A study on the kinetics of syngas production from glycerol over alumina-supported samarium–nickel catalyst

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/he

A study on the kinetics of syngas production from glycerol over alumina-supported samariumenickel catalyst Mohd Nasir Nor Shahirah a, Sureena Abdullah a,b, Jolius Gimbun a,b, Yun Hau Ng d, Chin Kui Cheng a,b,c,* a

Faculty of Chemical & Natural Resources Engineering, Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300 Gambang, Pahang, Malaysia b Centre of Excellence for Advanced Research in Fluid Flow (CARIFF), Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300 Gambang, Pahang, Malaysia c Rare Earth Research Centre (RERC), Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300 Gambang, Pahang, Malaysia d School of Chemical Engineering, The University of New South Wales, Sydney, NSW 2052, Australia

article info

abstract

Article history:

The current paper reports on the kinetics of syngas production from glycerol pyrolysis over

Received 22 January 2016

the alumina-supported nickel catalyst that was promoted with samarium, a rare earth

Received in revised form

element. The catalysts were synthesized via wet-impregnation method and its physico-

26 April 2016

chemical properties were subsequently characterized. Reaction studies were performed in

Accepted 27 April 2016

a 10 mm-ID stainless steel fixed bed reactor with reaction temperatures maintained at 973,

Available online xxx

1023

and

1073

K,

respectively,

employing

weight-hourly-space-velocity

of

4.5
104 ml g1 h1. The textural property examination showed that BET specific surface Keywords:

area was 2.09 m2 g1 for the unpromoted catalyst while the samarium promoted catalyst

Glycerol

has 2.68 m2 g1. Interestingly, the results were supported by the FESEM images which

Pyrolysis

showed that the promoted catalyst has smaller particle size compared to the unpromoted

Nickel

catalyst. Furthermore, the NH3- and CO2-TPD analyses proved that the strong and weak

Samarium

acid-basic sites were present. During glycerol pyrolysis, the syngas was produced directly

Kinetics

from the glycerol decomposition. This has created H2:CO ratios that were always lower

Syngas

than 2.0, which is suitable for Fischer-Tropsch synthesis. The activation energy based on power law modeling for the unpromoted catalyst was 35.8 kJ mol1 and 23.4 kJ mol1 for Sm-promoted catalyst with reaction order 1.20 and 1.10, respectively. Experimental data were also fitted to the LangmuireHinshelwood model. Upon subjected to both statistical and thermodynamics consistency criteria, it can be conclusively proved that single-site mechanisms with associative adsorption of glycerol best describe the glycerol pyrolysis over both unpromoted and Sm promoted catalyst in the current work, with regression coefficient values of more than 0.9. © 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

* Corresponding author. Faculty of Chemical & Natural Resources Engineering, Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300 Gambang, Pahang, Malaysia. Tel.: þ60 9 549 2896; fax: þ60 9 549 2889. E-mail address: [email protected] (C.K. Cheng). http://dx.doi.org/10.1016/j.ijhydene.2016.04.193 0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

2

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

Introduction Glycerol is co-produced as a byproduct of biodiesel synthesis from the transesterification of vegetable oils. Traditionally, glycerol is primarily employed as an additive in sectors such as food, cosmetics, pharmaceutical, etc.; nonetheless it requires purification step [1]. Purification of crude glycerol however, needs high refining cost. From commercialization perspective, in consideration of glycerol production and utilization plays important roles with regards to the economics and also sustainability of biodiesel industry, the expansion of novel processes for glycerol valorization is absolutely required. Hence, one of the most attractive ways to increase the value of the glycerol economy is by converting glycerol into syngas (mixture of H2 and CO) [1,2]. To date, most of the works in this area has utilized catalytic methods such as steam reforming [3] and also dry reforming [4]. The reactions involved are as shown in Eqs. (1) and (2):

C3H8O3 (v) þ H2O (v) / 3CO2 (g) þ 7H2 (g)

(1)

C3H8O3 (v) þ CO2 (g) / 4CO (g) þ 3H2 (g) þ H2O (v)

(2)

Although the theoretical chemical reaction equations indicate that there is a direct interaction between H2O(steam)/ CO2 with glycerol during the reforming reaction, it is still a subject of ambiguity whether these two agents (refers to H2O and CO2) really partake in the aforementioned reactions. Interestingly, previous works by Siew et al. [5] showed that during glycerol dry reforming, the glycerol substrate did not react with CO2. From their work, it is proposed that during glycerol dry reforming, glycerol molecule would decompose as in Eq. (3) to yield CO and H2. The roles of CO2 are restricted to reverse-water-gas-shift reaction as in Eq. (4) and also to gasify deposited carbon (cf. Eq. (5)).

C3H8O3 (v) / 3CO (g) þ 4H2 (g)

(3)

CO2 (g) þ H2 (g) / CO (g) þ H2O (v)

(4)

CO2 (g) þ C (s) / 2CO (g)

(5)

Therefore, direct catalytic glycerol decomposition/pyrolysis into syngas is the subject of current investigation in order to build a clear understanding of the kinetics of catalytic glycerol decomposition/pyrolysis. The kinetics of glycerol pyrolysis was normally investigated via power law and mechanistic model fittings. The power law model was widely used because of its simplicity in application and determination. However, this model only adequate for small range of partial pressure data [6]. Based on adsorption kinetics model at constant temperature, LangmuireHinshelwood (LH) is a developed model to relate the pressure with amount of adsorbed gas [7]. Indeed, there have been various mechanistic routes proposed for reforming

reaction in order to express reaction rate and kinetics parameter, referenced herein [8e12]. In particular, Cheng et al. [10] reported that LH model based on molecular adsorption of glycerol with surface reaction as the rate-controlling step, best described syngas production from glycerol over bimetallic CoeNi supported on alumina catalyst. It was determined that reaction rate was 0.25 with respect to partial pressure of glycerol with activation energy of 63.3 kJ mol1. Significantly, gas phase decomposition of glycerol, or also known as glycerol pyrolysis has been studied before. Pyrolysis which is also known as selective thermal processing [13] can be divided into three modes viz. fast, moderate and slow pyrolysis [14,15]. In order to obtain the maximum yield of liquid products, the fast pyrolysis necessitates either rapid heating rates or short residence times [16e18]. On the other hand, slow pyrolysis or also known as conventional pyrolysis is for solid product (char) preparation. On the other hand, the glycerol pyrolysis to syngas is carried out primarily in the moderate pyrolysis mode, accompanied by residence time of 1e5 s and temperature exceeding 873 K [19,20]. The catalytic glycerol decomposition in the current work differs from the pyrolysis as it involves heterogeneous catalysts. Most importantly, this work serves as a continuation of our previous work in this area, for the further understanding of direct syngas production from glycerol. Syngas production typically requires high reaction temperature (1123e1173 K). Therefore, support needs to be stabilized to minimize on-stream sintering. In view of stabilizing alumina support, the dopants based on alkaline-earth elements and/or rare earth metals can be employed to avoid conversion to any unwanted phases influenced by high working temperatures. According to the literature, K, Mg, Sr, Ba, Ca, Zr, Pr, Ce, La and Sm as promoters, act as the thermal stabilizers of alumina with capability to improve the textural properties [21e27]. Moreover, promoters are also normally introduced to the support to improve both the metal dispersion and to inhibit coke formation. In lieu of this, the present work reports for the first time, the catalytic glycerol decomposition over the Sm-promoted Ni/Al2O3 catalyst, aiming at understanding the kinetics of direct syngas production from glycerol.

Materials and method Catalyst synthesis The alumina support was pretreated in a muffle Carbolite furnace at 1073 K for 6 h employing a heating rate of 10 K/min. Then, it was sieved to 140e425 mm particle range. For the catalyst preparation, the sieved alumina was accurately weighed and transferred into a beaker that contained calculated amount of aqueous solution of Ni/(NO3)2.6H2O and Sm(NO3)3.6H2O for the preparation of 3wt%Sm-20wt%Ni/77wt %Al2O3 catalyst. For comparison, the unpromoted 20wt%Ni/ 80wt%Al2O3 catalyst was prepared via the same procedure. During the metals impregnation, the slurry was magneticstirred for 3 h at ambient condition. Subsequently, it was oven-dried at 393 K for overnight. The dried catalyst was then air-calcined at 1073 K for 5 h with a heating rate of 10 K min1.

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

Finally, the calcined catalyst was crushed and sieved to 140e250 mm particle range for reaction studies.

Catalyst characterization Thermogravimetric analyzer or TGA (Q500-model instrument) was employed to carry out non-isothermal temperature-programmed thermal studies, in order to determine the calcination profiles of the as-synthesized (dried) catalysts. During this procedure, nickel nitrate precursor would decompose to form nickel oxide. The ramping rate used was 10, 15 and 20 K min1 under air purging. Maximum calcination temperature was fixed at 1173 K. N2 physisorption isotherms of the catalysts were performed at 77 K using Thermo Scientific Surfer. Before physisorption, the catalysts were degassed overnight at 573 K and 1
104 Torr. In addition, FESEM of JEOL/JSM-7800F model was employed to perform surface imaging of the catalysts. The image was obtained at 5 kV of voltage acceleration with 30 kX magnifications. For elemental analysis, XRF (S8 Tiger model, Germany) was employed. The instrument comes with a 4.0 kW Rhodium (Rh) X-ray tube and four analyzer crystals, whilst Quant-express was used as the best detection method. Approximately 8.0 g of samples were used for each measurement. For the X-ray diffraction (XRD) measurement, Shimadzu diffractometer model XRD-6000 employing radiation of nickel filtered CuKa with a wavelength (l) of 1.5418  A at 40 mA and 45 kV was used. The scanning was performed from 10 to 80 . The crystalline size of the catalysts was determined from the Scherrer equation: Dp ¼

0:90
l b
cosq

3

two packed columns, i.e. a Supelco Molecular Sieve 13
(10 ft
1/8 in OD
2 mm ID, 60/80 mesh, Stainless Steel) and an Agilent Hayesep DB (30 ft
1/8 in OD
2 mm ID, 100/120 mesh, Stainless Steel). He gas was used as a carrier gas.

Results and discussion Temperature programmed calcination profiles Thermal calcination profiles were carried out under air blanket to investigate weight loss associated with decomposition of nitrate-precursor of the dried catalyst specimens. Figs. 1 and 2 show weight loss profile at different heating ramps and the resulting derivative weight profiles for both dried (uncalcined) 20wt%Ni/80wt%Al2O3 and 3wt%Sm-20wt%Ni/77wt%Al2O3, respectively. It can be observed from Fig. 1 that weight losses associated with water removal from the catalysts (designated as Zone I), as well as decomposition of metal-nitrate precursor (Zone II) occur from room temperature to 600 K. Beyond 600 K, the weight of solid specimen was stable. From this nonisothermal calcination profiles, the calcination temperature of the as-prepared catalyst for the ensuing catalytic reaction

(6)

where Dp is the crystallite size, l is the wavelength of the radiation, b is a half of the maximum intensity peak and q is the half of the diffraction angle. In addition, the acidity and basicity of the catalysts were determined via temperature programmed-desorption (TPD) using NH3 as a probe molecule for acid site and CO2 for basic site. The current work employed TPDRO 1100 Version 2.3. For CO2-TPD, sample weighing approximately 0.5 g was pretreated using 30 ml min1 of N2 gas from 323 to 423 K with a ramping rate of 10 K min1, followed by 30 min holding at the intended final temperature. Then, 30 ml min1 of CO2 was metered over the sample at 423 K for 60 min for adsorption purpose. The desorption then occurred under He-blanket (flow of 30 ml min1) accompanied simultaneously by ramping up to 900 K. For NH3-TPD, NH3 adsorption was carried out at room temperature for 1 h. Subsequently, the desorption was performed in the presence of 30 ml min1 of He gas, up to 900 K.

Catalytic activity evaluation Reaction study was performed in a stainless steel fixed bed reactor. For each runs, 0.20 g of catalyst was used. The reaction temperatures were set at 973, 1023 and 1073 K. The total inlet flow rate was set at 150 ml min1 employing weight hourly space velocity (WHSV) of 4.5
104 ml g1 h1. Agilent GC Model no. 6890 series with TCD capillary column was utilized to analyze gas compositions. The GC was equipped with

Fig. 1 e Weight loss profile for (a) unpromoted 20wt%Ni/ 80wt%Al2O3 (b) 3wt%Sm-20wt%Ni/77wt%Al2O3.

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

4

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

Fig. 2 e Derivative weight profiles of the unpromoted Ni/Al2O3 and 3 wt% Sm-Ni/Al2O3 for (a) 10 K min¡1 (b) 15 K min¡1 (c) 20 K min¡1. study was fixed at 1073 K (cf. Section Catalyst synthesis) to ensure complete decomposition of nitrate and also to accommodate the highest reaction temperature employed in the current study (cf. Section Activity evaluation). In addition, the derivative weight profiles also consistently show 2 zones of peak formations, viz. a series of shorter cum broader peaks at temperatures that ranged from 373 to 513 K representing both physical and hydration water eliminations, as well as distinct sharp peaks with temperature maxima at 575 K that can be attributed to a simultaneous decomposition of nickel nitrate and samarium nitrate into metal oxide counterparts, viz. NiO and Sm2O3. This decomposition process can be written as in Eqs. (7) and (8):

Ni(NO3)2$6H2O / Ni(NO3)2 þ 6H2O [ Zone I

7(a)

Sm(NO3)3$6H2O / Sm(NO3)3 þ 6H2O [ Zone I

7(b)

Ni(NO3)2 / NiO þ 2NO2 [ þ ½O2 [ Zone II

8(a)

2Sm(NO3)3 / Sm2O3 þ 6NO2 [ þ 3/2O2 [ Zone II

8(b)

It can be observed that there is a shift of maxima temperature to the lower temperature region for the Smpromoted catalyst compared to the unpromoted 20wt%Ni/ 80wt%Al2O3 catalyst. Moreover, the decomposition pattern for all the catalysts was similar at all different ramping rates, viz. 10, 15 and 20 K min1 indicating that decomposition

characteristics was independent of ramping rate. Furthermore, the calcination profiles representing gasesolid interaction for the sharpest peak (500e600 K) were also fitted to the “model-free” Kissinger-Akahira-Sunose (KAS) kinetic expression as in Eq. (9) to obtain calcination kinetic parameters. #   b AR Ea  ln 2 ¼ ln Ta Ea RTa "

(9)

where, b ¼ heating rate (K min1), A ¼ pre-exponential factor (s1), Ea ¼ activation energy (kJ mol1), Ta ¼ apparent temperature (K), R ¼ gas constant (8.314 J mol1 K1) From the KAS model, the activation energy (Ea) for the unpromoted 20wt%Ni/80wt%Al2O3 and 3wt%Sm-20wt%Ni/ 77wt%Al2O3 were obtained (refers to Fig. 3). It can be seen that the Ea equivalent to 195.46 kJ mol1 for 3wt%Sm-20wt%Ni/ 77wt%Al2O3 was lower than the Ea for the unpromoted 20wt% Ni/80wt%Al2O3 which was 283.69 kJ mol1. Therefore, it can be deduced that the presence of promoter has managed to reduce the interaction between Ni metal and the Al2O3 support. Consequently, lower activation energy was required to decompose the nitrate precursor. As reported before, the presence of rare earth metal can promote the calcination process and that the rare earth promoted Ni/Al2O3 catalyst has consistently exhibited smaller Ea compared to the unpromoted Ni/Al2O3 catalyst [28].

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

5

Fig. 3 e Linearized Kissinger-Akahira-Sunose (KAS) data for kinetics parameter determinations of the 20wt%Ni/80wt%Al2O3 and 3wt%Sm-20wt%Ni/77wt%Al2O3.

Surface morphology and textural property examinations

structure was well-preserved before and after Sm-inclusion.

Post-calcination, the obtained catalysts were subjected to further characterization. FESEM images detailing surface morphology of the as-synthesized catalysts are displayed in Fig. 4. For the unpromoted 20wt%Ni/80wt%Al2O3 catalyst (cf. Fig. 4(a)), the particle size was noticeably different from that of the 3wt%Sm-20wt%Ni/77wt%Al2O3 catalyst (refers to Fig. 4(b)). Indeed, the unpromoted catalyst showed irregular and bulkier particle, as well as less porous compared to the Sm-promoted catalyst. Obviously, the particles were finer and surface porosity has increased, in the presence of Sm. The alteration of surface morphology can be attributed to the “spacer effect” contributed by Sm metal which can prevent active metal (Ni) and support from merging during the calcination procedure. Textural promoters such as Sm play important roles to minimize sintering and improve metal dispersion which are essential to avoid formation of larger particles. The N2 physisorption isotherms as shown in Fig. 5 confirm different textural properties for both the unpromoted 20wt%Ni/80wt% Al2O3 and 3wt%Sm-20wt%Ni/77wt%Al2O3 catalysts. The physisorption isotherm for 20wt%Ni/80wt%Al2O3 as in Fig. 5(a) indicates that it is a physisorption isotherm of type-V isotherm. The isotherms displayed a shift towards type-IV isotherm (mesoporous range) upon Sm-promotion as illustrated in Fig. 5(b). Table 1 summarizes the textural properties of the fresh calcined catalysts. The BrunauereEmmetteTeller (BET) specific surface area of the 20wt%Ni/80wt%Al2O3 catalyst was relatively low, at 2.09 m2 g1. This is consistent with the high calcination temperature employed in the current work. Upon incorporation of Sm metal onto the Ni/Al2O3 catalyst, the BET specific surface area has improved to 2.68 m2 g1. This represents a near 30.0% surface area enhancement. In addition, the average pore size of 20wt%Ni/ 80wt%Al2O3 catalyst was 7.0 nm (mesoporous). These textural properties are in sync with the FESEM images shown earlier.

X-ray diffraction Fig. 6 shows XRD patterns for both sets of catalysts. Sharp peaks can be observed, demonstrating that the crystalline

Fig. 4 e Catalysts structure for calcined (a) unpromoted 20wt%Ni/80wt%Al2O3 (b) 3 wt% Sm-20wt%Ni/77wt%Al2O3.

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

6

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

Fig. 6 e XRD pattern for the calcined 20wt%Ni/80wt%Al2O3 and 3wt%Sm-20wt%Ni/77wt%Al2O3 catalysts (- NiAl2O4, C NiO).

face atoms of the metal crystallite may have occurred by the growth of metal particle size [19,29].

Acidity and basicity properties

Fig. 5 e N2-physisorption isotherms for calcined (a) 20wt% Ni/80wt%Al2O3 and (b) 3wt%Sm-20wt%Ni/77wt%Al2O3 catalysts.

Table 1 e Textural properties for 20wt%Ni/80wt%Al2O3 and 3wt%Sm-20wt%Ni/77wt%Al2O3 catalysts. Catalyst

20wt%Ni/80wt%Al2O3 3wt%-20wt%Ni/77wt%Al2O3

Pore diameter (nm)

BET specific surface area (m2 g1)

7.2 36.7

2.09 2.68

The CO2-TPD and NH3-TPD profiles are illustrated in Fig. 7. CO2-TPD showed two distinct peaks indicative of two different basic sites viz. Peak I that is located between 398 and 428 K and Peak II at 673e803 K. The lower temperature peak is indicative of the presence of weak basic site while Peak II is probably due to a strong basic site. The presence of two distinct peaks is also observed for the NH3-TPD analysis as shown in Fig. 7, namely a weak acid in the low temperature region (533e578 K) and a high temperature peak (663e708 K) indicative of a strong acid site. These similar peaks (based on acid-basic perspective) were also reported elsewhere [9]. Significantly, the strong acid site was most likely located at the interface of metalealumina support while basic site for Ni/Al2O3 catalyst may be due to the presence of surface hydroxyl and interstitial hydroxyl species in the alumina support.

Activity evaluation In addition, it can be observed from Fig. 6 that NiO is present judging by the 2q recorded at 37.51 , 43.58 , 63.13 and 75.63 . This NiO species has also interacted with the Al2O3-support to produce spinel NiAl2O4 crystal species, with 2q at 25.84 , 35.41 , 52.79 , 57.73 , 66.72 , 68.86 and 77.65 . The Sm-metal was however, not picked up in the XRD spectra, presumably due to its well-dispersion in the catalyst matrix. The crystal structure obtained was face-centered cubic with plane reflection of (200) at peak 43.58 . Table 2 summarizes the crystallite size for both catalysts, estimated for the sharpest peak which corresponds to the NiAl2O4 species. It can be seen that the Sm-promoted catalyst possessed slightly-larger average crystal size (42.0 nm) compared to the unpromoted 20wt%Ni/80wt%Al2O3 catalyst. This indicates that Sm introduction did not alter the crystalline structure. The increase in

The reaction study was conducted at temperatures that ranged from 973 to 1073 K for 3 h, employing 0.2 g of catalysts, culminating in weight-hourly-space velocity (WHSV) of 1 4.5
104 ml g1 cat h . The catalytic performance was evaluated in terms of conversion into gaseous product selectivity and yield which are defined as:

Table 2 e Crystallite size for the unpromoted and Smpromoted Ni/Al2O3 catalysts. Catalyst 20wt%Ni/80wt%Al2O3 3wt%Sm-20wt%Ni/77wt%Al2O3

Crystallite size (nm) 39.22 42.12

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

7

monoxide (CO) and methane (CH4) as a function of glycerol partial pressure at various reaction temperatures (973, 1023 and 1073 K). It can be seen that conversions attained by Smpromoted catalyst at various partial pressures (averaging 26.0%) are higher than the unpromoted catalyst (averaging 20.0%). Consequently, the YH2 and YCO for the Sm-promoted catalyst, as can be observed in Tables 3 and 4, are also higher than the 20wt%Ni/80wt%Al2O3 catalyst. This can be ascribed to the larger BET specific surface area possessed by the promoted catalyst. In addition, the H2 yield, CO yield and glycerol conversion increased with reaction temperature for both catalysts which concur with Arrhenius trend. It also can be seen that the yield of H2 (18e28%) and yield of CO (12e23%) always occupied the major gaseous product. In contrast, the yield of CH4 only accounts for an average of 2.0%. This demonstrates that during glycerol pyrolysis, reaction as in Eq. (3) is the major reaction, while some CH4 was being produced from the methanation reaction that has consumed both the H2 and CO. In addition, the product formation rates were also calculated using Eq. (12): ri ¼

Fig. 7 e (a) NH3-TPD profiles and (b) CO2-TPD profiles.

Glycerol conversion

XG ¼

FCO2 þ FCO þ FCH4
100 3
molar flow rate of glycerol in the feed

(10)

Yield of the carbon-containing products is defined as:

Fi
100 Yi ¼ 3
molar flow rate of glycerol in the feed

(11a)

Yield of H2 is defined as:

yi
Fi Wcatalyst

(12)

where yi is the molar composition, Fi represents the exit flow rate and Wcatalyst refers to catalyst weight. The resulting profiles are presented in Fig. 8 as a function of glycerol partial pressure for both unpromoted and promoted catalysts. It can be observed that for both sets of catalysts, the formation rates of H2 and CO increase with glycerol partial pressure, proving that these two species were directly produced from the glycerol. Furthermore, the product formation rates over the 3wt% Sm-20wt%Ni/77wt%Al2O3 catalyst can be observed from Fig. 8(b) to be higher than the one over the unpromoted 20wt% Ni/80wt%Al2O3 catalyst. As aforementioned, this could be attributed to the higher BET specific surface area and better metal dispersion in the presence of Sm-promoter. For CH4, the formation rate seems nearly invariant with partial pressure of glycerol, suggesting that it was produced via secondary route, viz. methanation. In addition, the H2:CO ratios were in the range of 1.30e1.60, nearly similar to the theoretical value of 1.33 for glycerol pyrolysis.

Kinetics rate modeling Power law model was used to estimate the kinetics parameters viz. apparent activation energy and the reaction order of the glycerol pyrolysis reaction. The non-linear regression of all collected reaction rates data was performed using Polymath solver version 6.0 for the following correlation: EA

rG ¼ Ae RT PbG

(13)

whereby YH2 ¼

2
FH2
100 8
molar feed flowrate of glycerol

(11b)

where Fi is the molar flow rate of species i. Tables 3 and 4 summarize the glycerol conversion, as well as the yield of products namely hydrogen (H2), carbon

rG ¼ rate of glycerol pyrolysis (mol g1 s1), A ¼ pre-exponential factor (s1), EA ¼ activation energy (kJ mol1), R ¼ gas constant (8.314 J mol1 K1), T ¼ reaction temperature (K),

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

8

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

Table 3 e Product formation yield over glycerol partial pressure for unpromoted 20wt%Ni/80wt%Al2O3 at various temperatures and partial pressures. Pgly (kPa)

4.5 9.0 13.5 18.0 22.5

YH2 (%)

YCH4 (%)

YCO (%)

XG (%)

973 K

1023 K

1073 K

973 K

1023 K

1073 K

973 K

1023 K

1073 K

973 K

1023 K

1073 K

22.64 27.51 17.91 17.71 17.52

24.00 19.76 18.00 18.56 21.74

18.97 26.01 21.66 16.29 20.70

1.89 1.88 1.67 1.97 1.37

0.54 1.34 1.37 1.14 0.97

0.78 0.37 1.07 1.55 1.10

16.98 17.87 12.77 11.02 11.94

19.69 13.84 14.40 11.80 13.53

18.97 22.20 18.19 13.96 16.01

25.47 20.33 20.41 20.19 20.66

24.81 21.77 20.05 20.27 23.19

20.14 22.57 23.27 18.62 22.36

Table 4 e Product formation yield over glycerol partial pressure for 3wt%Sm-20wt%Ni/77wt%Al2O3 at various temperatures and partial pressures. Pgly (kPa)

4.5 9.0 13.5 18.0 22.5

YH2 (%)

YCH4 (%)

YCO (%)

XG (%)

973 K

1023 K

1073 K

973 K

1023 K

1073 K

973 K

1023 K

1073 K

973 K

1023 K

1073 K

22.33 20.59 24.49 20.33 20.44

22.53 23.11 23.79 22.47 21.11

25.74 25.97 25.79 20.05 27.43

2.29 2.29 1.67 1.63 2.52

0.43 1.19 1.35 0.62 1.73

0.23 0.46 0.88 0.74 0.61

18.32 16.02 15.07 14.92 13.12

21.46 17.78 17.55 15.32 13.39

22.88 20.78 19.40 16.33 18.90

25.77 24.02 27.00 22.77 24.22

23.18 24.89 25.81 23.39 26.70

26.09 26.66 27.11 25.16 28.35

PG ¼ reactant/glycerol partial pressure (kPa), b ¼ reaction order The estimates of the kinetics parameters for both 20wt%Ni/ 80wt%Al2O3 and 3wt%Sm-20wt%Ni/77wt%Al2O3 catalysts are shown in Table 5. The activation energy for the unpromoted catalyst was 35.8 kJ mol1 whilst for the Sm-promoted catalyst, it was 23.4 kJ mol1, demonstrating the beneficial effect of Sm as a promoter in reducing the activation energy barrier. In addition, it can be surmised too that a direct catalytic reaction from glycerol alone could reduce the activation energy compared to reforming process, i.e. 63.3 kJ mol1 [10,30] and 44.0 kJ mol1 [31]. The reaction orders were estimated at 1.1 for unpromoted catalyst and 1.2 for Sm-promoted catalyst. The near similarity in the order of reaction suggests that addition of Sm did not alter the reaction mechanism apart from just reducing the energy barrier. Siew et al. [4] reported activation energy of 35.0 kJ mol1 with the order of reaction equals 0.72 for glycerol when 3wt%LaeNi/Al2O3 catalyst was employed during glycerol dry reforming. The presence of CO2 and also different rare earth metals being included in the catalyst formulation may have explain the slight variation in the kinetics parameters compared to the current work. The LangmuireHinshelwood (LH) models were also employed in order to determine a suitable reaction mechanism to represent the catalytic glycerol pyrolysis in the current work. Eley-Rideal (ER) model was not tested as blank runs

Table 5 e Kinetics data for reaction rate of glycerol pyrolysis. Catalysts Fig. 8 e Rate of products formation versus glycerol partial pressure for (a) unpromoted 20wt%Ni/80wt%Al2O3 (b) 3wt% Sm-20wt%Ni/77wt%Al2O3 at T ¼ 1073 K and ¡1 . WHSV ¼ 4.5 £ 104 ml g¡1 cat h

20wt%Ni/80wt%Al2O3 3wt%Sm-20wt%Ni/80wt%Al2O3

A (mol m2 s1 kPab)

EA (kJ mol1)

b

0.0034829 0.0231255

35.8 23.4

1.20 1.10

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

9

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

Table 6 e LangmuireHinshelwood reaction rate models. No.

Model krxn PG ð1þKG PG Þ

1

pffiffiffiffi krxn PG ffi pffiffiffiffiffiffiffiffi ð1þ KG PG Þ

2

Table 9 e Estimates of model 1 for BMW guidelines. 1/T (K1)

Description Single site associative adsorption of glycerol (G) with molecular surface reaction as r.d.s. Single site dissociative adsorption of glycerol (G) with molecular surface reaction as r.d.s.

Unpromoted

Sm promoted

(without catalyst) did not yield significant reaction, suggesting that the current reaction was purely catalytic effects and that the gas phase reaction was practically absent. Consequently, the LH models as in Table 6 were used for the mechanistic modeling. Model 1 represents an associative adsorption of glycerol with molecular surface reaction as the rate determining step (r.d.s.) whilst the Model 2 is on dissociative glycerol adsorption. The estimates kinetics parameters for both catalysts are summarized in Table 7. It can be seen that only Model 1 shows excellent fitting with R2 > 0.98, whilst the Model 2 was not able to fit the reaction rate data. Model 1 was subjected to further evaluation to evaluate its thermodynamic consistency as proposed by Boudart [32] and Vannice [33]. To satisfy thermodynamics evaluation, the entropy and enthalpy of glycerol adsorption onto catalysts were required which was obtained from:

ln K ¼ DH/(RT) þ DS/R

0.001028 0.000978 0.000932 0.001028 0.000978 0.000932

(14)

ln krxn

EA (kJ mol1)

R2

Model 1

Model 1

Model 1

8.603 8.112 7.792 8.118 7.961 7.593

70.64

0.9915

45.24

0.9362

10  DS and DS  12.2e0.0014DH

(15)

The computed thermodynamic parameters are shown in Table 8, whilst the estimated activation energy is presented in Table 9. Significantly, Model 1 also adheres to the thermodynamics consistency guidelines; therefore the Model 1 is a robust model to that can represent the catalytic glycerol pyrolysis. Moreover, the value of activation energy from the thermodynamic evaluation was 70.64 kJ mol1 (unpromoted) and 45.24 kJ mol1 (Sm promoted). In addition, both model showed excellent fits to the data based on high R2 values (>0.9). To summarize the results, single site mechanism with associative adsorption of glycerol were involved in glycerol pyrolysis on both unpromoted and Sm promoted catalysts in this study.

where,

Conclusions

K ¼ adsorption constant (kPa1), DH ¼ enthalpy (kJ mol1), DS ¼ entropy (J mol1 K1), R ¼ gas constant (8.314 J mol1 K1), T ¼ reaction temperature (K) In order to ensure the physical meaning of KG, the value of enthalpy and entropy must satisfy with the following guidelines:

The current work reports on the catalytic glycerol pyrolysis into syngas over 3wt%Sm-promoted Ni/Al2O3 catalyst. The products were comprised of H2 and CO with ratios between them averaging 1.60, in accordance to the glycerol decomposition reaction as the main reaction pathway. A slight formation of CH4 was noticeable too, consistent with a methanation reaction as a side reaction accompanying the primary reaction. Significantly, this low H2:CO ratio is even

Table 7 e LH model parameter estimations. Model no.

1

2

Temperature (K)

Unpromoted Ni/Al2O3

SmeNi/Al2O3

krxn (mol m2 s1 kPab)

KG (kPa1)

R2

krxn (mol m2 s1 kPab)

KG (kPa1)

R2

1.835
1004 2.998
1004 4.132
1004 N/A N/A N/A

0.0240 0.0121 0.0044 N/A N/A N/A

0.9765 0.9966 0.9952 N/A N/A N/A

2.98
1004 3.488
1004 5.038
1004 N/A N/A N/A

0.0099 0.0019 0.0025 N/A N/A N/A

0.9982 0.9976 0.9988 N/A N/A N/A

973 1023 1073 973 1023 1073

Table 8 e Estimates of model 1 for BMW guidelines. Catalysts Unpromoted Sm promoted

Parameter 1

1

DSG (J mol K ) DHG (kJ mol1) DSG (J mol1 K1) DHG (kJ mol1)

Value

R2

181.21 146.67 166.29 122.08

0.9808 0.9808 0.9326 0.9326

1st guideline

2nd guideline

181.21  10

181.21  193.14

166.29  10

166.29  172.72

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193

10

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 0

suitable for Fischer-Tropsch synthesis. The activation energy based on power law rate data for unpromoted catalyst was 35.8 kJ mol1 and 23.4 kJ mol1 for Sm promoted catalyst with reaction order 1.20 and 1.10, respectively. Moreover, reaction rate data were also fitted with LangmuireHinshelwood (LH) model. The resulting modeling exercise conclusively showed that a single site associate molecular adsorption can describe the catalytic glycerol pyrolysis for both unpromoted and Sm promoted catalyst with R2 values more than 0.9.

Acknowledgments Authors would like to thank the Ministry of Education Malaysia for funding this work via RACE fund (RDU151302). In addition, Nor Shahirah Mohd Nasir acknowledges the financial assistance (MyBrain15) from the Ministry of Education Malaysia and also the Universiti Malaysia Pahang for GRS1403174.

references

[1] Vaidya PD, Rodrigues AE. Glycerol reforming for hydrogen production: a review. Chem Eng Technol 2009;32:1463e9. [2] Freitas ACD, Guirardello R. Comparison of several glycerol reforming methods for hydrogen and syngas production using Gibbs energy minimization. Int J Hydrogen Energy 2014;39(31):17969e84. [3] Cheng CK, Foo SY, Adesina AA. Steam reforming of glycerol over Ni/Al2O3 catalyst. Catal Today 2011;178(1):25e33. [4] Siew KW, Lee HC, Gimbun J, Chin SY, Khan MR, TaufiqYap YH, et al. Syngas production from glycerol-dry (CO2) reforming over La-promoted Ni/Al2O3 catalyst. Renew Energy 2015;74(3):441e7. [5] Siew KW, Lee HC, Gimbun J, Cheng CK. Production of CO-rich hydrogen gas from glycerol dry reforming over La-promoted Ni/Al2O3 catalyst. Int J Hydrogen Energy 2014;39(13):6927e36. [6] Iyer MV, Norcio LP, Kugler EL, Dadyburjor DB. Kinetic modeling for methane reforming with carbon dioxide over a mixed-metal carbide catalyst. Ind Eng Chem Res 2003;42(12):2712e21. [7] Langmuir I. The constitution of fundamental properties of solids and liquids: part I. Solids J Am Chem Soc 1916;38:2221e95. [8] Benguerba Y, Dehimi L, Virginie M, Dumas C, Ernst B. Modelling of methane dry reforming over Ni/Al2O3 catalyst in a fixed-bed catalytic reactor. React Kinet Mech Catal 2014;114:109e19. [9] Foo SY, Cheng CK, Nguyen TH, Adesina AA. Evaluation of lanthanide-group promoters on Co-Ni/Al2O3 catalysts for CH4 dry reforming. J Mol Catal A Chem 2011;344(1e2):28e36. [10] Cheng CK, Foo SY, Adesina AA. Glycerol steam reforming over bimetallic Co-Ni/Al2O3. Ind Eng Chem Res 2010;49:10804e17. [11] Pichas C, Pomonis P, Petrakis D, Ladavos A. Kinetic study of the catalytic dry reforming of CH4 with CO2 over La2xSrxNiO4 perovskite-type oxides. Appl Catal A General 2010;386:116e23.

[12] Hou K, Hughes R. The kinetics of methane steam reforming over a Ni/a-Al2O catalyst. Chem Eng J 2001;82:311e28. [13] Lin YC, Huber GW. The critical role of heterogeneous catalysis in lignocellulosic biomass conversion. Energy Environ Sci 2009;2:68e80. [14] Klass DL. Biomass for renewable energy, fuels, and chemicals. San Diego: Academic Press; 1998. [15] Zhang L, Xu C, Champagne P. Overview of recent advances in thermo-chemical conversion of biomass. Energy Convers Manag 2010;51:969e82. [16] Bridgwater AV. Principles and practice of biomass fast pyrolysis processes for liquids. J Anal Appl Pyrolysis 1999;51:3e22. [17] Bridgwater AV, Meier D, Radlein D. An overview of fast pyrolysis of biomass. Org Geochem 1999;30:1479e93. [18] Bridgwater AV. Production of high grade fuels and chemicals from catalytic pyrolysis of biomass. Catal Today 1996;29:285e95. [19] Lin YC. Catalytic valorization of glycerol to hydrogen and syngas. Int J Hydrogen Energy 2013;38(6):2678e700. [20] Valliyappan T, Bakhshi NN, Dalai AK. Pyrolysis of glycerol for the production of hydrogen or syngas. Bioresour Technol 2008;99:4476e83. [21] Navarro RM, Alvarez-galvan MC, Rosa F, Fierro JLG. Hydrogen production by oxidative reforming of hexadecane over Ni and Pt catalysts supported on Ce/La-doped Al2O3. Appl Catal A General 2006;297:60e72. [22] Schaper H, Doesburg EBM, Van Reijen LL. The influence of lanthanum oxide on the thermal stability of gamma alumina catalyst supports. Appl Catal 1983;7:211e20. [23] Choudhary VR, Uphade BS, Mamman AS. Large enhancement in methane-to-syngas conversion activity of supported Ni catalysts due to precoating of catalyst supports with MgO, CaO or rare-earth oxide. Catal Lett 1995;32(3):387e90. [24] Osaki T, Mori T. Role of potassium in carbon-free CO2 reforming of methane on K-promoted Ni/Al2O3 catalysts. J Catal 2001;204(1):89e97. [25] Parmaliana A, Arena F, Frusteri F, Coluccia S, Marchese L, Martra G, et al. Magnesia-supported nickel catalysts: II. Surface properties and reactivity in methane steam reforming. J Catal 1993;141:34e47. [26] Souza MMVM, Schmal M. Production of synthesis gas from natural gas using ZrO2-supported platinum. Stud Surf Sci Catal 2004;147:133e8. [27] Bangala DN, Abatzoglou N, Chornet E. Steam reforming of naphthalene on NieCr/Al2O3 catalysts doped with MgO, TiO2, and La2O3. J AIChE 1998;44(4):927e36. [28] Siew KW, Lee HC, Khan MR, Gimbun J, Cheng CK. CO2 reforming of glycerol over La-Ni/Al2O3 catalyst: a longevity evaluative study. J Energy Chem 2015;24(3):366e73. [29] Lehnert K, Claus P. Influence of Pt particle size and support type on the aqueous-phase reforming of glycerol. Catal Commun 2008;9:2543e6. [30] Cheng CK, Foo SY, Adesina AA. Carbon deposition on bimetallic Co-Ni/Al2O3 catalyst during steam reforming of glycerol. Catal Today 2011;164(1):268e74. [31] Praharso, Adesina AA, Trimm DL, Cant NW. Kinetic study of iso-octane steam reforming over a nickel-based catalyst. Chem Eng J 2004;99:131e6. [32] Boudart M, Mears DE, Vannice MA. Kinetics of heterogeneous catalytic reactions. Ind Chim Belge 1967;32:281. [33] Vannice MA, Yun SH, Kalpakci B, Liauh WC. Entropies of adsorption in heterogeneous catalytic reactions. J Catal 1979;56:358.

Please cite this article in press as: Nor Shahirah MN, et al., A study on the kinetics of syngas production from glycerol over aluminasupported samariumenickel catalyst, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2016.04.193