- Email: [email protected]
CuMnOx catalysts for internal reforming methanol fuel cells: Application aspects
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
CuMnOx catalysts for internal reforming methanol fuel cells: Application aspects Joan Papavasiliou a, George Avgouropoulos a,b,*, Theophilos Ioannides a,* a
Foundation for Research and Technology-Hellas (FORTH), Institute of Chemical Engineering and High Temperature Chemical Processes (ICEHT), P.O. Box 1414, GR-26504 Patras, Greece b Department of Materials Science, University of Patras, GR-26504 Rio Patras, Greece
article info
abstract
Article history:
The Internal Reforming Methanol Fuel Cell (IRMFC) incorporates a methanol reforming
Received 15 November 2011
catalyst into the anodic compartment of a high temperature, polymer electrolyte
Received in revised form
membrane fuel cell (HT-PEMFC). The present work examines application aspects of the
17 February 2012
cell, including the operation temperature, the type of catalyst pretreatment, the effect of
Accepted 20 February 2012
phosphoric acid leaching and the time-on-stream behavior of CuMnOx catalysts.
Available online 23 March 2012
Combustion-synthesized structured catalysts can efficiently operate at 200 C with a 30% decline in MeOH conversion after 350 h on methanol/water stream in the presence of the
Keywords:
high temperature polymer electrolyte membrane electrode assembly. Differences observed
High temperature PEM fuel cell
in the catalytic activity of oxidized versus prereduced samples with respect to phosphoric
Internal reforming
acid poisoning are negligible after long operation time.
Methanol
Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
Hydrogen
reserved.
Copper-manganese catalysts Phosphoric acid
1.
Introduction
The internal reforming alcohol fuel cell (IRAFC) is a type of high temperature polymer electrolyte membrane fuel cell (HTPEMFC), in which the alcohol reforming catalyst is incorporated into the anode compartment [1e3]. Alcohol (methanol, for example) gets reformed by the catalyst inside the cell (internal reforming). Integration of the reformer into the fuel cell eliminates the need for additional heat exchangers and a separate fuel processor. Especially in the case of methanol as starting fuel, WGS and PrOx reactors are not necessary since (i) high temperature operation (i.e. 180e220 C) enables utilization of lower quality reformate feeds containing up to 2e3% CO and (ii) the “waste” heat produced by the fuel cell is in-situ
utilized to drive the endothermic reforming reaction. Thus, the design of the fuel processor-fuel cell system offers room for simplification, increase of efficiency and minimization of system weight and volume. Alcohol fuels and, more specifically methanol, are attractive chemical sources of hydrogen, since they can be easily stored, transported and dispensed, and, at the same time, they can be efficiently produced from a wide variety of sources including fossil fuels, but also agricultural products and municipal waste, wood and biomass, in general [3e7]. More importantly, they can be also made through chemical recycling of carbon dioxide. Methanol has 5e7 times higher energy density than compressed H2. Moreover, it has low sulfur content and can be reformed to hydrogen rich mixtures with
* Corresponding authors. Foundation for Research and Technology-Hellas (FORTH), Institute of Chemical Engineering and High Temperature Chemical Processes (ICE-HT), P.O. Box 1414, GR-26504 Patras, Greece. Tel.: þ30 2610965268; fax: þ30 2610965223. E-mail addresses: [email protected] (G. Avgouropoulos), [email protected] (T. Ioannides). 0360-3199/$ e see front matter Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2012.02.124
16740
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
low CO byproduct formation at the temperature level of IRAFC. H2 can be catalytically produced from methanol via steam reforming (SRM: CH3OH þ H2O # CO2 þ 3H2) at relatively low temperatures (200e300 C) with high selectivity [1,4,7e10]. In the case of the integrated PEM fuel cell system, the required heat for the SRM process will be supplied by the cell itself. Commercially available copper-based catalysts, typically with composition CueZnOe(Al2O3) have been widely used for generating hydrogen from methanol [4,5,7]. Even though these catalysts are widely used in H2 plants, they have not been designed for application in small stationary or portable/ mobile fuel processors. Thus, optimization of Cu-based catalysts for the envisaged application is of primary concern. It is well established that the activity and CO selectivity of Cubased catalysts is greatly dependent on their morphology (high copper dispersion is desirable) and the redox properties of the catalyst. Taking into account the need of the presence of well dispersed metallic (reduced) Cu species as active sites for methanol adsorption and reaction, significant enhancement of copper oxide reducibility -especially under methanol/water feedse is necessary at the cell operation level of around 200 C. Additionally, improvement of the catalytic activity will allow the reforming catalyst to operate more efficiently at the target IRAFC temperatures. In addition to catalyst activity, optimization of the catalyst formulation should take into account (i) the CO selectivity, (ii) the stability/deactivation in respect to coke formation or phosphoric acid leaching from polymer electrolyte membrane and (iii) pyrophoricity parameters. Recently, a single-cell (Internal Reforming Methanol Fuel Cell, IRMFC) incorporating ADVENT TPS phosphoric-aciddoped copolymer membrane and a CuMnOx/Cu foam methanol reforming catalyst in the anode was constructed and tested at 200 C demonstrating the functionality of the unit (Fig. 1) [1e3]. In view of the IRMFC implementation, important improvements of reforming catalysts are required regarding their efficient operation at temperatures around 200 C. In order to cope with the above operating framework we have reported on the application of combustion-synthesized CuMnOx catalysts for methanol reforming process. In the present study, CuMnOx catalyst synthesis in powder and monolithic form was carried out via the combustion method. Detailed screening of the catalysts was performed under conditions relevant to the proposed application taking into account the following parameters: temperature (focus at around 200 C), type of pretreatment, effect of phosphoric acid leaching and time-on-stream. Selected reforming catalysts were introduced in the anode compartment of an IRMFC and tested for their chemical stability.
2.
Experimental
2.1.
Synthesis of methanol reforming catalysts
The ureaenitrates combustion method (Fig. 2) was used for the synthesis of CuMnOx spinel oxide catalysts [11]. Urea [CH4N2O] with manganese nitrate [Mn(NO3)2$6H2O] and copper nitrate [Cu(NO3)2$3H2O] were mixed in an alumina crucible in the appropriate molar ratios (Cu/(Cu þ Mn) ¼ 0.30,
75% excess of urea) in a minimum volume of distilled water to obtain a transparent solution. The values for copper loading and initial amount of urea were selected from our previous studies on these catalytic systems prepared with the same technique and tested for the methanol reforming process [11,12]. The mixed solutions were heated for a few minutes at 80 C, so that excess water was removed. The resulting viscous gel was introduced in an open muffle furnace, preheated at 400e500 C, in a fuming cupboard. The gel started boiling with frothing and foaming, and in a couple of minutes ignited spontaneously with rapid evolution of a large quantity of gases, yielding a foamy, voluminous powder. In order to burnoff carbon residues and form well-crystalline structure, the powders were further heated at 550 C for 1 h (CuMn-ox samples). A half portion of the prepared batch was reduced at 350 C under H2 flow (CuMn-red samples). All the obtained powders were sieved to obtain the desired size, with diameter between 90 and 180 mm. CuMnOx catalysts supported on Cu metal foam (Changsha LYRUN New Material Co., Ltd.; 30 ppi (pores per linear inch) porosity; 7 mm thickness; 4000 g m2 density) were also prepared via the in-situ combustion method [1,13] in order to test the stability behavior with reaction time in the presence or absence of an ADVENT TPS high temperature polymer electrolyte membrane electrode assembly.
2.2. Treatment of combustion-synthesized CuMnOx catalysts with phosphoric acid vapors Since the reforming catalyst should be resistant toward deactivation caused by phosphoric acid (i.e. H3PO4 leaching from membrane electrode assembly), combustion-synthesized CuMnOx samples (both oxidized and reduced) were exposed to phosphoric acid vapors in a teflon autoclave vessel (Fig. 3). 750 mg of CuMn-ox and CuMn-red samples were placed on the bottom of a pyrex fritted disc assembly and treated hydrothermally in a teflon autoclave containing 15 ml of 85% H3PO4. The vessel was heated at a rate of 2 C min1 to the desired temperature level (150 or 200 C) where it was kept for 5 h. Following this treatment, the autoclave was opened and the samples were dried at 120 C for 12 h. In order to facilitate the presentation of results, the following encoding of catalysts is used: CuMn-ox or red-P150 or 200, denotes the oxidized or reduced sample treated with H3PO4 vapors either at 150 or 200 C. For example, the CuMn-red-P200 sample was prereduced at 350 C under H2 flow and treated at 200 C with H3PO4 vapors.
2.3.
Redox catalyst properties
In-situ XRD measurements (X-ray powder diffractometer: Bruker D8 Advance using Cu Ka radiation with l ¼ 0.15418 nm) were carried out on selected CuMnOx catalysts. Temperature programmed reduction (TPR) and oxidation (TPO) experiments were performed under H2 and O2 flows. Two different sets of experiments were performed. The applied experimental conditions were the following:
(A) Isothermal reduction with Η2 at 200 C , heating 30 C / 200 C (6 C min1), 5% Η2/Ν2, 60 ml min1
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
16741
Fig. 1 e Internal reforming methanol single cell. Upper left: CuMnOx catalyst supported on Cu foam; Upper right: POCO graphite bipolar plate with triple serpentine flow field arrangement; Bottom left: ADVENT TPS MEA, reformer, graphite bipolar plates and modified anode graphite plate; Bottom right: integrated single cell. , isothermal soak, 200 C, 15 h (B) Redox cycles , heating 30 C / 500 C (6 C min1), 5% Η2/Ν2, 60 ml min1 , free cooling, switch to Ν2 , oxidation 30 C / 500 C (6 C min1), 5% Ο2/Ν2, 60 ml min1
flow rate of the reaction mixture was 70 cm3 min1 (W/ F ¼ 0.257 g s cm3). The standard reaction feed was 5% MeOH, H2O/MeOH ¼ 1.5. Product and reactant analysis was carried out by a gas chromatograph (Shimadzu GC-14B) equipped with TCD and FID and He as carrier gas. The CO selectivity was calculated as CO selectivity ¼ [CO]out/([CO]out þ [CO2]out).
2.4.
2.5. Long term testing of CuMnOx/Cu foam reformer incorporated in an IRMFC
Catalyst screening
Activity and selectivity measurements for steam reforming of methanol were carried out at atmospheric pressure in a fixedbed reactor system, which has been described previously [11]. The standard catalyst weight was 0.3 g and the standard total
An internal reforming methanol single fuel cell was prepared for long term testing of CuMnOx catalysts. The employed single cell (Fig. 1) has the following features: (i) triple
Fig. 2 e Schematic view of the combustion process.
16742
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
set at 200 C. Reactants/products analysis was done by a gas chromatograph (Agilent Technologies, 6820 GC System) equipped with TCD and FID.
3. Fig. 3 e Experimental setup for catalysts treatment with phosphoric acid vapors. serpentine flow field arrangement, (ii) POCO graphite bipolar plates, (iii) copper current collectors, (iv) 25 cm2 electrode area with a platinum loading of 1.5 mg Pt/cm2, 2 g H3PO4/g Pt, (v) copolymer TPS electrolyte membrane of 180 mm thickness and 190 wt.% H3PO4 doping level. A standard single cell was purchased from Fuel Cell Technologies Inc. and was modified in such a way to fit the foam reforming catalyst (Fig. 1). The membrane electrode assembly comprising an ADVENT TPS H3PO4-doped copolymer, as the high temperature polymer electrolyte membrane, was sandwiched between the anodic electrode (9 g of CuMnOx catalyst supported on metallic copper foam (5 cm 5 cm 0.7 cm) and placed adjacent to the Pt/C anode electrocatalyst) and Pt/C cathodic electrode. A teflon sheet was employed in the experiment carried out in the absence of MEA in order to ensure anode compartment sealing. Vaporized methanol and water mixtures (10 vol.% MeOH, 15 vol.% H2O in He; 100 cm3 min1 total flow) were supplied to the anode compartment. The cell temperature was
3.1. Standard physicochemical characterization and catalytic properties of CuMnOx catalysts The structural and morphological properties of CueMn spinel oxide catalysts were investigated and discussed in detail in previous papers [11,12]. The following conclusions were drawn from these works: Fresh CueMn catalysts are composed of the spinel phase Cu1.5Mn1.5O4, as well as Mn2O3 and CuO, depending on the Cu/Mn ratio, and get reduced to Cu0 and MnO under methanol reforming conditions. The BET specific surface areas of the samples are less than 9 m2 g1. XPS analysis revealed the presence of two different oxidation states in both copper (Cu2þ and Cuþ) and manganese (Mn4þ and Mn3þ) in fresh catalysts and decomposition of the spinel in used catalysts. XRD and TPR measurements showed that at a high copper content (x ¼ 0.40, 0.50), the excess copper is not incorporated into the spinel, but is rather present as a separate CuO phase.
CuMn-ox
12
10
CuMn-ox-P150
10
8
CuMn-ox-P200 o open symbols: 300 C to RT
6
CO selectivity, %
CO selectivity, %
12
o
solid symbols: RT to 300 C
4 2 0 180
200
220
240
260
280
300
8
open symbols: 300 C to RT
6
solid symbols: RT to 300 C
o
o
4 2
180
200
220
240
260
280
300
60 40 20
A
0 200
220
240
260 o
T, C
280
300
MeOH conversion, %
100
80
180
CuMn-red CuMn-red-P150 CuMn-red-P200
0
100
MeOH conversion, %
Results and discussion
80 60 40 20
B
0 180
200
220
240
260
280
300
o
T, C
Fig. 4 e Comparison of catalytic activity/selectivity for the steam reforming of methanol of (A) oxidized and (B) prereduced samples, untreated and treated with H3PO4 vapors. Operating conditions: W/F [ 0.257 g s cmL3, 5% CH3OH, H2O/ CH3OH [ 1.5.
16743
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
The optimal catalyst was prepared with a Cu/(Cu þ Mn) ratio of 0.30.
have comparable activity to that of a commercial Cu/ZnO/ Al2O3 catalyst. The amount of CO produced during reforming was well below wateregas shift equilibrium (CO þ H2O # CO2 þ H2) and other byproducts such as formaldehyde, formic acid, methyl formate or dimethyl ether, which are often formed by reactions of methanol over Cubased catalysts, were not detected at the temperature level
12
CuMn-ox
10
CuMn-red
8
12 o
open symbols: 300 C to RT o
solid symbols: RT to 300 C
6 4
0 220
240
260
280
180
300
200
220
240
260
280
300
100
MeOH conversion, %
80 60 40 20
80 60 40 20
A
0 180
o
solid symbols: RT to 300 C
4
0 200
o
open symbols: 300 C to RT
6
2
100
MeOH conversion, %
8
2
180
CuMn-ox-P150 CuMn-red-P150
10
CO selectivity, %
CO selectivity, %
CuMnOx materials have been proven to be very efficient methanol steam reforming catalysts, able to selectively produce CO2 and H2 at temperatures lower than 240 C [11]. Despite their low surface areas (<9 m2 g1), these catalysts
200
220
240
260
280
B
0 180
300
200
220
260
280
300
T, C
T, C 12
CO selectivity, %
240 o
o
CuMn-ox-P200 CuMn-red-P200
10 8
o
open symbols: 300 C to RT o
solid symbols: RT to 300 C
6 4 2 0 180
200
220
240
260
280
300
MeOH conversion, %
100 80 60 40 20
C
0 180
200
220
240
260
280
300
o
T, C
Fig. 5 e Effect of H3PO4 and type of pretreatment (oxidized vs. reduced samples) on CuMnOx catalysts activity and selectivity for the steam reforming of methanol. (A) no H3PO4 treatment, (B) H3PO4 treatment at 150 C, (C) H3PO4 treatment at 200 C. Operating conditions: W/F [ 0.257 g s cmL3, 5% CH3OH, H2O/CH3OH [ 1.5.
16744
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
of IRMFC, i.e. 200 C. Previous mechanistic studies over combustion-synthesized CuMnOx catalysts lead to the conclusion that the SRM reaction pathway occurs exclusively via methyl formate hydrolysis [14]: Methanol dehydrogenation to formaldehyde and H2 1a. CH3OH / HCHO þ H2 1b. CH3OH þ HCHO / HCOOCH3 þ H2 or direct dehydrogenation of methanol to methyl formate and H2 1c. 2CH3OH / HCOOCH3 þ 2H2 Hydrolysis of methyl formate 2. HCOOCH3 þ H2O / CH3OH þ HCOOH Formic acid decomposition 3. HCOOH / H2 þ CO2
3.2. Effect of phosphoric acid and type of pretreatment (oxidation vs. reduction) on combustion-synthesized CuMnOx catalysts performance The CuMnOx catalyst behavior was examined in steam reforming of methanol following no pretreatment or a reductive one (CuMn-ox vs. CuMn-red). The goal is to determine whether: (i) pre-reduction leads to more active catalysts and (ii) direct exposure to the reaction feed leads to in-situ reduction with time-on-stream. The obtained activity/selectivity curves are shown in Figs. 4 and 5. It can be observed that the oxidized sample needs high temperature (>250 C) to get activated by the reaction mixture, otherwise there is no
difference in performance between the oxidized and the prereduced sample. The prereduced sample shows no difference in performance between the ascending and descending temperature sequence. Concerning the effect of phosphoric acid pretreatment, it can be concluded that: phosphoric acid does not affect the catalytic performance of oxidized samples while it lowers the catalytic activity of prereduced samples (almost 50% decrease in catalytic activity at 200 C, Figs. 4 and 5). Moreover these samples had lower activity during the descending temperature sequence. It has to be noted that surface concentration analysis of both prereduced and oxidized samples via X-ray photoelectron spectroscopy hardly revealed the presence of phosphorus-related species in trace amounts. However, the concentration of these species was well below the detection limit of atomic adsorption spectroscopy. Thus, the effect of phosphoric acid cannot be attributed to massive deposition on the catalyst surface and blockage of active sites. The literature of catalyst poisoning, especially regarding automotive catalysts, contains a number of studies which suggest the presence of a glassy material such as metal phosphate on the surface of the catalyst [15]. If formed, copper or manganese phosphates would be expected to bind the respective copper and/or manganese cations, thus preventing copperemanganese redox function required for methanol reforming [15]. The latter effect is more intense in the case of prereduced samples where more copper species are exposed to phosphoric acid, since decomposition/reduction of the spinel phase takes place along with metallic copper segregation on the catalyst surface [8,9].
Fig. 6 e In-situ XRD measurements of isothermal reduction of CuMnOx catalysts (CuMn-Ox vs. CuMn-ox-P200) at 200 C.
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
3.3.
Redox catalyst properties
The catalysts were subjected to successive TPR/TPO runs, monitored by in-situ XRD, in order to study catalyst reducibility and ability to withstand cycled reductioneoxidation
16745
without deterioration. In addition, this treatment can be considered as accelerated aging as it simulates conditions, which may be experienced in the fuel cell reformer. In-situ measurements are necessary in the case of spent catalysts since exposure to air after reaction may cause re-oxidation.
Fig. 7 e In-situ XRD measurements of redox behavior (3 TPR/TPO cycles; (a) 1st TPR, (b) 1st TPO, (c) 2nd TPR, (d) 2nd TPO, (e) 3rd TPR, (f) 3rd TPO) of CuMnOx catalyst (CuMn-ox sample).
16746
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
The pyrophoric behavior of reduced catalysts has to be especially avoided when the reduced catalyst is abruptly exposed to air after turning off the feed of reactants, since major local temperature increases can occur due to fast Cu oxidation, which may lead to deactivation. In-situ reduction of CuMn-ox and CuMn-ox-P200 samples was monitored in 15-h isothermal runs under hydrogen flow at 200 C (Fig. 6). The spinel phase of Cu1.5Mn1.5O4 is predominant initially in both samples, while a small amount of Mn2O3 appears to be present. (Partial) reduction of oxides is possible at 200 C after long exposure to H2 flow since the structure of both catalysts was changed and the spinel phase decomposed to MnO and Cu0 phases. Phosphoric acid appears to retard the spinel oxide decomposition/reduction. Three successive TPR/TPO runs were carried out over both samples in order to examine the reversibility of reduction/ oxidation cycles. In-situ XRD measurements of these runs are presented in Fig. 7 for the CuMn-ox catalyst. Similar patterns were also obtained for the H3PO4-treated sample. It can be observed that in the first redox cycle: (a) reduction/decomposition initiates between 250 and 300 C, (b) re-oxidation initiates between 200 and 250 C and (c) spinel oxide is formed above 450 C. In the second cycle, reduction/decomposition initiates between 250 and 300 C (w50 C higher than the first TPR), while re-oxidation initiates between 200 and 250 C and spinel oxide is formed above 450 C (no difference with the first TPO). The third cycle has similar characteristics with the second cycle, while decomposition/reduction of spinel oxide takes place at lower temperatures in agreement with previous TPR/ TPO measurements [11]. The results obtained in the TPR/TPO cycles indicate that the efficiency of spinel re-formation is diminished at successive reduction/oxidation cycles. It should be emphasized that the TPR profiles at high cycle numbers are quite similar to the one of pure CuO, which implies that during TPO Cu0 gets oxidized to Cu2þ, but copper ions do not get incorporated back to the spinel structure. This might be due to increasing difficulty of MnO re-oxidation at successive cycles. The amounts of O2 consumed in the last cycles correspond to full oxidation of Cu0 to CuO and oxidation of only a small percentage (20%) of Mn2þeMn3þ.
3.4. Long term testing of CuMnOx/Cu foam reformer in the presence and absence of MEA in the single cell Long term (350 h) catalytic runs (Fig. 8) were performed with CuMnOx/Cu foams (as-prepared and prereduced samples) in the presence and in the absence of ADVENT TPS MEA in the single cell, in order to investigate the effect of: (i) type of pretreatment and (ii) phosphoric acid leaching by MEA, on the catalytic performance with reaction time at 200 C. It can be observed that: U There is a 30% decline in MeOH conversion (with oxidized catalyst being less active than prereduced catalyst) after 350 h. U Oxidized catalyst needs 24 h to get activated in agreement with XRD measurements.
Fig. 8 e Stability of CuMnOx/Cu foam reforming catalyst in the presence and absence of ADVENT TPS MEA in the single cell.
U In the presence of MEA, MeOH conversion is reduced by 30%: negative effect of H3PO4 in agreement with fixed-bed reactor measurements of treated CuMnOx powders. U Differences among catalysts are less important at long operation times. It should noted that the oxidized sample had significantly inferior performance in the presence of the MEA (stability curve not shown) as compared with the corresponding curve in the absence of MEA. Although, the performance of oxidized catalysts remained “unaffected” after phosphoric acid treatment as shown in a previous section, the low reaction temperature of 200 C and phosphoric acid leaching hinder the effective in-situ reduction by the reaction mixture. Postmortem characterization of the samples attached in the anode side of the MEA was done by SEM/EDX and revealed the extended presence of phosphorus species at the top surface side of both oxidized and prereduced foam-based catalysts, i.e. at the MEA e foam interface. The concentration of phosphorus species diminished at a distance of 2e3 mm away from the interface, and became zero at the bottom side (i.e. 7 mm away from the interface) of the oxidized foam-based catalyst. Despite the fact that the open porous matrix of the foam allows the exposure of the whole volume of the reformer to phosphoric acid vapors, the poisoning effect is related with the oxidation state of the catalytic layer, albeit long operation times override the typical deactivation of the catalyst observed in the absence of MEA.
4.
Conclusions
The harsh conditions prevailing inside the IRMFC emainly due to phosphoric acide do affect the catalytic performance of the internal reforming catalyst. However, the observed effect is not drastic, since the CuMnOx catalyst maintains w50% of its initial activity. Prereduced catalysts are more active but also more sensitive to phosphoric acid attack. In any case, pre-
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
reduction can be accomplished by H2 to a certain degree at 200 C, albeit at quite long reduction times. Overall, the observed stability of the CuMnOx catalyst is satisfactory in the timeframe of the presented experiments.
Acknowledgment Financial support from The Fuel Cells and Hydrogen Joint Undertaking (FCH JU; Area SP1-JTI-FCH-2009.4.2; Grant agreement no. 245202) is gratefully acknowledged.
references
[1] Avgouropoulos G, Papavasiliou J, Daletou MK, Kallitsis JK, Ioannides T, Neophytides S. Reforming methanol to electricity in a high temperature PEM fuel cell. Appl Catal B Environ 2009;90:628. [2] Avgouropoulos G, Papavasiliou J, Daletou M, Geormezi M, Triantafyllopoulos N, Ioannides T, et al. Internal Reforming Alcohol High Temperature PEM Fuel Cell. US Patent 200861/ 095779; September 10 (2008). [3] Avgouropoulos G, Ioannides T, Kallitsis JK, Neophytides S. Development of an internal reforming alcohol fuel cell: concept, challenges and opportunities. Chem Eng J 2011;176e177:95. [4] Sa´ S, Silva H, Branda˜o L, Sousa JM, Mendesa A. Catalysts for methanol steam reformingeA review. Appl Catal B Environ 2010;99:43. [5] Patel S, Pant KK. Kinetic modeling of oxidative steam reforming of methanol over Cu/ZnO/CeO2/Al2O3 catalyst. Appl Catal A Gen 2009;356:189.
16747
[6] Muroyama H, Nakase R, Matsui T, Eguchi K. Ethanol steam reforming over Ni-based spinel oxide. Int J Hydrogen Energy 2010;35:1575. [7] Navarro RM, Pen˜a MA, Fierro JLG. Hydrogen production reactions from carbon feedstocks: fossil fuels and biomass. Chem Rev 2007;107:3952. [8] Valde´s-Solı´s T, Marba´n G, Fuertes AB. Nanosized catalysts for the production of hydrogen by methanol steam reforming. Catal Today 2006;116:354. [9] Frank B, Jentoft FC, Soerijanto H, Kro¨hnert J, Schlo¨gl R, Schoma¨cker R. Steam reforming of methanol over coppercontaining catalysts: Influence of support material on microkinetics. J Catal 2007;246:177. [10] Hong X, Ren S. Selective hydrogen production from methanol oxidative steam reforming over ZneCr catalysts with or without Cu loading. Int J Hydrogen Energy 2008;33: 700. [11] Papavasiliou J, Avgouropoulos G, Ioannides T. Combined steam reforming of methanol over CueMn spinel oxide catalysts. J Catal 2007;251:7. [12] Papavasiliou J, Avgouropoulos G, Ioannides T. Steam reforming of methanol over copperemanganese spinel oxide catalysts. Catal Commun 2005;6:497. [13] Papavasiliou J, Avgouropoulos G, Ioannides T. In situ combustion synthesis of structured CueCeeO and CueMneO catalysts for the production and purification of hydrogen. Appl Catal B Environ 2006;66:168. [14] Papavasiliou J, Avgouropoulos G, Ioannides T. Steady-state isotopic transient kinetic analysis of steam reforming of methanol over Cu-based catalysts. Appl Catal B Environ 2009;88:490. [15] Rokosz MJ, Chen AE, Lowe-Ma CK, Kucherov AV, Benson D, Paputa Peck MC, et al. Characterization of phosphoruspoisoned automotive exhaust catalysts. Appl Catal B 2001; 33:205.
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
CuMnOx catalysts for internal reforming methanol fuel cells: Application aspects Joan Papavasiliou a, George Avgouropoulos a,b,*, Theophilos Ioannides a,* a
Foundation for Research and Technology-Hellas (FORTH), Institute of Chemical Engineering and High Temperature Chemical Processes (ICEHT), P.O. Box 1414, GR-26504 Patras, Greece b Department of Materials Science, University of Patras, GR-26504 Rio Patras, Greece
article info
abstract
Article history:
The Internal Reforming Methanol Fuel Cell (IRMFC) incorporates a methanol reforming
Received 15 November 2011
catalyst into the anodic compartment of a high temperature, polymer electrolyte
Received in revised form
membrane fuel cell (HT-PEMFC). The present work examines application aspects of the
17 February 2012
cell, including the operation temperature, the type of catalyst pretreatment, the effect of
Accepted 20 February 2012
phosphoric acid leaching and the time-on-stream behavior of CuMnOx catalysts.
Available online 23 March 2012
Combustion-synthesized structured catalysts can efficiently operate at 200 C with a 30% decline in MeOH conversion after 350 h on methanol/water stream in the presence of the
Keywords:
high temperature polymer electrolyte membrane electrode assembly. Differences observed
High temperature PEM fuel cell
in the catalytic activity of oxidized versus prereduced samples with respect to phosphoric
Internal reforming
acid poisoning are negligible after long operation time.
Methanol
Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
Hydrogen
reserved.
Copper-manganese catalysts Phosphoric acid
1.
Introduction
The internal reforming alcohol fuel cell (IRAFC) is a type of high temperature polymer electrolyte membrane fuel cell (HTPEMFC), in which the alcohol reforming catalyst is incorporated into the anode compartment [1e3]. Alcohol (methanol, for example) gets reformed by the catalyst inside the cell (internal reforming). Integration of the reformer into the fuel cell eliminates the need for additional heat exchangers and a separate fuel processor. Especially in the case of methanol as starting fuel, WGS and PrOx reactors are not necessary since (i) high temperature operation (i.e. 180e220 C) enables utilization of lower quality reformate feeds containing up to 2e3% CO and (ii) the “waste” heat produced by the fuel cell is in-situ
utilized to drive the endothermic reforming reaction. Thus, the design of the fuel processor-fuel cell system offers room for simplification, increase of efficiency and minimization of system weight and volume. Alcohol fuels and, more specifically methanol, are attractive chemical sources of hydrogen, since they can be easily stored, transported and dispensed, and, at the same time, they can be efficiently produced from a wide variety of sources including fossil fuels, but also agricultural products and municipal waste, wood and biomass, in general [3e7]. More importantly, they can be also made through chemical recycling of carbon dioxide. Methanol has 5e7 times higher energy density than compressed H2. Moreover, it has low sulfur content and can be reformed to hydrogen rich mixtures with
* Corresponding authors. Foundation for Research and Technology-Hellas (FORTH), Institute of Chemical Engineering and High Temperature Chemical Processes (ICE-HT), P.O. Box 1414, GR-26504 Patras, Greece. Tel.: þ30 2610965268; fax: þ30 2610965223. E-mail addresses: [email protected] (G. Avgouropoulos), [email protected] (T. Ioannides). 0360-3199/$ e see front matter Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2012.02.124
16740
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
low CO byproduct formation at the temperature level of IRAFC. H2 can be catalytically produced from methanol via steam reforming (SRM: CH3OH þ H2O # CO2 þ 3H2) at relatively low temperatures (200e300 C) with high selectivity [1,4,7e10]. In the case of the integrated PEM fuel cell system, the required heat for the SRM process will be supplied by the cell itself. Commercially available copper-based catalysts, typically with composition CueZnOe(Al2O3) have been widely used for generating hydrogen from methanol [4,5,7]. Even though these catalysts are widely used in H2 plants, they have not been designed for application in small stationary or portable/ mobile fuel processors. Thus, optimization of Cu-based catalysts for the envisaged application is of primary concern. It is well established that the activity and CO selectivity of Cubased catalysts is greatly dependent on their morphology (high copper dispersion is desirable) and the redox properties of the catalyst. Taking into account the need of the presence of well dispersed metallic (reduced) Cu species as active sites for methanol adsorption and reaction, significant enhancement of copper oxide reducibility -especially under methanol/water feedse is necessary at the cell operation level of around 200 C. Additionally, improvement of the catalytic activity will allow the reforming catalyst to operate more efficiently at the target IRAFC temperatures. In addition to catalyst activity, optimization of the catalyst formulation should take into account (i) the CO selectivity, (ii) the stability/deactivation in respect to coke formation or phosphoric acid leaching from polymer electrolyte membrane and (iii) pyrophoricity parameters. Recently, a single-cell (Internal Reforming Methanol Fuel Cell, IRMFC) incorporating ADVENT TPS phosphoric-aciddoped copolymer membrane and a CuMnOx/Cu foam methanol reforming catalyst in the anode was constructed and tested at 200 C demonstrating the functionality of the unit (Fig. 1) [1e3]. In view of the IRMFC implementation, important improvements of reforming catalysts are required regarding their efficient operation at temperatures around 200 C. In order to cope with the above operating framework we have reported on the application of combustion-synthesized CuMnOx catalysts for methanol reforming process. In the present study, CuMnOx catalyst synthesis in powder and monolithic form was carried out via the combustion method. Detailed screening of the catalysts was performed under conditions relevant to the proposed application taking into account the following parameters: temperature (focus at around 200 C), type of pretreatment, effect of phosphoric acid leaching and time-on-stream. Selected reforming catalysts were introduced in the anode compartment of an IRMFC and tested for their chemical stability.
2.
Experimental
2.1.
Synthesis of methanol reforming catalysts
The ureaenitrates combustion method (Fig. 2) was used for the synthesis of CuMnOx spinel oxide catalysts [11]. Urea [CH4N2O] with manganese nitrate [Mn(NO3)2$6H2O] and copper nitrate [Cu(NO3)2$3H2O] were mixed in an alumina crucible in the appropriate molar ratios (Cu/(Cu þ Mn) ¼ 0.30,
75% excess of urea) in a minimum volume of distilled water to obtain a transparent solution. The values for copper loading and initial amount of urea were selected from our previous studies on these catalytic systems prepared with the same technique and tested for the methanol reforming process [11,12]. The mixed solutions were heated for a few minutes at 80 C, so that excess water was removed. The resulting viscous gel was introduced in an open muffle furnace, preheated at 400e500 C, in a fuming cupboard. The gel started boiling with frothing and foaming, and in a couple of minutes ignited spontaneously with rapid evolution of a large quantity of gases, yielding a foamy, voluminous powder. In order to burnoff carbon residues and form well-crystalline structure, the powders were further heated at 550 C for 1 h (CuMn-ox samples). A half portion of the prepared batch was reduced at 350 C under H2 flow (CuMn-red samples). All the obtained powders were sieved to obtain the desired size, with diameter between 90 and 180 mm. CuMnOx catalysts supported on Cu metal foam (Changsha LYRUN New Material Co., Ltd.; 30 ppi (pores per linear inch) porosity; 7 mm thickness; 4000 g m2 density) were also prepared via the in-situ combustion method [1,13] in order to test the stability behavior with reaction time in the presence or absence of an ADVENT TPS high temperature polymer electrolyte membrane electrode assembly.
2.2. Treatment of combustion-synthesized CuMnOx catalysts with phosphoric acid vapors Since the reforming catalyst should be resistant toward deactivation caused by phosphoric acid (i.e. H3PO4 leaching from membrane electrode assembly), combustion-synthesized CuMnOx samples (both oxidized and reduced) were exposed to phosphoric acid vapors in a teflon autoclave vessel (Fig. 3). 750 mg of CuMn-ox and CuMn-red samples were placed on the bottom of a pyrex fritted disc assembly and treated hydrothermally in a teflon autoclave containing 15 ml of 85% H3PO4. The vessel was heated at a rate of 2 C min1 to the desired temperature level (150 or 200 C) where it was kept for 5 h. Following this treatment, the autoclave was opened and the samples were dried at 120 C for 12 h. In order to facilitate the presentation of results, the following encoding of catalysts is used: CuMn-ox or red-P150 or 200, denotes the oxidized or reduced sample treated with H3PO4 vapors either at 150 or 200 C. For example, the CuMn-red-P200 sample was prereduced at 350 C under H2 flow and treated at 200 C with H3PO4 vapors.
2.3.
Redox catalyst properties
In-situ XRD measurements (X-ray powder diffractometer: Bruker D8 Advance using Cu Ka radiation with l ¼ 0.15418 nm) were carried out on selected CuMnOx catalysts. Temperature programmed reduction (TPR) and oxidation (TPO) experiments were performed under H2 and O2 flows. Two different sets of experiments were performed. The applied experimental conditions were the following:
(A) Isothermal reduction with Η2 at 200 C , heating 30 C / 200 C (6 C min1), 5% Η2/Ν2, 60 ml min1
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
16741
Fig. 1 e Internal reforming methanol single cell. Upper left: CuMnOx catalyst supported on Cu foam; Upper right: POCO graphite bipolar plate with triple serpentine flow field arrangement; Bottom left: ADVENT TPS MEA, reformer, graphite bipolar plates and modified anode graphite plate; Bottom right: integrated single cell. , isothermal soak, 200 C, 15 h (B) Redox cycles , heating 30 C / 500 C (6 C min1), 5% Η2/Ν2, 60 ml min1 , free cooling, switch to Ν2 , oxidation 30 C / 500 C (6 C min1), 5% Ο2/Ν2, 60 ml min1
flow rate of the reaction mixture was 70 cm3 min1 (W/ F ¼ 0.257 g s cm3). The standard reaction feed was 5% MeOH, H2O/MeOH ¼ 1.5. Product and reactant analysis was carried out by a gas chromatograph (Shimadzu GC-14B) equipped with TCD and FID and He as carrier gas. The CO selectivity was calculated as CO selectivity ¼ [CO]out/([CO]out þ [CO2]out).
2.4.
2.5. Long term testing of CuMnOx/Cu foam reformer incorporated in an IRMFC
Catalyst screening
Activity and selectivity measurements for steam reforming of methanol were carried out at atmospheric pressure in a fixedbed reactor system, which has been described previously [11]. The standard catalyst weight was 0.3 g and the standard total
An internal reforming methanol single fuel cell was prepared for long term testing of CuMnOx catalysts. The employed single cell (Fig. 1) has the following features: (i) triple
Fig. 2 e Schematic view of the combustion process.
16742
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
set at 200 C. Reactants/products analysis was done by a gas chromatograph (Agilent Technologies, 6820 GC System) equipped with TCD and FID.
3. Fig. 3 e Experimental setup for catalysts treatment with phosphoric acid vapors. serpentine flow field arrangement, (ii) POCO graphite bipolar plates, (iii) copper current collectors, (iv) 25 cm2 electrode area with a platinum loading of 1.5 mg Pt/cm2, 2 g H3PO4/g Pt, (v) copolymer TPS electrolyte membrane of 180 mm thickness and 190 wt.% H3PO4 doping level. A standard single cell was purchased from Fuel Cell Technologies Inc. and was modified in such a way to fit the foam reforming catalyst (Fig. 1). The membrane electrode assembly comprising an ADVENT TPS H3PO4-doped copolymer, as the high temperature polymer electrolyte membrane, was sandwiched between the anodic electrode (9 g of CuMnOx catalyst supported on metallic copper foam (5 cm 5 cm 0.7 cm) and placed adjacent to the Pt/C anode electrocatalyst) and Pt/C cathodic electrode. A teflon sheet was employed in the experiment carried out in the absence of MEA in order to ensure anode compartment sealing. Vaporized methanol and water mixtures (10 vol.% MeOH, 15 vol.% H2O in He; 100 cm3 min1 total flow) were supplied to the anode compartment. The cell temperature was
3.1. Standard physicochemical characterization and catalytic properties of CuMnOx catalysts The structural and morphological properties of CueMn spinel oxide catalysts were investigated and discussed in detail in previous papers [11,12]. The following conclusions were drawn from these works: Fresh CueMn catalysts are composed of the spinel phase Cu1.5Mn1.5O4, as well as Mn2O3 and CuO, depending on the Cu/Mn ratio, and get reduced to Cu0 and MnO under methanol reforming conditions. The BET specific surface areas of the samples are less than 9 m2 g1. XPS analysis revealed the presence of two different oxidation states in both copper (Cu2þ and Cuþ) and manganese (Mn4þ and Mn3þ) in fresh catalysts and decomposition of the spinel in used catalysts. XRD and TPR measurements showed that at a high copper content (x ¼ 0.40, 0.50), the excess copper is not incorporated into the spinel, but is rather present as a separate CuO phase.
CuMn-ox
12
10
CuMn-ox-P150
10
8
CuMn-ox-P200 o open symbols: 300 C to RT
6
CO selectivity, %
CO selectivity, %
12
o
solid symbols: RT to 300 C
4 2 0 180
200
220
240
260
280
300
8
open symbols: 300 C to RT
6
solid symbols: RT to 300 C
o
o
4 2
180
200
220
240
260
280
300
60 40 20
A
0 200
220
240
260 o
T, C
280
300
MeOH conversion, %
100
80
180
CuMn-red CuMn-red-P150 CuMn-red-P200
0
100
MeOH conversion, %
Results and discussion
80 60 40 20
B
0 180
200
220
240
260
280
300
o
T, C
Fig. 4 e Comparison of catalytic activity/selectivity for the steam reforming of methanol of (A) oxidized and (B) prereduced samples, untreated and treated with H3PO4 vapors. Operating conditions: W/F [ 0.257 g s cmL3, 5% CH3OH, H2O/ CH3OH [ 1.5.
16743
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
The optimal catalyst was prepared with a Cu/(Cu þ Mn) ratio of 0.30.
have comparable activity to that of a commercial Cu/ZnO/ Al2O3 catalyst. The amount of CO produced during reforming was well below wateregas shift equilibrium (CO þ H2O # CO2 þ H2) and other byproducts such as formaldehyde, formic acid, methyl formate or dimethyl ether, which are often formed by reactions of methanol over Cubased catalysts, were not detected at the temperature level
12
CuMn-ox
10
CuMn-red
8
12 o
open symbols: 300 C to RT o
solid symbols: RT to 300 C
6 4
0 220
240
260
280
180
300
200
220
240
260
280
300
100
MeOH conversion, %
80 60 40 20
80 60 40 20
A
0 180
o
solid symbols: RT to 300 C
4
0 200
o
open symbols: 300 C to RT
6
2
100
MeOH conversion, %
8
2
180
CuMn-ox-P150 CuMn-red-P150
10
CO selectivity, %
CO selectivity, %
CuMnOx materials have been proven to be very efficient methanol steam reforming catalysts, able to selectively produce CO2 and H2 at temperatures lower than 240 C [11]. Despite their low surface areas (<9 m2 g1), these catalysts
200
220
240
260
280
B
0 180
300
200
220
260
280
300
T, C
T, C 12
CO selectivity, %
240 o
o
CuMn-ox-P200 CuMn-red-P200
10 8
o
open symbols: 300 C to RT o
solid symbols: RT to 300 C
6 4 2 0 180
200
220
240
260
280
300
MeOH conversion, %
100 80 60 40 20
C
0 180
200
220
240
260
280
300
o
T, C
Fig. 5 e Effect of H3PO4 and type of pretreatment (oxidized vs. reduced samples) on CuMnOx catalysts activity and selectivity for the steam reforming of methanol. (A) no H3PO4 treatment, (B) H3PO4 treatment at 150 C, (C) H3PO4 treatment at 200 C. Operating conditions: W/F [ 0.257 g s cmL3, 5% CH3OH, H2O/CH3OH [ 1.5.
16744
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
of IRMFC, i.e. 200 C. Previous mechanistic studies over combustion-synthesized CuMnOx catalysts lead to the conclusion that the SRM reaction pathway occurs exclusively via methyl formate hydrolysis [14]: Methanol dehydrogenation to formaldehyde and H2 1a. CH3OH / HCHO þ H2 1b. CH3OH þ HCHO / HCOOCH3 þ H2 or direct dehydrogenation of methanol to methyl formate and H2 1c. 2CH3OH / HCOOCH3 þ 2H2 Hydrolysis of methyl formate 2. HCOOCH3 þ H2O / CH3OH þ HCOOH Formic acid decomposition 3. HCOOH / H2 þ CO2
3.2. Effect of phosphoric acid and type of pretreatment (oxidation vs. reduction) on combustion-synthesized CuMnOx catalysts performance The CuMnOx catalyst behavior was examined in steam reforming of methanol following no pretreatment or a reductive one (CuMn-ox vs. CuMn-red). The goal is to determine whether: (i) pre-reduction leads to more active catalysts and (ii) direct exposure to the reaction feed leads to in-situ reduction with time-on-stream. The obtained activity/selectivity curves are shown in Figs. 4 and 5. It can be observed that the oxidized sample needs high temperature (>250 C) to get activated by the reaction mixture, otherwise there is no
difference in performance between the oxidized and the prereduced sample. The prereduced sample shows no difference in performance between the ascending and descending temperature sequence. Concerning the effect of phosphoric acid pretreatment, it can be concluded that: phosphoric acid does not affect the catalytic performance of oxidized samples while it lowers the catalytic activity of prereduced samples (almost 50% decrease in catalytic activity at 200 C, Figs. 4 and 5). Moreover these samples had lower activity during the descending temperature sequence. It has to be noted that surface concentration analysis of both prereduced and oxidized samples via X-ray photoelectron spectroscopy hardly revealed the presence of phosphorus-related species in trace amounts. However, the concentration of these species was well below the detection limit of atomic adsorption spectroscopy. Thus, the effect of phosphoric acid cannot be attributed to massive deposition on the catalyst surface and blockage of active sites. The literature of catalyst poisoning, especially regarding automotive catalysts, contains a number of studies which suggest the presence of a glassy material such as metal phosphate on the surface of the catalyst [15]. If formed, copper or manganese phosphates would be expected to bind the respective copper and/or manganese cations, thus preventing copperemanganese redox function required for methanol reforming [15]. The latter effect is more intense in the case of prereduced samples where more copper species are exposed to phosphoric acid, since decomposition/reduction of the spinel phase takes place along with metallic copper segregation on the catalyst surface [8,9].
Fig. 6 e In-situ XRD measurements of isothermal reduction of CuMnOx catalysts (CuMn-Ox vs. CuMn-ox-P200) at 200 C.
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
3.3.
Redox catalyst properties
The catalysts were subjected to successive TPR/TPO runs, monitored by in-situ XRD, in order to study catalyst reducibility and ability to withstand cycled reductioneoxidation
16745
without deterioration. In addition, this treatment can be considered as accelerated aging as it simulates conditions, which may be experienced in the fuel cell reformer. In-situ measurements are necessary in the case of spent catalysts since exposure to air after reaction may cause re-oxidation.
Fig. 7 e In-situ XRD measurements of redox behavior (3 TPR/TPO cycles; (a) 1st TPR, (b) 1st TPO, (c) 2nd TPR, (d) 2nd TPO, (e) 3rd TPR, (f) 3rd TPO) of CuMnOx catalyst (CuMn-ox sample).
16746
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
The pyrophoric behavior of reduced catalysts has to be especially avoided when the reduced catalyst is abruptly exposed to air after turning off the feed of reactants, since major local temperature increases can occur due to fast Cu oxidation, which may lead to deactivation. In-situ reduction of CuMn-ox and CuMn-ox-P200 samples was monitored in 15-h isothermal runs under hydrogen flow at 200 C (Fig. 6). The spinel phase of Cu1.5Mn1.5O4 is predominant initially in both samples, while a small amount of Mn2O3 appears to be present. (Partial) reduction of oxides is possible at 200 C after long exposure to H2 flow since the structure of both catalysts was changed and the spinel phase decomposed to MnO and Cu0 phases. Phosphoric acid appears to retard the spinel oxide decomposition/reduction. Three successive TPR/TPO runs were carried out over both samples in order to examine the reversibility of reduction/ oxidation cycles. In-situ XRD measurements of these runs are presented in Fig. 7 for the CuMn-ox catalyst. Similar patterns were also obtained for the H3PO4-treated sample. It can be observed that in the first redox cycle: (a) reduction/decomposition initiates between 250 and 300 C, (b) re-oxidation initiates between 200 and 250 C and (c) spinel oxide is formed above 450 C. In the second cycle, reduction/decomposition initiates between 250 and 300 C (w50 C higher than the first TPR), while re-oxidation initiates between 200 and 250 C and spinel oxide is formed above 450 C (no difference with the first TPO). The third cycle has similar characteristics with the second cycle, while decomposition/reduction of spinel oxide takes place at lower temperatures in agreement with previous TPR/ TPO measurements [11]. The results obtained in the TPR/TPO cycles indicate that the efficiency of spinel re-formation is diminished at successive reduction/oxidation cycles. It should be emphasized that the TPR profiles at high cycle numbers are quite similar to the one of pure CuO, which implies that during TPO Cu0 gets oxidized to Cu2þ, but copper ions do not get incorporated back to the spinel structure. This might be due to increasing difficulty of MnO re-oxidation at successive cycles. The amounts of O2 consumed in the last cycles correspond to full oxidation of Cu0 to CuO and oxidation of only a small percentage (20%) of Mn2þeMn3þ.
3.4. Long term testing of CuMnOx/Cu foam reformer in the presence and absence of MEA in the single cell Long term (350 h) catalytic runs (Fig. 8) were performed with CuMnOx/Cu foams (as-prepared and prereduced samples) in the presence and in the absence of ADVENT TPS MEA in the single cell, in order to investigate the effect of: (i) type of pretreatment and (ii) phosphoric acid leaching by MEA, on the catalytic performance with reaction time at 200 C. It can be observed that: U There is a 30% decline in MeOH conversion (with oxidized catalyst being less active than prereduced catalyst) after 350 h. U Oxidized catalyst needs 24 h to get activated in agreement with XRD measurements.
Fig. 8 e Stability of CuMnOx/Cu foam reforming catalyst in the presence and absence of ADVENT TPS MEA in the single cell.
U In the presence of MEA, MeOH conversion is reduced by 30%: negative effect of H3PO4 in agreement with fixed-bed reactor measurements of treated CuMnOx powders. U Differences among catalysts are less important at long operation times. It should noted that the oxidized sample had significantly inferior performance in the presence of the MEA (stability curve not shown) as compared with the corresponding curve in the absence of MEA. Although, the performance of oxidized catalysts remained “unaffected” after phosphoric acid treatment as shown in a previous section, the low reaction temperature of 200 C and phosphoric acid leaching hinder the effective in-situ reduction by the reaction mixture. Postmortem characterization of the samples attached in the anode side of the MEA was done by SEM/EDX and revealed the extended presence of phosphorus species at the top surface side of both oxidized and prereduced foam-based catalysts, i.e. at the MEA e foam interface. The concentration of phosphorus species diminished at a distance of 2e3 mm away from the interface, and became zero at the bottom side (i.e. 7 mm away from the interface) of the oxidized foam-based catalyst. Despite the fact that the open porous matrix of the foam allows the exposure of the whole volume of the reformer to phosphoric acid vapors, the poisoning effect is related with the oxidation state of the catalytic layer, albeit long operation times override the typical deactivation of the catalyst observed in the absence of MEA.
4.
Conclusions
The harsh conditions prevailing inside the IRMFC emainly due to phosphoric acide do affect the catalytic performance of the internal reforming catalyst. However, the observed effect is not drastic, since the CuMnOx catalyst maintains w50% of its initial activity. Prereduced catalysts are more active but also more sensitive to phosphoric acid attack. In any case, pre-
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 7 ( 2 0 1 2 ) 1 6 7 3 9 e1 6 7 4 7
reduction can be accomplished by H2 to a certain degree at 200 C, albeit at quite long reduction times. Overall, the observed stability of the CuMnOx catalyst is satisfactory in the timeframe of the presented experiments.
Acknowledgment Financial support from The Fuel Cells and Hydrogen Joint Undertaking (FCH JU; Area SP1-JTI-FCH-2009.4.2; Grant agreement no. 245202) is gratefully acknowledged.
references
[1] Avgouropoulos G, Papavasiliou J, Daletou MK, Kallitsis JK, Ioannides T, Neophytides S. Reforming methanol to electricity in a high temperature PEM fuel cell. Appl Catal B Environ 2009;90:628. [2] Avgouropoulos G, Papavasiliou J, Daletou M, Geormezi M, Triantafyllopoulos N, Ioannides T, et al. Internal Reforming Alcohol High Temperature PEM Fuel Cell. US Patent 200861/ 095779; September 10 (2008). [3] Avgouropoulos G, Ioannides T, Kallitsis JK, Neophytides S. Development of an internal reforming alcohol fuel cell: concept, challenges and opportunities. Chem Eng J 2011;176e177:95. [4] Sa´ S, Silva H, Branda˜o L, Sousa JM, Mendesa A. Catalysts for methanol steam reformingeA review. Appl Catal B Environ 2010;99:43. [5] Patel S, Pant KK. Kinetic modeling of oxidative steam reforming of methanol over Cu/ZnO/CeO2/Al2O3 catalyst. Appl Catal A Gen 2009;356:189.
16747
[6] Muroyama H, Nakase R, Matsui T, Eguchi K. Ethanol steam reforming over Ni-based spinel oxide. Int J Hydrogen Energy 2010;35:1575. [7] Navarro RM, Pen˜a MA, Fierro JLG. Hydrogen production reactions from carbon feedstocks: fossil fuels and biomass. Chem Rev 2007;107:3952. [8] Valde´s-Solı´s T, Marba´n G, Fuertes AB. Nanosized catalysts for the production of hydrogen by methanol steam reforming. Catal Today 2006;116:354. [9] Frank B, Jentoft FC, Soerijanto H, Kro¨hnert J, Schlo¨gl R, Schoma¨cker R. Steam reforming of methanol over coppercontaining catalysts: Influence of support material on microkinetics. J Catal 2007;246:177. [10] Hong X, Ren S. Selective hydrogen production from methanol oxidative steam reforming over ZneCr catalysts with or without Cu loading. Int J Hydrogen Energy 2008;33: 700. [11] Papavasiliou J, Avgouropoulos G, Ioannides T. Combined steam reforming of methanol over CueMn spinel oxide catalysts. J Catal 2007;251:7. [12] Papavasiliou J, Avgouropoulos G, Ioannides T. Steam reforming of methanol over copperemanganese spinel oxide catalysts. Catal Commun 2005;6:497. [13] Papavasiliou J, Avgouropoulos G, Ioannides T. In situ combustion synthesis of structured CueCeeO and CueMneO catalysts for the production and purification of hydrogen. Appl Catal B Environ 2006;66:168. [14] Papavasiliou J, Avgouropoulos G, Ioannides T. Steady-state isotopic transient kinetic analysis of steam reforming of methanol over Cu-based catalysts. Appl Catal B Environ 2009;88:490. [15] Rokosz MJ, Chen AE, Lowe-Ma CK, Kucherov AV, Benson D, Paputa Peck MC, et al. Characterization of phosphoruspoisoned automotive exhaust catalysts. Appl Catal B 2001; 33:205.