Catalytic performance of manganese-promoted nickel catalysts for the steam reforming of tar from biomass pyrolysis to synthesis gas

Fuel 103 (2013) 122–129

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Catalytic performance of manganese-promoted nickel catalysts for the steam reforming of tar from biomass pyrolysis to synthesis gas Mitsuru Koike a, Chiaki Ishikawa b, Dalin Li b,c, Lei Wang b, Yoshinao Nakagawa b,c, Keiichi Tomishige a,b,c,⇑ a

Graduate School of Pure and Applied Sciences, University of Tsukuba, 1-1-1, Tennodai, Tsukuba, Ibaraki 305-8573, Japan Department of Applied Chemistry, School of Engineering, Tohoku University, 6-6-07, Aoba, Aramaki, Aoba-ku, Sendai 980-8579, Japan c Japan Science and Technology Agency, CREST, 6-6-07, Aoba, Aramaki, Aoba-ku, Sendai 980-8579, Japan b

a r t i c l e

i n f o

Article history: Received 22 December 2010 Received in revised form 6 April 2011 Accepted 6 April 2011 Available online 24 April 2011 Keywords: Biomass Steam reforming Tar Toluene Ni

a b s t r a c t We investigated the performance of Ni + MnOx/Al2O3 catalysts in the steam reforming of tar from the pyrolysis of cedar wood. Performance of Ni + MnOx/Al2O3 catalyst with optimum composition was much higher than those of the corresponding monometallic Ni and MnOx catalysts. This tendency is also supported by the activity test in the steam reforming of toluene as a model compound of tar. Based on the catalyst characterization results, the surface of Ni metal particles was partially covered with MnOx, and the interaction between Ni metal and MnOx can play an important role on the enhancement of the catalytic activity and the suppression of coke deposition in the steam reforming of tar. Excess MnOx addition decreased the catalytic activity by decreasing the number of the surface Ni atoms. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Conversion of biomass to synthesis gas is an important process for the production of liquid fuels by Fischer–Tropsch synthesis and chemicals by methanol synthesis, which is related to the network of C1 chemistry [1]. The production of synthesis gas from biomass is usually carried out in a non-catalytic system [2]. In the noncatalytic system, biomass is gasified with air at high temperature in order to decrease the tar formation [2]. When the catalyst is applied to the biomass gasification system, the catalysts should contribute to the decrease of tar concentration in the produced gas by catalytic conversion and the decrease of reaction temperature [3]. Our group has reported that Rh–CeO2 based catalysts exhibits high performance in the catalytic gasification of biomass with oxygen [4,5] using a fluidized bed reactor. On the Rh–CeO2 based catalysts, the interaction between Rh metal and CeO2 can be related to high catalytic performance [6]. In the case of the practical and industrial gasification process, air must be used as a gasifying agent because the production of oxygen needs very large and expensive facility. In the gasification of biomass with air, the product gas become diluted with nitrogen and is not suitable to the further conversion of synthesis gas to liquid fuels and chemicals, ⇑ Corresponding author at: Graduate School of Pure and Applied Sciences, University of Tsukuba, 1-1-1, Tennodai, Tsukuba, Ibaraki 305-8573, Japan. E-mail address: [email protected] (K. Tomishige). 0016-2361/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2011.04.009

where process needs to be pressurized. In contrast, the gasification of biomass with steam reforming of tar derived from the biomass pyrolysis can suppress the dilution of the product gas. In these reaction systems, catalysts should exhibit high activity and resistance to coke deposition in the steam reforming reaction [3], in particular, high resistance to coke deposition can be related to the enhancement of the catalyst stability [7]. As reported previously, Ni is one of the suitable components for the steam reforming of various organic compounds, and the effect of supports and additive oxides and metals with Ni has been investigated. We have reported that the addition of CeO2 to Ni catalysts enhanced the catalytic performance in terms of the activity and the suppression of coke formation. On the Ni + CeO2/Al2O3 catalysts prepared by the co-impregnation method, the formation of the Ni–CeO2 nanocomposite could give large interface between Ni metal and Ce oxide surface, and this was connected to high catalytic performance [8–12]. In addition, we have recently reported that the addition of Fe to Ni/Al2O3 catalysts promote the steam reforming reaction, and this property is caused by the Ni to Fe alloy formation [13]. On both Ni + CeO2/Al2O3 and Ni–Fe/Al2O3 catalysts, it is thought that an important point is that the presence of redox species such as cerium oxide and iron neighboring to the Ni surface. During the redox cycle the oxygen can easily transfer to the carbonaceous species and facilitates to form gas and thus it cleans the catalyst surface. In the present article, the modification of Ni/Al2O3 with Mn is attempted, because Mn species have high redox property like Ce

M. Koike et al. / Fuel 103 (2013) 122–129

species. In particular, we investigated the performance of Ni + MnOx/Al2O3 catalysts prepared by co-impregnation method in the steam reforming of tar derived from the pyrolysis of cedar wood. In addition, we tested the catalysts in the steam reforming of toluene as a model compound of tar in the biomass gasification process [14,15]. As reported previously, the reactivity of aromatic compounds is lower than that of other compounds contained in the biomass tar, and high catalytic activity of the steam reforming of toluene can be connected to high activity of the steam reforming of tar [16,17]. The catalysts were characterized by temperature programmed reduction with H2, X-ray diffraction, and H2 chemisorption, and the catalytic performance is discussed on the basis of characterization results.

2. Experimental 2.1. Catalyst preparation The support material of a-Al2O3 was prepared by the calcination of c-Al2O3 (KHO-24, Sumitomo Chemical Co., Ltd., 133 m2 g1, grain size 2–3 mm) in air at 1423 K. After calcination, it was crushed and sieved to particle sizes between 0.3 mm and 0.5 mm. The Ni + MnOx/Al2O3 catalysts were prepared by a coimpregnation method using the mixed aqueous solution of Ni(NO3)26H2O (WAKO Pure Chemical Industries, Ltd., 98.0%) and Mn(NO3)26H2O (Soekawa Chemical Co., Ltd., 99.9%). After the impregnation, the samples were dried at 383 K for 12 h followed by the calcination at 773 K for 3 h under air atmosphere. The loading amount of Ni was 12 wt.%. The additive amount of MnOx was represented as MnO2 in the range of 10–40 wt.% because the reduction degree of the Mn species was dependent on the additive amount. As a reference, the Ni + CeO2/Al2O3 catalyst with 12 wt.% Ni and 15 wt.% CeO2 was prepared by the co-impregnation method. The conditions for drying and calcining were the same as the case of Ni + MnOx/Al2O3, which were based on the previous report [9]. 2.2. Biomass Cedar wood was ground with a ball mill to about 0.1–0.3 mm size. The moisture content of the cedar wood was 7.2%. The dry basis composition by weight was C 50.8%, H 6.0%, O 41.8%, N 0.2%, and ash 1.1%. The elemental analysis was carried out by the Japan Institute of Energy. 2.3. Activity test in the steam reforming of tar derived from the pyrolysis of cedar wood The details of the procedures for the activity test in the steam reforming of tar and the diagram of the reactor have been described in our previous report [9]. Catalytic performance was evaluated using a laboratory-scale continuous feeding dual bed reactor. The biomass feeder consisted of a conical glass vessel with a screw valve at the bottom, allowing continuous feeding of biomass particles by vibrating the vessel with an electric vibrator. Nitrogen was used for transporting the biomass particles to the primary bed. Steam was supplied by the evaporation of water fed by a syringe pump. The cedar wood powder (0.1–0.3 mm) was supplied from the top of the reactor to the hot zone and thus it rapidly pyrolyzed mainly to gaseous tar and solid products. Solid products including char and ash were accumulated at the bottom of the primary bed. The dual bed reactor system prevented the contact of char and ash to catalyst. Gaseous products such as volatile tar and gases were introduced to the secondary catalyst bed together with steam in

123

order to proceed the steam reforming reactions on the catalyst surface. The details of the reaction conditions are as follow; feeding rate of biomass was 60 mg/min which provided the feeding rate of C, H, and O in 2360, 3350 and 1450 lmol/min, respectively excluding the moisture content. The feeding rate of added steam was 1110 lmol/min. The molar ratio of feeding steam including biomass moisture to feeding carbon was estimated to be 0.57. The feeding rate of N2 was 60 mL/min (2680 lmol/min). The tests were performed under atmospheric pressure. The amount of catalyst was 0.5 g. We evaluated the catalytic performance over the catalyst after the reduction pretreatment at 773 K for 0.5 h using 30 mL/min of H2. The sample of effluent gas was collected by a syringe and analyzed by a gas chromatograph (GC). The concentration of CO, CO2, and CH4 was measured by FID–GC (Gaskuropack 54) equipped with a methanator and that of H2 was determined by TCD–GC (MS-13X). The flow rate of the effluent gas was measured by a bubble flow meter. The formation rate of the gaseous products was based on the concentration obtained from the GC analyses and the flow rate of the effluent gas from the bubble flow meter. The yield of carbon-containing gaseous products (CO, CO2, CH4) was calculated by the formation rate normalized by the total carbon supplying rate of the biomass. The amount of char formed from the pyrolysis of the cedar, which was accumulated in the primary bed, can be measured after the activity test. Amount of coke deposited on the surface of catalysts can be also determined after each activity test. The amount of coke and char was determined by the amount of CO2 formed in the combustion after the reaction test, and the yield was calculated by the ratio to the total carbon amount in the fed biomass. As a result, the yields of gaseous products and solid products (coke and char) were available. In contrast, the amount of residual tar is difficult to determine because tar is easily condensable in the reactor system and a part of tar cannot be collected. Therefore, the yield of tar is obtained by the subtraction of carbon-based yield of gaseous and solid products from the total. We also evaluated the catalytic performance in terms of both the yield of residual tar and the formation rate of CO + H2 + 4CH4 for the comparison, where 4 times of CH4 is based on the reaction formula of CO hydrogenation to methane (CO + 3H2 ? CH4 + H2O). The performance was evaluated for 15 min in the activity test. The formation rate of the gaseous products was almost stable during 15 min on all the catalysts used in this study. 2.4. Steam reforming of toluene as model compound of tar derived from biomass Steam reforming of toluene was carried out using a fixed-bed reactor. The reactor was made of a quartz tube (4 mm i.d.). The reaction temperature was monitored by a thermocouple, which was inserted into the outlet of the catalyst bed. The catalyst weight was 0.1 g, which provided the bed thickness of 7 mm. Before experiment the catalysts were reduced in flowing H2 and N2 (H2/ N2 = 30/30) at 773 K for 0.5 h. After the reactor was purged with N2, the steam and toluene were supplied through vaporizing chamber. In the vaporization chamber, the water and toluene were vaporized and the carrier gas N2 sweeps out the vapor to the catalyst bed. The molar ratio of steam to carbon was S/C = 3.4 and W/ F = 0.03–1.1 g h/mol. In this case, F represents the total flow rate of the reactant gas and N2 carrier gas. The reaction temperature was 873 K. The sample of effluent gas was collected by a syringe and analyzed by a gas chromatograph (GC). The concentrations of CO, CO2, and CH4 were measured by FID–GC (Gaskuropack 54) equipped with a methanator and that of benzene and toluene was measured by FID–GC (DB-1) and that of H2 was determined by TCD–GC (MS-13X). The flow rate of the effluent gas was measured by a bubble flow meter. Toluene conversion is calculated

M. Koike et al. / Fuel 103 (2013) 122–129

2.5. Measurement of the amount of carbon deposition Thermogravimetric analyses (TGA) for the estimation of the amount of the deposited carbon were carried out by using DTA60 (Shimadzu). TGA profiles of a part of the catalysts (10 mg) after the catalytic use were obtained under air flowing (50 mL/min) at the heating rate of 10 K/min. Exothermic weight loss was observed at the temperature range between 600 and 900 K. This can be assigned to the combustion of deposited carbon [18]. It is possible to estimate the amount of carbon deposition on the basis of this weight loss. 2.6. Catalyst characterization The fresh catalysts were characterized by the measurement of H2 adsorption, which was carried out in a high-vacuum system by a volumetric method (dead volume: 65 cm3). The 0.15 g catalysts in the measurement cell were reduced with H2 at 773 K for 0.5 h and evacuated at the same temperature for 0.5 h. After the pretreatment, the adsorption amount of H2 was measured at room temperature; the equilibrium pressure was about 2.7 kPa. Temperature programmed reduction (TPR) with H2 was performed in the fixed-bed flow reactor. The TPR profile of each sample was recorded from room temperature to 1123 K under a flow of 5.0% H2/Ar, and the flow rate was 30 mL/min. The catalyst weight was 50 mg. The heating rate was 10 K/min and it was kept at final temperature 1123 K for 30 min. The consumption of H2 was monitored continuously with TCD gas chromatograph equipped with frozen acetone trap in order to remove H2O from the effluent gas. The amount of H2 consumption was estimated from the peak area in the TPR profiles. Powder X-ray diffraction (XRD) patterns of the samples after the reduction (H2 flow, 773 K, 0.5 h) were collected on a Philips X’pert diffraction-meter using Cu Ka (k = 0.154 nm) generated at 40 kV and 20 mA. 3. Results and discussion 3.1. Catalytic performance of Ni + MnOx/Al2O3 in the steam reforming of tar derived from the pyrolysis of cedar wood Fig. 1 shows the catalytic performance of Ni + MnOx/Al2O3 in steam reforming of tar at 823 K. In the case of Ni/Al2O3, the amount of the residual tar was large and the ratio of H2/CO (H2/CO) was rather low, indicating the low reforming activity of Ni/Al2O3. The addition of MnOx to Ni/Al2O3 promoted the steam reforming reaction monotonously in the range of MnO2 6 20 wt.%, and the amount of tar and coke decreased simultaneously. In contrast, the excess addition of MnOx (MnO2 > 20 wt.%) decreased the formation rate of gaseous products significantly and increased the tar amount. The catalytic activity of Ni + MnOx/Al2O3 in the steam reforming of tar, which is reflected by the residual tar amount and the formation rate of CO, H2 and CH4, is maximum at MnO2 = 20 wt.%. The volcano-type dependence of the steam reforming activity indicates that the addition of MnOx on the steam reforming activity has both promoting and suppressing effects. It should be noted that MnOx/Al2O3 exhibited much lower activity

H2/CO 1.1

1.8

1.8

2.1

2.1

0.5 3500

100 Char 80

Coke

60

3000

Tar 2500

40

CO2 CH4

2000

20 CO

Formation rate (CO+H 2+4CH4) /µmol min-1

by 1-(unreacted toluene)/(fed toluene). The selectivity of CO, CO2 and CH4 is calculated by CO/(CO + CO2 + CH4), CO2/(CO + CO2 + CH4), and CH4/(CO + CO2 + CH4), respectively. In all the experiments, the benzene formation rate (<1.2%-C) and the carbon deposition rate (<0.3%-C) were much lower than that of CO + CO2 + CH4 formation rate and these are negligible in the selectivity calculation.

Carbon based yield/%-C

124

1500

0 0

10

20

30

Ni+MnOx/Al2O3

40

20

MnOx/Al2O3

Additive amount of MnO 2 / wt% Fig. 1. Catalytic performance in steam reforming of tar over Ni + MnOx/Al2O3 catalysts at 823 K.

than Ni/Al2O3, and it is concluded that high activity of Ni + MnOx/ Al2O3 can be caused by synergy between Ni and MnOx. Fig. 2 shows the reaction temperature dependence of steam reforming of tar over Ni + MnOx/Al2O3 (MnO2 = 20 wt.%), Ni + CeO2/Al2O3 [9] and Ni/Al2O3. The activity was higher at higher reaction temperature and the amount of coke became smaller effectively at higher reaction temperature. In case of Ni/Al2O3, the tar was not removed even at 923 K. However, tar was removed almost entirely at 873 K by addition of the optimum amount of MnOx and CeO2. Regarding lower reaction temperature (823 K), the activity of Ni + MnOx/Al2O3 was so high that almost all the tar could be converted by the steam reforming reaction. In contrast, the activity of Ni + CeO2/Al2O3 catalyst was not so high as that of the Ni + MnOx/Al2O3 catalyst, and the residual tar was detected at 823 K, although almost no tar was observed on Ni + CeO2/Al2O3 catalyst at 873 K. Furthermore, Ni/Al2O3 catalysts showed lower activity than that of the Ni + MnOx/Al2O3 and Ni + CeO2/Al2O3 catalysts and it gave the residual tar even at 923 K. Another important point is that the additive effect of MnOx and CeO2 on the suppression of the coke deposition. At each reaction temperature, the order of the resistance to coke formation was as follows: Ni + MnOx/Al2O3 > Ni + CeO2/Al2O3 > Ni/Al2O3. On the basis that the suppression of coke deposition can be related to enhancement of catalyst stability in the reforming of hydrocarbons [7], it is suggested that the stability of Ni + MnOx/Al2O3 can be higher than that of the Ni + CeO2/Al2O3 and Ni/Al2O3. 3.2. Catalytic performance of Ni/Al2O3 and Ni + MnOx/Al2O3 in the steam reforming of toluene as a model compound of tar derived from the biomass pyrolysis The reaction temperature for the catalytic steam reforming and the biomass pyrolysis is not high (823–973 K) in our experiment. We reported that the tar content was as high as 30%-C at 823 K in the non-catalytic system under the same conditions [19]. These reaction conditions were closed to the pyrolysis for the production of bio-oil [20] and the temperature was too low for the non-catalytic gasification [21]. This indicates that the tar contains various kinds of compounds such as aromatic compound, aliphatic one, carboxylic one, monomeric sugars, aldehydes, furans and so on. In particular, it has been known that the reactivity of aromatic compounds is lower than other compounds, and this means that aromatic compounds tend to be difficult to remove [22]. Therefore,

125

M. Koike et al. / Fuel 103 (2013) 122–129

H2/CO 1.8

1.6

1.7

1.4

2.2

2.1

1.1

2.0

1.8

60

CO2

40

CH4

4000 3500 3000 2500

20

CO 0

2000

/µmol min

80

-1

4500

Char Coke Tar

Formation rate (CO+H2+4CH4)

Carbon based yield/%-C

100

1500 823

873

923

823

873

923

823

923

873

Catalysts and temperature / K

Ni+MnOx/Al2O3 (MnO2=20 wt%)

Ni+CeO2/Al2O3 (CeO2=15 wt%)

Ni/Al2O3

Fig. 2. Reaction temperature dependence of steam reforming of tar over Ni/Al2O3, Ni + CeO2/Al2O3 (CeO2 = 15 wt.%) and Ni + MnOx/Al2O3 (MnO2 = 20 wt.%) after H2 reduction.

(b)

15

100

40 5

Conversion or selectivity/%

10 60

H2 /CO

Conversion or selectivity/%

15

80

80

20

0

100

10 60

H2 /CO

(a)

40 5 20

0

0.2

0.4

0.6

0.8

1

0 1.2

0

0

0.2

0.4

-1

0.6

0.8

0 1.2

1

-1

W/F /g h mol

W/F /g h mol

Fig. 3. Dependence of the conversion and H2/CO in stream reforming of toluene over (a) Ni/Al2O3 and (b) Ni + MnOx/Al2O3 on contact time (W/F). h: toluene conversion. N: CO selectivity. d: CO2 selectivity. j: CH4 selectivity. }: H2/CO Reaction conditions: steam/carbon = 3.4 (toluene/H2O/N2 = 1/24/57), reaction temperature 873 K. reaction time 40 min. W/F = 0.03–1.1 g h mol1.

80 1500 60 1000 40 500 20

0 0

10

20

30

40

50

Time on stream/min

60

70

0 80

4000

100

80

3000

60 2000 40 1000 20

0 0

10

20

30

40

50

60

70

Toluene conversion/%-C

(b)

100

Formation rate/µmol min-1

2000

Toluene conversion/%-C

Formation rate/µmol min-1

(a)

0 80

Time on stream/min

Fig. 4. Reaction time dependence on the formation rates of products and toluene conversion in steam reforming of toluene at 873 K over (a) Ni/Al2O3 and (b) Ni + MnOx/Al2O3 (MnO2 = 20 wt.%). h: toluene conversion. N: CO, d: CO2, j: CH4, }: H2 Reaction conditions: steam/carbon = 3.4 (toluene/H2O/N2 = 1/24/57), reaction temperature 873 K. reaction time 80 min. W/F = 0.22 g h mol1 (a), 0.11 g h mol1 (b).

M. Koike et al. / Fuel 103 (2013) 122–129

high activity in toluene steam reforming can support the high activity in the steam reforming of the biomass tar. Fig. 3 shows the W/F dependence of toluene conversion, selectivity to CO, CO2, CH4, and the H2/CO ratio in steam reforming of toluene over Ni/Al2O3 and Ni + MnOx/Al2O3 (MnO2 = 20 wt.%) catalysts. On both of the catalysts, in the low W/F conditions, toluene conversion and the CO selectivity increased with increasing W/F, and CO2 selectivity and H2/CO ratio decreased with increasing W/ F. In this reaction, steam reforming of toluene can give CO and H2 (C7H8 + 7H2O?7CO + 11H2), where CO is converted to CO2 by water gas shift reaction (CO + H2O?CO2 + H2). The amount of the unreacted H2O decreased with increasing the W/F and thus the contribution of water gas shift reaction decreased, which resulted in decreased CO2 selectivity. From comparison between Ni/Al2O3 and Ni + MnOx/Al2O3 at low W/F region, Ni + MnOx/Al2O3 gave higher toluene conversion than Ni/Al2O3, indicating that the Ni + MnOx/Al2O3 exhibited higher activity than that of Ni/Al2O3. On the other hand, in the high W/F region, all the toluene is converted, and the selectivity of products approached almost to the same level, which corresponds to the equilibrium. Fig. 4 shows the reaction time dependence of Ni/Al2O3 and Ni + MnOx/Al2O3. The formation rate of the products and conversion decreased gradually on Ni/Al2O3. In contrast, the formation rate and the conversion were stable on Ni + MnOx/Al2O3 indicating that the addition of MnOx can be effective to the enhancement of the catalyst stability. In this experiment, since the activity of Ni/ Al2O3 and Ni + MnOx/Al2O3 was different, we chose the different W/F in order to adjust the same conversion level. This is because the carbon deposition has two formation routes: decomposition of toluene (C7H8?7C + 4H2) and CO disproportionation (2CO?C + CO2). When the conversion level is almost the same by adjusting the W/F condition, the partial pressure of the reactant and products in the catalyst bed on each catalyst is also similar. Fig. 5 shows the TG profiles on Ni/Al2O3 and Ni + MnOx/Al2O3 after the catalytic use shown in Fig. 4. Weight loss in the range of 700–900 K assigned to the combustion of coke was observed on both catalysts, and the deposited coke amount was determined by 110 mgg1 cat and 60 mg gg1 cat on Ni/Al2O3 and Ni + MnOx/Al2O3, respectively. This indicates that the MnOx addition enhanced the resistance to coke deposition. Considering that the W/F on Ni + MnOx/Al2O3 was half of that on Ni/Al2O3, Ni + MnOx/Al2O3 has much higher resistance to coke deposition. This can be related to higher stability of Ni + MnOx/Al2O3. Furthermore, the tendency of the results in steam reforming of toluene is similar to that in the steam reforming of tar.

5

300 -5 200 -10 100 -15

0

500

700

900

Temperature/K

1100

-100

5

500

400

0

Weight change/wt%

Weight change/wt%

Fig. 6. XRD patterns of catalysts: after calcination at 773 K for 3 h (a) Ni/Al2O3 (b) Ni + MnOx/Al2O3 (MnO2 = 10 wt.%) (c) Ni + MnOx/Al2O3 (MnO2 = 20 wt.%) (d) Ni + MnOx/Al2O3 (MnO2 = 30 wt.%) (e) Ni + MnOx/Al2O3 (MnO2 = 40 wt.%). h = Al2O3, N = NiO, O = Mn2O3, s = NiMnO3.

400

0

-20 300

Fig. 6 shows the XRD patterns of Ni/Al2O3 and Ni + MnOx/Al2O3 after the calcination at 773 K for 3 h. The peak intensity was normalized by the peak assigned to a-Al2O3 at 2h = 43.4°. After calcination, on Ni/Al2O3, the peak at 2h = 37.3° was assigned to NiO. The peak due to NiMnO3 (2h = 33.7°, 36.6°) was grew gradually with increasing the additive amount of Mn. These results show that added MnO2 mainly interacted with NiO to form NiMnO3, in particular, NiO and NiMnO3 was formed on the Ni + MnOx/Al2O3 catalyst (MnO2 < 40 wt.%). Furthermore, the addition of excess Mn gave Mn2O3 (2h = 33.0°) at MnO2 = 40 wt.%.

(b)

500

DTA/µV

(a)

3.3. Catalyst characterization

300 -5 200 -10

DTA/µV

126

100 -15

-20 300

0

500

700

900

1100

-100

Temperature/K

Fig. 5. TG–DTA profiles of (a) Ni/Al2O3 and (b) Ni + MnOx/Al2O3 (MnO2 = 20 wt.%) after the catalytic use. Measurement conditions: air flow, heating rate 10 K/min, sample weight 10 mg. Reaction conditions: steam/carbon = 3.4 (toluene/H2O/N2 = 1/24/57), reaction temperature 873 K. reaction time 80 min. W/F = 0.22 g h mol1 (a), 0.11 g h mol1 (b). Catalysts were used in Fig. 4.

127

M. Koike et al. / Fuel 103 (2013) 122–129

Fig. 7 shows the temperature programmed reduction (TPR) profiles of Ni + MnOx/Al2O3 together with those of MnOx/Al2O3 and Ni/ Al2O3 for the comparison. Since the temperature for the H2 reduction pretreatment in the catalytic reaction is 773 K, the amount of the H2 consumption below 773 K was considered, and the results are listed in Table 1. On MnOx/Al2O3, two H2 consumption peaks were observed with maximum at 600 K and 700 K, which can be attributed to the reduction of MnO2 to Mn3O4 and Mn3O4 to MnO, respectively [23]. On Ni/Al2O3, the H2 consumption with maximum at approximately 700 K was typically due to the reduction of NiO [24]. Almost all the Ni species on Ni/Al2O3 were reduced on the basis of H2 consumption amount and the reduction degree (Table 1). From comparison between MnOx/Al2O3 and Ni/Al2O3, the reduction of Mn species on the support proceeds at lower temperature than NiO on the support which was observed by XRD as shown in Fig. 6. On the Ni + MnOx/Al2O3 (MnO2 5 20 wt.%) cata-

Ni+MnOx/Al2O3 [MnO2/wt%]

[40]

H2 consumption/a.u.

[30] [20] [10] [0]

MnOx/Al2O3 273

473

673

873

1073

Temperature / K Fig. 7. TPR profiles of Ni + MnOx/Al2O3, Ni/Al2O3 and MnOx/Al2O3. Conditions: heating rate 10 K/min, 5% H2/Ar flow rate 30 ml/min. Sample weight: 50 mg. Loading amount of Mn on MnOx/Al2O3 was 20 wt.% MnO2.

lysts, H2 consumption at low temperature, which is observed on MnOx/Al2O3, was not observed at all. This behavior indicates that almost all the added Mn interacted with NiO, which was also supported by the XRD results (Fig. 6). The H2 consumption peak on the Ni + MnOx/Al2O3 (MnO2 5 20 wt.%) catalysts in the temperature range of 573–850 K can be assigned to the reduction of NiMnO3 and NiO. On the Ni + MnOx/Al2O3 (MnO2 > 20 wt.%), the H2 consumption peak is not so different from that on Ni + MnOx/Al2O3 (MnO2 5 20 wt.%). This can be explained by the formation of Mn2O3 oxide detected in the XRD results. Fig. 8 shows the XRD patterns of Ni/Al2O3 and Ni + MnOx/Al2O3 after the reduction at 773 K for 30 min. On the reduced Ni/Al2O3, the peak at 2h = 44.4° assigned to Ni metal was observed, and an average size of Ni metal particles was determined to be 27.1 nm calculated from the peak width [25]. On the Ni + MnOx/Al2O3 catalysts, the peak assigned to MnO increased gradually with increasing Mn amount. The average size of Ni metal particles determined from XRD on the Ni + MnOx/Al2O3 catalysts (MnO2 = 10–30 wt.%) was about 20 nm, which was smaller than that on Ni/Al2O3, indicating the addition of Mn promoted the dispersion of Ni metal. Too much Mn addition suppressed the dispersion on the Ni + MnOx/Al2O3 catalyst (MnO2 = 40 wt.%). The intensity of the diffraction peak due to Ni metal was minimized by the addition of 20 wt.% MnO2, however, the reason is not clear at present and further investigation is necessary. The amount of H2 adsorption on the fresh catalysts after the reduction at 773 K is listed in Table 1. The amount of H2 adsorption on Ni + MnOx/Al2O3 decreased almost monotonously with increasing MnOx additive amount. The dispersion calculated from the H2 adsorption amount and that from the particle size based on the XRD are also listed in Table 1. In the case of Ni/Al2O3, two dispersion values agreed well. On the other hand, in the case of Ni + MnOx/Al2O3, the dispersion from the H2 adsorption amount is smaller than that from the XRD results, suggesting the suppression of H2 adsorption by the coverage of MnOx on the Ni particle surface. The similar interpretation has been applied to various metal particles modified with metal oxide species such as Rh + CeO2 [26], Rh + ReOx [27], Pt + ReOx [28], Rh + MoOx [29] and Ni + CeO2 [9] catalyst systems. Another important point is that the catalytic performance of Ni + MnOx/Al2O3 (MnO2 = 20 wt.%) in steam reforming of tar from biomass pyrolysis and toluene in terms of catalytic activity and resistance to coke deposition was higher than that of Ni/Al2O3, although the H2 adsorption amount on Ni + MnOx/Al2O3 (MnO2 = 20 wt.%) smaller than that on Ni/Al2O3. Therefore, the promoting effect of MnOx addition on Ni/Al2O3 cannot be interpreted

Table 1 Properties of Ni and Ni–MnOx catalysts. Catalyst

Loading amount MnOx (as MnO2)

Ni

Ni/Al2O3 Ni + MnOx/ Al2O3

MnOx/ Al2O3 a b c d e f

wt.%

mmol g1 cat

wt.%

mmol g1 cat

12 12 12 12 12 –

2.0 2.0 2.0 2.0 2.0 –

– 10 20 30 40 20

– 1.0 2.0 3.0 4.0 2.0

H2 consumptionain TPR/mmol g1 cat

2.0 3.0 3.9 4.6 4.9 2.1

Ni-based reduction degree b/%

100 150 195 230 245 –

H2 adsorption c/ lmolg1 cat

32 21 21 20 17 –

H2 consumption below 773 K in TPR profiles shown in Fig. 4. H2 consumption/Ni, the stoichiometry of the Ni reduction is Ni2+ + H2 ? Ni0 + 2H+. Total hydrogen adsorption at 298 K. Calculated from the Scherrer equation25. Dispersion of Ni from XRD is calculated by the relation: (dispersion/%) = 9.71/(particle size/nm)10 [30]. 2  (H2 adsorption)/Ni metal  100.

Particle size of Ni from XRD d/nm

27.1 20.8 21.4 20.6 27.9 –

Dispersion/%

XRDe

H2 adsorptionf

3.6 4.7 4.5 4.7 3.5 –

3.2 2.1 2.1 2.0 1.7 –

128

M. Koike et al. / Fuel 103 (2013) 122–129

4. Conclusions (i) The addition of MnOx to Ni/Al2O3 at the optimum amount enhanced the catalytic performance regarding the activity and suppression of coke deposition in the steam reforming of tar produced from the pyrolysis of cedar wood. (ii) The optimized Ni + MnOx/Al2O3 catalyst exhibited higher catalytic performance than Ni/Al2O3 in the steam reforming of toluene as a model compound in terms of the catalytic activity, the suppression of coke deposition, and the catalyst stability. (iii) Catalyst characterization by hydrogen adsorption and XRD indicated that the surface of Ni metal particles was partially covered with MnOx. (iv) High performance in the steam reforming activity of the optimized Ni + MnOx/Al2O3 catalyst is caused by the synergy between the activation of tar on the Ni species and the oxygen atom supplied to the carbonaceous intermediate from neighboring MnOx.

Acknowledgements

Fig. 8. XRD patterns of catalysts: after H2 reduction at 773 K for 30 min. (a) Ni/ Al2O3 (b) Ni + MnOx/Al2O3 (MnO2 = 10 wt.%) (c) Ni + MnOx/Al2O3 (MnO2 = 20 wt.%) (d) Ni + MnOx/Al2O3 (MnO2 = 30 wt.%) (e) Ni + MnOx/Al2O3 (MnO2 = 40 wt.%). h = Al2O3, 4 = Ni, d = MnO.

by the increase of the active sites, but by the synergy between Ni and MnOx on MnOx-modified Ni surface.

3.4. Promoting effect of MnOx addition to Ni catalysts in the steam reforming It has been known that monometallic Ni catalysts exhibited much higher reforming activity of hydrocarbons than monometallic MnOx catalysts. This property is caused by the high activation ability of C–H and C–C bond in the hydrocarbon molecules on the Ni metal surface. Therefore, it seems that the additive effect of MnOx is the increase of the number of active Ni surface, but in fact, the surface Ni atoms are not increased by the Mn addition caused by the site blocking of surface Ni atoms with MnOx. Another possibility is that MnOx has cocatalytic function with high redox property. It is suggested that the MnOx species are present on the surface of the Ni particles, the oxygen atoms derived from MnOx species can be supplied to the Ni species to promote the reaction between carbonaceous species on Ni and oxygen species. Similar role on the supply of oxygen has been reported over CeO2 [9]. The Ni metal–CeO2 nanocomposite catalyst prepared by the co-impregnation method has large interface between Ni metal and CeO2 and it can be connected to high reforming activity of tar from biomass pyrolysis [9]. In addition, the Rh + CeO2 catalysts have been reported to be effective to the biomass gasification [4– 6], and the catalytic partial oxidation of methane at short contact time, where CeO2 play an important role in the supply of oxygen species to Rh metal surface [16]. Promoting effect of MnOx addition can be explained by this role of the cocatalyst. On the other hand, the suppressing effect of excess MnOx addition can be interpreted by the increase of MnOx-covering Ni atoms. The decrease of the activity cannot be compensated by the promoting effect of MnOx addition on the catalysts with the excess MnOx addition.

This work was in part supported by Japan Science and Technology Agency (JST), Ministry of Education, Culture, Sports, Science and Technology, Japan, and Showa-Shell oil company. References [1] Huber GW, Iborra S, Corma A. Synthesis of transportation fuels from biomass: Chemistry, catalysts, and engineering. Chem Rev 2006;106:4044–98. [2] Bridgwater AV. Catalysis in thermal biomass conversion. Appl Catal A: Gen 1994;116:5–47. [3] Xu C, Donald J, Byambajav E, Ohtsuka Y. Recent advances in catalysts for hotgas removal of tar and NH3 from biomass gasification. Fuel 2010;89:1784–95. [4] Asadullah M, Ito S, Kunimori K, Yamada M, Tomishige K. Biomass gasification to hydrogen and syngas at low temperature: Novel catalytic system using fluidized-bed reactor. J Catal 2002;208:255–9. [5] Tomishige K, Asadullah M, Kunimori K. Syngas production by biomass gasification using Rh/CeO2/SiO2 catalysts and fluidized bed reactor. Catal Today 2004;89:389–403. [6] Miyazawa T, Okumura K, Kunimori K, Tomishige K. Promotion of oxidation and reduction of Rh species by interaction of Rh and CeO2 over Rh/CeO2/SiO2. J Phys Chem C 2008;112:2574–83. [7] Asadullah M, Miyazawa T, Ito S, Kunimori K, Yamada M, Tomishige K. Gasification of different biomasses in a dual-bed gasifier system combined with novel catalysts with high energy efficiency. Appl Catal A: Gen 2004;267:95–102. [8] Tomishige K, Kimura T, Miyazawa T, Nishikawa J, Kunimori K. Promoting effect of the interaction between Ni and CeO2 on steam gasification of biomass. Catal Commun 2007;8:1074–9. [9] Kimura T, Miyazawa T, Nishikawa J, Miyao T, Naito S, Okumura K, et al. Development of Ni catalysts for tar removal by steam gasification of biomass. Appl Catal B: Environ 2006;68:160–70. [10] Nishikawa J, Nakamura K, Asadullah M, Miyazawa T, Kunimori K, Tomishige K. Catalytic performance of Ni/CeO2/Al2O3 modified with noble metals in steam gasification of biomass. Catal Today 2008;131:146–55. [11] Nishikawa J, Miyazawa T, Nakamura K, Asadullah M, Kunimori K, Tomishige K. Promoting effect of Pt addition to Ni/CeO2/Al2O3 catalyst for steam gasification of biomass. Catal Commun 2008;9:195–201. [12] Nakamura K, Miyazawa T, Sakurai T, Miyao T, Naito S, Begum N, et al. Promoting effect of MgO addition to Pt/Ni/CeO2/Al2O3 in the steam gasification of biomass. Appl Catal B: Environ 2009;86:36–44. [13] Wang L, Li D, Koike M, Koso S, Nakagawa Y, Xu Y, et al. Catalytic performance and characterization of Ni–Fe catalysts for the steam reforming of tar from biomass pyrolysis to synthesis gas. Appl Catal A: Gen 2011;392:248–55. [14] Swierczynski D, Libs S, Courson C, Kiennemann A. Steam reforming of tar from a biomass gasification process over Ni/olivine catalyst using toluene as a model compound. Appl Catal B: Environ 2007;74:211–22. [15] Srinakruang J, Sato K, Vitidsant T, Fujimoto K. A highly efficient catalyst for tar gasification with steam. Catal Commun 2005;6:437–40. [16] Bona S, Guillen P, Alcalde JG, Garcıa L, Bilbaoa R. Toluene steam reforming using coprecipitated Ni/Al catalysts modified with lanthanum or cobalt. Chem Eng J 2008;137:587–97. [17] Duprez D. Selective steam reforming of aromatic compounds on metal catalysts. Appl Catal A: Gen 1992;82:111–57.

M. Koike et al. / Fuel 103 (2013) 122–129 [18] Tomishige K, Himeno Y, Matsuo Y, Yoshinaga Y, Fujimoto K. Catalytic performance and carbon deposition behavior of a NiOMgO solid solution in methane reforming with carbon dioxide under pressurized conditions. Ind Eng Chem Res 2000;39:1891–7. [19] Miyazawa T, Kimura T, Nishikawa J, Kado S, Kunimori K, Tomishige K. Catalytic performance of supported Ni catalysts in partial oxidation and steam reforming of tar derived from the pyrolysis of wood biomass. Catal Today 2006;115:254–62. [20] Putun AE, Ozcan A, Putun E. Pyrolysis of hazelnut shells in a fixed-bed tubular reactor: yields and structural analysis of bio-oil. J Anal Appl Pyrolysis 1999;52:33–49. [21] Zhou J, Chen Q, Zhao H, Cao X, Mei Q, Luo Z, et al. Biomass–oxygen gasification in a high-temperature entrained-flow gasifier. Biotechnol Adv 2009;27:606–11. [22] Coll R, Salvado J, Farriol X, Montane D. Steam reforming model compounds of biomass gasification tars: conversion at different operating conditions and tendency towards coke formation. Fuel Process Technol 2001;74:19–31. [23] Kapteijn F, Singoredjo L, Moulijn JA. Activity and selectivity of pure manganese oxides in the selective catalytic reduction of nitric oxide with ammonia. Appl Catal B: Environ 1994;3:173–89.

129

[24] Yoshida K, Begum N, Ito S, Tomishige K. Oxidative steam reforming of methane over Ni/a-Al2O3 modified with trace noble metals. Appl Catal A: Gen 2009;358:186–92. [25] Sashital SR, Cohen JB, Burwell Jr RL, Butt JB. Pt/SiO2: II. Characterization of the gel and the platinum particles by X-Ray diffraction. J Catal 1977;50:479–93. [26] Tanaka H, Kaino R, Okumura K, Kizuka T, Tomishige K. Catalytic performance and characterization of Rh–CeO2/MgO catalysts for the catalytic partial oxidation of methane at short contact time. J Catal 2009;268:1–8. [27] Shinmi Y, Koso S, Kubota T, Nakagawa Y, Tomishige K. Modification of Rh/SiO2 catalyst for the hydrogenolysis of glycerol in water. Appl Catal B: Environ 2010;94:318–26. [28] Ebashi T, Ishida Y, Nakagawa Y, Ito S, Kubota T, Tomishige K. Preferential CO Oxidation in a H2-Rich Stream on PtReOx/SiO2: Catalyst Structure and Reaction Mechanism. J Phys Chem C 2010;114:6518–26. [29] Koso S, Ueda N, Shinmi Y, Okumura K, Kizuka T, Tomishige K. Promoting effect of Mo on the hydrogenolysis of tetrahydrofurfuryl alcohol to 1,5-pentanediol over Rh/SiO2. J Catal 2009;267:89–92. [30] Mustard DG, Bartholomew CH. Determination of metal crystallite size and morphology in supported nickel catalysts. J Catal 1981;67:186–206.