Biogas reforming using renewable wind energy and induction heating

G Model

ARTICLE IN PRESS

CATTOD-9119; No. of Pages 10

Catalysis Today xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Catalysis Today journal homepage: www.elsevier.com/locate/cattod

Biogas reforming using renewable wind energy and induction heating María Natividad Pérez-Camacho a , Jehad Abu-Dahrieh a , David Rooney a,∗ , Kening Sun b a b

CenTACat, School of Chemistry and Chemical Engineering, Queen’s University Belfast, Belfast, Northern Ireland, BT9 5AG, UK Department of Chemistry, Harbin Institute of Technology, Harbin 150001, PR China

a r t i c l e

i n f o

Article history: Received 28 February 2014 Received in revised form 25 May 2014 Accepted 3 June 2014 Available online xxx Keywords: Biogas Induction heating Reforming Perovskites Wind power

a b s t r a c t While the benefits of renewable energy are well known and used to influence government policy there are a number of problems which arise from having significant quantities of renewable energies on an electricity grid. The most notable problem stems from their intermittent nature which is often out of phase with the demands of the end users. This requires the development of either efficient energy storage systems, e.g. battery technology, compressed air storage etc. or through the creation of demand side management units which can utilise power quickly for manufacturing operations. Herein a system performing the conversion of synthetic biogas to synthesis gas using wind power and an induction heating system is shown. This approach demonstrates the feasibility of such techniques for stabilising the electricity grid while also providing a robust means of energy storage. This exemplar is also applicable to the production of hydrogen from the steam reforming of natural gas. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Greener sources of energy are continually being investigated as alternatives to fossil fuels and of these, biogas which is derived from the anaerobic digestion of biomass, is interesting from both a financial and environmental viewpoint. Biogas is a secondary energy carrier which can be manufactured from a range of organic materials resulting in the production of methane that can be used for the generation of on-site energy or potentially through gas to liquid processes (GTL) to methanol, dimethyl ether (DME) or higher hydrocarbons by Fischer–Tropsch processes [1,2]. As the main components of natural biogas are methane (55–70% in volume) and carbon dioxide (30–45% in volume) with ppm levels of H2 S, NH3 , H2 , N2 , O2 and H2 O [3], scrubbed mixtures of this are thus suitable feed for the dry reforming of methane (DRM) reaction (CH4 + CO2 ↔ 2CO + 2H2 , H◦ = 247 kJ mol−1 ). Dry reforming of methane represents the lowest cost route to the production of syngas [4] from biogas whereas steam reforming (CH4 + H2 O ↔ CO + 3H2 , H◦ = 206 kJ mol−1 ) is the most industrially applied catalytic process used to date [5]. The main obstacle to the industrial implementation of DRM is the absence of commercial catalysts with a proven high activity and stability and a resistance to carbon accumulation [6]. The most active metals for the DRM are those from groups 8, 9 and 10 with

∗ Corresponding author. Tel.: +44 2890974050; fax: +44 2890974687. E-mail address: [email protected] (D. Rooney).

Ni being the most commonly used because of its availability and price [7]. Recent studies have focused on the development of complex mixed metallic oxides with a perovskite structure as they are suitable for methane conversion in reactions such as steam reforming, partial oxidation and dry reforming. It has been reported that the perovskite-type oxides precursors fulfil the requirements of high metal dispersion and thermal stability in the dry reforming of methane [8]. The utilization of ABO3 perovskites oxides, where the A-site cation is a rare earth and/or alkaline earth, and the Bsite cation is a transition metal, has increased lately [9]. A common structure for DRM is LaNiO3 and it has been found that the addition of alkali metals or rare earth elements helps reduce carbon deposition [10,11] while the dilution of Ni in the metallic particles by a second metal such as Mn, Fe, Cu, or Al helps prevent sintering [12]. A perovskite-type mixed metal oxide material with the composition Na0.5 La0.5 Ni0.3 Al0.7 O2.5 is one such catalyst which has shown very good performance [13]. This particular catalyst forms the basis of the work herein and was selected from a range of previously tested catalysts the results of which are not included here. As discussed above this reaction is highly endothermic and thus requires a significant energy input. Conventional heating using standard furnaces results in slow start-up times due to the overall heat capacity of the system. While this has certain advantages it is considered to be too slow in order to adjust to the changes in wind speed which this paper is targeting. Hence the feasibility of energising the reaction by induction heating will be shown. This method can also be used as a strategy for lowering carbon rates and improving stability when compared to traditional heating

http://dx.doi.org/10.1016/j.cattod.2014.06.010 0920-5861/© 2014 Elsevier B.V. All rights reserved.

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model CATTOD-9119; No. of Pages 10

ARTICLE IN PRESS M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx

2

systems [14,15]. According to theory, three types of electromagnetic heating can be described depending on the frequency of the electromagnetic field used: (a) induction heating (IH) or radiofrequency (RF), (b) dielectric or (c) microwave [16]. IH presents the lowest frequency being in the range of 1000 Hz which presents several advantages as it is not harmful to human body and is uniform and this flexibility gives the possibility of adapting it to renewable energy fluctuations. For example IH has shown its ability to quickly and homogeneously heat metallic surfaces [17] similar to the stainless steel reactor is used in the present work. To our knowledge induction heating has never been applied to the dry reforming of methane reaction before and its direct application in combination with wind power is described in this paper. While IH techniques for the DMR are considered novel Dieckmann [14] has previously demonstrated an enhanced performance of a Ni-based catalysts for diesel reforming using electromagnetic fields, reporting an approximate 4 fold reduction in the carbon formation. When studying the dry reforming of methane, coke formation and catalyst deactivation must be considered and keeping these two issues under control is critical for long term operation. Additionally the reverse water gas shift (RWGS), methanation reaction and steam reforming of methane may also affect the equilibrium [18,19]. Hence the effect of IH on the catalyst stability as well as product distribution is important.

2. Experimental 2.1. Preparation of perovskite catalytic precursors A sol–gel method was applied to prepare the mixed-oxide Na0.5 La0.5 Ni0.3 Al0.7 O2.5 [13] used in the induction heating experiments. Initially precursor solutions were prepared by dissolving the raw materials in previously heated propionic acid. The precursor reagents; lanthanum nitrate hexahydrate La(NO3 )3 ·6H2 O, aluminum nitrate nonahydrate Al(NO3 )3 ·9H2 O, nickel nitrate hexahydrate Ni(NO3 )2 ·6H2 O and sodium nitrate NaNO3 were separately dissolved in the minimum volume of hot propionic acid at atmospheric pressure and 90 ◦ C. All reagents had a purity >99% and were supplied by Sigma–Aldrich Ltd. After dissolution, each solution was mixed and stirred at atmospheric pressure and 130 ◦ C for 120 min. After that, the propionic acid was distilled in a reflux process until the formation of a gel was observed, which was dried at 90 ◦ C overnight and then calcined in a static air atmosphere at 725 ◦ C for 4 h. After calcination, the catalyst was tested.

2.2. Experimental setup A system using a feed composed of a mixture of methane and carbon dioxide, i.e. the main components of natural biogas was used. The dry reforming tests were carried out with a synthetic mixture of CH4 :CO2 :Ne in a ratio of 9.5:9.5:1 with a GHSV of 30,000 cm3 gcat −1 h−1 when using the gas chromatographer for analysis of the gas outlet or with a mixture of CH4 :CO2 : 5% Kr/Ar in a ratio of 9.5:9.5:1 when using the mass spectrometer for analysis of the gas outlet. The total flow used was 50 mL min−1 . This is a suitable approximation to natural biogas while allowing for an inert gas (Ne or Kr) to be used for calibration. All gases were supplied by BOC and had greater than 99% purity. The bed of catalyst was set up with 0.1 g of catalyst, in the form of 250–425 ␮m diameter pellets and was placed in the centre of the stainless steel reactor (inner diameter of 6.35 mm) and secured with quartz wool at both sides in order to prevent movement of the catalyst during reaction.

A schematic diagram of the system is shown in Fig. 1. Here a 2 kV Ambrell Easyheat unit (Fig. S1) was used to produce the induction heating through a 9-turn copper solenoid coil applicator. Cooling water was used inside the coil at 20 ◦ C. The stainless steel fixed bed reactor was situated inside the induction coil. The gases were fed into the reactor and the products analysed by an on-line gas chromatograph (GC) and/or mass spectrometer (MS) where appropriate. The GC utilised was a Perkin Elmer Clarus 500 provided with a Hayesep DB column and equipped with FID and TCD detectors. A quadrupole MS HidenTM HPR-20 operated by MASsoft software was also used. A post reaction temperature programme oxidation (TPO) was performed using the MS and a Carbolite® furnace model MTF 10/25/130, this was done in situ in the stainless steel reactor. After the reforming reaction, the sample was not taken out, but the inlet and outlet of the reactor were disconnected to remove the induction coil. The coil was then substituted by the Carbolite® MTF 10/25/130 furnace. The outlet of the reactor was connected to the mass spectrometer for the analysis of the gas products. For the temperature programmed oxidation experiments typically, 2, 15, 16, 18, 20, 22, 28, 32, 44, 82, and 84 amu were followed while for the catalytic experiments 13, 28, 44, 82 and 84 amu were registered. The TPOs were conducted from room temperature to a maximum of 800 ◦ C using a heating ramp of 5 ◦ C/min. The feed used was 50 mL min−1 of total flow in a ratio of 20% O2 /Ar: 5% Kr/Ar: He = 2.5:1:1.5, this is 25 mL min−1 of 20% O2 /Ar, 10 mL min−1 of 5% Kr/Ar (using Kr as internal standard for the TPO) and balanced with He (15 mL min−1 ). The completed 2 kV Ambrell Easyheat unit used to provide the electromagnetic field and the system operating under biogas reforming is shown in the supplementary information. The reaction temperature was measured using a high speed fibre optic infrared transmitter OS4000 from Omega UK which operates in the temperature range from 300 to 1200 ◦ C. It was positioned such that no contact with the reactor took place and the signal was recorded online using the OMEGASOFTTM infrared temperature software interface. In order to provide additional temperature data inside the reactor, a K-type thermocouple supplied by Omega UK was used. This temperature was also recorded online using the PicoLog data acquisition software under reaction conditions. Figures of the real system set up used are shown in the supplementary information (Figs. S1 and S2). When steam reforming of methane experiments were performed, slight modification of the rig to allow the introduction of steam in the feed of gases was required. The steam reforming experiment was carried out with an S/C ratio of 1.1 to allow all the methane to be converted. The total flow rate used was 50 mL min−1 with 5% of internal standard (Ne or 5% Kr/Ar) depending if GC or MS was used. Methane and the internal standard were fed through a saturator whose temperature was controlled by a GrantTM GP 200 thermostatic bath. All the gas lines from the mass flow controllers to just before the induction system and after the outlet of it to the GC/MS entrance were heated to prevent water condensation using regulators controlled by tapes from Barnstead-ElectrothermalTM . 2.3. Characterisation of the mixed metallic oxides X-ray diffraction (XRD), energy-dispersive x-ray spectroscopy (EDX/EDS), analysis, Brunauer–Emmett–Teller Analysis (BET) and scanning electron microscope (SEM) measurements were used over the fresh and used samples. The XRD analysis was carried out using a PANalytical XpertPro equipment, which operates at 40 kV and 40 mA, using Cu K␣ radiation ( = 0.154 nm). Xpert data viewer software was used in conjunction with the instrument. The SEM used was a FEI Quanta 250 FEG MKII with a high resolution environmental microscope (ESEM) using XT microscope

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model

ARTICLE IN PRESS

CATTOD-9119; No. of Pages 10

M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx

3

IR probe for temperature T > 300 ˚C IH Coil Safety relief valve CH4

P

SS Tubular Reactor

P CO2

P

20% O2 /Ar

Effluent for Analysis to GC/ MS

Ne

He

Vent Fig. 1. Schematic drawing of the experimental set up.

3. Results and discussion 3.1. Blank heating tests Before the catalytic experiments, it was necessary to determine the performance of the induction heating system. A preliminary test was carried out to calibrate the heating rate of the catalyst by comparing the temperature on the surface as measured using the infrared probe and the actual bed temperature using the thermocouple installed in the middle of the reactor. This was carried out using the same gas hourly space velocity to be used in the reaction and was performed by ramping up the current (A) of the induction heating system. Fig. 2 demonstrates the rapid response of the system to changes in the applied current from room temperature to ∼555 ◦ C. The same rate cannot be plotted using data obtained from the infrared probe due to the lower temperature limitation of these measurements (∼300 ◦ C) therefore this data is not shown in Fig. 2. At a current of 69.5 A this gives a heating rate of ∼65 ◦ C/min for the first cycle with the cooling rate being slower (∼41 ◦ C/min) with a second heating up rate of ∼100 ◦ C/min. It is suggested that the second heating rate is higher due to changes in the properties of the reactor under sustained heating. Several experiments performed with the same reactor showed that the heating rates found afterwards are approximately similar to the second rate.

When heated to a given temperature, ferromagnetic materials firstly undergo a reduction of their magnetic properties, and can lose them altogether. The temperature at which the permeability drops to  = 1 is called the magnetic transformation temperature or the Curie point (here  represents the permeability or degree of magnetization of a material which is equal to unity for non-magnetic materials and between 100 and 500 for materials which are easy to heat with IH) [20]. The Curie point or temperature would be the maximum temperature to which the material retains its magnetic properties and after this point it would become paramagnetic. For stainless steel this temperature is close to 900 ◦ C [20] although it is influenced by the composition of the alloy. It is notable that when a workpiece is heated under induction heating, the temperature on the surface (measured using the infrared probe) is lower than the temperature of the workpiece itself. The surface temperature is reduced by (1) the finite depth of current penetration and (2) surface radiation [21]. The heat loss in the reactor can be modelled by the skin effect equations (Eqs. (1) and (2)) [22] and by the Stefan-Boltzmann law which describes the radiation loss (in Watts per square metre) as shown in Eq. (3) [21]. From the skin effect equations, it can be easily deduced that the

900 T catalytic bed (ºC)

800 700

Temperature/ ºC

Control software and linked to the EDX detector. The EDX used was a 10 mm2 SDD Detector-x-act from Oxford Instruments which utilizes Aztec® EDS analysis software. Both systems used the same chamber. The Brunauer–Emmett–Teller Analysis (BET) was obtained using a Micromeritics ASAP 2010 system. This was achieved by placing the sample under vacuum for 3 h to remove any adsorbed species, such as water, followed by flushing with He for 2 min. This dry sample was then weighed and placed under vacuum to complete the purification process. The sample was then subjected to varying pressures of N2 gas at liquid nitrogen temperature, and the absorption of N2 at these pressures was recorded to obtain an adsorption isotherm. The specific surface area was measured by using the BET equation.

600 500 400 300 200 100 0 0

10

20

30

40

50

60

70

80

90

100

Time / min Fig. 2. Heating up and down of the inducted system catalytic bed measured by thermocouple (Current 69.5 A, 17 V, 202 kHz).

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model

ARTICLE IN PRESS

CATTOD-9119; No. of Pages 10

M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx T catalytic bed (ºC)

900

863

9

300

800

8

785

250 707

700

Power/ kW

Temperature/ ºC

10

350

T surface (ºC)

624

600 557

552 505

500

70

80

4 3

50

90

100

110

120

1

0 0:00

300 60

5 Grid connection

150

2

423 382

50

6

200

100

463 400

7

130

2:24

4:48

7:12

Time / h

heat energy converted from electric energy is concentrated on the surface of the object [22]. ix = i0 e−x/d0

(1)

2 ω

(2)

PR =

5.67 8

10

εS4

(3)

Here ix is the current density at the position x (distance from the skin or surface of the object), i0 is the current density on the skin depth where x is 0, do is a constant determined by the frequency (current penetration depth or skin depth),  is the resistivity of the object,  is the permeability of the object and ω is the frequency of the current flowing through the object, ε is the emissivity coefficient of surface (dimensionless),  S is the absolute surface temperature (Kelvin) and PR is the radiated loss per square meter of surface (W m−2 ). The constant 5.67 × 10−8 W m−2 K−4 is known as Stefan’s constant [21,22]. Due to the above another test was carried out to determine the difference between the temperature reached in the bed of catalyst and the surface of the reactor for the different values of current applied in the induction heating system. The temperature on the surface was measured using the infrared probe and the actual bed temperature using the thermocouple installed in the middle of the reactor. This test was done using the same gas hourly space velocity to be used in the posterior reactions and performed by ramping up the current (A) of the induction heating system, i.e. the set point (SP). Fig. 3 represents these results where temperature on the surface and temperature in the catalytic bed are plotted vs. the applied current. The values of current were chosen as they were needed for the model with wind power as it will be explained in Section 3.2. There are clear differences between these two temperature measurement methods and a significant factor in this difference can be attributed to errors in the emissivity of the steel. It is known that the emissivity of steel will change as a function of temperature and composition [23]. Herein an emissivity of 0.94 was initially chosen however this is high and the real emissivity may be between 0.1 and 0.6 depending on the true surface temperature. However it should be noted that this effect alone cannot account for the differences shown i.e. the difference should become smaller as a function of temperature as the emissivity approaches a constant value. However here these differences become larger and hence this serves to show the importance of heat transfer and surface cooling in the

Fig. 4. Power curve ( ) calculated for the experiment using a Nordex/N80 turbine and wind speed vs time profile from Belfast Harbour ( ).

system which is a key factor for the future development of an improved reactor design with higher thermal efficiency. It should be noted that the system used here was uninsulated thus allowing for high radiative and convective heat losses from the surface, both of which would increase with temperature. Overall this discussion highlights the importance, and difficulties, with obtaining accurate temperature measurements in such systems however similar temperatures in the surface were measured for all the experiments performed thus showing the repeatability of the method. 3.2. Wind power approach-catalytic test As stated above the dry reforming of methane reaction was performed in the inducted heated system due to its quick response (up to 100 ◦ C/min). This means the system has a very low thermal inertia which aids control of the heating rate [24]. For the main test the wind speed of a typical day in Belfast, UK, was taken, over a 12 h period. Using these data the actual power that can be obtained was calculated for a general wind turbine and converted to the current used in the solenoid coil applicator. Power and wind speed are presented in Fig. 4. By applying this current the resulting profile of CH4 conversion was obtained as was the syngas production (Fig. 5). The wind speed was taken for the 17th of May 2013, as measured at Belfast City Harbour (http://www.windfinder.com/report/belfast harbour). The wind turbine chosen for the calculations was a Nordex/N80 unit (2.5 MW) because of its extensive application on wind farms across Northern Ireland. The power curve of this type of turbine is 100

863 ºC

90

785 ºC

80

CH4 conversion/ %

Fig. 3. Comparison of temperature reached in the catalytic bed and on the surface of the reactor for the different values of current applied in the inducted heated system with GHSV = 30,000 cm3 gcat −1 h−1 and CH4 :CO2 :Ne ratio of 9.5:9.5:1. ( ) Temperature in catalytic bed (K-type thermocouple); ( ) Temperature in surface of reactor (infrared probe).

d0 =

0 12:00 14:24 16:48 19:12 21:36 0:00

9:36

Current / A



Wind speed/ ms -1

4

70

707 ºC 707 ºC

60

707 ºC

50 40

624 ºC

30

557 ºC

557 ºC

20

624 ºC 624 ºC

557 ºC

10

557 ºC 624 ºC

557 ºC 557 ºC 557 ºC

0 0

100

200

300

400

500

600

700

TOS / min Fig. 5. Methane conversion profile obtained in the induction heating experiment using a perovskite catalyst and several currents in the dry reforming of methane. Atmospheric pressure, CH4 :CO2 :inert ratio of 9.5:9.5:1 and GHSV = 30,000 cm3 gcat −1 h−1 .

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model

ARTICLE IN PRESS

CATTOD-9119; No. of Pages 10

M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx Table 1 Power and current applied in the experiment in the induction heating unit. Time (h)

Actual P (kW)

Current (A)

Time (h)

Actual P (kW)

Current (A)

06:20 06:50 07:20 07:50 08:20 08:50 09:20 09:50 10:20 10:50 11:20 11:50 12:20 12:50 13:20 13:50

12.95 12.95 87.33 121.92 121.92 189.28 245.74 142.21 121.92 189.28 121.92 59.99 59.99 59.99 121.92 121.92

0 0 0 69.5 69.5 86.4 98.1 75.4 69.5 86.4 69.5 0 0 0 69.5 69.5

14:20 14:50 15:20 15:50 16:20 16:50 17:20 17:50 18:20 18:50 19:20 19:50 20:20 20:50 21:20

87.33 121.92 142.21 142.21 121.928 121.92 142.21 142.21 189.28 312.43 142.21 87.33 87.33 121.92 59.99

0 69.5 75.4 75.4 69.5 69.5 75.4 75.4 86.4 110.5 75.4 0 0 69.5 0

available in the company documentation. Of relevance here is that there is a minimum wind speed before the turbine can produce power i.e. 3 m/s furthermore the maximum velocity of the wind which can be used is 23 m/s. Using the equation [25], P (W) = Cp 0.5  A v3 , the resulting power can be calculated for the wind turbine. This power has to be corrected using the Betz Limit [25], which only allows a 30% of the power to be actually used, where P is the maximum power of the turbine (W), Cp is the dimensionless power coefficient (0.5),  is the density of air (1.1 kg/m3 ), A is the area of the rotor (0.25 ␲• D2 ) where D is the diameter of the turbine (80 m) and v is the velocity of the wind (m/s). The values of the actual calculated power and the current that the induction unit uses are presented in Table 1, while Fig. 4 shows the actual calculated power over the same time period. The methane conversion profile over this time span is shown in Fig. 5 in which the current was varied as specified in Table 1. Clearly due to the difference of scales (full scale turbine vs laboratory reactor) a scaling factor is required. For this reason a value 1000 times lower than the turbine power was taken. For example during the first 30 min of the experiment the power from the turbine is 12.95 kW, however a value of 12.95 W was used in the induction heating system. Using a MS in this case, mass 13 was followed, from where the conversion of methane was calculated for the catalytic experiment. The red points ( ) indicate a stop in the experiment, which corresponds to a situation when the wind speed was considered to be too low to keep the wind turbine running. As it is observed in Fig. 5, the conversion of methane is directly dependant on the current applied to the system. A higher current also means a higher temperature which results in a higher equilibrium conversion. The thermodynamic equilibrium of the dry reforming of methane reaction was calculated using the Model Analysis Tool in Aspen Plus® (not shown here) which shows that conversion of methane increases when temperature increases. Furthermore the conversion of carbon dioxide also increases with temperature and it is always slightly higher than methane conversion because of the effect of the RWGS reaction which consumes carbon dioxide [19]. The H2 /CO ratio is also enhanced with increasing temperature and finally reaches a value of unity after 900 ◦ C. Over the 12 h of experiment considered, the induction heating system was on during 10.5 h as the wind was below the required range for the remaining 1.5 h. From the data it is evident that during 5 h the system worked at 69.5 A which gave a conversion of methane of ∼20% corresponding to a temperature of the bed of catalyst of ∼550 ◦ C. This conversion is not very high but it has to be noted that the equilibrium conversion under these conditions is

5

Table 2 Summary of methane and carbon dioxide conversions and H2 /CO ratio for the catalytic experiment of dry reforming of methane steady-state. Actual P (kW)

Current (A)

bed

Tcatalytic (◦ C)

CH4 conversion

CO2 conversion

H2 /CO ratio

121.92 142.21 189.28 245.74 312.43

69.5 75.4 86.4 98.1 110.5

557 624 707 785 863

20% 30% 60% 80% 90%

25% 40% 70% 85% 94%

0.48 0.50 0.90 0.95 0.98

28%. For 3 h the current applied was 75.4 A resulting in an average conversion of methane of ∼30% (equilibrium in this case is 49%). The system was held at 86.4 A for another 1.5 h giving a ∼60% of average methane conversion (74.5% in the equilibrium) while for half an hour the system was under 98.1 A which provided a conversion of ∼80% (while equilibrium is 88%). Conversion of methane reaches a maximum of ∼90% for the highest value of current applied, this is 110.5 A which corresponds to a temperature of around 860 ◦ C in the bed of catalyst and this value is very close to the equilibrium (94.5%) again showing that the system performance is similar to the traditional thermal system with the advantages already mentioned. Similar experiments using a GC and steady state temperature conditions revealed the H2 /CO ratio and carbon dioxide conversion shown in Table 2. Here the necessity of high response in order to analyse the product obligated the use of MS instead of GC. However the conditions which use a realistic biogas feed and high conversions obtained did not allow for recording of H2 and CO masses in the MS. As commented already the effect of RWGS over the dry reforming of methane increases the conversion of carbon dioxide which is always higher than the registered methane conversion. The H2 /CO ratios obtained are very close to equilibrium data and are almost equal to 1 for higher values of current applied. In the last two lines of Table 2, it is reported that H2 /CO ratios are very close to but below unity which can explain the slightly higher CO2 conversion when compared with CH4 conversion. The same result has been obtained in the equilibrium simulation with Aspen Plus® as shown in the supplementary information equilibrium graph (Fig. S3) and Table S1. There was no formation of any other hydrocarbon gaseous product. Hydrogen and carbon monoxide dominated although carbon formation was registered as will be discussed later. 3.3. Steam reforming of methane The same type of experiment was applied to the steam reforming of methane reaction in order to check the application of the system for a higher hydrogen production. The steam test was conducted for a shorter period of time. Fig. 6 shows the comparison profiles of methane conversion obtained. As shown in this catalytic experiment the performance of the system improves with respect to the dry case. It is clear that under steam reforming conditions, the drop and increment of conversion are less dramatic with a slower thermal response. As observed in Fig. 6, the conversion of methane is always higher than in the dry case and a maximum of 95% was reached when the current applied was 98.1 A, the maximum in this experiment. This value is close to thermodynamic equilibrium since at the temperature reached in the catalytic bed at this stage (853 ◦ C) the methane conversion is almost 100%. It should be noted that the steam reforming of methane is less endothermic than the dry reforming of methane, hence for the same current applied the conversions should be higher. Table 3 shows the comparison between the two cases.

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model

ARTICLE IN PRESS

CATTOD-9119; No. of Pages 10

M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx

6

120000

100

900 TPO dry reforming 700

Temperature profile (ºC)

70 60

DRM

50

SRM

CO2 / ppm

CH4 conversion/ %

800

TPO steam reforming

100000

80

40

600

80000

500 60000 400 300

40000

30 20

200 20000

10

100

0 0

50

100

150

200

250

300

350

TOS / min

The smoother behaviour detected in the SRM without a sharp increment or decrement of conversion can also be due to the thermal capacity of the reactants, mainly from the steam which was generated using the saturator system. The steam formed allows the conservation of some of the thermal energy and in consequence, the catalytic bed temperature. However this effect is likely to be very small. GC also results showed that the H2 /CO ratio was ∼2.8 for the highest conversion situation. It is notable that in general when using a catalyst prepared by metallic nickel dispersed on a support or in this case, a mixture of different oxides, pre-reduction is required. In the present case, an in-situ reduction under the reaction conditions was used. The feed was sent to the catalytic bed and the induction heating applied. As it can be observed in Figs. 5 and 6, after around 5 min under reaction conditions, the conversion of methane starts to increase from zero to 20% in case of DRM and 50% for SRM. These 5 min under reaction conditions and high temperature therefore served to act as the in situ reduction of the catalyst. 3.4. Temperature programme oxidation experiments TPO experiments were conducted over the perovskite catalysts previously tested in the induction heating system. The CO2 profiles evolved from both TPO experiments are shown in Fig. 7. There are two main peaks formed. The first one started at ∼380 ◦ C for both experiments (dry and steam reforming), and reached the maximum when the temperature was 560 and 610 ◦ C for steam and dry reforming experiments respectively. The second peak was obtained around 680 ◦ C for the catalyst tested under steam reforming conditions and it was completely removed before reaching 800 ◦ C. The TPO after the DRM test indicates higher carbon deposition which was eliminated after 800 ◦ C in any case. The high temperature and

Table 3 Comparison of temperatures for dry and steam reforming experiments depending of the current applied. Current (A)

Dry reforming Tcatalytic 557 624 707 785 863

bed

(◦ C)

0

0 0

50

100

150

200

250

Time / min

Fig. 6. Methane conversion profile obtained in the induction heating experiment ) or steam ( using a perovskite catalyst and several currents in the dry ( ) reforming of methane. Rest of conditions: atmospheric pressure, CH4 :CO2 : inert ratio of 9.5: 9.5: 1 for dry reforming and CH4 :H2 O:inert ratio of 10.5:8.5:1 for steam reforming (GHSV = 30,000 cm3 gcat −1 h−1 ).

69.5 75.4 86.4 98.1 110.5

Temperature / ºC

90

Steam reforming Tsurface (◦ C)

Tcatalytic

382 423 463 505 552

630 670 772 853 –

bed

(◦ C)

Tsurface (◦ C) 420 471 489 539 –

Fig. 7. CO2 profile from TPO after experiments for dry ( ) or steam ( ) reform). Conditions of TPO: 50 mL min−1 total ing of methane. Temperature profile ( flow, 20% O2 /Ar: 5% Kr/Ar: He = 2.5: 1: 1.5.

large area of the second peak suggests the formation of graphitic carbon [26]. In general literature shows that there are three different types of coke: (1) carbon, C␣ being evolved as CO2 at low temperature, 150–220 ◦ C, which corresponds to the coke deposited on the metallic sites (this type has not been found on these samples) (2) intermediate temperature carbon, C␤ which CO2 peak appears under 500–600 ◦ C, related to the coke deposited on the interface between the metallic sites and the support close to the metalsupport; and (3) a highest temperature peak of CO2 -carbon, C␥ that corresponds to graphitic coke deposited on the support which cannot be catalysed by metallic centres and it is eliminated at T > 650 ◦ C. The longer the exposure to reaction conditions produces an increment of the C␤ and C␥ species because of their relative inertness [27]. The coke found correspond to types (2) and (3) and it was completely removed from the sample after reaching 800 ◦ C in the TPO. The average coking rate was calculated from the integration of the area under the CO2 profile resulting from the TPO experiment using Origin© . The integration of the curve gives a value of ppm of CO2 ·min from which the number of mol of CO2 per weight of catalyst is calculated for a given time on stream. After the DRM experiment, the rate was 1.513 gCarbon gcat −1 or 0.136 gCarbon gcat −1 h−1 while for the SRM it was 0.506 gCarbon gcat −1 or 0.084 gCarbon gcat −1 h−1 . Clearly higher carbon formation was found on the dry reforming of methane as expected. Also it was easy to remove the carbon formed during the steam experiment. It is also worth noting that the above experiments were carried out using non-diluted feeds (apart from a Ne or Kr tracer). Recent research working with non-dilution of the feed use a lower gas hourly space velocity (GHSV) which helps lower carbon formation rates. Serrano-Lotina and Daza [28] registered a rate of 0.002 gCarbon gcat −1 h−1 for an experiment which lasted for 300 h at 750 ◦ C using CH4 :CO2 = 1:1 while their gas hourly space velocity was 4800 cm3 gcat −1 h−1 in our case was 30,000 cm3 gcat −1 h−1 . Xu et al. [29] obtained 0.0946 mgCarbon gcat −1 h−1 and Al-Fatesh et al. [30] 0.12 mgCarbon gcat −1 h−1 for 800 ◦ C and 850 ◦ C respectively. While it is not the purpose of this paper to illustrate a whole carbon study comparison, it is interesting to mention these related cases. True comparison is very difficult to do as each group worked with different reaction conditions. Literature suggests that more important than carbon/coke quantification is their location and structure in affecting the catalytic activity. The mechanism of formation is also different depending of the catalyst type [31]. At low medium temperature, thus at T < 800 ◦ C, the main source of carbon formation is the Boudouard

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model

ARTICLE IN PRESS

CATTOD-9119; No. of Pages 10

M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx

7

Table 4 Comparison results obtained in thermal-IH systems. Tbed (◦ C)

TIR (◦ C)

xCH4 (%)

xCO2 (%)

H2 /CO ratio

Rate gCarbon /gperovsk /h

Thermal IH

800 ∼800

– 470

75 76

80 83

0.96 0.95

0.4626 0.1723

Table 5 Structural properties for fresh perovskites.

800 700

La (%)

Ni (%)

Al (%)

80000

7.03 3.7

44.10 42.8

11.18 11.0

12.00 8.5

70000

1.6

90 80

600

60000

500

50000 400

40000

300

30000

70

1.2

60 50 0.8 40 30

H2/CO Ratio

CH4, CO2 conversion / %

900

90000

Na (%)

CO2 / ppm

Theoretical (from chemical equation) Catalyst prepared in QUB

100000

0.4

20 10

20000

200

10000

100

Temperature / ºC

System

0

0 0

50

100

150

200

250

300

350

400

450

Time / min Fig. 9. CO2 profile from TPO after experiments for dry reforming of methane ) or induction system ( ). Temperaperformed in the thermal system ( ture profile ( ). Conditions of TPO: 50 mL min−1 total flow, 20%O2 /Ar: 5%Kr/Ar: He = 2.5: 1: 1.5.

0

0 0

50

100

150

200

250

300

350

Time / min Fig. 8. CH4 and CO2 conversion and H2 /CO ratio for Na0.5 La0.5 Ni0.3 Al0.7 O2.5 perovskite catalyst tested in the IH system, P = 1 bar, 100 mg of catalyst, FT = 50 mL min−1 , CH4 :CO2 :Ne = 9.5: 9.5: 1, ( ) CH4 conversion, ( ) CO2 conversion and ( ) H2 /CO ratio.

reaction while at very high temperature, T > 800 ◦ C, this reaction is supressed but the methane thermolisis can be very active [32]. In practice, carbon formation regions have to be prevented because once the carbon formation is initiated, the rates can be high enough to plug the pores which leads to catalyst failure within a few hours or days [31]. 3.5. Comparison of thermally-inducted heating experiments One of the main hypotheses of the present work is that the use of induction heating provides a reduction of the carbon formation [14]. Several experiments were performed to compare the carbon formation rates in the thermal and induction heated systems for the dry reforming of methane reaction. Tests performed in both systems provided very similar conversions and H2 /CO ratios which make them suitable for the carbon rate comparison. A first experiment was performed in the thermal system at 800 ◦ C and atmospheric pressure using the perovskite-type mixed metal-oxides material. The time on stream was 6 h with CH4 :CO2 :Ne = 9.5:9.5:1 and GHSV = 30,000 cm3 gcat −1 h−1 . Results analysed by means of the gas chromatographer showed that methane and carbon dioxide conversions were 75 and 80% respectively, with an H2 /CO ratio of 0.96 as it is reported in Table 4. Fig. 8 shows the conversions of methane and carbon dioxide and H2 /CO ratios for the experiment performed in the IH system using the perovskite-type catalyst. This serves to show the stability of the IH system for the DRM reaction as well. The set point used in this case was 91.6 A of current which was found to produce the same conversion as obtained in the thermal reactor at 800 ◦ C. The temperature in the IH system was measured using both, the infrared probe for the surface temperature and the type-K thermocouple for

the catalytic bed temperature. These temperatures were 457 and 801 ◦ C respectively. After the experiment, a TPO experiment was also carried out to elucidate the carbon formation finding a rate of 0.1723 gcarbon gperovskite −1 h−1 for the inducted heated catalyst while in the thermal system it was 0.4626 gcarbon gperovskite −1 h−1 . The results are also summarised in Table 4 and the TPO profiles are shown in Fig. 9 where is clearly observed the higher amount of carbon accumulation in the traditional system which took longer to be completely oxidised. The carbon oxidized at higher temperature decreases in the IH experiment. The experimental results showed a reduction in the carbon formation of 2.7 times when using IH while literature showed a reduction of ∼ 4 times in the reforming of diesel streams [14]. The effect of the IH field (50 kHz at 92 V/cm and 350 kHz at 98 V/cm) was studied [14] on the olefin formation in the reforming of 50 ppmw sulphur diesel fuel when using a Ni-based catalyst. According to Joensen and Nostrup-Nielsen [32], olefins are precursors of the carbon formation and as reported [14] a lower tendency for its formation is obtained when IH is applied. 3.6. Characterisation of the perovskite-type catalyst The perovskite-type mixed metal-oxide catalyst Na0.5 La0.5 Ni0.3 Al0.7 O2.5 used in the dry and steam reforming experiments were also characterised before and after reaction using XRD, BET, SEM and EDX. The XRD scan was carried out between 2 equal to 15.0 and 80.0, with a time/step rate of 59.690 s. Fresh and used perovskites (after catalytic experiment plus TPO) showed a main diffraction peak in the range 30◦ < 2 < 35◦ . The trimetallic structure was not observed, instead it appears that LaNiO3 , LaAlO3 and NaNiO2 were formed. The XRD spectrums of the perovskites in the fresh and used states are shown in Fig. 10 where it can be seen that NaNiO2 (PDF 740920) was observed at 16, 27 and 40◦ values of 2. Also LaAlO3 (PDF 851071) and LaNiO3 (PDF 792448) were identified: LaAlO3 in the peaks 24, 41, 48 and 70◦ while LaNiO3 was found at 34, 54, 59 and 75◦ . The catalyst showed the same basic structure as reported in literature [13].

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model

ARTICLE IN PRESS

CATTOD-9119; No. of Pages 10

M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx

8

0

10

20

30

40

50

60

70

80

90

2θ// degrees Fig. 10. XRD patterns comparing the Na0.5 La0.5 Ni0.3 Al0.7 O2.5 catalysts: fresh state ), used in the DRM model ( ), used in the SRM model ( ), used in ( ). Species identified: 䊉 NaNiO2 PDF 740920,  LaAlO3 PDF DRM (before TPO) ( 851071,  LaNiO3 PDF 792448 and. PDF 89-8487.

Khalesi et al. [13] suggested that the absence of trimetallic structures is due to the initial presence of the nitrates and to the competition between the bimetallic LaNiO3 and LaAlO3 and the formation of the trimetallic perovskite compounds. On the other hand, they did not find specific differences between the XRD patterns before and after the reactivity tests, although in some cases the intensity of the alkali metal oxide peaks increased. In our study, this does not seem to occur for the NaNiO2 , only alkali oxide was identified, although these peaks are so small that it is very difficult to determine exactly.

The samples shown in the XRD analysis are fresh, used after DRM and SRM and TPO test and used after DRM experiment only (Fig. 10). Using the Scherrer equation the particle size of the samples was calculated. For the fresh perovskites the average particle size is 30.38 nm while for the used catalyst after TPO is 26.46 and 24.14 nm for the samples tested under DRM and SRM conditions respectively. The calculated particle size of the material after reaction (previous TPO) is 28.30 nm which is also very close to the value for the fresh state. These values have been calculated using the phase at 2 = 33◦ which corresponds to the LaNiO3 phase. The slightly smaller particle size of the used samples is attributed to experimental error. For all the used catalysts (even after TPO), the coke graphite PDF 89-8487 peak was identified in 2 equal to 44◦ although the whole structure of the catalyst seems to be kept after the reactions and/or TPO. The XRD before the TPO could not be performed for the samples of the model wind energy-DRM or SRM because the TPO was carried out in situ after the catalytic experiment and without taking the samples out of the reactor. This allowed a more accurate quantification of the carbon accumulated in each reaction. The structure of the perovskite-type precursors was only characterised after calcination. Prior to this step no precipitation or reaction between the propionic acid solvent and the nitrates of Na, La, Ni and Al were observed. Literature has shown FT-IR spectroscopy characterisation done over the same type of sample [13]. In their case, the gel produced from the individual starting salts and mixture thereof was investigated prior to calcination. The gel showed peaks at around 1433 cm−1 (bidentate propionate peak) and 1570 cm−1 (monodentate propionate peak). Nitrate peaks at ∼830 and 1380 cm−1 and a propionic acid peak at around 1468 cm−1 . The nonexistence of precipitation or reaction between

Fig. 11. SEM analysis of the Na0.5 La0.5 Ni0.3 Al0.7 O2.5 solid oxide (A) before, (B) after dry reforming and (C) after steam reforming.

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model CATTOD-9119; No. of Pages 10

ARTICLE IN PRESS M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx

the propionic acid and the starting salts demonstrates the successful production of the perovskite structure according to Khalesi et al. [13]. This type of characterisation was not done in the current work and the single perovskite structure Na0.5 La0.5 Ni0.3 Al0.7 O2.5 was not obtained. It is important to note that Na0.5 La0.5 Ni0.3 Al0.7 O2.5 represents a quantitative description of the material composition. Future work will include looking at this important fact. The current structure showed a mixture of different bimetallic oxides as commented in XRD results. Roger et al. [33] demonstrated that the use of salts (dissolved in propionic acid, as in the used method of preparation of the catalysts) is responsible for the deficiency of certain compounds in the perovskite structure, causing low stability of the samples. The alternative would be the direct use of oxides [33]. The Ni crystallite size has been reported as a determining factor for the carbon deposition and thus for the reforming activity [34]. The dissociation of CO, one of the steps of the Boudouard reaction, responsible for carbon deposition, involves Ni Ni ensembles, formed by 4–6 atoms of Ni as active sites. Also the C H bond activation of the methane molecule is considered to be the determining step on dry/steam reforming of methane. The variation of the reaction rate as a function of the particle size has to be noted, methane activation is the prototype reaction of class II type behaviour [34]. In this case, the reaction rate increases when the particle size decreases. The higher reactivity tendency must be balanced to the higher deposition of carbon in small particles leading to a compromise situation which has been reported as an optimum size of Ni particles ca. 10–20 nm [34]. The calculated sizes from the XRD analysis were higher than 20 nm as commented (∼30 nm). The metal composition of the fresh perosvkite was analysed using an EDX system linked to the SEM apparatus. The specific surface area was measured by using the BET equation [35] applied to the results from the Micromeritics ASAP 2010 instrument. The results are for powders of fresh catalysts. The catalyst prepared had a similar content in metals as the theoretical content (calculated from the chemical equation Na0.5 La0.5 Ni0.3 Al0.7 O2.5 in weight percentage) as reported in Table 5. Only the Na content was much lower than the expected value (3.7 vs. 7%). The surface area was 4.9 m2 g−1 which is lower than the value reported in literature [13], 7.2 m2 g−1 . In the specific case of dry reforming of methane, it was reported that a low surface area seems to be highly favourable because when the surface area is high it can favour the secondary reactions by allowing re-adsorption [13]. Similar perovskites catalysts have also been used for steam reforming of methane with a good performance. The SEM micrographs of the Na0.5 La0.5 Ni0.3 Al0.7 O2.5 were used to analyse the surface homogeneity of the powdered materials. The results before and after the experiments are shown in Fig. 11. It is believed that the formation of a porous structure during the calcination step is due to the decomposition of nitrates when NO2 is eliminated. This means that the nitrates present in the initial salt can be responsible for the generation of the porous structure [13]. Fig. 11(A) shows the fresh structure which is in accordance with similar materials from literature. As it is seen in Fig. 11(B) and (C), the used structure looks different, i.e. the surface has changed. Hence SEM showed sintering evidence not found with other techniques. TPOs and posterior experimental tests (not reported in this paper) showed the ability of the catalyst to recover its activity for the DRM, although sintering could not be avoided as the SEM analysis demonstrated. It was observed that the perovskite-type catalyst although very good in terms of activity (as shown in Figs. 5, 6 and 8 where conversions close to equilibrium were reported), the thermal stability was not as expected. SEM showed the sintering effect after long time on stream.

9

4. Conclusions This is the first publication dealing with reforming of biogas heated by an induction device. Within this study we have shown the potential for biogas to be combined with wind energy to act as an energy storage medium. Biogas like wind is a rapidly growing renewable energy source within Northern Ireland and Europe. Herein experiments using induction heating have been performed in order to demonstrate the potential of using rapid heating techniques (up to ∼100 ◦ C/min) to convert biogas to a higher heating value syngas when electricity from the grid is available during peaks of wind energy generation. The feasibility of the system has been demonstrated using a metallic reactor made of stainless steel. The IH system was used because of its very quick response times and low inertia which manifested its application here for the reforming of biogas using a mix of oxide type catalyst. High and stable conversions over the operating period were observed for the dry reforming of biogas case. Steam reforming experiments exhibited the same applicability for higher hydrogen production using IH and wind power with less carbon deposition and slower thermal response. Future applications of the system are being built to maximize the energy storage of the unit with improved activity and lower deactivation due to excessive carbon formation on the perovskite catalyst. Experimental results for the DRM reaction compared with the traditional thermal heated reactor have been compared in terms of carbon deposition. Lower carbon formation rates were encountered when using the IH system. The understanding of the mechanisms involved in the IH, which leads to lower carbon formation, still needs to be addressed. Although high conversions were found with the samples prepared, SEM showed the sintering of the perovskite-type catalyst. The catalyst was calcined at 725 ◦ C which is lower than some of the temperatures reached during the experimental tests. An increment of this temperature of calcination has been proposed as future work to increment the stability of the catalyst against sintering. XRD showed that the structure obtained is not a single phase perovskite but a mixture of oxides. A recommendation is the study of further methods of preparation using different precursors [34]. Acknowledgements The authors would like to thank the EPSRC for funding the work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cattod. 2014.06.010. References [1] Perspectives for Biogas in Europe. Floris van Foreest. NG 70. December 2012. The Oxford Institute for Energy Studies. [2] J. Abu-Dahrieh, D. Rooney, A. Goguet, Y. Saih, Chem. Eng. J (2012) 201. [3] J. Abu-Dahrieh, A. Orozco, E. Groom, D. Rooney, Bioresour. Technol. 102 (23) (2011) 10922. [4] Y. Asencios, C.B. Rodella, E.M. Assaf, Appl. Catal. B: Environ. 132–133 (2013) 1. [5] M. Coffey, Energy and power generation: maximising biogas yields from sludge, Filtr. Separat. 46 (1) (2009) 12. [6] B. Fidalgo, J.A. Menéndez, Fuel Process. Technol. 95 (2012) 55. [7] J.R. Rostrup-Nielsen, J. Sehested, J.K. Norskov, Adv. Catal. 47 (2002) 65. [8] M.E. Rivas, J.L.G. Fierro, M.R. Goldwasser, E. Pietri, M.J.A. Pérez-Zurita, C. Griboval, G. Leclercq, Appl. Catal. A: Gen. 344 (1–2) (2008) 10. [9] M.C.J. Bradford, M. Albert, Vannice, J. Catal. 173 (1) (1998) 157. [10] A. Slagtern, U.R. Olsbye, Blom. Appl. Catal. A 145 (1–2) (1996) 375. [11] R. Blom, I.M. Dahl, A. Slagtern, B. Sortland, A. Spjelkavik, E. Tangstad, Catal. Today 21 (2–3) (1994) 535. [12] Z. Cheng, Q. Wu, J. Li, Q.M. Zhu, Catal. Today 30 (1–3) (1996) 147.

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010

G Model CATTOD-9119; No. of Pages 10 10

ARTICLE IN PRESS M.N. Pérez-Camacho et al. / Catalysis Today xxx (2014) xxx–xxx

[13] A. Khalesi, H.R. Arandiyan, M. Parvari, Ind. Eng. Chem. Res. 47 (2008) 5892–5898. [14] G. Dieckmann Development of Ni-based sulfurresistant catalyst for diesel reforming. Office of fossil Energy Fuel Cell Program, FY2006 Annual Report 2006: 160-3. [15] A. Ovenston, J.R. Walls, J. Chem. Soc.-Faraday Trans. I 79 (1983) 1073–1084. [16] T.K. Houlding, E.V. Rebrov, Green Process Synth. 1 (2012) 19, http://dx.doi. org/10.1515/greenps-2011-0502. [17] J.P. Joly, F. Gaillard, E. Peillex, M. Romand, Vacuum 59 (2000) 854. [18] X.E. Verykios, Int. J. Hydrogen Energ. 28 (2003) 1045. [19] H. Arbag, S. Yasyerli, N. Yasyerli, G. Dogu, Int. J. Hydrogen Energ. 35 (2010) 2296. [20] M.G. Lozinskii, Industrial applications of induction heating, Pergamon, Oxford, 1969. [21] .J. Davies, P. Simpson, Induction heating handbook, McGraw-Hill, 1979. [22] Induction Heating System Topology Review. Fairchild Semiconductor. AN 9012, 2000.

[23] Y.F. Liu, Z.L. Hu, D.H. Shi, K. Yu, Int. J. Thermophys. 34 (2013) 496–506, http://dx.doi.org/10.1007/s10765-013-1421-3. [24] J. Leclercq, F. Giraud, D. Bianchi, K. Fiaty, F. Gaillard, Appl. Catal. B: Environ. 146 (2014) 131. [25] T. Ackermann, Wind Power in Power Systems, John Wiley and sons, 2005. [26] N. Martin, M. Viniegra, E. Lima, G. Espinosa, Ind. Eng. Chem. Res. 43 (2004) 1206. [27] D. Liu, X. Yang Quek, W.N.E. Cheo, R. Lau, A. Borgna, Y. Yang, J. Catal. 266 (2009) 380. [28] A. Serrano-Lotina, L. Daza, J. Power Sources 238 (2013) 81. [29] J. Xu, W. Zhou, Z. Li, J. Wang, J. Ma, Int. J. Hydrogen Energ. 34 (2009) 6646. [30] A.S.A. Al-Fatesh, A.H. Fakeeha, A.E. Abasaeed, Chin. J. Catal. 32 (2011) 1604. [31] C.H. Bartholomew, Appl. Catal. A: Gen. 212 (2001) 17–60. [32] L.A. Arkatova, Catal. Today 157 (2010) 170. [33] A.C. Roger, C. Petit, A. Kiennemann, J. Catal. 167 (1997) 447–459. ˜ [34] R. Pereniguez, V.M. Gonzalez-de la Cruz, A. Caballero, J.P. Holgado, Appl. Catal. B: Environ. 123–124 (2012) 324. [35] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309.

Please cite this article in press as: M.N. Pérez-Camacho, et al., Biogas reforming using renewable wind energy and induction heating, Catal. Today (2014), http://dx.doi.org/10.1016/j.cattod.2014.06.010