- Email: [email protected]
Micro reactor hydrogen production from ethylene glycol reforming using Rh catalysts supported on CeO2 and La2O3 promoted α-Al2O3
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Micro reactor hydrogen production from ethylene glycol reforming using Rh catalysts supported on CeO2 and La2O3 promoted a-Al2O3 U. Izquierdo a,*, M. Wichert b, G. Kolb b, V.L. Barrio a, R. Zapf b, A. Ziogas b, S. Neuberg b, P.L. Arias a, J.F. Cambra a a b
Faculty of Engineering, University of the Basque Country, Bilbao, Spain Institut fu¨r Mikrotechnik Mainz GmbH, Mainz, Germany
article info
abstract
Article history:
In this work the validation of ethylene glycol (EG) as raw material for hydrogen production
Received 30 September 2013
by steam reforming (SR) and oxidative steam reforming (OSR) reactions was studied using
Received in revised form
microchannel testing reactors. The experiments were carried out at a steam to carbon ratio
23 December 2013
(S/C) of 4.0, several temperatures and atmospheric pressure. Rh based catalysts were
Accepted 27 December 2013
designed using a-Al2O3 modified with different contents of CeO2 or La2O3 oxides as their
Available online 27 January 2014
supports. Different temperatures were tested for SR experiments (725, 675 and 625 C) at a
Keywords:
100 NL/h gcat) at a constant temperature of 675 C. In the case of OSR experiments, only the
volume hourly space velocity (VHSV) of 200 NL/h gcat as well as different VHSV (300 and Ethylene glycol
effect of the VHSV was studied at 675 C. A long term experiment was carried out with the
Cerium
2.5Rhe20Ce catalyst which lasted over 115 h in stable conditions. The catalysts physico-
Lanthanum
chemical properties were studied by their characterization using the following techniques:
Hydrogen
ICP-AES, N2 physisorption, TPR, TEM, XRD and XPS.
Rhodium
Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Micro reactor
1.
Introduction
The green hydrogen production is still a big challenge for the scientific community. Among the different possible pathways, this paper focuses on a heterogeneously catalyzed green production process, with emphasis on catalyst development. According to previous work of the authors in which different Rh catalysts supported on a-Al2O3 were prepared and tested under conditions of ethylene glycol (EG) steam and oxidative steam reforming (SR and OSR), new catalytic formulations have been designed to be tested under the same operating
conditions in order to improve their activity and stability and reduce the by-product formation. The EG is considered a renewable and available energy carrier because it is the most abundant molecule between the products obtained from the catalytic conversion of cellulose (accounting for more than 70% of cellulose derivatives [1,2]). The most important reactions that involve the EG reforming processes under investigation are the following: EG SR :
C2 H6 O2 þ 2H2 O ¼ 2CO2 þ 5H2
1
DH ¼ 91 kJ mol
* Corresponding author. Tel.: þ34 946017297. E-mail address: [email protected] (U. Izquierdo). 0360-3199/$ e see front matter Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.12.170
(1)
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
1
EGOSR : C2 H6 O2 þ1=2O2 þH2 O ¼ 2CO2 þ4H2 DH ¼ 150kJ mol
(2) EG decomposition :
C2 H6 O2 ¼ 2CO þ 3H2 DH ¼ 168 kJ mol
1
(3) r-WGS :
By-products formation :
1
DH ¼ 41 kJ mol
(4)
CO þ 3H2 ¼ CH4 þ H2 O
(5)
CO2 þ H2 ¼ CO þ H2 O
C2H6O2 þ H2 ¼ C2H4 þ 2H2O
(6)
C2H6O2 ¼ C2H4O þ H2O
(7)
C2H6O2 þ 2H2 ¼ C2H6 þ 2H2O
(8)
Apart from the reactions that produce hydrogen (1)e(3) and the reversible water-gas shift (r-WGS) reaction, the main possible routes through which different by-products can be formed are listed (5)e(8) [3]. Therefore, new catalytic formulations in order to avoid the by-product formation and the well known catalytic surface carbon deposition are looked for. With this aim, and basing on the previous experiments, 2.5 wt.% Rh on a-alumina (named 2.5Rh) was used as a reference which was modified by using CeO2 and La2O3 promoters. The used of CeO2 as additive was decided owing to the unique features attributed to this oxide when it is incorporated into the catalytic support. On one hand, under reductive atmosphere Ce4þ ions can be reduced to Ce3þ forming Ce2O3 and as a consequence, the corresponding oxygen vacancy [4]. In addition the ceria-based catalysts may enhance the active metal phase dispersion into the support surface, increasing the reforming capacity [5e7]. Furthermore, the ceria is a well-suited material to suppress coke formation in reforming reactions [8]. Regarding the use of lanthana oxide, it has also been applied to improve the resistance against coke formation of reforming catalysts [9]. Additionally, it has been reported that La2O3 reduces the acidity of the alumina, prevents metal sintering and avoids catalyst deactivation [10].
2.
Experimental procedure
2.1.
Catalysts preparation and impregnation
Rh based catalysts and supported on modified a-alumina were designed based on a common wash-coating procedure with subsequent impregnation [11]. The corresponding Rh amount (RhCl3$ H2O, Alfa Aesar, Johnson Matthey Company) to achieve the intended metal load of 2.5 wt.% was dissolved in distilled water. Then, certain amounts of a-alumina and modifiers were added in order to achieve the intended load of 10 wt.% (2.5Rhe10Ce) and 20 wt.% (2.5Rhe20Ce) of CeO2 (Ce(NO3)3*6H2O, Alfa Aesar) and 5 wt.% (2.5Rhe5La) and
5249
10 wt.% (2.5Rhe10La) of La2O3 (La(NO3)3*6H2O, FLUKA). After several minutes of subsequent stirring, the mixtures were left for three hours without stirring at room temperature. Finally, they were calcined under oxidative conditions at 450 C for 6 h. The catalysts coating was performed applying a procedure described by Zapf et al. [12]. previously, in which Polyvinyl alcohol (PVA), distilled water and acetic acid were added to the previously calcined catalysts. The resulting suspensions were coated into the micro reactors used for the activity tests. The reactors are composed of two platelets in a sandwich arrangement carrying 14 channels each with 500 mm width and 250 mm depth introduced by wet chemical etching, which have been described in previous studies [13]. After thermal pre-treatment of the platelets the previously prepared catalyst suspension was coated twice onto them followed by calcination at 600 C for 2 h after each coating. After calcinations, the platelets were attached face to face including inlet and outlet capillaries and sealed by laser welding. They were put into a stainless steel block equipped with thermocouples and heating cartridges to adjust the desired reaction temperature as described previously [13].
2.2.
Activity measurements
The activity tests were carried out in a lab scale set up. For all the experiments a steam to carbon mixture (S/C) of 4.0 was used. This mixture was placed in a tank under N2 pressure as driving force for feeding a liquid mass flow controller (LMFC), and using an evaporator, the mixture was introduced into the system. In the case of OSR experiments, a synthetic air stream consisted of 21% O2 and 79% N2 (vol.) was added by a gas MFC to the previously vaporized ethylene glycol-water mixture in order to reach an intended atomic Oxygen to Carbon (O/C) ratio of 0.15. The S/C and O/C values were chosen owing to previous, unpublished experience with polyalcohol reforming and diesel steam reforming [14] Before starting the experiment, the reactor was by-passed until a stable feed composition and the desired reactor temperature were reached. The obtained product composition was analyzed using an online Mass Spectrometer (MS, InProcess Instruments Online Mass Spectrometer 400 GAM) and an online Micro gas chromatograph (m-GC, Varian CP-4900). Only permanent gases were analyzed with the m-GC for which a cold trap was installed. All the catalysts were tested at atmospheric pressure. Different temperatures were investigated under conditions of SR experiments (725, 675 and 625 C) at a volume hourly space velocity (VHSV) of 200 L/h gcat as well as different VHSV (300 and 100 L/h gcat) at a constant temperature of 675 C. In the case of OSR experiments, only the effect of the VHSV was studied at 675 C. Each experimental test lasted 1 h. After these investigations, a long term experiment was conducted with the catalysts showing best overall selectivity and stability. The characterization techniques used, (surface area and pore volume and diameter (Brunauer Emmett Teller, BET), Temperature Programmed Reduction (TPR), Transmission Electron Microscope (TEM), X-Ray Diffraction (XRD), and X-Ray Photoelectron Spectroscopy (XPS)) were also described in previous work of the authors [15].
5250
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
The conversion of C2H6O2 and the selectivity towards different carbon containing species CiHjOk and H2 were calculated from the following equations:
SCi Hj Ok ¼ 2 i
SH2 ¼ Ph
c1;Ci Hj Ok c0;C2 H6 O2 c1;C2 H6 O2 c1;H2
j=2 c1;Ci Hj Ok
i
100½%
100½%
172
368
(9)
(10)
(11)
gas
903
145
Intensity (arb. units)
c0;C2 H6 O2 c1;C2 H6 O2 100½% c0;C2 H6 O2
XC2 H6 O2 ¼
135
175 904 102
570 903 460
885
161 94
where C0 and C1 are the concentration of the species at the reactor inlet (as determined by bypass measurements) and outlet, respectively. 100
200
300
400
500
600
700
800
900
1000
T (ºC)
3.
Results and discussion
2.5Rh
3.1.
Characterization results
Fig. 1 e TPR profiles of the calcined catalysts.
2.5Rh-10Ce
2.5Rh-20Ce
2.5Rh-5La
2.5Rh-10La
3.1.1. Catalysts textural properties and chemical compositions In Table 1 catalysts chemical composition obtained by ICPAES measurements are outlined. The measured values are in general slightly higher than intended. Textural properties of the fresh calcined catalysts are also summarized in Table 1. Taking into account that a surface area (SA) of 10.6 m2/g was measured for the bare a-alumina, this SA increased when Rh was incorporated. In addition, higher SA was measured when support modifiers were added to the corresponding catalysts. These textural property changes could be due to the interactions between Rh and modifiers with the alumina [16,17]. Comparing the catalysts containing ceria and lanthana additives, the incorporation of CeO2 contributed more than La2O3 to the surface area increase. Regarding the total pore volume (PV) with respect to the catalysts with modified support, higher total PV was measured for the CeO2 containing catalysts than for the La2O3 ones, and all of them reached higher values than 2.5Rh catalyst. This could be explained by the larger SA measured in the other catalysts prepared. The average pore size radius (PR) was measured. While pores of 11.8 nm were measured for the 2.5Rh catalyst, the CeO2 containing samples showed similar PR in the range of 10 nm, while the La2O3 containing samples showed PR in the range of 15 nm both independent of the corresponding content of CeO2 and La2O3.
3.1.2.
Temperature programmed reduction, TPR
In Fig. 1, TPR profiles of the fresh calcined catalysts are represented. For the 2.5Rh catalyst, three main peaks were observed with maximums at 94, 161 and around 570 C. The peaks at 94 and 161 C are assigned to the reduction of isolated RhOx species, corresponding to a three-dimensional RhOx phase (large particles) and two-dimensional surface RhOx phase species at low and high temperatures, respectively [17,18]. In the case of the peak at higher temperature, it has been reported that it corresponds to the reduction of the formed Rh-support strong interactions [19]. Finally, in spite of rhodium aluminates species formation being reported at very high temperatures, 800e1000 C [19], they were not detected for this catalyst. As far as the remaining catalysts are concerned, a significant change in the reducibility of the formed species was observed. For the CeO2 containing catalysts, a unique reduction peak appeared at low temperature centered at 102 C, which is wider than the peak measured for the 2.5Rh catalyst at similar temperature (see Table 2), but covered almost the same temperature range, from 60 C to 200 C, approximately. A new peak appeared at very high temperature, with maximum between 885 and 903 C. Comparing the CeO2 containing catalysts with the 2.5Rh catalyst, this peak cannot
Table 1 e Catalysts surface area, SA, pore volume, PV, pore radius, PR, and chemical compositions. Catalyst
a-Al2O3 2.5Rh 2.5Rhe10Ce 2.5Rhe20Ce 2.5Rhe5La 2.5Rhe10La
Surface area SA (m2/g)
Pore volume PV (cm3/g)
Pore radius PR ( A)
10.6 11.4 19.9 30.1 11.7 12.4
0.013 0.067 0.097 0.149 0.093 0.089
49 118 98 99 158 145
Chemical compositions (wt.%) Rh
Ce or La
e 2.6 2.3 2.5 2.6 2.2
e e 9.2 17.1 4.5 8.6
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 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Table 2 e Deconvolution of the peaks detected by TPR. Catalyst 2.5Rh
2.5Rhe10Ce
2.5Rhe20Ce
2.5Rhe5La
2.5Rhe10La
Peaks ( C)
Area
Width
95 161 550 102 460 570 885 102 460 570 903 145 175 904 135 172 903
443.4 238.8 115.0 538.7 156.8 84.2 149.3 338.2 65.9 119.4 566.4 399.4 923.9 136.7 497.2 585.7 60.1
20.5 27.2 87.5 42.7 82.7 58.3 42.0 46.7 82.3 76.0 44.9 30.2 59.6 68.5 27.1 60.9 31.1
be attributed to a rhodium aluminates species. Therefore, this peak could be assigned to surface and bulk reduction of Ce4þCe3þ species [6,20]. For both CeO2 containing catalysts, the small peak assigned to a rhodiumesupport interaction detected at intermediate temperature, between 460 and 570 C, remained (see Table 3). In the case of the La2O3 doped catalysts, similar reduction profiles as for the CeO2 doped samples were obtained, but the reduction of the RhOx species occurred at higher temperatures. In addition, the detected broad peak is clearly formed by the contribution of two smaller ones, with maxima at 135 C and 175 C. Regarding the peaks formed at intermediate temperature, (between 368 and 570 C) they were also detected for the La2O3 containing catalysts for which the highest contribution was measured at 368 C. Finally, the peaks observed at the highest temperatures are smaller, lower area, compared with the CeO2 containing samples.
3.1.3.
Transmission electron microscope TEM
In Fig. 2 the TEM micrographs for the fresh catalysts are shown. In general, the Rh particles are well dispersed and independent and no aggregates can be observed. Small particles can clearly be observed for all CeO2 and La2O3 containing samples. For the 2.5Rhe10Ce catalyst all the Rh particles observed are smaller than 3 nm and they are very well dispersed. In addition, CeO2 small particles can also be observed, which are larger, between 4 and 6 nm, than the Rh particles. In the case of the 2.5Rhe20Ce catalyst, similar Rh particles were detected. However, bigger CeO2 particles were formed in this catalyst, bigger than 5 nm, which might be due to the higher amount of additive [17]. Regarding the La2O3 containing catalysts, apart from the Rh particles no any La
5251
structure was observed for the 2.5Rhe5La catalyst. However, big crystals were detected for the 2.5Rhe10La catalyst, around 10 nm, which some authors identified possible different lanthanum oxide species [21e23]. However, as the samples were only calcined, the La2O3 structure seems to be one most likely present, despite it could not be identified. For both La2O3 containing catalysts well dispersed Rh particles with a particle sizes smaller than 4 nm were observed.
3.1.4.
X-ray diffraction XRD
XRD spectra of the fresh calcined catalysts are shown in the Fig. 3. In those spectra the a-Al2O3 was perfectly indentified (Marked by the symbol þ in the figure) by the corresponding peaks among the measured 2q positions of 34.46, 38.05, 43.62, 52.89, 57.76, 61.57, 66.74 and 68.44 (Corundum, Powder Diffraction File (PDF): 01-075-1862). In the case of Rh species, the rhodium oxide contributions were supposed to appear at least at 48.97 and 53.89 (PDF: 01- 071-2084) in the measured 2q region. However, no peaks were detected. This only could be explained due to the limitations of the technique regarding the crystal sizes and the low amount of this compound in the sample. Regarding the CeO2 (marked by the asterisk symbol in the figure), peaks were detected in both catalysts at 2q ¼ 28.76, 33.52, 47.66 and 56.65 , identified according to the 00-004-0593 PDF. Finally, in the case of the La2O3 containing catalysts, the peaks contributions should be detected at least at 2q ¼ 30.81, 40.64 and 46.36 but no peaks appeared as it also happened for the Rh crystals. Through this technique, the crystal sizes calculated according to the Scherrer equation were calculated. For the CeO2 containing catalysts, these results agreed with the ones obtained by TEM; even though the minimum of 200 TEM images would be needed in order to obtain a good particle size distribution, similar conclusions can be obtained; smaller CeO2 particles were detected for the 2.5Rhe10Ce catalyst than for the 2.5Rhe20Ce one (4 and 7 nm respectively). Regarding the Rh, particles were detected by TEM but not crystals by XRD which could be due to the small size and the concentration present in the catalysts. Regarding the lanthanum oxide especies, as big particles were observed by TEM, around 10 nm, the reason for which they were not detected by XRD could be related with the amorphous structure formed.
3.1.5.
XPS
XPS results of the fresh calcined catalysts summarized in Table 4 have been analyzed in order to determine the chemical composition of the catalyst surface. Comparing the obtained results, higher metallic Rh was detected for the 2.5Rh catalyst but the total Rh3d proportion was the lowest. This means that the Rh could be preferentially located in the pores. In the case of the CeO2 containing catalysts, the metallic Rh
Table 3 e Catalysts particle sizes measured by TEM and crystal sizes measured by XRD. Technique TEM
XPS
Specie
2.5Rhe10Ce
2.5Rhe20Ce
2.5Rhe5La
2.5Rhe10La
Rh (nm) CeO2 (nm) La especie (nm) CeO2 (nm)
<4 4e6 e 4
<4 4e6 e 7
<4 e n.d. n.d.
<4 e 10 n.d.
5252
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Fig. 2 e TEM micrographs of the calcined catalysts. Scale bar [ 5 nm.
proportion was lower, but the total Rh3d proportion higher. Finally, for the La2O3 containing catalysts the total Rh3d amount increased even more, but the measured metallic Rh was the lowest one. Thus, for the sample containing additives the Rh could be preferentially located on the surface and the modifiers addition could oxidize the initial metallic Rh. In addition, these results agreed with the textural properties +: Al2O3
Intensity (arb. units)
*: CeO2
*
*
* +
+
*
3.2.
+ + +
+ +
20
30 2.5Rh
2.5Rh-10Ce
40
50 2.5Rh-20Ce
presented in Table 1 due to the larger surface areas and higher total pore volumes measured for the samples containing additives. Regarding the detected surface proportions for the additives and their atomic concentrations, the obtained results revealed that they also were preferably located on the surface of the catalysts. Finally, atomic ratios of the most relevant species were calculated in order to confirm their enrichment on the surface. As expected, higher proportions were measured by XPS, and all the calculated Ce/Al and La/Al proportions increased with the amount used for the catalysts preparation. In the case of Rh/Al proportion, it increased due to the lower Al content used in the catalysts containing modifiers (see Table 5).
+
+
60 2.5Rh-5La
Fig. 3 e XRD spectra of the Rh catalysts.
70 2.5Rh-10La
Results from activity testing
The results presented in Figs. 4 and 5 correspond to the test carried out at different temperatures and a constant VHSV of 200 LN/(gcat h). In the case of Fig. 4, SR activity results obtained for the catalysts under investigation are shown in terms of EG conversion (left) and hydrogen yield (right). Almost full conversion was achieved by all tested catalysts and only the 2.5Rhe10Ce catalyst obtained slightly lower conversion at the temperature of 725 C. Attending to the hydrogen selectivity plot, the 2.5Rh catalyst obtained lower values at the highest temperature which can be related to the by-products formation as shown in Fig. 5. As far as the remaining catalysts are
5253
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Table 4 e Fresh calcined catalysts binding energies, BE, of Al 2p, Rh 3d5/2, Ce 3d5/2, La 3d5/2 core levels, atomic concentrations, AC, and Rh and Rh3D proportions. Theoretical and fresh catalysts surface atomic ratios, AR. Catalyst
BE and AC [eV, (%)] Al 2p
2.5Rh 2.5Rhe10Ce 2.5Rhe20Ce 2.5Rhe5La 2.5Rhe10La
73.0 73.2 73.0 72.9 73.0
(33.1) (22.0) (17.8) (27.4) (21.0)
Rh 3d
BE and prop. [eV, (%)]
Ce 3d5/2
La 3d5/2
[%]
Rh
e 882.6 (10.2) 883.1 (14.9) e e
e e e 835.2 (6.7) 835.0 (11.0)
1.98 2.76 2.89 2.90 3.74
307.2 (31.2) 307.2 (8.0) 307.2 (17.5) 307.2 (4.1) 307.2 (4.0)
concerned, very similar hydrogen selectivity values were measured. The selectivity towards the main by-products formed is represented in Fig. 5. The measured by-products are CO, CO2, CH4 and CH3CHO (Fig. 5a, b, c and d respectively). The catalysts did not show any selectivity towards C2H4 or C2H6. In general, all the catalysts containing additives behaved similarly. The only difference observed between them referred to the increase CH3CHO selectivity observed for the CeO2 containing catalysts, which increased with the temperature. However, the 2.5Rh catalyst showed different results; the CO selectivity was closer to equilibrium conversion and the CO2 selectivity much lower than the equilibrium conversion. On the other hand, attending to the CH4 selectivity results (Fig. 5c), this catalyst showed very low selectivity towards CH4. This effect might increase the calculated H2 selectivity but it did not, because the high CH3CHO selectivity (See Fig. 5d) consumed the hydrogen. In Fig. 6 the results obtained during operation conditions of OSR at a reaction temperature of 675 C and different VHSV are presented in terms of the EG conversion (left) and hydrogen yield (right). As observed before, almost full ethylene glycol conversion was achieved by all tested catalysts under OSR process. Attending the existing small differences, the 2.5Rh catalyst achieved a bit lower conversion, and quite small differences can be observed among the catalysts containing additives. In addition, only the 2.5Rh catalyst was smoothly affected by the VHSV increase. In Fig. 7 a comparison between the main by-product formation at different VHSV for both SR and OSR processes are
3þ
Rh 308.5 308.7 308.9 308.9 308.9
(78.8) (91.9) (82.5) (95.9) (96.0)
Rh/Al
Ce/Al
La/Al
Theor.
Fresh
Theor.
Fresh
Theor.
Fresh
0.013 0.014 0.016 0.013 0.014
0.060 0.125 0.162 0.106 0.178
e 0.057 0.129 e e
e 0.462 0.837 e e
e e e 0.054 0.114
e e e 0.244 0.525
presented. In general, higher CH4 and CH3CHO selectivity was measured operating under SR conditions. However, although higher CH4 selectivity was measured for the 2.5Rh catalyst under OSR conditions, this catalyst showed the lowest CH4 selectivity of all samples tested. Regarding the CH3CHO selectivity, it was negligible for all catalysts except to the 2.5Rh sample. For this catalyst were obtained the highest concentrations of this by-product, especially under SR conditions at the VHSV of 200 LN/(gcat h). Thus, in spite of the small differences between CeO2 and La2O3 containing catalysts regarding the methanation reaction, higher CH3CHO selectivity was measured for the sample containing less CeO2, while the addition of lanthana suppressed the CH3CHO selectivity almost completely. In Table 4 the hydrogen production rates measured at the different conditions tested are summarized. The 2.5Rh catalyst obtained the highest hydrogen production rates at the lowest temperatures under SR conditions. This effect could be due to the higher selectivity of this catalyst to the r-WGS reaction, which it is explained with the low CO2 selectivity measured (Fig. 5b). In the case of the OSR process, the 2.5Rh catalyst reached the highest hydrogen production at low VHSV tested but the 2.5Rhe5La catalyst was the one obtaining the highest production rates at the highest VHSV. Finally, a long term experiment was conducted under SR conditions, at a VHSV of 200 LN/(gcat h) and the lowest reaction temperature of 625 C. For this experiment the 2.5Rhe20Ce catalyst was selected. The results, presented in the Fig. 8, showed complete conversion (>99.9%), and a constant and
Table 5 e Hydrogen production rates at the tested conditions. Catalyst
2.5Rh
2.5Rhe10Ce
2.5Rhe20Ce
2.5Rhe5La
2.5Rhe10La
Catalyst amount (g) 0.0171
0.0190
0.0189
0.0172
0.0173
C2H6O2 SR VHSV ¼ 200 NL/h gcat T ( C)
H2 flow (L/gcat h)
725 675 625 725 675 625 725 675 625 725 675 625 725 675 625
94.93 99.91 101.68 96.99 98.59 98.53 97.63 98.91 98.61 97.68 98.98 98.71 93.45 98.93 98.85
C2H6O2 SR T ¼ 675 C VHSV (L/gcat h)
H2 flow (L/gcat h)
300 100
150.69 50.83
300 100
148.47 49.50
300 100
148.69 49.50
300 100
148.86 49.58
300 100
148.71 49.53
C2H6O2 OSR T ¼ 675 C VHSV (L/gcat h)
H2 flow (L/gcat h)
300 200 100 300 200 100 300 200 100 300 200 100 300 200 100
146.63 119.38 48.92 105.79 87.71 48.06 141.15 90.72 48.06 148.79 95.37 42.88 139.33 85.59 49.61
5254
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Fig. 4 e Ethylene glycol conversion and hydrogen selectivity vs. reaction temperature as obtained for all catalysts under investigation under conditions of SR and VHSV of 200 L/(gcat h).
stable process during 115 h, though the hydrogen concentration declined slightly. However, in this case a low amount of C2H4 was observed (below 0.1%), which was not measured at the same operation conditions in the previous experiments. In the case of CH3CHO, the highest concentration measured by the m-GC was 0.05927% (vol.).
4.
Conclusions
The valorization of ethylene glycol as a raw material for syngas production using Rh based catalysts supported on a-Al2O3
modified with La2O3 or CeO2 oxides was the main objective of this work focusing on stable catalyst performance and suppression of the formation of by-products such as acetaldehyde. The 2.5Rhe20Ce catalyst showed the highest surface area and pore volume. In addition, the reduction peaks measured for this catalyst were the highest at the highest temperatures. This can be due to stronger metal-modifier and supportemodifier interactions. The XRD analyses revealed well dispersed rhodium oxide particles with a size lower than 4 nm for all the prepared catalysts. In the case of XPS results, they confirmed an enrichment of the active species on the catalyst
(a)
(b)
(c)
(d)
Fig. 5 e Selectivities vs. reaction temperature as obtained for the studied catalysts under conditions of SR; (a) top left: CO selectivity; (b) top right: CO2 selectivity; (c) bottom left: CH4 selectivity; (d) bottom right: CH3CHO selectivity; the values as calculated for the thermodynamic equilibrium of the reaction mixture are included at VHSV of 200 L/(gcat h).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
5255
Fig. 6 e Ethylene glycol conversion and hydrogen selectivity vs. space velocity as obtained for all catalysts under investigation under conditions of OSR.
surface although initially (after calcination) the proportion of Rh in a metallic state was lower for the catalysts with the modified support. Regarding the activity results, complete conversion and selectivities close to the equilibrium were achieved by all catalysts tested for both SR and OSR processes. In addition, C2H4 and C2H6 by-products were only detected in minor amounts during the long-term stability test and the
concentrations of the measured by-products, CH4 and CH3CHO, were very low. The 2.5Rh catalyst showed the lowest CH4 selectivity but the highest towards CH3CHO. Thus, the measured hydrogen selectivity for this catalyst was the lowest in both SR and OSR processes. The addition of La2O3 and CeO2 suppressed the formation of acetaldehyde, which is harmful for downstream processing such as water-gas shift catalysts and PEM fuel cells, while this is not the case for methane.
(a)
(b)
(c)
(d)
Fig. 7 e Selectivities at different VHSV as obtained for all catalysts at 675 C reaction temperature under conditions of SR and OSR; (a) top left: CH4 selectivity (SR); (b) top right: CH4 selectivity (OSR); (c) bottom left: CH3CHO selectivity (SR); (d) bottom right: CH3CHO selectivity (OSR).
5256
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
2.5Rh-20Ce Stability test 60
HO
50
Molar composition (%)
40
H
30 6
CO
4
CO
0,6
CH
0,4
CH
0,2
CHO CH
0,0 0
20
40
60
80
100
120
Time (h)
Fig. 8 e Obtained product composition for the 2.5Rhe20Ce catalyst stability test carried out at 625 C and VHSV of 200 L/(gcat h). However, due to a high selectivity observed for this catalyst to the r-WGS reaction, it reached the highest hydrogen production rate at low temperatures for SR. A long term experiment was conducted using the 2.5Rhe20Ce catalyst which lasted more than 115 h showing almost stable behavior.
Acknowledgments This work has been supported by the Basque Government GV/ EJ (US11/04), Saiotek (S-PE11UN064), the Ministry of Science and Innovation (OIL2H2-ENE2011-23950) and the University of the Basque Country UPV/EHU (S-PE12UN072).
references
[1] Kim HD, Kim TW, Park HJ, Jeong KE, Chae HJ, Jeong SY. Hydrogen production via the aqueous phase reforming of ethylene glycol over platinum-supported ordered mesoporous carbon catalysts: effect of structure and framework-configuration. Int J Hydrogen Energy 2012;37:12187e97. [2] Law HY, Kung MC, Kung HH. Low temperature NOx removal from diesel exhaust by coupling ethylene glycol reforming with SCR. Catal Today 2008;136:40e5. [3] Dauenhauer PJ. Millisecond autothermal catalytic reforming of carbohydrates for synthetic fuels by reactive flash volatilization. University of Minnesota; 2008. [4] Yee A, Morrison SJ, Idriss H. A study of ethanol reactions over Pt/CeO2 by temperature-programmed desorption and in situ FT-IR Spectroscopy: evidence of benzene formation. J Catal 2000;191:30e45. [5] Kim G. Ceria-promoted three-way catalysts for auto exhaust emission control. Ind Eng Chem Prod Res Dev 1982;21:267e74. [6] Yao HC, Yu Yao YF. Ceria in automotive exhaust catalysts: oxygen storage. J Catal 1984;86:254e65.
[7] Gu DM, Chu YY, Wang ZB, Jiang ZZ, Yin GP, Liu Y. Methanol oxidation on Pt/CeO2eC electrocatalyst prepared by microwave-assisted ethylene glycol process. Appl Catal B Environ 2011;102:9e18. [8] Lin KH, Tsai CY, Chang ACC. Preparation of high surface area mesoporous Ni2xCe1xO2 (x ¼ 0, 0.05, 0.13, 0.2) and its applications in methane steam reforming. Mod Res Catal 2013;2:42e9. [9] Chen Y, Xu H, Jin X, Xiong G. Integration of gasoline prereforming into autothermal reforming for hydrogen production. Catal Today 2006;116(3):334e40. [10] Iriondo A, Barrio VL, Cambra JF, Arias PL, Gu¨emez MB, Navarro RM, et al. Influence of La2O3 modified support and Ni and Pt active phases on glycerol steam reforming to produce hydrogen. Catal Commun 2009;10:1275e8. [11] Kolb G. Review: microstructured reactors for distributed and renewable production of fuels and electrical energy. Chem Eng Proc 2013;65:1e44. [12] Zapf R, Becker-Willinger C, Berresheim K, Bolz H, Gnaser H, Hessel V, et al. Detailed characterization of various porous alumina-based catalyst coatings within microchannels and their testing for methanol steam reforming. Chem Eng Res Des 2003;81:721e9. [13] Kolb G, Zapf R, Hessel V, Lo¨we H. Propane steam reforming in micro-channels-results from catalyst screening and optimization. Appl Catal A 2004;277:155e66. [14] Kolb G, Hofmann C, O’Connell M, Schu¨rer J. Micro-structured reactors for diesel steam reforming, water-gas shift and preferential oxidation in the kilowatt power range. Catal Today 2009;147:176e84. [15] Izquierdo U, Barrio VL, Bizkarra K, Gutierrez AM, Arraibi JR, Gartzia L, et al. Ni and Rh-Ni catalysts supported on zeolites L for hydrogen and syngas production by biogas reforming processes. Chem Eng J 2014;238:178e88. [16] Ghelamallah M, Granger P. Impact of barium and lanthanum incorporation to supported Pt and Rh on a-Al2O3 in the dry reforming of methane. Fuel 2012;97:269e76. [17] Ocsachoque M, Pompeo M, Gonzalez G. RheNi/CeO2eAl2O3 catalysts for methane dry reforming. Catal today 2011;172:226e31. [18] Li Y, Wang X, Xie C, Song C. Influence of ceria and nickel addition to alumina-supported Rh catalyst for propane steam reforming at low temperatures. Appl Catal A Gen 2009;357:213e22. [19] Karatzas X, Dawody J, Grant A, Svensson EE, Pettersson LJ. Zone-coated Rh-based monolithic catalyst for autothermal reforming of diesel. Appl Catal B Environ 2011;101:226e38. [20] Terribile D, Trovarelli A, Leitenburg C, Primavera A, Dolcetti G. The synthesis and characterization of mesoporous high-surface area ceria prepared using a hybrid organic/inorganic route. Catal Today 1999;47:133e40. ´ lvarez-Galva´n MC, Fierro JLG, [21] Constantinou DA, A Efstathiou AM. Low-temperature conversion of phenol into CO, CO2 and H2 by steam reforming over La-containing supported Rh catalysts. Appl Catal B Environ 2012;117118:81e95. [22] Chen H, Yu H, Peng F, Wang H, Yang J, Pan M. Efficient and stable oxidative steam reforming of ethanol for hydrogen production: effect of in situ dispersion of Ir over Ir/La2O3. J Catal 2010;269:281e90. [23] Rivas ME, Fierro JLC, Goldwasser MR, Pietri E, Pe´rez-Zurita JM, Griboval-Constant A, et al. Structural features and performance of LaNi1xRhxO3 system for the dryreforming of methane. Appl Catal A Gen 2008;344:10e9.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Micro reactor hydrogen production from ethylene glycol reforming using Rh catalysts supported on CeO2 and La2O3 promoted a-Al2O3 U. Izquierdo a,*, M. Wichert b, G. Kolb b, V.L. Barrio a, R. Zapf b, A. Ziogas b, S. Neuberg b, P.L. Arias a, J.F. Cambra a a b
Faculty of Engineering, University of the Basque Country, Bilbao, Spain Institut fu¨r Mikrotechnik Mainz GmbH, Mainz, Germany
article info
abstract
Article history:
In this work the validation of ethylene glycol (EG) as raw material for hydrogen production
Received 30 September 2013
by steam reforming (SR) and oxidative steam reforming (OSR) reactions was studied using
Received in revised form
microchannel testing reactors. The experiments were carried out at a steam to carbon ratio
23 December 2013
(S/C) of 4.0, several temperatures and atmospheric pressure. Rh based catalysts were
Accepted 27 December 2013
designed using a-Al2O3 modified with different contents of CeO2 or La2O3 oxides as their
Available online 27 January 2014
supports. Different temperatures were tested for SR experiments (725, 675 and 625 C) at a
Keywords:
100 NL/h gcat) at a constant temperature of 675 C. In the case of OSR experiments, only the
volume hourly space velocity (VHSV) of 200 NL/h gcat as well as different VHSV (300 and Ethylene glycol
effect of the VHSV was studied at 675 C. A long term experiment was carried out with the
Cerium
2.5Rhe20Ce catalyst which lasted over 115 h in stable conditions. The catalysts physico-
Lanthanum
chemical properties were studied by their characterization using the following techniques:
Hydrogen
ICP-AES, N2 physisorption, TPR, TEM, XRD and XPS.
Rhodium
Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Micro reactor
1.
Introduction
The green hydrogen production is still a big challenge for the scientific community. Among the different possible pathways, this paper focuses on a heterogeneously catalyzed green production process, with emphasis on catalyst development. According to previous work of the authors in which different Rh catalysts supported on a-Al2O3 were prepared and tested under conditions of ethylene glycol (EG) steam and oxidative steam reforming (SR and OSR), new catalytic formulations have been designed to be tested under the same operating
conditions in order to improve their activity and stability and reduce the by-product formation. The EG is considered a renewable and available energy carrier because it is the most abundant molecule between the products obtained from the catalytic conversion of cellulose (accounting for more than 70% of cellulose derivatives [1,2]). The most important reactions that involve the EG reforming processes under investigation are the following: EG SR :
C2 H6 O2 þ 2H2 O ¼ 2CO2 þ 5H2
1
DH ¼ 91 kJ mol
* Corresponding author. Tel.: þ34 946017297. E-mail address: [email protected] (U. Izquierdo). 0360-3199/$ e see front matter Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.12.170
(1)
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
1
EGOSR : C2 H6 O2 þ1=2O2 þH2 O ¼ 2CO2 þ4H2 DH ¼ 150kJ mol
(2) EG decomposition :
C2 H6 O2 ¼ 2CO þ 3H2 DH ¼ 168 kJ mol
1
(3) r-WGS :
By-products formation :
1
DH ¼ 41 kJ mol
(4)
CO þ 3H2 ¼ CH4 þ H2 O
(5)
CO2 þ H2 ¼ CO þ H2 O
C2H6O2 þ H2 ¼ C2H4 þ 2H2O
(6)
C2H6O2 ¼ C2H4O þ H2O
(7)
C2H6O2 þ 2H2 ¼ C2H6 þ 2H2O
(8)
Apart from the reactions that produce hydrogen (1)e(3) and the reversible water-gas shift (r-WGS) reaction, the main possible routes through which different by-products can be formed are listed (5)e(8) [3]. Therefore, new catalytic formulations in order to avoid the by-product formation and the well known catalytic surface carbon deposition are looked for. With this aim, and basing on the previous experiments, 2.5 wt.% Rh on a-alumina (named 2.5Rh) was used as a reference which was modified by using CeO2 and La2O3 promoters. The used of CeO2 as additive was decided owing to the unique features attributed to this oxide when it is incorporated into the catalytic support. On one hand, under reductive atmosphere Ce4þ ions can be reduced to Ce3þ forming Ce2O3 and as a consequence, the corresponding oxygen vacancy [4]. In addition the ceria-based catalysts may enhance the active metal phase dispersion into the support surface, increasing the reforming capacity [5e7]. Furthermore, the ceria is a well-suited material to suppress coke formation in reforming reactions [8]. Regarding the use of lanthana oxide, it has also been applied to improve the resistance against coke formation of reforming catalysts [9]. Additionally, it has been reported that La2O3 reduces the acidity of the alumina, prevents metal sintering and avoids catalyst deactivation [10].
2.
Experimental procedure
2.1.
Catalysts preparation and impregnation
Rh based catalysts and supported on modified a-alumina were designed based on a common wash-coating procedure with subsequent impregnation [11]. The corresponding Rh amount (RhCl3$ H2O, Alfa Aesar, Johnson Matthey Company) to achieve the intended metal load of 2.5 wt.% was dissolved in distilled water. Then, certain amounts of a-alumina and modifiers were added in order to achieve the intended load of 10 wt.% (2.5Rhe10Ce) and 20 wt.% (2.5Rhe20Ce) of CeO2 (Ce(NO3)3*6H2O, Alfa Aesar) and 5 wt.% (2.5Rhe5La) and
5249
10 wt.% (2.5Rhe10La) of La2O3 (La(NO3)3*6H2O, FLUKA). After several minutes of subsequent stirring, the mixtures were left for three hours without stirring at room temperature. Finally, they were calcined under oxidative conditions at 450 C for 6 h. The catalysts coating was performed applying a procedure described by Zapf et al. [12]. previously, in which Polyvinyl alcohol (PVA), distilled water and acetic acid were added to the previously calcined catalysts. The resulting suspensions were coated into the micro reactors used for the activity tests. The reactors are composed of two platelets in a sandwich arrangement carrying 14 channels each with 500 mm width and 250 mm depth introduced by wet chemical etching, which have been described in previous studies [13]. After thermal pre-treatment of the platelets the previously prepared catalyst suspension was coated twice onto them followed by calcination at 600 C for 2 h after each coating. After calcinations, the platelets were attached face to face including inlet and outlet capillaries and sealed by laser welding. They were put into a stainless steel block equipped with thermocouples and heating cartridges to adjust the desired reaction temperature as described previously [13].
2.2.
Activity measurements
The activity tests were carried out in a lab scale set up. For all the experiments a steam to carbon mixture (S/C) of 4.0 was used. This mixture was placed in a tank under N2 pressure as driving force for feeding a liquid mass flow controller (LMFC), and using an evaporator, the mixture was introduced into the system. In the case of OSR experiments, a synthetic air stream consisted of 21% O2 and 79% N2 (vol.) was added by a gas MFC to the previously vaporized ethylene glycol-water mixture in order to reach an intended atomic Oxygen to Carbon (O/C) ratio of 0.15. The S/C and O/C values were chosen owing to previous, unpublished experience with polyalcohol reforming and diesel steam reforming [14] Before starting the experiment, the reactor was by-passed until a stable feed composition and the desired reactor temperature were reached. The obtained product composition was analyzed using an online Mass Spectrometer (MS, InProcess Instruments Online Mass Spectrometer 400 GAM) and an online Micro gas chromatograph (m-GC, Varian CP-4900). Only permanent gases were analyzed with the m-GC for which a cold trap was installed. All the catalysts were tested at atmospheric pressure. Different temperatures were investigated under conditions of SR experiments (725, 675 and 625 C) at a volume hourly space velocity (VHSV) of 200 L/h gcat as well as different VHSV (300 and 100 L/h gcat) at a constant temperature of 675 C. In the case of OSR experiments, only the effect of the VHSV was studied at 675 C. Each experimental test lasted 1 h. After these investigations, a long term experiment was conducted with the catalysts showing best overall selectivity and stability. The characterization techniques used, (surface area and pore volume and diameter (Brunauer Emmett Teller, BET), Temperature Programmed Reduction (TPR), Transmission Electron Microscope (TEM), X-Ray Diffraction (XRD), and X-Ray Photoelectron Spectroscopy (XPS)) were also described in previous work of the authors [15].
5250
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
The conversion of C2H6O2 and the selectivity towards different carbon containing species CiHjOk and H2 were calculated from the following equations:
SCi Hj Ok ¼ 2 i
SH2 ¼ Ph
c1;Ci Hj Ok c0;C2 H6 O2 c1;C2 H6 O2 c1;H2
j=2 c1;Ci Hj Ok
i
100½%
100½%
172
368
(9)
(10)
(11)
gas
903
145
Intensity (arb. units)
c0;C2 H6 O2 c1;C2 H6 O2 100½% c0;C2 H6 O2
XC2 H6 O2 ¼
135
175 904 102
570 903 460
885
161 94
where C0 and C1 are the concentration of the species at the reactor inlet (as determined by bypass measurements) and outlet, respectively. 100
200
300
400
500
600
700
800
900
1000
T (ºC)
3.
Results and discussion
2.5Rh
3.1.
Characterization results
Fig. 1 e TPR profiles of the calcined catalysts.
2.5Rh-10Ce
2.5Rh-20Ce
2.5Rh-5La
2.5Rh-10La
3.1.1. Catalysts textural properties and chemical compositions In Table 1 catalysts chemical composition obtained by ICPAES measurements are outlined. The measured values are in general slightly higher than intended. Textural properties of the fresh calcined catalysts are also summarized in Table 1. Taking into account that a surface area (SA) of 10.6 m2/g was measured for the bare a-alumina, this SA increased when Rh was incorporated. In addition, higher SA was measured when support modifiers were added to the corresponding catalysts. These textural property changes could be due to the interactions between Rh and modifiers with the alumina [16,17]. Comparing the catalysts containing ceria and lanthana additives, the incorporation of CeO2 contributed more than La2O3 to the surface area increase. Regarding the total pore volume (PV) with respect to the catalysts with modified support, higher total PV was measured for the CeO2 containing catalysts than for the La2O3 ones, and all of them reached higher values than 2.5Rh catalyst. This could be explained by the larger SA measured in the other catalysts prepared. The average pore size radius (PR) was measured. While pores of 11.8 nm were measured for the 2.5Rh catalyst, the CeO2 containing samples showed similar PR in the range of 10 nm, while the La2O3 containing samples showed PR in the range of 15 nm both independent of the corresponding content of CeO2 and La2O3.
3.1.2.
Temperature programmed reduction, TPR
In Fig. 1, TPR profiles of the fresh calcined catalysts are represented. For the 2.5Rh catalyst, three main peaks were observed with maximums at 94, 161 and around 570 C. The peaks at 94 and 161 C are assigned to the reduction of isolated RhOx species, corresponding to a three-dimensional RhOx phase (large particles) and two-dimensional surface RhOx phase species at low and high temperatures, respectively [17,18]. In the case of the peak at higher temperature, it has been reported that it corresponds to the reduction of the formed Rh-support strong interactions [19]. Finally, in spite of rhodium aluminates species formation being reported at very high temperatures, 800e1000 C [19], they were not detected for this catalyst. As far as the remaining catalysts are concerned, a significant change in the reducibility of the formed species was observed. For the CeO2 containing catalysts, a unique reduction peak appeared at low temperature centered at 102 C, which is wider than the peak measured for the 2.5Rh catalyst at similar temperature (see Table 2), but covered almost the same temperature range, from 60 C to 200 C, approximately. A new peak appeared at very high temperature, with maximum between 885 and 903 C. Comparing the CeO2 containing catalysts with the 2.5Rh catalyst, this peak cannot
Table 1 e Catalysts surface area, SA, pore volume, PV, pore radius, PR, and chemical compositions. Catalyst
a-Al2O3 2.5Rh 2.5Rhe10Ce 2.5Rhe20Ce 2.5Rhe5La 2.5Rhe10La
Surface area SA (m2/g)
Pore volume PV (cm3/g)
Pore radius PR ( A)
10.6 11.4 19.9 30.1 11.7 12.4
0.013 0.067 0.097 0.149 0.093 0.089
49 118 98 99 158 145
Chemical compositions (wt.%) Rh
Ce or La
e 2.6 2.3 2.5 2.6 2.2
e e 9.2 17.1 4.5 8.6
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 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Table 2 e Deconvolution of the peaks detected by TPR. Catalyst 2.5Rh
2.5Rhe10Ce
2.5Rhe20Ce
2.5Rhe5La
2.5Rhe10La
Peaks ( C)
Area
Width
95 161 550 102 460 570 885 102 460 570 903 145 175 904 135 172 903
443.4 238.8 115.0 538.7 156.8 84.2 149.3 338.2 65.9 119.4 566.4 399.4 923.9 136.7 497.2 585.7 60.1
20.5 27.2 87.5 42.7 82.7 58.3 42.0 46.7 82.3 76.0 44.9 30.2 59.6 68.5 27.1 60.9 31.1
be attributed to a rhodium aluminates species. Therefore, this peak could be assigned to surface and bulk reduction of Ce4þCe3þ species [6,20]. For both CeO2 containing catalysts, the small peak assigned to a rhodiumesupport interaction detected at intermediate temperature, between 460 and 570 C, remained (see Table 3). In the case of the La2O3 doped catalysts, similar reduction profiles as for the CeO2 doped samples were obtained, but the reduction of the RhOx species occurred at higher temperatures. In addition, the detected broad peak is clearly formed by the contribution of two smaller ones, with maxima at 135 C and 175 C. Regarding the peaks formed at intermediate temperature, (between 368 and 570 C) they were also detected for the La2O3 containing catalysts for which the highest contribution was measured at 368 C. Finally, the peaks observed at the highest temperatures are smaller, lower area, compared with the CeO2 containing samples.
3.1.3.
Transmission electron microscope TEM
In Fig. 2 the TEM micrographs for the fresh catalysts are shown. In general, the Rh particles are well dispersed and independent and no aggregates can be observed. Small particles can clearly be observed for all CeO2 and La2O3 containing samples. For the 2.5Rhe10Ce catalyst all the Rh particles observed are smaller than 3 nm and they are very well dispersed. In addition, CeO2 small particles can also be observed, which are larger, between 4 and 6 nm, than the Rh particles. In the case of the 2.5Rhe20Ce catalyst, similar Rh particles were detected. However, bigger CeO2 particles were formed in this catalyst, bigger than 5 nm, which might be due to the higher amount of additive [17]. Regarding the La2O3 containing catalysts, apart from the Rh particles no any La
5251
structure was observed for the 2.5Rhe5La catalyst. However, big crystals were detected for the 2.5Rhe10La catalyst, around 10 nm, which some authors identified possible different lanthanum oxide species [21e23]. However, as the samples were only calcined, the La2O3 structure seems to be one most likely present, despite it could not be identified. For both La2O3 containing catalysts well dispersed Rh particles with a particle sizes smaller than 4 nm were observed.
3.1.4.
X-ray diffraction XRD
XRD spectra of the fresh calcined catalysts are shown in the Fig. 3. In those spectra the a-Al2O3 was perfectly indentified (Marked by the symbol þ in the figure) by the corresponding peaks among the measured 2q positions of 34.46, 38.05, 43.62, 52.89, 57.76, 61.57, 66.74 and 68.44 (Corundum, Powder Diffraction File (PDF): 01-075-1862). In the case of Rh species, the rhodium oxide contributions were supposed to appear at least at 48.97 and 53.89 (PDF: 01- 071-2084) in the measured 2q region. However, no peaks were detected. This only could be explained due to the limitations of the technique regarding the crystal sizes and the low amount of this compound in the sample. Regarding the CeO2 (marked by the asterisk symbol in the figure), peaks were detected in both catalysts at 2q ¼ 28.76, 33.52, 47.66 and 56.65 , identified according to the 00-004-0593 PDF. Finally, in the case of the La2O3 containing catalysts, the peaks contributions should be detected at least at 2q ¼ 30.81, 40.64 and 46.36 but no peaks appeared as it also happened for the Rh crystals. Through this technique, the crystal sizes calculated according to the Scherrer equation were calculated. For the CeO2 containing catalysts, these results agreed with the ones obtained by TEM; even though the minimum of 200 TEM images would be needed in order to obtain a good particle size distribution, similar conclusions can be obtained; smaller CeO2 particles were detected for the 2.5Rhe10Ce catalyst than for the 2.5Rhe20Ce one (4 and 7 nm respectively). Regarding the Rh, particles were detected by TEM but not crystals by XRD which could be due to the small size and the concentration present in the catalysts. Regarding the lanthanum oxide especies, as big particles were observed by TEM, around 10 nm, the reason for which they were not detected by XRD could be related with the amorphous structure formed.
3.1.5.
XPS
XPS results of the fresh calcined catalysts summarized in Table 4 have been analyzed in order to determine the chemical composition of the catalyst surface. Comparing the obtained results, higher metallic Rh was detected for the 2.5Rh catalyst but the total Rh3d proportion was the lowest. This means that the Rh could be preferentially located in the pores. In the case of the CeO2 containing catalysts, the metallic Rh
Table 3 e Catalysts particle sizes measured by TEM and crystal sizes measured by XRD. Technique TEM
XPS
Specie
2.5Rhe10Ce
2.5Rhe20Ce
2.5Rhe5La
2.5Rhe10La
Rh (nm) CeO2 (nm) La especie (nm) CeO2 (nm)
<4 4e6 e 4
<4 4e6 e 7
<4 e n.d. n.d.
<4 e 10 n.d.
5252
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Fig. 2 e TEM micrographs of the calcined catalysts. Scale bar [ 5 nm.
proportion was lower, but the total Rh3d proportion higher. Finally, for the La2O3 containing catalysts the total Rh3d amount increased even more, but the measured metallic Rh was the lowest one. Thus, for the sample containing additives the Rh could be preferentially located on the surface and the modifiers addition could oxidize the initial metallic Rh. In addition, these results agreed with the textural properties +: Al2O3
Intensity (arb. units)
*: CeO2
*
*
* +
+
*
3.2.
+ + +
+ +
20
30 2.5Rh
2.5Rh-10Ce
40
50 2.5Rh-20Ce
presented in Table 1 due to the larger surface areas and higher total pore volumes measured for the samples containing additives. Regarding the detected surface proportions for the additives and their atomic concentrations, the obtained results revealed that they also were preferably located on the surface of the catalysts. Finally, atomic ratios of the most relevant species were calculated in order to confirm their enrichment on the surface. As expected, higher proportions were measured by XPS, and all the calculated Ce/Al and La/Al proportions increased with the amount used for the catalysts preparation. In the case of Rh/Al proportion, it increased due to the lower Al content used in the catalysts containing modifiers (see Table 5).
+
+
60 2.5Rh-5La
Fig. 3 e XRD spectra of the Rh catalysts.
70 2.5Rh-10La
Results from activity testing
The results presented in Figs. 4 and 5 correspond to the test carried out at different temperatures and a constant VHSV of 200 LN/(gcat h). In the case of Fig. 4, SR activity results obtained for the catalysts under investigation are shown in terms of EG conversion (left) and hydrogen yield (right). Almost full conversion was achieved by all tested catalysts and only the 2.5Rhe10Ce catalyst obtained slightly lower conversion at the temperature of 725 C. Attending to the hydrogen selectivity plot, the 2.5Rh catalyst obtained lower values at the highest temperature which can be related to the by-products formation as shown in Fig. 5. As far as the remaining catalysts are
5253
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Table 4 e Fresh calcined catalysts binding energies, BE, of Al 2p, Rh 3d5/2, Ce 3d5/2, La 3d5/2 core levels, atomic concentrations, AC, and Rh and Rh3D proportions. Theoretical and fresh catalysts surface atomic ratios, AR. Catalyst
BE and AC [eV, (%)] Al 2p
2.5Rh 2.5Rhe10Ce 2.5Rhe20Ce 2.5Rhe5La 2.5Rhe10La
73.0 73.2 73.0 72.9 73.0
(33.1) (22.0) (17.8) (27.4) (21.0)
Rh 3d
BE and prop. [eV, (%)]
Ce 3d5/2
La 3d5/2
[%]
Rh
e 882.6 (10.2) 883.1 (14.9) e e
e e e 835.2 (6.7) 835.0 (11.0)
1.98 2.76 2.89 2.90 3.74
307.2 (31.2) 307.2 (8.0) 307.2 (17.5) 307.2 (4.1) 307.2 (4.0)
concerned, very similar hydrogen selectivity values were measured. The selectivity towards the main by-products formed is represented in Fig. 5. The measured by-products are CO, CO2, CH4 and CH3CHO (Fig. 5a, b, c and d respectively). The catalysts did not show any selectivity towards C2H4 or C2H6. In general, all the catalysts containing additives behaved similarly. The only difference observed between them referred to the increase CH3CHO selectivity observed for the CeO2 containing catalysts, which increased with the temperature. However, the 2.5Rh catalyst showed different results; the CO selectivity was closer to equilibrium conversion and the CO2 selectivity much lower than the equilibrium conversion. On the other hand, attending to the CH4 selectivity results (Fig. 5c), this catalyst showed very low selectivity towards CH4. This effect might increase the calculated H2 selectivity but it did not, because the high CH3CHO selectivity (See Fig. 5d) consumed the hydrogen. In Fig. 6 the results obtained during operation conditions of OSR at a reaction temperature of 675 C and different VHSV are presented in terms of the EG conversion (left) and hydrogen yield (right). As observed before, almost full ethylene glycol conversion was achieved by all tested catalysts under OSR process. Attending the existing small differences, the 2.5Rh catalyst achieved a bit lower conversion, and quite small differences can be observed among the catalysts containing additives. In addition, only the 2.5Rh catalyst was smoothly affected by the VHSV increase. In Fig. 7 a comparison between the main by-product formation at different VHSV for both SR and OSR processes are
3þ
Rh 308.5 308.7 308.9 308.9 308.9
(78.8) (91.9) (82.5) (95.9) (96.0)
Rh/Al
Ce/Al
La/Al
Theor.
Fresh
Theor.
Fresh
Theor.
Fresh
0.013 0.014 0.016 0.013 0.014
0.060 0.125 0.162 0.106 0.178
e 0.057 0.129 e e
e 0.462 0.837 e e
e e e 0.054 0.114
e e e 0.244 0.525
presented. In general, higher CH4 and CH3CHO selectivity was measured operating under SR conditions. However, although higher CH4 selectivity was measured for the 2.5Rh catalyst under OSR conditions, this catalyst showed the lowest CH4 selectivity of all samples tested. Regarding the CH3CHO selectivity, it was negligible for all catalysts except to the 2.5Rh sample. For this catalyst were obtained the highest concentrations of this by-product, especially under SR conditions at the VHSV of 200 LN/(gcat h). Thus, in spite of the small differences between CeO2 and La2O3 containing catalysts regarding the methanation reaction, higher CH3CHO selectivity was measured for the sample containing less CeO2, while the addition of lanthana suppressed the CH3CHO selectivity almost completely. In Table 4 the hydrogen production rates measured at the different conditions tested are summarized. The 2.5Rh catalyst obtained the highest hydrogen production rates at the lowest temperatures under SR conditions. This effect could be due to the higher selectivity of this catalyst to the r-WGS reaction, which it is explained with the low CO2 selectivity measured (Fig. 5b). In the case of the OSR process, the 2.5Rh catalyst reached the highest hydrogen production at low VHSV tested but the 2.5Rhe5La catalyst was the one obtaining the highest production rates at the highest VHSV. Finally, a long term experiment was conducted under SR conditions, at a VHSV of 200 LN/(gcat h) and the lowest reaction temperature of 625 C. For this experiment the 2.5Rhe20Ce catalyst was selected. The results, presented in the Fig. 8, showed complete conversion (>99.9%), and a constant and
Table 5 e Hydrogen production rates at the tested conditions. Catalyst
2.5Rh
2.5Rhe10Ce
2.5Rhe20Ce
2.5Rhe5La
2.5Rhe10La
Catalyst amount (g) 0.0171
0.0190
0.0189
0.0172
0.0173
C2H6O2 SR VHSV ¼ 200 NL/h gcat T ( C)
H2 flow (L/gcat h)
725 675 625 725 675 625 725 675 625 725 675 625 725 675 625
94.93 99.91 101.68 96.99 98.59 98.53 97.63 98.91 98.61 97.68 98.98 98.71 93.45 98.93 98.85
C2H6O2 SR T ¼ 675 C VHSV (L/gcat h)
H2 flow (L/gcat h)
300 100
150.69 50.83
300 100
148.47 49.50
300 100
148.69 49.50
300 100
148.86 49.58
300 100
148.71 49.53
C2H6O2 OSR T ¼ 675 C VHSV (L/gcat h)
H2 flow (L/gcat h)
300 200 100 300 200 100 300 200 100 300 200 100 300 200 100
146.63 119.38 48.92 105.79 87.71 48.06 141.15 90.72 48.06 148.79 95.37 42.88 139.33 85.59 49.61
5254
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
Fig. 4 e Ethylene glycol conversion and hydrogen selectivity vs. reaction temperature as obtained for all catalysts under investigation under conditions of SR and VHSV of 200 L/(gcat h).
stable process during 115 h, though the hydrogen concentration declined slightly. However, in this case a low amount of C2H4 was observed (below 0.1%), which was not measured at the same operation conditions in the previous experiments. In the case of CH3CHO, the highest concentration measured by the m-GC was 0.05927% (vol.).
4.
Conclusions
The valorization of ethylene glycol as a raw material for syngas production using Rh based catalysts supported on a-Al2O3
modified with La2O3 or CeO2 oxides was the main objective of this work focusing on stable catalyst performance and suppression of the formation of by-products such as acetaldehyde. The 2.5Rhe20Ce catalyst showed the highest surface area and pore volume. In addition, the reduction peaks measured for this catalyst were the highest at the highest temperatures. This can be due to stronger metal-modifier and supportemodifier interactions. The XRD analyses revealed well dispersed rhodium oxide particles with a size lower than 4 nm for all the prepared catalysts. In the case of XPS results, they confirmed an enrichment of the active species on the catalyst
(a)
(b)
(c)
(d)
Fig. 5 e Selectivities vs. reaction temperature as obtained for the studied catalysts under conditions of SR; (a) top left: CO selectivity; (b) top right: CO2 selectivity; (c) bottom left: CH4 selectivity; (d) bottom right: CH3CHO selectivity; the values as calculated for the thermodynamic equilibrium of the reaction mixture are included at VHSV of 200 L/(gcat h).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
5255
Fig. 6 e Ethylene glycol conversion and hydrogen selectivity vs. space velocity as obtained for all catalysts under investigation under conditions of OSR.
surface although initially (after calcination) the proportion of Rh in a metallic state was lower for the catalysts with the modified support. Regarding the activity results, complete conversion and selectivities close to the equilibrium were achieved by all catalysts tested for both SR and OSR processes. In addition, C2H4 and C2H6 by-products were only detected in minor amounts during the long-term stability test and the
concentrations of the measured by-products, CH4 and CH3CHO, were very low. The 2.5Rh catalyst showed the lowest CH4 selectivity but the highest towards CH3CHO. Thus, the measured hydrogen selectivity for this catalyst was the lowest in both SR and OSR processes. The addition of La2O3 and CeO2 suppressed the formation of acetaldehyde, which is harmful for downstream processing such as water-gas shift catalysts and PEM fuel cells, while this is not the case for methane.
(a)
(b)
(c)
(d)
Fig. 7 e Selectivities at different VHSV as obtained for all catalysts at 675 C reaction temperature under conditions of SR and OSR; (a) top left: CH4 selectivity (SR); (b) top right: CH4 selectivity (OSR); (c) bottom left: CH3CHO selectivity (SR); (d) bottom right: CH3CHO selectivity (OSR).
5256
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 5 2 4 8 e5 2 5 6
2.5Rh-20Ce Stability test 60
HO
50
Molar composition (%)
40
H
30 6
CO
4
CO
0,6
CH
0,4
CH
0,2
CHO CH
0,0 0
20
40
60
80
100
120
Time (h)
Fig. 8 e Obtained product composition for the 2.5Rhe20Ce catalyst stability test carried out at 625 C and VHSV of 200 L/(gcat h). However, due to a high selectivity observed for this catalyst to the r-WGS reaction, it reached the highest hydrogen production rate at low temperatures for SR. A long term experiment was conducted using the 2.5Rhe20Ce catalyst which lasted more than 115 h showing almost stable behavior.
Acknowledgments This work has been supported by the Basque Government GV/ EJ (US11/04), Saiotek (S-PE11UN064), the Ministry of Science and Innovation (OIL2H2-ENE2011-23950) and the University of the Basque Country UPV/EHU (S-PE12UN072).
references
[1] Kim HD, Kim TW, Park HJ, Jeong KE, Chae HJ, Jeong SY. Hydrogen production via the aqueous phase reforming of ethylene glycol over platinum-supported ordered mesoporous carbon catalysts: effect of structure and framework-configuration. Int J Hydrogen Energy 2012;37:12187e97. [2] Law HY, Kung MC, Kung HH. Low temperature NOx removal from diesel exhaust by coupling ethylene glycol reforming with SCR. Catal Today 2008;136:40e5. [3] Dauenhauer PJ. Millisecond autothermal catalytic reforming of carbohydrates for synthetic fuels by reactive flash volatilization. University of Minnesota; 2008. [4] Yee A, Morrison SJ, Idriss H. A study of ethanol reactions over Pt/CeO2 by temperature-programmed desorption and in situ FT-IR Spectroscopy: evidence of benzene formation. J Catal 2000;191:30e45. [5] Kim G. Ceria-promoted three-way catalysts for auto exhaust emission control. Ind Eng Chem Prod Res Dev 1982;21:267e74. [6] Yao HC, Yu Yao YF. Ceria in automotive exhaust catalysts: oxygen storage. J Catal 1984;86:254e65.
[7] Gu DM, Chu YY, Wang ZB, Jiang ZZ, Yin GP, Liu Y. Methanol oxidation on Pt/CeO2eC electrocatalyst prepared by microwave-assisted ethylene glycol process. Appl Catal B Environ 2011;102:9e18. [8] Lin KH, Tsai CY, Chang ACC. Preparation of high surface area mesoporous Ni2xCe1xO2 (x ¼ 0, 0.05, 0.13, 0.2) and its applications in methane steam reforming. Mod Res Catal 2013;2:42e9. [9] Chen Y, Xu H, Jin X, Xiong G. Integration of gasoline prereforming into autothermal reforming for hydrogen production. Catal Today 2006;116(3):334e40. [10] Iriondo A, Barrio VL, Cambra JF, Arias PL, Gu¨emez MB, Navarro RM, et al. Influence of La2O3 modified support and Ni and Pt active phases on glycerol steam reforming to produce hydrogen. Catal Commun 2009;10:1275e8. [11] Kolb G. Review: microstructured reactors for distributed and renewable production of fuels and electrical energy. Chem Eng Proc 2013;65:1e44. [12] Zapf R, Becker-Willinger C, Berresheim K, Bolz H, Gnaser H, Hessel V, et al. Detailed characterization of various porous alumina-based catalyst coatings within microchannels and their testing for methanol steam reforming. Chem Eng Res Des 2003;81:721e9. [13] Kolb G, Zapf R, Hessel V, Lo¨we H. Propane steam reforming in micro-channels-results from catalyst screening and optimization. Appl Catal A 2004;277:155e66. [14] Kolb G, Hofmann C, O’Connell M, Schu¨rer J. Micro-structured reactors for diesel steam reforming, water-gas shift and preferential oxidation in the kilowatt power range. Catal Today 2009;147:176e84. [15] Izquierdo U, Barrio VL, Bizkarra K, Gutierrez AM, Arraibi JR, Gartzia L, et al. Ni and Rh-Ni catalysts supported on zeolites L for hydrogen and syngas production by biogas reforming processes. Chem Eng J 2014;238:178e88. [16] Ghelamallah M, Granger P. Impact of barium and lanthanum incorporation to supported Pt and Rh on a-Al2O3 in the dry reforming of methane. Fuel 2012;97:269e76. [17] Ocsachoque M, Pompeo M, Gonzalez G. RheNi/CeO2eAl2O3 catalysts for methane dry reforming. Catal today 2011;172:226e31. [18] Li Y, Wang X, Xie C, Song C. Influence of ceria and nickel addition to alumina-supported Rh catalyst for propane steam reforming at low temperatures. Appl Catal A Gen 2009;357:213e22. [19] Karatzas X, Dawody J, Grant A, Svensson EE, Pettersson LJ. Zone-coated Rh-based monolithic catalyst for autothermal reforming of diesel. Appl Catal B Environ 2011;101:226e38. [20] Terribile D, Trovarelli A, Leitenburg C, Primavera A, Dolcetti G. The synthesis and characterization of mesoporous high-surface area ceria prepared using a hybrid organic/inorganic route. Catal Today 1999;47:133e40. ´ lvarez-Galva´n MC, Fierro JLG, [21] Constantinou DA, A Efstathiou AM. Low-temperature conversion of phenol into CO, CO2 and H2 by steam reforming over La-containing supported Rh catalysts. Appl Catal B Environ 2012;117118:81e95. [22] Chen H, Yu H, Peng F, Wang H, Yang J, Pan M. Efficient and stable oxidative steam reforming of ethanol for hydrogen production: effect of in situ dispersion of Ir over Ir/La2O3. J Catal 2010;269:281e90. [23] Rivas ME, Fierro JLC, Goldwasser MR, Pietri E, Pe´rez-Zurita JM, Griboval-Constant A, et al. Structural features and performance of LaNi1xRhxO3 system for the dryreforming of methane. Appl Catal A Gen 2008;344:10e9.