.S. Li, A.M. Zhu, K.J. Wang, Y. Xu and Z.M. Song, Catal. Today 98 (2004) 617.

[9]

W.R.A.M. Robinson and J.C. Mol, Appl. Catal. 76 (1991) 117.

[10] J.C.J. Bart and R.P.A. Sneeden, Catal. Today 2 (1987) 1. [11] J.M. Thomas and W.J. Thomas, Principles and Practice of Heterogeneous Catalysis (VCH Publishers, Weinheim, 1996). [12] A. Indarto, J.W. Choi, H. Lee and H.K. Song, in: Proc. Spring 2006 Korean Chem. Eng. Symp. [13] A. Indarto, J.W. Choi, H. Lee and H.K. Song, J. Nat. Gas Chem. 14 (2005) 13. [14] A. Indarto, J.W. Choi, H. Lee and H.K. Song, J. Nat. Gas Chem. 15 (2006) 87. [15] A. Indarto, J.W. Choi, H. Lee and H.K. Song, Energy 31 (2006) 2650. [16] A.B. Anderson and J.J. Maloney, J. Phys. Chem. 92 (1998) 809. [17] S. Wang, G. Q. Lu and G.J. Millar, Energy Fuels 10 (1996) 896. [18] F. Gudmundson, J.L. Persson, M. Försth, F. Behrendt, B. Kasemo and A. Rośen, J. Catal. 179 (1998) 420. [19] A. Indarto, J.W. Choi, L. Hwaung, H.K. Song and J. Palgunadi, J. Rare Earths 24 (2006) 513. [20] P.K. de Bokx, A.R. Balkenende and J.W. Geus, J. Catal. 117 (1989) 467. [21] F.L. Baudais, A.J. Borschke, J.D. Fedyk and M.J. Dignam, Surf. Sci. 100 (1980) 210.

15

CH4 He O2

MFC AC power supply

MFC

MFC blow heater

Plasma Reactor

heating tape

Figure 1. Experimental setup

16

GC

hydrotalcite malachite

CuO ZnO

Intensity (a.u.)

oven dried

after calcination

after reaction 10

20

30

40

50

60

70

80

o

2θ ( )

(a)

Intensity (a.u.)

hydrotalcite malachite

CuO ZnO

CZA

1%Pt/CZA

3%Ni/CZA 3%Fe/CZA 3%Y/CZA 10

20

30

40

50

60

70

80

o

2θ ( ) (b)

Figure 2. The XRD patterns of the catalyst (a) the XRD phases change of CZA at various catalyst treatments (b) the uncalcined XRD pattern of CZA as function of various types of metal loading.

17

(a)

(b)

18

(c)

Zn Cu

Cu

Al

Zn Cu Zn

(d) Figure 3. (a-c) SEM surface picture of CZA (a-b) after calcinations (c) after 6 hours plasma reaction (d) EDS elemental analysis

19

3.0 2.5 2.0 1.5 1.0

CZA CZAi CZAb Catalyst loading: 0.5 gram

0.5 0.0

0

40

80

120

160

200

Time (minute) Figure 4. Methanol increment selectivity and its stability as comparison between prepared CZA catalyst and commercialized CZA catalysts.

20

Methanol selectivity (%)

30 25 20 15 10 5 0

blank

0.5 gram

1 gram

1.5 gram

Catalyst loading Figure 5. Methanol selectivity as a function of catalyst loading.

21

50

CH4 conversion or CH3OH selectivity (%)

45 40

CH4 conversion

35 30 25

CH3OH selectivity (CZA)

20 15 10

CH3OH selectivity (blank )

5 0

50

100

150

200

Power (Watt) Figure 6. Effect of power supply on the conversion of methane and methanol selectivity

22

Methanol selectivity (%)

35 30 25 20 15 10 5 0

blank

CZA

1%Pt/CZA 3%Ni/ CZA 3%Fe/ CZA 3%Y/ CZA

Catalyst Figure 7. Effect of various types of metal loading of CZA catalyst on the methanol selectivity

23

Table 1. Products distribution as a function of various CZA catalyst catalyst blank CZA CZAi CZAb

conversion CH4 (%) 28.1 25.3 31.0 29.8

H2 17.0 10.9 11.2 14.6

CO 45.5 36.8 35.7 34.8

selectivity (%) CO2 C2 21.5 9.3 12.0 5.6 10.6 6.7 12.3 7.5

C3 1.3 0.0 0.0 0.0

C4 1.7 0.0 0.0 0.0

Table 2. Products distribution as a function of catalyst loading in the reactor metal loading (gram) blank 0.5 gram 1 gram 1.5 gram

conversion CH4 (%) 28.1 25.3 30.6 33.9

H2 17.0 10.9 13.6 12.0

CO 45.5 36.8 34.8 32.9

selectivity (%) CO2 C2 21.5 9.3 12.0 5.6 9.7 3.6 6.6 6.9

C3 1.3 0.0 0.0 0.0

C4 1.7 0.0 0.0 0.0

Table 3. Products distribution as a function of supplied power power (Watt) 50 80 110 140

conversion CH4 (%) 28.5 38.0 39.5 45.0

H2 10.9 14.7 18.9 20.5

CO 36.8 39.5 40.0 42.7

selectivity (%) CO2 C2 12.0 5.6 13.3 4.8 15.2 6.2 18.3 4.0

C3 0.0 0.0 0.0 0.0

C4 0.0 0.0 0.0 0.0

Table 4. Products distribution as a function of various metals loading doped on CZA catalyst metal loading blank CZA 1% Pt/ CZA 3% Ni/ CZA 3% Fe/ CZA 3% Y/ CZA

conversion CH4 (%) 28.1 25.3 24.6 26.0 30.8 25.3

H2 17.0 10.9 15.4 18.9 13.7 10.1

CO 45.5 36.8 35.6 35.0 34.4 33.8

24

selectivity (%) CO2 C2 21.5 9.3 12.0 5.6 9.5 3.5 15.7 7.5 12.1 9.5 9.0 5.6

C3 1.3 0.0 0.0 2.2 0.0 0.0

C4 1.7 0.0 0.0 0.0 0.0 0.0