Scholarly article on topic 'Microstructured Catalytic Hollow Fiber Reactor for Methane Steam Reforming'

Microstructured Catalytic Hollow Fiber Reactor for Methane Steam Reforming Academic research paper on "Chemical engineering"

0
0
Share paper
OECD Field of science
Keywords
{""}

Academic research paper on topic "Microstructured Catalytic Hollow Fiber Reactor for Methane Steam Reforming"

I&EC

, pubs.acs.org/IECR

resea rch

Industrial & Engineering Chemistry Research

Microstructured Catalytic Hollow Fiber Reactor for Methane Steam Reforming

Ana Gouveia Gil, Zhentao Wu, David Chadwick,* and K. Li*

Department of Chemical Engineering, Imperial College London, London SW7 2AZ, United Kingdom

Article

ABSTRACT: Microstructured alumina hollow fibers, which contain a plurality of radial microchannels with significant openings on the inner surface, have been fabricated in this study and used to develop an efficient catalytic hollow fiber reactor. Apart from low mass-transfer resistance, a unique structure of this type facilitates the incorporation of Ni-based catalysts, which can be with or without the aged secondary support, SBA-15. In contrast to a fixed bed reactor, the catalytic hollow fiber reactor shows similar methane conversion, with a gas hourly space velocity that is approximately 6.5 times higher, a significantly greater CO2 selectivity, and better productivity rates. These results demonstrate the advantages of dispersing the catalyst inside the microstructured hollow fiber as well as the potential to reduce the required quantity of catalyst.

■ INTRODUCTION

Hydrogen has been widely used as a critical feedstock for chemical processes (e.g., petro-chemical and ammonia production) and it is an attractive and promising energy carrier.1,2 The state-of-art in hydrogen production is mainly based on steam reforming of fossil fuels, for which methane is preferable because of its availability and high H/C ratio.1 The water-gas shift reaction can be performed simultaneously to produce additional hydrogen, reduce operating costs, and facilitate CO2 capture at the source. The main reactions involved in this process are thus steam reforming (SR) (reaction I) and water-gas shift (WGS) (reaction II), with the overall reaction III given below:3,4

CH4 + H2O ^ CO + 3H2 AH298 = +206.1 kJ-mol-1

CO + H2O ^ CO2 + H2 AH298 = -41.15 kJ-mol-1

CH4 + 2H2O ^ CO2 + 4H2 AH298 = +165.0 kJ-mol-1

Reforming reactions of this type are usually performed in separated catalytic packed-bed reactors (PBRs) because of significant differences in operating temperatures and catalyst composition.5 The composition and geometry of catalysts for SR and WGS have been extensively studied in order to achieve more efficient heat and mass transfer, lower pressure drop along the reactor, and less deactivation due to temperature and concentration profiles. However, the whole process is still energy intensive and costly because of its multistep character and operating conditions.6 Process intensification, such as the use of microreactors, and integration, which combines reaction and product purification into a membrane reactor, can potentially improve the overall efficiency of the SR of methane process. Microreactors, as miniaturizations of reaction units, have been receiving increasing attention because of their compact design; improved heat- and mass-transfer efficiencies; reduced weight and enhanced lifetime of catalyst; and improved conversion, yield, and selectivity.7-10 All these features come at

lower capital and operating costs.11 The microreactor configuration provides mostly a laminar (1< Re < 1000), directed, and highly symmetric hydrodynamic flow and reduces the interparticle mass-transfer resistance, allowing a better and faster contact between reactants and catalyst.12'13 Meanwhile, short diffusion lengths for heat and mass transfer contribute to the small temperature and concentration profiles inside microreactors. The precise control of temperature, pressure, residence time and flow rate achieved in microreactors reduces the intrinsic risks of performing explosive and highly exothermic reactions. In contrast, the PBR configuration features a turbulent and asymmetric flow with preferential pathways for gases, promoting massive heat- and mass-transfer resistance and, consequently more significant temperature and concentration profiles.14

Ceramic hollow fibers, which can be of a symmetric or asymmetric structure, have been employed as a catalyst substrate for catalytic hollow fibers and catalytic hollow fiber membrane reactors.15 20 Symmetric hollow fibers (SHFs) consist of a fully spongelike structure with relatively small pore sizes, low porosity, and high mass-transfer resistance. The incorporation of catalyst into SHFs is usually carried out by using a solution technology, such as sol-gel and impregnation, leading to a homogeneous distribution of a highly dispersed catalyst.19'20 On the other hand, asymmetric hollow fibers (AHFs) contain one or more spongelike layers and a plurality of self-organized microvoids or channels, which greatly increase the actual surface area-to-volume ratio over SHF. The catalyst is mainly deposited inside the microvoids or channels, working as thousands of parallel microreactors with enhanced mass- and heat-transfer efficiency and, consequently, greater catalytic performance. Despite the high surface-area-to-volume ratio, the ceramic hollow fiber normally presents very limited specific surface area. Therefore, secondary supports are usually

Received: December 21, 2014 Revised: April 28, 2015 Accepted: May 5, 2015

ACS Publications © XXXX American Chemical Society

employed to increase the specific surface area and, as a result, to promote a more uniform distribution of higher loadings of catalyst.21 Santa Barbara Amorphous 15 (SBA-15) silica emerges as a good secondary support for nickel-based catalyst because it has a high surface area, large pore volume, uniform pore size distribution, thick pore wall, and good thermal stability.22 The nickel particles can be dispersed inside SBA-15 pores and consequently prevent the sintering of nickel, the metal loss, and the formation of large ensembles.23

The catalytic performance of catalytic hollow fibers is affected by several factors such as the morphology of the hollow fiber substrate, the distribution and loading of the catalyst, the configuration, and operating conditions. GarciaGarcia and Li19 investigated the effects of Al2O3 hollow fiber morphology on catalyst particle size and on the performance of catalytic membrane reactors for the water-gas shift reaction. For this purpose, a Cu-based catalyst was incorporated into symmetric and asymmetric hollow fibers with a spongelike layer of 300 and 100 ^m, respectively, using a sol-gel Pechini method. A uniform catalyst distribution was observed in both symmetric and asymmetric catalytic hollow fibers. However, the particle size of the catalyst presented significant differences. For the symmetric catalytic hollow fiber, a uniform particle size of less than 50 nm was observed throughout the spongelike layer. On the other hand, the asymmetric catalytic hollow fiber presented catalyst particles of approximately 350 nm and <50 nm in the conical microchannels region and spongelike layer, respectively. The formation of larger particles was caused by the growth of particles during the polymerization step, as the microchannels were filled with sol-gel solution. In contrast, for the sol-gel solution penetrated into the spongelike layer by capillarity forces, the volume of solution was too little to allow the growth of the particles. Moreover, it was proven that the catalyst with a smaller particle size presented higher catalytic activity. Furthermore, the enhanced catalytic performance of the asymmetric catalytic hollow fibers is attributed to the improved heat- and mass-transfer efficiencies and higher internal area provided by the conical microchannels.

Additionally, the incorporation of catalyst into the conical microvoids of asymmetric hollow fibers have been proven to significantly reduce the amount of catalyst required for reaction.24 The microvoids contribute to higher geometric surface area and, as a result, lead to more uniformly dispersed catalyst. However, previous studies were mainly focused on single reactions, e.g., ethanol steam reforming, ' methanol steam reforming,1 ' and dehydrogenation reaction.21'27 Meanwhile, both ends of the microvoids are submerged between spongelike layers with a packed-pore network (average pore size of approximately 0.1-0.2 j«m), which limits the methods of incorporating catalysts and is still not ideal for an efficient mass transfer.

As a result, this study addresses the incorporation of Ni-based catalysts, with or without mesoporous SBA-15 as a secondary support, into a microstructured alumina hollow fiber, forming a highly compact porous reactor for SR and WGS at relatively low operating temperatures (375-550 °C). In contrast to ceramic hollow fibers used previously for similar purposes, the one employed in this study provides a plurality of radial microchannels, instead of microvoids, with one end fully open on the inner surface of the hollow fiber. Apart from facilitating the incorporation of catalyst, such microchannels are more efficient regarding mass transfer. Moreover, the porous reactor of this proof-of-principle study can be considered as an

advanced catalytic hollow fiber substrate for further coating of Pd-based membranes, leading to a highly compact membrane reactor design for precombustion carbon capture.

2. EXPERIMENTAL SECTION

2.1. Preparation of Microstructured Al2O3 Hollow Fibers. The microstructured Al2O3 hollow fibers were prepared by a viscous-fingering-induced phase inversion technique followed by high-temperature sintering, allowing great flexibility in morphology control.28 30 In this study, a uniform ceramic suspension consisting of approximately 62.7 wt % Al2O3 powder (1 j«m, VWR), 30.6 wt % solvent (dimethyl sulfoxide, DMSO, VWR), 0.4 wt % dispersant (Arlacel P135), and 6.3 wt % polymer binder (poly(ether sulfone), PESf) was prepared via ball milling. The suspension was then degassed under vacuum to remove air bubbles, transferred into a 200 mL stainless steel syringe, and extruded through a tube-in-orifice spinneret (o.d., 3 mm; i.d., 1.2 mm) at a rate of 7 mLmin-1 into a water bath, with a bore fluid flow rate of 5 mL-min-1 and a zero air gap (0 cm). The formed hollow fiber precursors were then dried and sintered at 1300 °C for 4 h (Elite TSH17/75/450).

2.2. Preparation of 25 wt % NiO/SBA-15 Catalyst Powder. To prepare the SBA-15 sol, 10.0 g of Pluronic P123 (PEG-PPG-PEG block copolymer, average MW = 5800, Aldrich) was first dissolved in a mixture of absolute ethanol (50 g, VWR) and hydrochloric acid (2.0 g, 1 M) using a magnetic stirrer (RCT basics, Ika), followed by the addition of 20.8 g of TEOS (tetraethyl orthosilicate, Aldrich). The phase transition of SBA-15 from sol to gel was performed in a fan oven (Salvislab Thermocenter) at 40 °C for 48 h. The SBA-15 gel was then calcinated at 600 °C for 5 h with a heating rate of 1 °C-min-1. The powder thus obtained was impregnated with Ni by wet incipient impregnation, using an ethanol-based nickel nitrate solution (25 wt % Ni), followed by a calcination at 550 °C for 6 h with a heating rate of 1 °C-min-1.

2.3. Development of Catalytic Hollow Fiber (CHF). CHFs were prepared by depositing NiO/SBA-15 catalyst into the microstructured Al2O3 hollow fiber. The catalyst was synthesized in situ via a two-step process, i.e., incorporation of SBA-15 via a sol-gel method followed by wet impregnation of nickel nitrate.

2.3.1. Incorporation of SBA-15. The SBA-15 sol was prepared as mentioned above. The formulation was heat-treated in a fan oven (Salvislab Thermocenter) at 40 °C for 0, 6, or 12 h to adjust its viscosity, which largely determines the quantity and distribution of SBA-15 inside the Al2O3 hollow fiber. The resultant SBA-15 sol was then incorporated into the hollow fibers by immersing the fibers in the SBA-15 precursor solution under vacuum. After the remaining SBA-15 sol inside the hollow fiber lumen was expelled, using compressed air, the evaporation of the remaining ethanol (solvent) in the sol was carried out at 40 °C overnight. The fibers were then calcinated at 600 °C for 5 h with a heating rate of 1 °C-min-1.

2.3.2. Impregnation of Ni. Ni as the catalytic active phase was incorporated into the hollow fiber with or without SBA-15 via impregnation of ethanol-based nickel nitrate solution (25 wt % Ni) using the same vacuum-assisted process mentioned above. The nickel nitrate was then converted to nickel oxide at 550 °C for 6 h with a heating rate of 1 °C-min-1.

The loadings of SBA-15 and Ni were determined by comparing the weights of the hollow fibers before and after the impregnation and calcination steps. The presence of

precursors was taken into account when determining the loadings.

2.4. Characterization. The microstructures of the hollow fiber and CHFs were analyzed by scanning electron microscopy (SEM) instruments (JEOL JSM-5610LV and LEO Gemini 1525 FEG). Samples were first coated with gold in a vacuum chamber for 2 min at 20 mA (EMITECH Model K550) and connected with the metal support by brushing silver paint. A detailed assessment of the pore structure and porosity was carried out by mercury intrusion porosimetry (Autopore IV 9500, Micrometrics), over a pressure range from 1.5 X 103 to 2.3 X 108 Pa and with a set stabilization time of 10 s.

The mechanical strength of the Al2O3 hollow fiber was assessed using a tensile tester (Instron Model 5544) with a load cell of 5 kN.

The Brunauer-Emmett-Teller (BET) method was used to investigate the specific surface area of hollow fibers and catalytic hollow fibers. N2 adsorption isotherms were measured at 77 K using a Tristar 3000 volumetric system. Prior to gas adsorption, samples were degassed at 120 °C overnight and under nitrogen atmosphere.

X-ray diffraction patterns of crushed hollow fiber and CHFs (fine powder) were obtained using an X'celerator detector (X'Pert PRO model), with a radiation source of Cu Ka, a set voltage of 40 kV, a set current of 40 mA, and in a 20 degree range from 5 to 80°.

2.5. Catalytic Performance. The catalytic performance of CHFs was evaluated using the experimental apparatus illustrated in Figure 1. The flow rate of each gas (reactants and inert sweep gas) was controlled by individual mass flow controllers (Brooks Instrument, model 5800) with a collective reader (Brooks Instrument, model 0254). The liquid water was fed into a heating coil (1/16 in. stainless steel tube) by a syringe pump (Chemix N5000). The temperature of the tubular

horizontal furnace (Vecstar SP HVT) was controlled by temperature controller (CAL 9400) and monitored by a thermocouple located at the central position of the uniform heating zone (7 cm). The inlet and outlet pressure of the reactor were monitored by a digital pressure sensor (Sick, 10 bar). The composition of the effluent from the reactor was analyzed online by a gas chromatograph (Varian 3900) with a packed column (ShinCarbon, part 19808) after a stabilization time of 1 h at each operating temperature. The total gas flow rate was monitored using a bubble flow meter.

The CHF module was assembled by inserting a 30 cm catalytic hollow fiber into a 3/8 in. stainless steel tube and sealing both ends with epoxy resin. The module was inserted in the furnace and centered in order to keep the sealed end outside the heating zone. After the module was connected to inlet and outlet streams, the system was purged thoroughly with argon (50 mL-min-1) and monitored by oxygen analyzer (model 572 with ±0.01% resolution, Servomex). The reduction of nickel-based catalyst was carried out under H2 atmosphere (10% H2/Ar of 50 mL-min-1) at 400 °C for 2 h. The catalytic performance of the CHFs was investigated at set temperatures between 375 and 550 °C under atmospheric pressure.

For the purpose of comparison, a 25 wt % Ni/SBA-15 catalyst in powder form was also prepared by a sol-gel process followed by wet incipient impregnation and tested in a packed bed reactor configuration with a radial dimension of approximately 1 cm. A 113 mg sample of the catalyst were dispersed in 2.00 g of fine silica carbide (SiC), giving a packed-bed length of approximately 2 cm. The quantity of catalyst was defined considering the highest Ni loading achieved for CHFs (27.7 mg of Ni in CHFs of 7 cm). The conditioning and operating procedures adopted were identical to those described above for CHFs. The catalytic performance at each temperature point was evaluated three times, having a deviation of less than 2%.

The overall performance was evaluated based on effective CH4 conversion (xch4), CO2 selectivity (Sco2), and productivity rate (Y), which are defined by the equations

(FCO + FCO,)

X 100%

CH,Inlet

(fco + fco2) Co + fco2)

X 100%

Figure 1. Schematic representation of experimental apparatus.

mNi (3)

where FCH4, FCO, and Fqo2 are the flow rates (mol-min-1) of methane, carbon monoxide, and carbon dioxide, respectively, and mNi is the mass of catalyst (g). Gas hourly space velocity (GHSV) was defined as volumetric flow rate of reactants (m3-min-1) over total volume of reactor (m3).

3. RESULTS AND DISCUSSION

3.1. Microstructured Alumina Hollow Fiber. Figure 2 shows SEM images of Al2O3 hollow fibers prepared by a viscous-fingering-induced phase inversion technique and sintered at 1300 °C. The unique asymmetric microstructured Al2O3 hollow fiber consists of two distinct regions: an outer spongelike layer covering less than 20% of the cross-sectional thickness of the hollow fiber, as can be seen in Figure 2A,A1;

and a plurality of self-organized radial microchannels penetrating through the inner surface, as shown in Figure 2B,B1. Moreover, there is a thin wall separating the radial microchannels which is part of the spongelike structure throughout the hollow fiber. The spongelike structure possesses a packed-pore network, similar to that of the outer surface, as shown in Figure 2C,C1.

Figure 3 provides further details of the asymmetric pore structure of the alumina hollow fiber. As can be seen, the lower

0.01 0.1 1 10 100 Pore Size Diameter (jim)

Figure 3. Mercury intrusion porosimetry of Al2O3 hollow fiber.

and wider peak from 5 to 25 ^m represents the radial microchannels, and the second peak at approximately 0.18 ^m represents the average pore size of the spongelike structure, agreeing well with previous studies.30

The properties of the spongelike structure, such as thickness, porosity, and pore size, largely determine the mechanical strength and mass transfer across and inside the fiber. Usually, a higher sintering temperature leads to better mechanical strength, but at the expense of lower porosity and higher mass-transfer resistance. The mechanical strength of such fibers is approximately 56 MPa, which is 34% higher than the value reported in a previous study using fibers made of the same material, with similar morphology and sintered at the same temperature.30 The higher mechanical strength might be caused by a slight difference in the thickness of the outer spongelike

layer as well as number and dimension of the radial microchannels.

The large number of radial microchannels inside the hollow fiber contributes to significantly increase the geometric surface area, to ease catalyst incorporation, and to improve interaction between reactants and catalyst. The high porosity of the hollow fiber (57%), which is provided not only by the large number of microchannels but also by the interparticle spaces inside the spongelike structure, significantly reduces mass-transfer resistance. Garcia-Garcia et al.16 investigated the application of asymmetric Al2O3 hollow fibers for catalytic reactions. The asymmetric Al2O3 hollow fiber, with 50 MPa of mechanical strength and a nitrogen permeation of 200 L-m-2-s-1 at 200 KPa, was used as a support for a palladium membrane and as a substrate for a copper-based catalyst. Despite the significant length (approximately 80% of cross-section) of petal shaped microvoids, a hollow fiber of this type still represents a considerable barrier to mass transfer, mainly because of its lower overall porosity and much less porous inner surface. In contrast, the highly porous microstructured hollow fiber in this study presents a high nitrogen permeability of 425 L*m-2*s-1, measured at room temperature, 200 kPa, and with a membrane area of 0.8 cm2. This highlights the low mass-transfer resistance of the substrate, which will greatly promote the contact between reactants and catalyst. Moreover, the dimensions and volume of the radial microchannels play an important role in catalyst deposition and loading, especially for catalyst wash coating in which the size of catalyst particles is suggested to be at least three times smaller than the microchannel entran-ces.16,29 On the other hand, incorporating catalyst via a solution technique, such as depositing the secondary support via a sol-gel-based method followed by impregnating the catalytic active phase, has been proven to be efficient for both symmetric and asymmetric ceramic hollow fiber supports.15,16 Because of the less porous surface of the previous hollow fiber substrates, concentration and viscosity of the precursor solutions need to be carefully controlled to avoid blocking the surface, which can increase mass-transfer resistance, thus limiting the quantity of catalyst that can be deposited. Benefiting from the large microchannel openings, more concentrated precursor solution can thus be employed for higher loadings of the secondary

support as well as the catalyst active phase without affecting the mass-transfer efficiency of the hollow fiber substrates.

3.2. Catalytic Hollow Fiber. 3.2.1. Incorporation of SBA-15. The overall performance of catalytic reactors for methane-reforming reactions is greatly affected by the specific surface area provided by the substrate in which the active metal, nickel, is deposited.21 Secondary supports, such as SBA-15, are commonly required in order to provide additional surface area and, consequently, reduce the deactivation of metal-based catalyst due to carbon formation and sintering. After high-temperature sintering, the specific surface area of the alumina hollow fiber is considerably low. As a result, SBA-15 was incorporated prior to impregnation of the catalytic phase. The deposition of SBA-15 was performed by a sol-gel-based process because this technique has been widely used for incorporation of catalyst into hollow fibers and can be used in virtually any type of microstructured support.20 The quantity and distribution of SBA-15 inside the microstructured Al2O3 hollow fiber affects both Ni impregnation and catalytic performance and can be controlled by varying the aging time and hence the viscosity of the SBA-15 sol. Table 1 and Figure 4

Table 1. Loading and Specific Surface Area of Different

Ageing Time SBA-15/Al2O3 HF

composition of SBA-15 loading specific surface area (m2-g 1)

CHF (wt %)

Al2O3 HF n/a 2.6

0 h SBA-15 2.61 14.8

6 h SBA-15 5.84 34.4

12 h SBA-15 7.27 50.1

show the properties and SEM images of the hollow fibers without and with SBA-15 at different aging times. SBA-15 loadings of 2.61%, 5.84%, and 7.27% were obtained for the aging time of 0, 6, and 12 h, respectively, which is reasonable because the adhesion of sols to a porous substrate increases with viscosity.31 The incorporation of SBA-15 into hollow fibers largely enhanced the specific surface area by 6, 13, and 19 times for the aging times of 0, 6, and 12 h, respectively. These results are in agreement with a previous study.21 Apart from increasing the specific surface area, aging time affected the distribution of SBA-15 inside the hollow fiber. Thus, as can be seen in Figure 4A1-D1, with longer aging time, and consequently more viscous SBA-15 sol, the penetration of SBA-15 sol into the packed-pore network of the spongelike structure is more difficult. As a result, the SBA-15 is preferentially distributed along the microchannels, as shown in Figure 4 A2-D2, and it covers more small pores on the inner surface (packed-pore network) among microchannel entrances, as can be seen in Figure 4A3-D3. Moreover, the incorporation of SBA-15, for all aging time sols, did not block, fully or partially, the microchannel entrances on the inner surface of the fiber.

3.2.2. Wet Impregnation of Ni. The nickel nitrate, precursor of catalytic active nickel, was successfully impregnated into the Al2O3 hollow fibers with or without SBA-15. As can be seen in Figure 5, crystals of nickel oxide were formed along the radial microchannels of all the samples; however, neither crystal nor particles of NiO could be clearly identified in SEM images of spongelike structure. Despite this fact, a homogeneous distribution of NiO on the packed-pore structure was visually observed by a light gray color, as shown in Figure 6. Although

the adsorption capacity of SBA-15 is significantly higher than that of alumina, the actual loading of NiO in the samples with and without SBA-15 was quite similar (varying from 4 to 5.7 wt % as determined by weight difference). This is mainly because the amount of SBA-15 incorporated is considerably small (Table 1) when compared with the significant pore volume from both the radial microchannels and spongelike structure of the alumina hollow fiber.

The X-ray diffraction patterns of the CHFs and pure Al2O3 hollow fiber are presented in Figure 7. The distinctive peaks of Al2O3 and NiO can be identified in all CHF samples. However, the intensity of the NiO peaks was low because of the reduced amount and high dispersion. Furthermore, the amorphous nature and the small quantity of the SBA-15 prevented the detection of the characteristic peak of SBA-15 at 20 value of 21.79°.

The porosity of the hollow fiber was considerably affected by the incorporation of NiO or NiO and SBA-15, as can be seen in Table 2. A porosity loss of 3.7, 16.5, 24.1, and 24.3% was observed for NiO, NiO/0h SBA-15, NiO/6h SBA-15, and NiO/12h SBA-15 CHFs, respectively.

Figure 8 shows a more detailed analysis of the pore structure and reveals that the radial microchannels (between 5 and 25 y«m) are little affected by the incorporation of NiO and SBA-15, which is in agreement with Figures 4 and 5. In contrast, the pore size of the spongelike structure was highly affected because of the partial filling of the interparticles spaces. As the aging time of SBA-15 increases, the intensity of the peak at 0.18 ^m decreases and a third peak, with smaller diameter (0.07 ^m), becomes more manifest and intense. The deposition of SBA-15 particles in the packed-pore structure and formation of SBA-15 layer along the microchannels resulted in the partial filling and blockage of pores and, consequently, a reduction of their size. As shown in Figure 4B1,B2, the majority of the SBA-15 incorporated in 0h SBA-15-CHF was distributed in the interparticle spaces of the packed-pore spongelike structure, leading to a reduction of the overall porosity in the spongelike structure. This statement is supported by mercury intrusion porosimetry results, in which the peak corresponding to packed-pore structure suffered a small drop in intensity and slight shift in position toward a smaller pore size. Such a phenomenon becomes more significant with SBA-15 aged at longer time because of the formation of a SBA-15 layer along the microchannel.

3.3. Catalytic Performance. The catalytic performance of the CHFs and 25 wt % NiO/SBA-15 catalyst was evaluated based on effective methane conversion, CO2 selectivity, and productivity rate (Figure 9). Table 2 lists the amount of Ni (in milligrams) involved in the different reactors.

Figure 9A shows the methane conversion of CHFs and PBR as a function of temperature. Apparently, effective methane conversion for both CHFs and PBR increase with temperature, showing similar trends among all different configurations and reaching values close to the equilibrium conversion. Although the methane conversion of PBR is slightly higher than that of CHFs at 540 °C, it started to drop in a slow manner at approximately 560 °C. This may be caused by encapsulation of nickel due to instability of SBA-15 structure, as reported by McMinn et al.32 In contrast, such a drop in methane conversion was not observed in CHFs. One of the possible reasons is that CHFs provide an ideal flow of the reactants, with the widely opened radial microchannels significantly reducing mass- and heat-transfer resistance and consequently allowing a more

Figure 4. SEM images of Al2O3 hollow fiber (A) and SBA-15 in Al2O3 hollow fiber with aging time of 0 h (B), 6 h (C), and 12 h (D): A1, B1, C1, and D1 spongelike layer between microchannels; A2, B2, C2, and D2 microchannel entrance; and A3, B3, C3, and D3 inner surface.

efficient interaction between the reactants and the catalyst. At 465 °C, the effective methane conversion of all CHFs, including CHF without SBA-15 (Ni-CHF), is rather close to the thermodynamic equilibrium. This indicates great efficiencies when dispersing catalyst inside the microstructured alumina hollow fiber. On the other hand, at elevated operating temperatures (>500 °C), CHFs containing SBA-15 showed higher methane conversion over the Ni-CHF, which may be linked to the additional surface area provided by SBA-15. As the operating temperature increases and reaches values similar or above catalyst calcination temperature (540 °C), the difference between actual methane conversion and thermodynamic equilibrium value increases, possibly as a result of sintering of the catalyst, loss of active sites, and encapsulation of Ni by both SBA-15 and Al2O3.32

Figure 9B compares CO2 selectivity of different reactor designs. CO2 selectivity keeps decreasing with increasing operating temperatures, as expected from the thermodynamic point of view, because steam methane reforming (SMR) is highly endothermic and WGS is slightly exothermic. Evidently, CO2 selectivity of PBR is much lower than the CHF counterparts within the selected range of operating temperatures. The benefits of dispersing the catalyst inside micro-

structured alumina hollow fibers are then further proven to enhance reaction performance by reducing mass-transfer resistance and, consequently, achieving a more uniform concentration profile. A uniform and highly dispersed catalyst also increases the accessibility of the catalyst to reactants, which increases the possibility of CO (produced in SMR) to be further converted into CO2 via WGS.

Because of major differences in reactor configuration, such as flow patterns, catalyst distribution and loading, reactor volume, and contact/residence time, the overall performance of the different reactor configurations cannot be precisely assessed by exclusively analyzing effective methane conversion and CO2 selectivity. Although the methane conversions are quite similar, the nickel loading is considerably different and gas hourly space velocity (GHVS) in CHF is approximately 6.5 times higher than in the PBR configuration (27 255 h-1 and 4053 h-1, respectively). Consequently, productivity rate, which is calculated based on the formation of CO and CO2 over the mass of Ni involved in each reactor configuration (eq 3), is employed for a more sensible comparison. As seen in Figure 9C, the productivity rate of CHFs can be much greater than that of the PBR. However, at intermediate temperatures the effective methane conversion of PBR is rather close to

Figure 5. SEM images of (A) NiO/Al2O3 HF, (B) NiO/0h SBA-15-CHF, (C) NiO/6h SBA-15-CHF, and (d) NiO/12h SBA-15-CHF.

Table 2. Mass of Nickel Available for Reaction and Porosity of the Substrate after Catalyst Incorporation

reactor configuration

Ni-CHF

Ni/0hSBA-15-CHF Ni/6hSBA-15-CHF Ni/12hSBA-15-CHF

mass of Ni (n 27.7 27.7

16.7 23.4

porosity (%) n/a 54.4 47.2 43.9 43.2

Figure 6. Image of Al2O3 hollow fiber (A) and NiO/12h SBA-15-CHF (B).

Figure 7. XRD patterns of assembled CHFs and Al2O3 HF and NiO characteristic peaks (dashed lines).

Figure 8. MIP results of the Al2O3 HF and different catalytic hollow fibers.

thermodynamic equilibrium value and, as a result, it is not clear if the catalyst is operating at full capacity. Therefore, the comparison between catalytic performance of CHFs and PBR must be assessed at 540 °C. The productivity rate of PBR is considerably lower than that of CHFs containing SBA-15, which can be an advantage of dispersing catalyst inside the microstructured hollow fibers. The discrepancy between the

95-S 90- 851

£ 758"

6560 J-.-,-.-,-■-1-.-,-.-

350 400 450 500 550 600 100

'el > 80

3 60 u

B 20 s

^ Oi-.-■-.-,---r---

350 400 450 500 550 600

Temperature (°C)

—■—25wt%Ni/SBA-15 PBR—•—Ni-CHF—A—Ni/ohSBA-15 -CHI' —T—Ni/ 6hSBA-15 - CHF—*—Ni/12h SBA-15 - CHF

Figure 9. Temperature-dependent (A) effective methane conversion, (B) CO

2 selectivity, and (C) productivity rate of different CHFs and PBR configurations.

CHFs with SBA-15 might be due to the difference of actual mass of Ni, as the effective methane conversion values are very similar. This suggests the amount of nickel available is higher than necessary and requires further tuning.

4. CONCLUSIONS

A microstructured catalytic hollow fiber was successfully developed by incorporating Ni/SBA-15 catalyst into alumina hollow fibers with low mass- and heat-transfer resistance. The unique and widely opened radial microchannels (5-25 ^m) on the inner surface of the hollow fiber facilitate catalyst incorporation and contribute to higher interaction between the reactants and catalyst, resulting in a higher catalytic performance. The methane conversion achieved from both CHF and PBR configuration was similar and close to thermodynamic equilibrium values (around 25% at 465 °C); however, the space velocity in CHF was considerably higher (approximately 6.5 times). Meanwhile, greater CO2 selectivity was obtained from CHFs because of more uniform concentration and temperature profiles when comparing to those of PBR. Furthermore, the great productivity rate of CHFs containing SBA-15 indicates their potential for further reducing the catalyst required.

■ AUTHOR INFORMATION Corresponding Author

*Tel.: +44 207 5945676. Fax: +44 207 5945629. E-mail: kang. li@imperial.ac.uk.

The authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

The authors gratefully acknowledge the research funding provided by EPSRC in the United Kingdom (Grant: EP/ I010947/1)

■ REFERENCES

(1) Angeli, S. D.; Monteleone, G.; Giaconia, A.; Lemonidou, A. A. State-of-the-art catalysts for CH4 steam reforming at low temperature. Int. J. Hydrogen Energy 2014, 39 (5), 1979-1997.

(2) Chaubey, R.; Sahu, S.; James, O. O.; Maity, S. A review on development of industrial processes and emerging techniques for production of hydrogen from renewable and sustainable sources. Renewable Sustainable Energy Rev. 2013, 23 (0), 443-462.

(3) Holladay, J. D.; Hu, J.; King, D. L.; Wang, Y. An overview of hydrogen production technologies. Catal. Today 2009, 139 (4), 244260.

(4) Hydrogen and Syngas Production and Purification Technologies; Liu, K., Song, C., Subramani, V., Eds.; Wiley: Hoboken, NJ, 2009.

(5) Mendes, D.; Mendes, A.; Madeira, L. M.; Iulianelli, A.; Sousa, J. M.; Basile, A. The water-gas shift reaction: from conventional catalytic systems to Pd-based membrane reactors—A review. Asia-Pac. J. Chem. Eng. 2010, 5 (1), 111-137.

(6) Aasberg-Petersen, K.; Nielsen, C. S.; Jorgensen, S. L. Membrane reforming for hydrogen. Catal. Today 1998, 46 (2-3), 193-201.

(7) Tonkovich, A. L. Y.; Yang, B.; Perry, S. T.; Fitzgerald, S. P.; Wang, Y. From seconds to milliseconds to microseconds through tailored microchannel reactor design of a steam methane reformer. Catal. Today 2007, 120 (1), 21-29.

(8) Tonkovich, A. Y.; Perry, S.; Wang, Y.; Qiu, D.; LaPlante, T.; Rogers, W. A. Microchannel process technology for compact methane steam reforming. Chem. Eng. Sci. 2004, 59 (22-23), 4819-4824.

(9) Zanfir, M.; Gavriilidis, A. Catalytic combustion assisted methane steam reforming in a catalytic plate reactor. Chem. Eng. Sci. 2003, 58 (17), 3947-3960.

(10) Zhai, X.; Ding, S.; Cheng, Y.; Jin, Y.; Cheng, Y. CFD simulation with detailed chemistry of steam reforming of methane for hydrogen production in an integrated micro-reactor. Int. J. Hydrogen Energy 2010, 35 (11), 5383-5392.

(11) Charpentier, J. C. Process Intensification by Miniaturization. Chem. Eng. Technol. 2005, 28 (3), 255-258.

(12) Jahnisch, K.; Hessel, V.; Lowe, H.; Baerns, M. Chemistry in Microstructured Reactors. Angew. Chem., Int. Ed. 2004, 43 (4), 406446.

(13) Kolb, G.; Hessel, V. Micro-structured reactors for gas phase reactions. Chem. Eng. J. (Amsterdam, Neth.) 2004, 98 (1-2), 1-38.

(14) Wang, F.; Qi, B.; Wang, G.; Li, L. Methane steam reforming: Kinetics and modeling over coating catalyst in micro-channel reactor. Int. J. Hydrogen Energy 2013, 38 (14), 5693-5704.

(15) García-García, F. R.; Rahman, M. A.; Kingsbury, B. F. K.; Li, K. A novel catalytic membrane microreactor for COx free H2 production. Catal. Commun. 2010, 12 (3), 161-164.

(16) García-García, F. R.; Kingsbury, B. F. K.; Rahman, M. A.; Li, K. Asymmetric ceramic hollow fibres applied in heterogeneous catalytic gas phase reactions. Catal. Today 2012, 193 (1), 20-30.

(17) Rahman, M. A.; García-García, F. R.; Li, K. Development of a catalytic hollow fibre membrane microreactor as a microreformer unit for automotive application. J. Membr. Sci. 2012, 390-391, 68-75.

(18) García-García, F. R.; Torrente-Murciano, L.; Chadwick, D.; Li, K. Hollow fibre membrane reactors for high H2 yields in the WGS reaction. J. Membr. Sci. 2012, 405-406 (0), 30-37.

(19) García-García, F. R.; Li, K. New catalytic reactors prepared from symmetric and asymmetric ceramic hollow fibres. Appl. Catal., A 2013, 456 (0), 1-10.

(20) Gallegos-Suarez, E.; García-García, F. R.; Gonzaiez-Jimenez, I. D.; Rodríguez-Ramos, I.; Guerreo-Ruiz, A.; Li, K. Ceramic hollow fibres catalytic enhanced reactors for glycerol steam reforming. Catal. Today 2014, 233 (0), 21-30.

(21) Gbenedio, E.; Wu, Z.; Hatim, I.; Kingsbury, B. F. K.; Li, K. A multifunctional Pd/alumina hollow fibre membrane reactor for propane dehydrogenation. Catal. Today 2010, 156 (3-4), 93-99.

(22) Zhao, D.; Sun, J.; Li, Q.; Stucky, G. D. Morphological Control of Highly Ordered Mesoporous Silica SBA-15. Chem. Mater. 2000, 12 (2), 275-279.

(23) Wan, H.; Li, X.; Ji, S.; Huang, B.; Wang, K.; Li, C. Effect of Ni Loading and CeIZri-IO2 Promoter on Ni-Based SBA-15 Catalysts for Steam Reforming of Methane. J. Nat. Gas Chem. 2007, 16 (2), 139147.

(24) Rahman, M. A.; García-García, F. R.; Hatim, M. D. I.; Kingsbury, B. F. K.; Li, K. Development of a catalytic hollow fibre membrane micro-reactor for high purity H2 production. J. Membr. Sci. 2011, 368 (1-2), 116-123.

(25) Rahman, M. A.; García-García, F. R.; Li, K. On-board H2 generation by a catalytic hollow fibre microreactor for portable device applications. Catal. Commun. 2011, 16 (1), 128-132.

(26) García-García, F. R.; Tsang, S. C.; Li, K. Hollow fibre based reactors for an enhanced H2 production by methanol steam reforming. J. Membr. Sci. 2014, 455 (0), 92-102.

(27) Hatim, M. D. I.; Fazara, M. A. U.; Syarhabil, A. M.; Riduwan, F. Catalytic Dehydrogenation of Methylcyclohexane (MCH) to Toluene in a Palladium/Alumina Hollow Fibre Membrane Reactor. Procedia Eng. 2013, 53 (0), 71-80.

(28) Kingsbury, B. F. K.; Li, K. A morphological study of ceramic hollow fibre membranes. J. Membr. Sci. 2009, 328 (1-2), 134-140.

(29) Kingsbury, B. F. K.; Wu, Z.; Li, K. A morphological study of ceramic hollow fibre membranes: A perspective on multifunctional catalytic membrane reactors. Catal. Today 2010, 156 (3-4), 306-315.

(30) Lee, M.; Wu, Z.; Wang, R.; Li, K. Micro-structured alumina hollow fibre membranes - Potential applications in wastewater treatment. J. Membr. Sci. 2014, 461 (0), 39-48.

(31) Chen, K.-Y.; Shen, C.-C.; Lee, C.-Y.; Lee, S.-J.; Leu, C.-H.; Wang, J.-H.; Yeh, C.-T. Coating powdered copper catalyst with yttria sol. Mater. Chem. Phys. 2011, 128 (1-2), 57-61.

(32) McMinn, T. E.; Moates, F. C.; Richardson, J. T. Catalytic steam reforming of chlorocarbons: Catalyst deactivation. Appl. Catal., B 2001, 31 (2), 93-105.