- Email: [email protected]
reformer system utilizing monolith catalysts for hydrogen production
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Combined pre-reformer/reformer system utilizing monolith catalysts for hydrogen production Jung-Il Yang a, Jae-Hong Ryu b, Kwan-Young Lee b, Nam-Jo Jung a, Ji Chan Park a, Dong Hyun Chun a, Hak-Joo Kim a, Jung Hoon Yang a, Ho-Tae Lee a, Inho Cho c, Heon Jung a,* a
Korea Institute of Energy Research, 71-2 Jang-Dong, Yuseong-Gu, Daejeon 305-343, South Korea Department of Chemical and Biological Engineering, Korea University, 5-1 Anam-Dong, Seongbuk-Gu, Seoul 136-701, South Korea c SK innovation, 140-1 Wonchon-Dong, Yuseong-Gu, Daejeon 305-712, South Korea b
article info
abstract
Article history:
The pre-reforming of higher hydrocarbon, propane, was performed to generate
Received 8 January 2011
hydrogen from LPG without carbon deposition on the catalysts. A Ru/Ni/MgAl2O4
Received in revised form
metallic monolith catalyst was employed to minimize the pressure drop over the
1 April 2011
catalyst bed. The propane pre-reforming reaction conditions for the complete conver-
Accepted 17 April 2011
sion of propane with no carbon formation were identified to be the following: space
Available online 24 May 2011
velocities over 2400 h1 and temperatures between 400 and 450 C with a H2O/C1 ratio of 3. The combined pre-reformer and the main reformer system with the Ru/Ni/MgAl2O4
Keywords:
metallic monolith catalyst was employed to test the conversion propane to syngas
Pre-reformer
where the reaction heat was provided by catalytic combustors. Propane was converted
Reformer
in the pre-reformer to 52.5% H2, 27.0% CH4, 17.5% CO, and 3.0% CO2 with a 331 C inlet
C catalyst outlet temperature. The main steam reforming
Metallic monolith catalyst
temperature and a 482
Catalytic combustor
reactor converted the methane from the pre-reformer with a conversion of higher than
Pilot scale
99.0% with a 366 C inlet temperature and an 824 C catalyst outlet temperature. With
Hydrogen production
a total of 912 cc of the Ru/Ni/MgAl2O4 metallic monolith catalyst in the main reformer, the H2 production from the propane reached an average of 3.25 Nm 3h1 when the propane was fed at 0.4 Nm3h1. Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Pre-reforming technology has been used for many years as the first stage in the steam reforming of heavy hydrocarbons, e.g. LPG or naphtha, to synthesize gas or H2 [1]. This process is a valuable tool for improving the efficiency and solving problems related to the conventional tubular steam reforming process (main reforming) [2]. During
pre-reforming, heavy hydrocarbons are first converted to methane and carbon oxides (CO, CO2) at relatively low temperatures, typically from 400 to 550 C, and finally H2rich streams are produced in the main reforming process at higher temperatures of >800 C [3e5]. Pre-reforming operated at low temperatures is an inevitable process because the steam reforming of higher hydrocarbons, which generally occurs at high temperatures, easily results in
* Corresponding author. Tel.: þ82 42 860 3663; fax: þ82 42 860 3134. E-mail address: [email protected] (H. Jung). 0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2011.04.130
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
coke deposition on the walls of the reformer and catalyst surface [6]. The pre-reforming process can be expressed with the following reactions [7]. Cn Hm þ nH2 O/nCO þ ðn þ ðm=2ÞÞH2 1 ¼ 498:1 kJmol ; for n ¼ 3 CO þ 3H2 ¼ CH4 þ H2 O CO þ H2 O ¼ CO2 þ H2
DH298K
1 DH298K ¼ 206:2 kJmol 1 DH298K ¼ 41:2 kJmol
(1)
(2)
(3)
First, a mixture of CO and H2 is produced from irreversible steam reforming (1) of higher hydrocarbons (n > 1). CH4 and CO2 are also formed by the subsequent methanation reaction (2) and water-gas shift reaction (3), respectively. The recently introduced steam reforming process is generally composed of 2 stages, main steam reformer and pre-reformer, due to the many benefits of using an additional pre-reformer [8,9]. Significant energy savings can be obtained by reducing the steam-to-carbon ratio. Moreover, carbon formation caused by the cracking of higher hydrocarbons in the top of the reformer tubes is avoided by the installation of a prereformer [2]. A nickel containing steam reforming catalyst can also be used in the pre-reforming process. However, a nickel based catalyst is easily deactivated during reforming processes, especially due to poisoning of the catalyst surface by coke deposition (coking), when the operations are performed at low steam-to-carbon ratios and high temperatures. To alleviate the coking problem, the addition of noble metals to the Ni catalysts, such as Ru, have resulted in significant improvements in both the activity and coke-resisting ability of silica supported Ni catalysts in the CO2 reforming reaction [10]. The steam reforming reaction over pelletized Ni catalysts is a well-known heat-transfer limited reaction [11]. We washcoated Ru-doped Ni catalyst powders on metallic monoliths and they had an enhanced heat-transfer capability. There are increased needs for small-scale hydrogen production units, especially for hydrogen refueling stations [12,13]. When the feed for the hydrogen product is LPG or naphtha, compact pre-reformer/reformer combination units are necessary [3]. Therefore, we developed a novel selfsustaining 5 kWe-class LPG reformer, which mainly consists of a very active metallic monolith catalyst in a highly endothermic reforming reaction, a combined compact system of a pre-reformer/reformer, and a catalytic combustor to supply the endothermic reaction heat. That is, there are three new ideas in this work. Firstly, we investigated the behavior of a Ru-doped Ni/MgAl2O4 catalyst washcoated on the metallic monolith for both the propane pre-reforming reaction and the main steam reforming reaction. Secondly, a combined unit of a pre-reformer/reformer at a pilot scale was developed for the LPG-fed hydrogen production unit. Finally, the endothermic reaction heat was supplied by a combustion reactor equipped with Pd catalysts that were supported on the ceramic monolith.
2.
Experimental
2.1.
Catalyst preparation
8851
The commercial Ni-based steam reforming catalyst (SBET ¼ 19.2 m2g1) was washcoated on the metallic monolith. The preparation procedure can be found elsewhere [14]. Fecralloy plate (Fe 72.8%/Cr 22%/Al 5%/Y 0.1%/Zr 0.1%, Goodfellow Co.) was used to make the monolith (640 cpi) with dimensions of 2.2 cm (diameter) 2 cm (height). Once the washcoating of the Ni catalysts was completed, a small amount of noble metal (0.12 wt.% Ru) was doped into the Ni catalyst to serve as both the reduction-enhancing agent and the coke-resisting agent. RuCl3 (RuCl3$xH2O, Aldrich Chem. Co.) was used as a noble metal precursor and the wet impregnation method was used.
2.2.
Catalyst characterization
The surface area, pore size, and pore volume of the catalyst were measured by N2 adsorptionedesorption isotherms using a BELSORP mini-II (BEL Japan Inc.) apparatus. Before measurements, the samples were degassed in a vacuum at 573 K for 6 h. XRD data of the sample was collected in a Rigaku D-Max 2500 (18 kW) diffractometer with Cu Ka radiation. The stepscans were taken over the range of 2q from 20 to 80 in steps of 0.02 /s. For the TPR experiments, the catalyst sample was purged with 50 ml/min of Ar for 1 h at 500 C, and then cooled to room temperature. The level of hydrogen consumption was measured using a BELCAT Catalyst Analyzer (BEL Japan Inc.) by heating the sample from ambient temperature to 1000 C with a heating rate of 10 C/min in a flow of 5% H2/Ar at 30 ml/ min. The real content of Ru was measured in the samples by inductively coupled plasma-atomic emission spectroscopy (ICP-AES), with a PolyScan 60E (Hewlett Packard, USA). The metal (Ni, MgO, Al2O3) loading amounts of the samples were measured by a MiniPal2 energy dispersive X-ray fluorescence (EDXRF) spectrometer. The carbon formation on the catalyst after the prereforming reaction was investigated by Cs-corrected STEM using a JEOL microscope that was operated at 200 kV (National Nanofeb Center, Korea). The samples were prepared by putting a few drops of the corresponding colloidal solutions on carbon coated copper grids (Ted Pellar, Inc.).
2.3. Catalytic reaction and pre-reformer/reformer system To evaluate the behavior of the Ru-doped Ni/MgAl2O4 catalyst washcoated on the metallic monolith for the pre-reforming reaction, steam reforming of propane was performed in a tubular stainless steel reactor (inside diameter of 2.22 cm) operated at atmospheric pressure. Propane was supplied from a cylinder of C3H8 (99.999%) by a mass flow controller (Brooks, 5850E). Steam was supplied by a steam generator. One cylindrical metallic monolith catalyst (2.2 cm diameter and 2 cm
8852
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
Fig. 1 e Schematic of diagram of combined pre-reformer and main reformer system.
height, volume of 7.6 cm3) was used for the test. A monolith without a washcoat was placed in front of the washcoated catalyst for the even distribution of the reactant gas. A gas mixture of C3H8 and steam was passed through the metallic monolith catalyst at various gas hourly space velocities (GHSV). The ratio of H2O/C1 was also varied between 1 and 3. The compositions of the products were analyzed by a gas chromatograph (HP5890, TCD) equipped with a Porapak Q column (8.2 ft 1/8 inch). Based on the flow rate and the composition of the product gas, the conversion of C3H8 was calculated with the following equation: C3 H8 conversionð%Þ ¼ ½inlet flow rate of C3 H8 outlet flow rate of C3 H8 = ðinlet flow rate of C3 H8 Þ 100
(4)
And conversion of C1 is defined as how much carbon in the inlet C3H8 is converted to CO and CO2 and is calculated as: C1 conversionð%Þ ¼ ½3 inlet flow rate of C3 H8 3 outlet flow rate of C3 H8 outlet flow rate of CH4 = 3 inlet flow rate of C3 H8 100
The pre-reformer and the reformer were shell-and-tube heatexchanger type reactors made of stainless steel. There were 10 tubes (2.2 cm ID 50 cm height) and 6 tubes in the reactors of the pre-reformer and the reformer, respectively. Each tube was filled with 20 cylindrical metallic monolith catalysts with 2.2 cm diameters and 2 cm heights (total volume of the catalyst: pre-reformer ¼ 1520 cc, reformer ¼ 912 cc). As shown in Fig. 1, two catalytic combustors were used to supply heat to the pre-reformer and the reformer. Pd washcoated ceramic monolith catalysts (6.2 cm diameters 2 cm heights) were used in the catalytic combustors. The propane and air were used as feed gases for combustor operation. Propane was fed to the combustor for the reformer in two nozzles, with the primary fuel combusted over the catalyst bed to raise the gas to a high enough temperature to burn the secondary fuel without the catalyst (Fig. 2). The combustor for the pre-reformer was a single-fuel-nozzle combustor with the nozzle placed in front of the catalyst bed, since the temperature requirement for the pre-reformer was lower than that of the reformer.
(5)
We designed and installed a combined pre-reformer/ reformer system to produce 4 Nm3h1 of hydrogen (Fig. 1).
3.
Results and discussion
3.1.
Catalyst characterizations
Surface area, pore size, and pore volume of the catalyst sample are summarized in Table 1. It can be seen that the surface area of the Ru-doped Ni/MgAl2O4 catalyst decreased compared to that of the commercial catalyst due to the two
Table 1 e Physical properties of the Ru-doped Ni/MgAl2O4 catalyst. BET surface area (m2g1) Fig. 2 e Schematic diagram of catalytic combustor for main reformer with duel fuel injection.
14.6
Ave. pore diameter (nm)
Ave. pore volume (cm3g1)
29.4
0.107
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
8853
Fig. 4 e TPR curve of the Ru-doped Ni/MgAl2O4 catalyst calcined at 900 C. Fig. 3 e X-ray diffraction pattern of the Ru-doped Ni/ MgAl2O4 catalyst calcined at 900 C.
calcination procedures at 900 C followed by catalyst washcoating on the metallic monolith and Ru doping. The X-ray diffraction (XRD) pattern of the Ru-doped Ni/ MgAl2O4 catalyst is shown in Fig. 3 and it matches with the MgAl2O4 (JCPDS No. 05-0672), NiAl2O4 (JCPDS No. 10-0339), MgNiO2 (JCPDS No. 24-0712), Al2O3 (JCPDS No. 46-1212), and RuO2 (JCPDS No. 43-1027) phases [15]. The TPR profile of the catalyst is shown in Fig. 4 and two hydrogen consumption peaks are observed in the TPR profile. The first shoulder peak at 580 C was assigned to the reduction of NiO interacting weakly with the support (MgO and/or Al2O3) or to small Ni particles [16e18]. The second peak at 770 C was assigned to the reduction of stable Ni2þ compounds where Ni2þ is strongly interacting with MgO and/or Al2O3, which was clearly identified in the XRD result [16e18]. The loading content of Ru was measured to be 0.1% by using inductively coupled plasma-atomic emission spectrometry (ICP-AES).
From the elemental analysis by EDXRF, the weight percentages of Ni, MgO, and Al2O3 were measured to be 18, 17, and 65%, respectively. TEM results of the Ru-doped Ni/MgAl2O4 catalyst before and after the pre-reforming reaction are illustrated on Fig. 5A and B, respectively. Ni particles were observed to have various sizes, and the filamentous carbons were not observed for the fresh catalyst run from the TEM analysis (Fig. 5A). However, it can be seen (Fig. 5B) that the TEM image of the catalyst powders after the reaction shows the generation of bright branches on particle surface which clearly indicate carbon formation during the pre-reforming reaction.
3.2.
Propane pre-reforming
Avoiding carbon formation is the most important aspect of operating pre-reformers. Although some information relating to the operating conditions for the possible carbon formation can be understood from thermodynamics, experimental
Fig. 5 e TEM images of the Ru-doped Ni/MgAl2O4 catalyst (A) before pre-reforming reaction and (B) after the reaction: T [ 425 ± 25 C, SV [ 1800 h-1, and S/C1 [ 1.
8854
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
Table 2 e Pre-reforming of propane with various space velocities (S/C1 [ 1,T [ 425 ± 25 C).
Table 3 e Pre-reforming of propane according to H2O/C1 ratio (SV [ 1800 hL1, T [ 425 ± 25 C).
Space velocity (h1)
Steam/Carbon ()
1800
2400
42 yes
24 no
C3H8 conversion (%) Coking
C3H8 conversion (%) Coking
1
2
3
42 yes
100 yes (partially)
100 no
C1 ratio of 2 (Table 3). However, a small amount of coke was formed on the catalyst. Thus, the safe operating conditions of propane pre-reforming over the Ru-doped Ni metallic monolith catalyst are: GHSV of about 1800 h1, temperatures higher than 425 C and H2O/C1 ratio higher than 3.
verification by the use of a real catalyst and a test with a larger scale unit are still very important. The pre-reforming of propane was performed at two different space velocities with the H2O/C1 ratio set to 1 over a single cylindrical Ru-doped Ni metallic monolith catalyst (7.6 cm3). Since the pre-reforming reaction is the combination of the endothermic steam reforming reaction and the exothermic methanation reaction, the temperature of the catalyst fluctuated during the run and the average temperature at the exit of the catalyst was 425 C with a high of 450 C and a low of 400 C. As listed in Table 2, the propane conversion was 24% at a GHSV of 2400 h1. After the run, the catalyst was examined and no carbon was found on the catalyst. At the lower GHSV of 1800 h1, the propane conversion was increased to 42%, but coke was formed on the catalyst (Fig. 5B). Thus, the H2O/C1 ratio of 1 is too low both for the complete conversion of propane and for the avoidance of coke formation. While maintaining the GHSV at 1800 h1 and the average temperature at 425 C, we increased the H2O/C1 ratio and the result is listed in Table 3. As the H2O/C1 ratio was increased to 3, the propane conversion reached 100%. At the same time, there was no coke formation. At 425 C and a GHSV of 1800 h1, the H2O/C1 ratio of 3 is required to obtain the complete conversion of propane with no danger of carbon formation. As the temperature was increased to 500 C, the complete conversion of propane was achieved even at a H2O/
3.3. scale
Combined pre-reformer/reformer test at the pilot
A pre-reformer and a main reformer were designed to produce about 4 Nm3h1 of hydrogen from propane. Two independent catalytic combustors were employed to supply the heat for both the initial heating and the reaction. Fig. 6 shows the temperature profile during the pre-reformer start-up. The hot gas from the catalytic combustor was passed through the sell side of the shell-and-tube-type reactor counter-currently to supply heat to the pre-reformer. Once both the inlet temperature and the outlet temperature of the catalyst bed reached above 350 C, steam and propane were sent to the prereformer. Since the hot combustion gas passed through the pre-reformer co-currently, there was a difference between the inlet temperature and the outlet temperature of the catalyst bed. At the beginning of the pre-reforming reaction, the inlet catalyst bed temperature dropped from 615 to about 150 C. The endothermic steam reforming reaction must have proceeded to a large extent at this stage with a less extent of the exothermic methanation reaction, resulting in a temperature
Pre-reforming reaction start-up 1200
100 90 80 Flue gas Temp. in catalytic combustor Inlet Temp. in pre-reformer Outlet Temp. in pre-reformer C3H8 conversion
800 600
70 60 50 40
400
30
C3H8 conversion (%)
Temperature (oC)
1000
20
200
10 0
0
1
2
3
4
5
6
7
8
9
0 10 11 12 13 14 15 16
Tme (h) Fig. 6 e Temperature profiles of pre-reformer (C3H8 6.64 LminL1, H2O 59.73 LminL1, S/C1 [ 3, GHSV [ 2620 hL1).
8855
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
1600
Table 4 e Pre-reformer product compositions and propane (C3H8) conversion: Reactants; 6.64 LminL1 C3H8 (99.999%), 59.73 LminL1 H2O (g), S/C1 [ 3, GHSV [ 2620 hL1.
Time (h) C3H8 conv. (%) H2 vol. (%) CO vol. (%) CH4 vol. (%) CO2 vol. (%)
365/505 133/495 309/472 338/482 331/482 4.5
5
9
11
12
100 63.7 13.4 16.8 6.3
100 57.0 7.9 18.5 16.8
100 53.4 19.2 23.8 4.0
100 53.0 18.0 27.1 2.9
100 52.5 17.5 27.0 3.0
1200
Temperature (oC)
Temp. ( C)a
1400
1000 800 TC 1 TC 2
600 400 200
a Inlet temperature in pre-reformer/outlet temperature in prereformer.
0
0
1
2
3
4
5
6
7
8
Time (h)
drop at the inlet catalyst bed. As the reaction time increased, the inlet temperature also increased to the steady state value of about 330 C, due to the increased extent of the methanation reaction. Table 4 lists the gas composition at the exit of the pre-reformer that showed the increase of the methane concentration as the reaction proceeded. As long as the exit temperature of the pre-reformer was maintained at a temperature higher than 470 C, the propane conversion was 100%. At the steady state all propane was reformed to CO and H2, and then about 48% of the CO was converted to CH4 by methanation and 5% of the CO was converted to H2 by the water-gas shift reaction. When the steady state of the pre-reformer was reached, all the products from the pre-reformer were introduced to the main reformer for further conversion to methane. Before the temperature of the main reformer was raised to the required target, the pre-reforming products were vented through the by-pass line. As mentioned, the catalytic combustor with dual-fuelnozzles was used to supply heat to the main reformer, since the steam reforming of methane requires high reaction temperatures. Fig. 7 shows the temperature profile of the gas within the combustor. The temperature of the gas right after the combustion catalyst was below 1000 C, but the subsequent non-catalytic combustion raised the gas temperature further. The temperature of the combustor exit gas supplied to the reformer was about 1150 C. This way we could maintain the bed temperature of the combustion catalyst at a low level to achieve a longer catalyst life, while achieving the high temperature necessary to supply heat to the reformer by
Fig. 7 e Temperature profiles of dually fuel injected catalytic combustor (Primary C3H8 3.3 LminL1, Air 175 LminL1, Secondary C3H8 3.3 LminL1, Cat. vol. 58.9 cc, GHSV [ 615,000 hL1).
burning the secondary fuel stably. During the heating of the reformer, the catalyst bed of the reformer was continuously purged with nitrogen. The main reforming reaction was initiated by introducing the pre-reformer product gas to the reformer when the temperature of the catalyst bed inlet and that of the bed outlet were 615 and 780 C, respectively. No reduction pretreatment of the Ru-doped Ni/MgAl2O4 reforming catalyst that was washcoated on the metallic monolith was performed because Ru facilitates the reduction of NiO to Ni by reforming products [14]. Table 5 lists compositions of reformed products from propane by the combined pre-reformer/main reformer system. The reactants introduced to the main reformer were the products of the pre-reformer. The flow rate of methane was calculated from the total flow rate of the pre-reformer product and its methane concentration. The steam reforming reaction was performed at H2O/C1 ¼ 4 over 120 pieces of metallic monoliths that were washcoated by the Ru-doped Ni/MgAl2O4 reforming catalyst (each piece had a volume of 7.6 cc). Because of the lower temperatures of the pre-reformer outlet gas and the endothermic nature of the steam reforming reaction, the temperature at the inlet of the main reformer was maintained at about 360 C. However, the hot gas from the catalytic
Table 5 e Main reformer product compositions and methane (C1) conversion: Reactants; CH4 10.54 LminL1, H2O 42.07 LminL1 (S/C1 [ 4), GHSV [ 3461 hL1. Temp. ( C)a Time (h) C1 conv. (%) H2 vol. (%) CO vol. (%) CH4 vol.(%) CO2 vol. (%) H2 (Nm3h1)
416/805
331/886
340/921
356/927
367/867
362/839
364/834
365/832
365/826
366/825
366/824
6
7
8
9
10
11
12
13
14
15
16
95.7 72.4 19.8 0.62 7.80 3.14
100 72.8 17.6 0 9.50 3.22
100 72.7 17.5 0 9.60 3.22
99.8 72.9 17.4 0.03 9.70 3.23
99.3 73.4 15.3 0.10 11.3 3.31
99.2 73.0 16.8 0.12 10.1 3.25
99.1 73.1 16.3 0.13 10.5 3.27
99.1 73.1 16.6 0.13 10.3 3.26
99.1 73.6 14.3 0.13 12.1 3.35
99.0 73.1 16.6 0.14 10.3 3.26
99.0 72.9 17.7 0.14 9.90 3.24
a Inlet temperature in main reformer/outlet temperature in main reformer.
Ave.
99.0 73.0 16.9 0.14 10.1 3.25
8856
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
combustor was able to sustain the outlet temperature of the reformer at 824 C or higher because hot gas entered the heatexchanger type reformer counter-currently. At those temperatures, C1 conversions of higher than 99% were obtained with a methane slip concentration of less than 0.14%. A hydrogen concentration of higher than 72% was obtained. The H2 production rate was calculated based on the product flow rate at room temperature and its concentration. The total H2 production rate was about 3.25 Nm3h1 with the propane fed at 0.4 Nm3h1.
4.
Conclusions
The pre-reforming of propane combined with the main steam reforming reaction was performed for syngas production over a Ru/Ni/MgAl2O4 metallic monolith catalyst. Propane prereforming runs showed that the carbon formation on the catalyst can be avoided with the complete conversion of propane with the following conditions: space velocities of over 2400 h1 and temperatures between 400 and 450 C with an H2O/C1 ratio of 3. Based on the pre-reforming results, the pilot scale prereformer/reformer system was designed using heatexchanger type reactors charged with the Ru/Ni/MgAl2O4 metallic monolith catalyst and the reaction heat was provided by catalytic combustors. Propane was converted in the prereformer to 52.5% H2, 27.0% CH4, 17.5% CO and 3.0% CO2 with a 331 C inlet temperature and a 482 C catalyst outlet temperature. Pre-reformer products were fed to the main steam reforming reactor and the methane conversion was higher than 99.0% with a 366 C inlet temperature and a 824 C catalyst outlet temperature. With 912 cc of the Ru/Ni/MgAl2O4 metallic monolith catalyst, the total H2 production from propane reached an average of 3.25 Nm3h1.
references
[1] Chen Y, Xu H, Jin X, Xiong G. Integration of gasoline prereforming into autothermal reforming for hydrogen production. Catal Today 2006;116:334e40. [2] http://www.topsoe.com/business_areas/synthesis_gas/ w/media/PDF%20files/ Prereforming/Topsoe_%
[3]
[4] [5]
[6]
[7] [8] [9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17] [18]
20prereforming%20_cat_%20brochure.ashx/ (Pre-reforming catalysts). Christensen TS. Adiabatic prereforming of hydrocarbons e an important step in syngas production. Appl Catal A 1996; 138:285e309. Twigg MV. Catalyst handbook. 2nd ed. London: Wolfe Publishing; 1989. Rakib MA, Grace JR, Lim CJ, Elnashaie SSEH. Steam reforming of heptane in a fluidized bed membrane reactor. J Power Sources 2010;195:5749e60. Nagaoka K, Sato K, Nishiguchi H, Takita Y. Influence of support on catalytic behavior of nickel catalysts in oxidative steam prereforming of n-butane for fuel cell applications. Appl Catal A 2007;327:139e46. Rostrup-Nielsen JR. Activity of nickel catalysts for steam reforming of hydrocarbons. J Catal 1973;31:173e99. Rostrup-Nielsen JR, Hansen JHB. CO2-reforming of methane over transition metals. J Catal 1993;144:38e49. Masuda M, Tabata T, Okada O, Wood BJ, Tong GT, Mccarty JG. The role of sulfur in the deposition of carbon on ruthenium and nickel steam reforming catalysts. Stud Surf Sci Catal 1991;68:185e93. Crisafulli C, Scire` S, Maggiore R, Minico` S, Galvagno S. CO2 reforming of methane over Ni-Ru and Ni-Pd bimetallic catalysts. Catal Lett 1999;59:21e6. Flytzani-Stephanopoulos M, Voecks GE, Charng T. Modelling of heat transfer in non-adiabatic monolith reactors and experimental comparisons of metal monoliths with packed beds. Chem Eng Sci 1986;41:1203e12. Calo E, Giannini A, Monteleone G. Small stationary reformers for H2 production from hydrocarbons. Int J Hydrogen Energy 2010;35:9828e35. http://www.getenergysmart.org/files/hydrogeneducation/ 6hydrogenproduction steammethanereforming.pdf/ (Hydrogen fact sheet). Ryu JH, Lee KY, La H, Kim HJ, Yang JI, Jung H. Ni catalyst wash-coated on metal monolith with enhanced heattransfer capability for steam reforming. J Power Sources 2007;171:499e505. Guo Y, Zhou L, Kameyama H. Steam reforming reactions over a metal-monolithic anodic alumina-supported Ni catalyst with trace amounts of noble metal. Int J Hydrogen Energy in press. Djaidja A, Libs S, Kiennemann A, Barama A. Characterization and activity in dry reforming of methane on NiMg/Al and Ni/ MgO catalysts. Catal Today 2006;113:194e200. Maluf SS, Assaf EM. Ni catalysts with Mo promoter for methane steam reforming. Fuel 2009;88:1547e53. Park YC, Oh ES, Rhee HK. Characterization and catalytic activity of WNiMo/Al2O3 catalysts for hydrodenitrogenation of pyridine. Ind Eng Chem Res 1997;36:5083e9.
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Combined pre-reformer/reformer system utilizing monolith catalysts for hydrogen production Jung-Il Yang a, Jae-Hong Ryu b, Kwan-Young Lee b, Nam-Jo Jung a, Ji Chan Park a, Dong Hyun Chun a, Hak-Joo Kim a, Jung Hoon Yang a, Ho-Tae Lee a, Inho Cho c, Heon Jung a,* a
Korea Institute of Energy Research, 71-2 Jang-Dong, Yuseong-Gu, Daejeon 305-343, South Korea Department of Chemical and Biological Engineering, Korea University, 5-1 Anam-Dong, Seongbuk-Gu, Seoul 136-701, South Korea c SK innovation, 140-1 Wonchon-Dong, Yuseong-Gu, Daejeon 305-712, South Korea b
article info
abstract
Article history:
The pre-reforming of higher hydrocarbon, propane, was performed to generate
Received 8 January 2011
hydrogen from LPG without carbon deposition on the catalysts. A Ru/Ni/MgAl2O4
Received in revised form
metallic monolith catalyst was employed to minimize the pressure drop over the
1 April 2011
catalyst bed. The propane pre-reforming reaction conditions for the complete conver-
Accepted 17 April 2011
sion of propane with no carbon formation were identified to be the following: space
Available online 24 May 2011
velocities over 2400 h1 and temperatures between 400 and 450 C with a H2O/C1 ratio of 3. The combined pre-reformer and the main reformer system with the Ru/Ni/MgAl2O4
Keywords:
metallic monolith catalyst was employed to test the conversion propane to syngas
Pre-reformer
where the reaction heat was provided by catalytic combustors. Propane was converted
Reformer
in the pre-reformer to 52.5% H2, 27.0% CH4, 17.5% CO, and 3.0% CO2 with a 331 C inlet
C catalyst outlet temperature. The main steam reforming
Metallic monolith catalyst
temperature and a 482
Catalytic combustor
reactor converted the methane from the pre-reformer with a conversion of higher than
Pilot scale
99.0% with a 366 C inlet temperature and an 824 C catalyst outlet temperature. With
Hydrogen production
a total of 912 cc of the Ru/Ni/MgAl2O4 metallic monolith catalyst in the main reformer, the H2 production from the propane reached an average of 3.25 Nm 3h1 when the propane was fed at 0.4 Nm3h1. Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Pre-reforming technology has been used for many years as the first stage in the steam reforming of heavy hydrocarbons, e.g. LPG or naphtha, to synthesize gas or H2 [1]. This process is a valuable tool for improving the efficiency and solving problems related to the conventional tubular steam reforming process (main reforming) [2]. During
pre-reforming, heavy hydrocarbons are first converted to methane and carbon oxides (CO, CO2) at relatively low temperatures, typically from 400 to 550 C, and finally H2rich streams are produced in the main reforming process at higher temperatures of >800 C [3e5]. Pre-reforming operated at low temperatures is an inevitable process because the steam reforming of higher hydrocarbons, which generally occurs at high temperatures, easily results in
* Corresponding author. Tel.: þ82 42 860 3663; fax: þ82 42 860 3134. E-mail address: [email protected] (H. Jung). 0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2011.04.130
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
coke deposition on the walls of the reformer and catalyst surface [6]. The pre-reforming process can be expressed with the following reactions [7]. Cn Hm þ nH2 O/nCO þ ðn þ ðm=2ÞÞH2 1 ¼ 498:1 kJmol ; for n ¼ 3 CO þ 3H2 ¼ CH4 þ H2 O CO þ H2 O ¼ CO2 þ H2
DH298K
1 DH298K ¼ 206:2 kJmol 1 DH298K ¼ 41:2 kJmol
(1)
(2)
(3)
First, a mixture of CO and H2 is produced from irreversible steam reforming (1) of higher hydrocarbons (n > 1). CH4 and CO2 are also formed by the subsequent methanation reaction (2) and water-gas shift reaction (3), respectively. The recently introduced steam reforming process is generally composed of 2 stages, main steam reformer and pre-reformer, due to the many benefits of using an additional pre-reformer [8,9]. Significant energy savings can be obtained by reducing the steam-to-carbon ratio. Moreover, carbon formation caused by the cracking of higher hydrocarbons in the top of the reformer tubes is avoided by the installation of a prereformer [2]. A nickel containing steam reforming catalyst can also be used in the pre-reforming process. However, a nickel based catalyst is easily deactivated during reforming processes, especially due to poisoning of the catalyst surface by coke deposition (coking), when the operations are performed at low steam-to-carbon ratios and high temperatures. To alleviate the coking problem, the addition of noble metals to the Ni catalysts, such as Ru, have resulted in significant improvements in both the activity and coke-resisting ability of silica supported Ni catalysts in the CO2 reforming reaction [10]. The steam reforming reaction over pelletized Ni catalysts is a well-known heat-transfer limited reaction [11]. We washcoated Ru-doped Ni catalyst powders on metallic monoliths and they had an enhanced heat-transfer capability. There are increased needs for small-scale hydrogen production units, especially for hydrogen refueling stations [12,13]. When the feed for the hydrogen product is LPG or naphtha, compact pre-reformer/reformer combination units are necessary [3]. Therefore, we developed a novel selfsustaining 5 kWe-class LPG reformer, which mainly consists of a very active metallic monolith catalyst in a highly endothermic reforming reaction, a combined compact system of a pre-reformer/reformer, and a catalytic combustor to supply the endothermic reaction heat. That is, there are three new ideas in this work. Firstly, we investigated the behavior of a Ru-doped Ni/MgAl2O4 catalyst washcoated on the metallic monolith for both the propane pre-reforming reaction and the main steam reforming reaction. Secondly, a combined unit of a pre-reformer/reformer at a pilot scale was developed for the LPG-fed hydrogen production unit. Finally, the endothermic reaction heat was supplied by a combustion reactor equipped with Pd catalysts that were supported on the ceramic monolith.
2.
Experimental
2.1.
Catalyst preparation
8851
The commercial Ni-based steam reforming catalyst (SBET ¼ 19.2 m2g1) was washcoated on the metallic monolith. The preparation procedure can be found elsewhere [14]. Fecralloy plate (Fe 72.8%/Cr 22%/Al 5%/Y 0.1%/Zr 0.1%, Goodfellow Co.) was used to make the monolith (640 cpi) with dimensions of 2.2 cm (diameter) 2 cm (height). Once the washcoating of the Ni catalysts was completed, a small amount of noble metal (0.12 wt.% Ru) was doped into the Ni catalyst to serve as both the reduction-enhancing agent and the coke-resisting agent. RuCl3 (RuCl3$xH2O, Aldrich Chem. Co.) was used as a noble metal precursor and the wet impregnation method was used.
2.2.
Catalyst characterization
The surface area, pore size, and pore volume of the catalyst were measured by N2 adsorptionedesorption isotherms using a BELSORP mini-II (BEL Japan Inc.) apparatus. Before measurements, the samples were degassed in a vacuum at 573 K for 6 h. XRD data of the sample was collected in a Rigaku D-Max 2500 (18 kW) diffractometer with Cu Ka radiation. The stepscans were taken over the range of 2q from 20 to 80 in steps of 0.02 /s. For the TPR experiments, the catalyst sample was purged with 50 ml/min of Ar for 1 h at 500 C, and then cooled to room temperature. The level of hydrogen consumption was measured using a BELCAT Catalyst Analyzer (BEL Japan Inc.) by heating the sample from ambient temperature to 1000 C with a heating rate of 10 C/min in a flow of 5% H2/Ar at 30 ml/ min. The real content of Ru was measured in the samples by inductively coupled plasma-atomic emission spectroscopy (ICP-AES), with a PolyScan 60E (Hewlett Packard, USA). The metal (Ni, MgO, Al2O3) loading amounts of the samples were measured by a MiniPal2 energy dispersive X-ray fluorescence (EDXRF) spectrometer. The carbon formation on the catalyst after the prereforming reaction was investigated by Cs-corrected STEM using a JEOL microscope that was operated at 200 kV (National Nanofeb Center, Korea). The samples were prepared by putting a few drops of the corresponding colloidal solutions on carbon coated copper grids (Ted Pellar, Inc.).
2.3. Catalytic reaction and pre-reformer/reformer system To evaluate the behavior of the Ru-doped Ni/MgAl2O4 catalyst washcoated on the metallic monolith for the pre-reforming reaction, steam reforming of propane was performed in a tubular stainless steel reactor (inside diameter of 2.22 cm) operated at atmospheric pressure. Propane was supplied from a cylinder of C3H8 (99.999%) by a mass flow controller (Brooks, 5850E). Steam was supplied by a steam generator. One cylindrical metallic monolith catalyst (2.2 cm diameter and 2 cm
8852
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
Fig. 1 e Schematic of diagram of combined pre-reformer and main reformer system.
height, volume of 7.6 cm3) was used for the test. A monolith without a washcoat was placed in front of the washcoated catalyst for the even distribution of the reactant gas. A gas mixture of C3H8 and steam was passed through the metallic monolith catalyst at various gas hourly space velocities (GHSV). The ratio of H2O/C1 was also varied between 1 and 3. The compositions of the products were analyzed by a gas chromatograph (HP5890, TCD) equipped with a Porapak Q column (8.2 ft 1/8 inch). Based on the flow rate and the composition of the product gas, the conversion of C3H8 was calculated with the following equation: C3 H8 conversionð%Þ ¼ ½inlet flow rate of C3 H8 outlet flow rate of C3 H8 = ðinlet flow rate of C3 H8 Þ 100
(4)
And conversion of C1 is defined as how much carbon in the inlet C3H8 is converted to CO and CO2 and is calculated as: C1 conversionð%Þ ¼ ½3 inlet flow rate of C3 H8 3 outlet flow rate of C3 H8 outlet flow rate of CH4 = 3 inlet flow rate of C3 H8 100
The pre-reformer and the reformer were shell-and-tube heatexchanger type reactors made of stainless steel. There were 10 tubes (2.2 cm ID 50 cm height) and 6 tubes in the reactors of the pre-reformer and the reformer, respectively. Each tube was filled with 20 cylindrical metallic monolith catalysts with 2.2 cm diameters and 2 cm heights (total volume of the catalyst: pre-reformer ¼ 1520 cc, reformer ¼ 912 cc). As shown in Fig. 1, two catalytic combustors were used to supply heat to the pre-reformer and the reformer. Pd washcoated ceramic monolith catalysts (6.2 cm diameters 2 cm heights) were used in the catalytic combustors. The propane and air were used as feed gases for combustor operation. Propane was fed to the combustor for the reformer in two nozzles, with the primary fuel combusted over the catalyst bed to raise the gas to a high enough temperature to burn the secondary fuel without the catalyst (Fig. 2). The combustor for the pre-reformer was a single-fuel-nozzle combustor with the nozzle placed in front of the catalyst bed, since the temperature requirement for the pre-reformer was lower than that of the reformer.
(5)
We designed and installed a combined pre-reformer/ reformer system to produce 4 Nm3h1 of hydrogen (Fig. 1).
3.
Results and discussion
3.1.
Catalyst characterizations
Surface area, pore size, and pore volume of the catalyst sample are summarized in Table 1. It can be seen that the surface area of the Ru-doped Ni/MgAl2O4 catalyst decreased compared to that of the commercial catalyst due to the two
Table 1 e Physical properties of the Ru-doped Ni/MgAl2O4 catalyst. BET surface area (m2g1) Fig. 2 e Schematic diagram of catalytic combustor for main reformer with duel fuel injection.
14.6
Ave. pore diameter (nm)
Ave. pore volume (cm3g1)
29.4
0.107
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
8853
Fig. 4 e TPR curve of the Ru-doped Ni/MgAl2O4 catalyst calcined at 900 C. Fig. 3 e X-ray diffraction pattern of the Ru-doped Ni/ MgAl2O4 catalyst calcined at 900 C.
calcination procedures at 900 C followed by catalyst washcoating on the metallic monolith and Ru doping. The X-ray diffraction (XRD) pattern of the Ru-doped Ni/ MgAl2O4 catalyst is shown in Fig. 3 and it matches with the MgAl2O4 (JCPDS No. 05-0672), NiAl2O4 (JCPDS No. 10-0339), MgNiO2 (JCPDS No. 24-0712), Al2O3 (JCPDS No. 46-1212), and RuO2 (JCPDS No. 43-1027) phases [15]. The TPR profile of the catalyst is shown in Fig. 4 and two hydrogen consumption peaks are observed in the TPR profile. The first shoulder peak at 580 C was assigned to the reduction of NiO interacting weakly with the support (MgO and/or Al2O3) or to small Ni particles [16e18]. The second peak at 770 C was assigned to the reduction of stable Ni2þ compounds where Ni2þ is strongly interacting with MgO and/or Al2O3, which was clearly identified in the XRD result [16e18]. The loading content of Ru was measured to be 0.1% by using inductively coupled plasma-atomic emission spectrometry (ICP-AES).
From the elemental analysis by EDXRF, the weight percentages of Ni, MgO, and Al2O3 were measured to be 18, 17, and 65%, respectively. TEM results of the Ru-doped Ni/MgAl2O4 catalyst before and after the pre-reforming reaction are illustrated on Fig. 5A and B, respectively. Ni particles were observed to have various sizes, and the filamentous carbons were not observed for the fresh catalyst run from the TEM analysis (Fig. 5A). However, it can be seen (Fig. 5B) that the TEM image of the catalyst powders after the reaction shows the generation of bright branches on particle surface which clearly indicate carbon formation during the pre-reforming reaction.
3.2.
Propane pre-reforming
Avoiding carbon formation is the most important aspect of operating pre-reformers. Although some information relating to the operating conditions for the possible carbon formation can be understood from thermodynamics, experimental
Fig. 5 e TEM images of the Ru-doped Ni/MgAl2O4 catalyst (A) before pre-reforming reaction and (B) after the reaction: T [ 425 ± 25 C, SV [ 1800 h-1, and S/C1 [ 1.
8854
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
Table 2 e Pre-reforming of propane with various space velocities (S/C1 [ 1,T [ 425 ± 25 C).
Table 3 e Pre-reforming of propane according to H2O/C1 ratio (SV [ 1800 hL1, T [ 425 ± 25 C).
Space velocity (h1)
Steam/Carbon ()
1800
2400
42 yes
24 no
C3H8 conversion (%) Coking
C3H8 conversion (%) Coking
1
2
3
42 yes
100 yes (partially)
100 no
C1 ratio of 2 (Table 3). However, a small amount of coke was formed on the catalyst. Thus, the safe operating conditions of propane pre-reforming over the Ru-doped Ni metallic monolith catalyst are: GHSV of about 1800 h1, temperatures higher than 425 C and H2O/C1 ratio higher than 3.
verification by the use of a real catalyst and a test with a larger scale unit are still very important. The pre-reforming of propane was performed at two different space velocities with the H2O/C1 ratio set to 1 over a single cylindrical Ru-doped Ni metallic monolith catalyst (7.6 cm3). Since the pre-reforming reaction is the combination of the endothermic steam reforming reaction and the exothermic methanation reaction, the temperature of the catalyst fluctuated during the run and the average temperature at the exit of the catalyst was 425 C with a high of 450 C and a low of 400 C. As listed in Table 2, the propane conversion was 24% at a GHSV of 2400 h1. After the run, the catalyst was examined and no carbon was found on the catalyst. At the lower GHSV of 1800 h1, the propane conversion was increased to 42%, but coke was formed on the catalyst (Fig. 5B). Thus, the H2O/C1 ratio of 1 is too low both for the complete conversion of propane and for the avoidance of coke formation. While maintaining the GHSV at 1800 h1 and the average temperature at 425 C, we increased the H2O/C1 ratio and the result is listed in Table 3. As the H2O/C1 ratio was increased to 3, the propane conversion reached 100%. At the same time, there was no coke formation. At 425 C and a GHSV of 1800 h1, the H2O/C1 ratio of 3 is required to obtain the complete conversion of propane with no danger of carbon formation. As the temperature was increased to 500 C, the complete conversion of propane was achieved even at a H2O/
3.3. scale
Combined pre-reformer/reformer test at the pilot
A pre-reformer and a main reformer were designed to produce about 4 Nm3h1 of hydrogen from propane. Two independent catalytic combustors were employed to supply the heat for both the initial heating and the reaction. Fig. 6 shows the temperature profile during the pre-reformer start-up. The hot gas from the catalytic combustor was passed through the sell side of the shell-and-tube-type reactor counter-currently to supply heat to the pre-reformer. Once both the inlet temperature and the outlet temperature of the catalyst bed reached above 350 C, steam and propane were sent to the prereformer. Since the hot combustion gas passed through the pre-reformer co-currently, there was a difference between the inlet temperature and the outlet temperature of the catalyst bed. At the beginning of the pre-reforming reaction, the inlet catalyst bed temperature dropped from 615 to about 150 C. The endothermic steam reforming reaction must have proceeded to a large extent at this stage with a less extent of the exothermic methanation reaction, resulting in a temperature
Pre-reforming reaction start-up 1200
100 90 80 Flue gas Temp. in catalytic combustor Inlet Temp. in pre-reformer Outlet Temp. in pre-reformer C3H8 conversion
800 600
70 60 50 40
400
30
C3H8 conversion (%)
Temperature (oC)
1000
20
200
10 0
0
1
2
3
4
5
6
7
8
9
0 10 11 12 13 14 15 16
Tme (h) Fig. 6 e Temperature profiles of pre-reformer (C3H8 6.64 LminL1, H2O 59.73 LminL1, S/C1 [ 3, GHSV [ 2620 hL1).
8855
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
1600
Table 4 e Pre-reformer product compositions and propane (C3H8) conversion: Reactants; 6.64 LminL1 C3H8 (99.999%), 59.73 LminL1 H2O (g), S/C1 [ 3, GHSV [ 2620 hL1.
Time (h) C3H8 conv. (%) H2 vol. (%) CO vol. (%) CH4 vol. (%) CO2 vol. (%)
365/505 133/495 309/472 338/482 331/482 4.5
5
9
11
12
100 63.7 13.4 16.8 6.3
100 57.0 7.9 18.5 16.8
100 53.4 19.2 23.8 4.0
100 53.0 18.0 27.1 2.9
100 52.5 17.5 27.0 3.0
1200
Temperature (oC)
Temp. ( C)a
1400
1000 800 TC 1 TC 2
600 400 200
a Inlet temperature in pre-reformer/outlet temperature in prereformer.
0
0
1
2
3
4
5
6
7
8
Time (h)
drop at the inlet catalyst bed. As the reaction time increased, the inlet temperature also increased to the steady state value of about 330 C, due to the increased extent of the methanation reaction. Table 4 lists the gas composition at the exit of the pre-reformer that showed the increase of the methane concentration as the reaction proceeded. As long as the exit temperature of the pre-reformer was maintained at a temperature higher than 470 C, the propane conversion was 100%. At the steady state all propane was reformed to CO and H2, and then about 48% of the CO was converted to CH4 by methanation and 5% of the CO was converted to H2 by the water-gas shift reaction. When the steady state of the pre-reformer was reached, all the products from the pre-reformer were introduced to the main reformer for further conversion to methane. Before the temperature of the main reformer was raised to the required target, the pre-reforming products were vented through the by-pass line. As mentioned, the catalytic combustor with dual-fuelnozzles was used to supply heat to the main reformer, since the steam reforming of methane requires high reaction temperatures. Fig. 7 shows the temperature profile of the gas within the combustor. The temperature of the gas right after the combustion catalyst was below 1000 C, but the subsequent non-catalytic combustion raised the gas temperature further. The temperature of the combustor exit gas supplied to the reformer was about 1150 C. This way we could maintain the bed temperature of the combustion catalyst at a low level to achieve a longer catalyst life, while achieving the high temperature necessary to supply heat to the reformer by
Fig. 7 e Temperature profiles of dually fuel injected catalytic combustor (Primary C3H8 3.3 LminL1, Air 175 LminL1, Secondary C3H8 3.3 LminL1, Cat. vol. 58.9 cc, GHSV [ 615,000 hL1).
burning the secondary fuel stably. During the heating of the reformer, the catalyst bed of the reformer was continuously purged with nitrogen. The main reforming reaction was initiated by introducing the pre-reformer product gas to the reformer when the temperature of the catalyst bed inlet and that of the bed outlet were 615 and 780 C, respectively. No reduction pretreatment of the Ru-doped Ni/MgAl2O4 reforming catalyst that was washcoated on the metallic monolith was performed because Ru facilitates the reduction of NiO to Ni by reforming products [14]. Table 5 lists compositions of reformed products from propane by the combined pre-reformer/main reformer system. The reactants introduced to the main reformer were the products of the pre-reformer. The flow rate of methane was calculated from the total flow rate of the pre-reformer product and its methane concentration. The steam reforming reaction was performed at H2O/C1 ¼ 4 over 120 pieces of metallic monoliths that were washcoated by the Ru-doped Ni/MgAl2O4 reforming catalyst (each piece had a volume of 7.6 cc). Because of the lower temperatures of the pre-reformer outlet gas and the endothermic nature of the steam reforming reaction, the temperature at the inlet of the main reformer was maintained at about 360 C. However, the hot gas from the catalytic
Table 5 e Main reformer product compositions and methane (C1) conversion: Reactants; CH4 10.54 LminL1, H2O 42.07 LminL1 (S/C1 [ 4), GHSV [ 3461 hL1. Temp. ( C)a Time (h) C1 conv. (%) H2 vol. (%) CO vol. (%) CH4 vol.(%) CO2 vol. (%) H2 (Nm3h1)
416/805
331/886
340/921
356/927
367/867
362/839
364/834
365/832
365/826
366/825
366/824
6
7
8
9
10
11
12
13
14
15
16
95.7 72.4 19.8 0.62 7.80 3.14
100 72.8 17.6 0 9.50 3.22
100 72.7 17.5 0 9.60 3.22
99.8 72.9 17.4 0.03 9.70 3.23
99.3 73.4 15.3 0.10 11.3 3.31
99.2 73.0 16.8 0.12 10.1 3.25
99.1 73.1 16.3 0.13 10.5 3.27
99.1 73.1 16.6 0.13 10.3 3.26
99.1 73.6 14.3 0.13 12.1 3.35
99.0 73.1 16.6 0.14 10.3 3.26
99.0 72.9 17.7 0.14 9.90 3.24
a Inlet temperature in main reformer/outlet temperature in main reformer.
Ave.
99.0 73.0 16.9 0.14 10.1 3.25
8856
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 8 8 5 0 e8 8 5 6
combustor was able to sustain the outlet temperature of the reformer at 824 C or higher because hot gas entered the heatexchanger type reformer counter-currently. At those temperatures, C1 conversions of higher than 99% were obtained with a methane slip concentration of less than 0.14%. A hydrogen concentration of higher than 72% was obtained. The H2 production rate was calculated based on the product flow rate at room temperature and its concentration. The total H2 production rate was about 3.25 Nm3h1 with the propane fed at 0.4 Nm3h1.
4.
Conclusions
The pre-reforming of propane combined with the main steam reforming reaction was performed for syngas production over a Ru/Ni/MgAl2O4 metallic monolith catalyst. Propane prereforming runs showed that the carbon formation on the catalyst can be avoided with the complete conversion of propane with the following conditions: space velocities of over 2400 h1 and temperatures between 400 and 450 C with an H2O/C1 ratio of 3. Based on the pre-reforming results, the pilot scale prereformer/reformer system was designed using heatexchanger type reactors charged with the Ru/Ni/MgAl2O4 metallic monolith catalyst and the reaction heat was provided by catalytic combustors. Propane was converted in the prereformer to 52.5% H2, 27.0% CH4, 17.5% CO and 3.0% CO2 with a 331 C inlet temperature and a 482 C catalyst outlet temperature. Pre-reformer products were fed to the main steam reforming reactor and the methane conversion was higher than 99.0% with a 366 C inlet temperature and a 824 C catalyst outlet temperature. With 912 cc of the Ru/Ni/MgAl2O4 metallic monolith catalyst, the total H2 production from propane reached an average of 3.25 Nm3h1.
references
[1] Chen Y, Xu H, Jin X, Xiong G. Integration of gasoline prereforming into autothermal reforming for hydrogen production. Catal Today 2006;116:334e40. [2] http://www.topsoe.com/business_areas/synthesis_gas/ w/media/PDF%20files/ Prereforming/Topsoe_%
[3]
[4] [5]
[6]
[7] [8] [9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17] [18]
20prereforming%20_cat_%20brochure.ashx/ (Pre-reforming catalysts). Christensen TS. Adiabatic prereforming of hydrocarbons e an important step in syngas production. Appl Catal A 1996; 138:285e309. Twigg MV. Catalyst handbook. 2nd ed. London: Wolfe Publishing; 1989. Rakib MA, Grace JR, Lim CJ, Elnashaie SSEH. Steam reforming of heptane in a fluidized bed membrane reactor. J Power Sources 2010;195:5749e60. Nagaoka K, Sato K, Nishiguchi H, Takita Y. Influence of support on catalytic behavior of nickel catalysts in oxidative steam prereforming of n-butane for fuel cell applications. Appl Catal A 2007;327:139e46. Rostrup-Nielsen JR. Activity of nickel catalysts for steam reforming of hydrocarbons. J Catal 1973;31:173e99. Rostrup-Nielsen JR, Hansen JHB. CO2-reforming of methane over transition metals. J Catal 1993;144:38e49. Masuda M, Tabata T, Okada O, Wood BJ, Tong GT, Mccarty JG. The role of sulfur in the deposition of carbon on ruthenium and nickel steam reforming catalysts. Stud Surf Sci Catal 1991;68:185e93. Crisafulli C, Scire` S, Maggiore R, Minico` S, Galvagno S. CO2 reforming of methane over Ni-Ru and Ni-Pd bimetallic catalysts. Catal Lett 1999;59:21e6. Flytzani-Stephanopoulos M, Voecks GE, Charng T. Modelling of heat transfer in non-adiabatic monolith reactors and experimental comparisons of metal monoliths with packed beds. Chem Eng Sci 1986;41:1203e12. Calo E, Giannini A, Monteleone G. Small stationary reformers for H2 production from hydrocarbons. Int J Hydrogen Energy 2010;35:9828e35. http://www.getenergysmart.org/files/hydrogeneducation/ 6hydrogenproduction steammethanereforming.pdf/ (Hydrogen fact sheet). Ryu JH, Lee KY, La H, Kim HJ, Yang JI, Jung H. Ni catalyst wash-coated on metal monolith with enhanced heattransfer capability for steam reforming. J Power Sources 2007;171:499e505. Guo Y, Zhou L, Kameyama H. Steam reforming reactions over a metal-monolithic anodic alumina-supported Ni catalyst with trace amounts of noble metal. Int J Hydrogen Energy in press. Djaidja A, Libs S, Kiennemann A, Barama A. Characterization and activity in dry reforming of methane on NiMg/Al and Ni/ MgO catalysts. Catal Today 2006;113:194e200. Maluf SS, Assaf EM. Ni catalysts with Mo promoter for methane steam reforming. Fuel 2009;88:1547e53. Park YC, Oh ES, Rhee HK. Characterization and catalytic activity of WNiMo/Al2O3 catalysts for hydrodenitrogenation of pyridine. Ind Eng Chem Res 1997;36:5083e9.