Catalytic Oxidation of Methanol by Aqueous ... - Wiley Online Library

ORIGINAL RESEARCH PAPER

DOI: 10.1002/fuce.201300126

Catalytic Oxidation of Methanol by Aqueous POM on Al2O3 Supported Catalysts and Electrochemical Performance of POM P. Li1, W. Mi1, 2*, Q. Su1, 2*, C. Luo1 1

2

Department of Thermal Science and Energy Engineering, School of Mechanical Engineering, University of Science and Technology Beijing, Beijing 100083, China Beijing Key Laboratory of Energy Saving and Emission Reduction for Metallurgical Industry, University of Science and Technology Beijing, Beijing 100083, China

Received June 20, 2013; accepted September 10, 2013; published online November 25, 2013

Abstract Phosphomolybdic acid (H3PMo12O40, POM) was attempted to be used as the energy-storage agent in this paper to avoid some problems of the direct methanol fuel cell (DMFC), such as catalyst poisoning and methanol permeation. Catalytic oxidation of methanol by aqueous POM on Al2O3 supported catalysts with Pt and Ru active metal was evaluated in the presence of liquid water. The process takes advantage of the high catalytic activities of platinum for methanol oxidation. The effects of temperature, reaction time, and methanol concentration on activity were observed. The catalytic activity

1 Introduction Extensive research has recently been launched into the development of direct methanol fuel cells (DMFCs), as methanol is an abundant, inexpensive liquid fuel that can be easily handled, stored, and transported, as compared to hydrogen [1]. DMFCs are the strong candidates for new portable power devices, such as future laptop computers or cellular phones [2]. Despite the obvious advantages of the DMFC, the practical process is still slow. The main reason is that CO produced in the intermediate of the methanol process is strongly absorbed on the catalyst surface, which causes the catalyst poisoning, activity decreasing and the decline of methanol oxidation current. In addition, the performance of DMFCs descends due to methanol permeation through the membrane from the anode to the cathode, which leads to the formation of mixed potentials in the cathode and direct oxidation of methanol in the cathode catalyst layer [3]. Dumesic and coworkers [4, 5] have reported a new process that directly utilizes CO as a source of energy by converting it

56

© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

of Pt/Al2O3 is better than that of Ru/Al2O3 for the oxidation of methanol by POM. The methanol conversion rate reached 93.55% on the Pt/Al2O3 at 80 °C after reaction for 1 h. The electrochemical experiments indicate that POM shows a larger current density in redox processes on an Au electrode than methanol. The redox process of reduced POM is a reversible multi-electron transfer process. Keywords: Au Electrode, DMFC, Electrochemical, Fuel Cell, Methanol, Phosphomolybdic Acid (POM)

with water to CO2 via an aqueous solution containing a reducible polyoxometalate (POM) compound, H3PMo12O40, over gold catalysts. Then, they developed process that can use COcontaining gas streams from the catalytic reforming of hydrocarbons to produce an aqueous solution of reduced polyoxometalate compounds that can be used to generate power. The reduced polyoxometalate can be reoxidized in fuel cells that contain simple carbon anodes. The POM serves as a strong oxidizing agent for CO and as an energy-storage agent for electrons and protons in above processes. After reaction, the solution stores energy in the form of protons and electrons associated with reduced metal cations that can be reoxidized readily at a fuel-cell anode by the transfer of electrons from the POM to the electrode and the transport of protons through the PEM to the cathode. Heteropoly acid, such as POM as a kind of macromolecular, can store and transmit electrons, and has good redox

– [*] Corresponding author, [email protected]

FUEL CELLS 14, 2014, No. 1, 56–63

Li et al.: Catalytic Oxidation of Methanol by Aqueous POM on Al2O3 process was condensed in the liquid nitrogen. The gas after removing H2O flowed into a thermal conductivity detector. 2.2.2 H2 Chemisorption The sample pretreatment was same as that in TPR analysis. After that process, H2–Ar mixture gases (ratio of H2 and Ar volume is 10:90) were injected into the U-shape tube by pulse ways until the signals intensity was stable. According to the amount of H2 chemisorbed on the catalysts, the metal dispersion, area, and particle size could be calculated. The metal dispersion D is defined as the ratio of the metal atoms surface and the total number of metal atoms, which can be calculated by H2 adsorption amount. For supported catalyst, the dispersion D (%) is: D

K × 2 × Nads × 10 LM

6

× 100

(1)

where, Nads (lmol gcat–1) is the amount of H2 adsorption and K is the correction factor. As one Ru or Pt atom adsorbs one hydrogen atom, respectively, so K = 1. LM (wt.%) is the loading of catalyst and M (g mol–1) is the metal atomic weight. Ru is 101.1 g mol–1, and Pt is 195.1 g mol–1. The metal area is the plot of metal atom cross-sectional area and the number of surface metal atoms. For supported catalyst, the metal area MA (m2 gcat–1) is: MA K × 2 × Nads × 10

2 Materials and Methods

×M

6

× NA × S

(2)

2

2.1 Catalyst Preparation The catalyst support was c-Al2O3 powder (50 lm, Shandong Aluminum Industry Group). Active metal precursor was chloroplatinic acid and RuCl3 (AR, Shanghai Jiu Yue Chemical Co, Ltd.). The supported catalyst was prepared by impregnation method. The loading of active metal was 20 wt.%. Firstly, chloroplatinic acid solution with a certain concentration was impregnated on the c-Al2O3 support, and the precursor was dried in the oven at 110 °C for 2 h. Secondly, the same operations were repeated until the active metal loading reached 20 wt.%. Next, it was reduced by hydrogen for 2 h at 380 °C. Finally, the catalyst was cooled and stored in nitrogen atmosphere. 2.2 Catalyst Characterization 2.2.1 Temperature-Programmed Reduction (TPR) Analysis TPR was conducted in a chemisorption instrument (Micrometric Autochem 2910, USA). First, approximately 100 mg of catalyst sample was heated to 150 °C at 10 °C min–1 ramp rate in a flow of Helium (He) gas of 50 ml min–1 and maintained for 60 min when temperature reached 150 °C in order to purge the water vapor (H2O) and other impurities adsorbed on the sample. Then the sample was cooled down to 40 °C. The catalysts were heated from 40 to 380 °C with a heating rate of 10 °C min–1 in a flow of 50 cm3 min–1 of 10 vol.% H2 in Argon (Ar). The produced H2O in the reduction

FUEL CELLS 14, 2014, No. 1, 56–63

www.fuelcells.wiley-vch.de

where, S (m ) is the cross sectional area of atom. Ru is 5.64 × 10–20 m2, and Pt is 6.07 × 10–20 m2. NA is the Avogadro constant which is 6.02 × 1023. For different catalysts, metal particle sizes are different. For supported catalyst, the particle size d (nm) is: d

6 000 × LM q × MA

(3)

where, q is the metal density. Ru density is 12.4 g cm–3, and Pt density is 21.4 g cm–3. 2.2.3 BET Specific Surface Area and Pore Structure The specific surface area and pore size distribution of catalyst were measured on Quantachrome Autosorb-1 Nitrogen adsorption analyzer using BET method. First, the sample was degassed by heating from room temperature to 110 °C at the rate of 10 °C min–1, and maintained at 110 °C for 12 h. Then it was weighed, cooled in liquid N2. The adsorption–desorption isotherms of the sample for N2 at liquid nitrogen temperature could be obtained through the computer by logging the amount of nitrogen gas adsorbed in the sample at the pre-set different pressures. The date of relative pressure P/P0 between 0.05–0.30 was used. Specific surface area was calculated according to the Brunauer–Emmett–Teller (BET) equation. Pore volume was calculated by the data of relative pressure P/P0. Using the BJH (Barrett–Joyner–Halenda) method the pore size distribution was calculated. The cross-sectional area for nitrogen molecules was 0.162 nm2.


57


property [6, 7]. Using a-Keggin type PW12O403––TiO2 system in the acidic environment of low pH value at 35 °C, oxidation of methanol can occur through photocatalysis reaction, and more than 40 mA cm–2 current density could be obtained. Even when using POM without TiO2, the methanol could be effectively oxidized [8]. Liquid formic acid could be degraded and mineralized into CO2 and H2O at the temperature of 25 °C by using [Xn+W11O39](12–n)–/SiO2 (X = Si, Ge, and P) composite membrane by UV radiation [8]. The aqueous solution of reduced polyoxometalate compounds could generate power by the oxidation of reduced polyoxometalate in fuel cells that contain simple carbon anodes, which has been proved to be feasible by Dumestic’s experiments. Here, we developed a process for DMFC by employing the POM as a strong oxidizing agent for methanol and as an energy-storage agent for electrons and protons to attempt to avoid some disadvantages of DMFC. The oxidation of methanol by POM is the key problem for the new process. Our research results showed that the methanol could be oxidized by POM on catalysts. Catalytic oxidation of methanol by aqueous phosphomolybdic acid (POM) over different active metal catalysts was evaluated in the presence of liquid water. And the electrochemical property of the reduced POM was tested comparing with methanol in this paper.


Li et al.: Catalytic Oxidation of Methanol by Aqueous POM on Al2O3 2.3 Catalyst Performance Evaluation for the Reaction of Methanol and Phosphomolybdic Acid (POM) The reaction of POM (AR, China Tianjin Jinke Fine Chemical Institute) and methanol was carried out in a slurry reactor. The reaction equipment is shown in Figure 1. Before the reaction process, the blank experiments without catalysts were carried out at different conditions (including different methanol concentration, reaction temperature, and time) in order to consider the amount of methanol evaporation and carbon produced by the reaction reagents. Then 100 ml mixed solution of POM (0.5 mol L–1), methanol (0.01, 0.03, and 0.05 mol L–1 separately for different reaction conditions) and catalyst (2 g) were added into the reactor at room temperature. Then the solution was stirred vigorously in order to eliminate diffusion influence and heated up to the required temperature. The reaction time was maintained for 10, 20, 30, and 60 min, respectively. The quantitative analysis of the carbon content was measured by carbon analyzer (TOC-VCPH, Japan SHIMADZU). The volatilized methanol was cooled and refluxed into the reactor through the serpentine condenser to reduce the experimental error. The influences of temperature, reaction time, and catalyst type on methanol conversion were investigated. Methanol conversion is calculated according to Eq. (4): X

n1 V1 n2 V2 × 100% n1 V1

(4)

where, X is the conversion of methanol, which is equal to the carbon change if the methanol is converted into carbon dioxide completely. n1 (mg L–1) is carbon total content of solution before reaction and n2 (mg L–1) is the carbon total content of

Fig. 1 Schematic diagram of the reaction equipment for phosphomolybdic acid (POM) and methanol: (1) electromagnetic mixer; (2) two openings flask; (3) electromagneton agitation; (4) water inlet; (5) water outlet; (6) thermometer; (7) serpentine reflux tube; (8) cooling water inlet; (9) cooling water outlet.

58


the solution after reaction. V1 is the volume of solution before reaction and V2 is the volume of solution after reaction. The difference of solution volume before and after the reaction is very small after testing, so V1 ≈ V2. For the production of methanol oxidation, usually the product includes formaldehyde, formic acid, and CO2 according to the oxidation depth of methanol. In order to determine the oxidation product of methanol, we test the product after the reaction using several indicators as following process. First, the solution after reaction was vaporized and condensed by ice bath to remove POM in the solution to avoid the color influence. The obtained solution was tested by Fehling reagent, potassium permanganate reagent, and silver mirror reagent (mixture solution of silver nitrate and ammonia) separately to detect the presence of HCHO, HCOOH, CH3OH in the solution. The only reaction products of methanol oxidized by POM are CO2 and water as former experiments have shown [9]. So the amount of methanol oxidized by POM can be calculated out through detecting the change of carbon content in the solution before and after reaction with TOC instrument. 2.4 Electrode Preparation and Electrochemical Measurement Electrochemical performance of methanol and reduced POM solution was tested in a three electrode system using cyclic voltammetry (CV), I–t curves and chronopotentiometric on the Electrochemical Workstation (CHI660D, Shanghai Chen-Hua Instruments). The schematic diagram for electrochemical test is shown in Figure 2. The working electrode was a self-made gold flake (1 cm × 1 cm) electrode. The auxiliary electrode usually use Pt (1.5 cm × 1.5 cm in this paper) electrode because it is only for the current passing to study the polarization of the electrode, which is not substantially polarization in the measurement process. Reference electrode is the Ag/AgCl electrode. The mixture of epoxy resin and ethylenediamine with the mass ratio of 9:1 was used to seal the gold flake in a plastic tube, making it only one side exposed outside to form the working electrode. The working electrode was placed in H2SO4 (0.5 mol L–1) solution and scanned within the potential range of –0.1 to 1.2 V until stable and reproducible cyclic voltammogram curves appearing, then we initialized the poten-

Fig. 2 Schematic diagram for electrochemical experiment: (1) water bath; (2) phosphomolybdic acid (POM) solution; (3) working electrode; (4) reference electrode; (5) counter electrode; (6) electrochemical workstation; (7) PC.


FUEL CELLS 14, 2014, No. 1, 56–63

Li et al.: Catalytic Oxidation of Methanol by Aqueous POM on Al2O3

3 Results and Discussion 3.1 Catalyst Characterization Results 3.1.1 H2-TPR Figure 3 shows the H2-TPR spectrum of Ru/Al2O3 and Pt/ Al2O3 catalyst with the same metal loading amount (20 wt.%). There are no peaks in the TPR profiles for c-Al2O3, which means that c-Al2O3 support could not be easily reduced. This is consistent with the results of Mitsui et al. [11]. A little peak in the TPR profiles for Pt/Al2O3 catalyst indicates that the Pt was not easily oxidized by the oxygen in air after preparation. The reduction peak of ruthenium oxide at 100 °C showed that metallic ruthenium was easily oxidized. 3.1.2 H2 Chemisorptions Table 1 gives the amount of H2 chemisorptions, dispersion, metal area, and particle size of Ru/Al2O3 and Pt/Al2O3. The amount of H2 chemisorption has a great extent increase after loading the active metal component. With the same loading, the amount of H2 chemisorption on Pt/Al2O3 is twice more than that on Ru/Al2O3, thus the dispersion is also correspondingly larger and the particle size of the metal particles is smaller. With the increasing of H2 chemisorptions, the number of active sites on Pt/Al2O3 catalyst is increased, thus the activity of Pt/Al2O3 is improved in the reaction.

Fig. 3 H2-TPR profiles of Ru/Al2O3, Pt/Al2O3 catalysts, and Al2O3 support. Table 1 H2 chemisorption, metal area, dispersion, and particle size for two kinds of catalysts.

3.1.3 The Pore Structure Generally the pores structure is an important factor to affect the performance of catalytic reaction [9]. Table 2 shows the pore features of Pt/Al2O3 and Ru/Al2O3, including specific surface area, pore volume, and average pore diameter. After loading the active metal component, the pore volume, specific surface area, and average pore diameter of Al2O3 support are reduced accordingly. The reduction of specific surface area is caused by the embedding and aggregating of active metal species, such as Pt and Ru during the process of impregnation, drying, and calcinations. Figure 4 shows the pore size distribution of Al2O3, Pt/ Al2O3, and Ru/Al2O3. As shown in Figure 4, mesoporous pores mainly exist in two catalysts. The pores are mainly concentrated at about 4–5 nm for both catalysts. In comparison, pore distribution of Ru/Al2O3 is slightly uniform at smaller pores, about 4 nm. 3.2 Activity Evaluation of Catalyst Dumesic measured the reaction rate by monitoring the color change in the POM solution with a UV/vis spectrometer operating at 500 nm. The color of pure POM solution is yellow. After POM is reduced, the solution color turns to be blue or navy blue. Figure 5 shows the color change of POM solution after reaction with methanol over Ru/Al2O3 catalyst for different time, which indicates that POM has been reduced by methanol. According to the reaction phenomena shown in Tables 3–5, there is no HCHO and HCOOH existing in the solution, but CH3OH. This may due to that CH3OH conversion does not reach the 100%, some residue CH3OH exists in the solution. In fact, gases released from the solution during reaction

Fig. 4 The pore size distribution of Al2O3, Pt/Al2O3, and Ru/Al2O3.

Table 2 Catalyst pore structure.

Catalyst

H2 chemisorption Metal area (m2 gcat–1) (lmol gcat–1)

Dispersion (%)

Particle size (nm)

Catalyst

Pore volume (ml gcat–1)

Specific surface area (m2 gcat–1)

Average pore diameter (nm)

Al2O3 Pt/Al2O3 Ru/Al2O3

0.14 63.68 28.32

– 2.42 1.92

– 2.05 50.32

Al2O3 Pt/Al2O3 Ru/Al2O3

0.43 0.28 0.21

255 198 180

6.6 6.0 4.6

FUEL CELLS 14, 2014, No. 1, 56–63

– 4.65 2.80



59


tial at –0.1 V for electrode polarization for 3 min, in order to completely remove the surface oxide which may exist [10]. The voltage scanning speed was 50 mV s–1.



Fig. 5 Color change of POM solution after reaction with methanol over Ru/Al2O3 catalyst for 10, 20, 30, and 60 min (from left to right), respectively.

Fig. 6 Methanol conversion over Pt/Al2O3 and Ru/Al2O3 catalyst for different reaction time.

Table 3 Test of solution by Fehling reagent (mixture solution of copper sulfate, potassium sodium tartrate, and sodium hydroxide). Sample

Phenomenon

HCHO Vaporized solution

Red precipitate immediately appears after heating No change in solution color, no precipitate

Table 4 Test of solution by potassium permanganate reagent. Sample

Phenomenon

HCHO CH3OH

Yellow precipitate immediately appears Initially solution color turns to be blue. After 3 min, yellow precipitate appears Initially solution turns to be blue. After 3 min, yellow precipitate appears

Vaporized solution

Table 5 Test of solution by silver mirror reagent (mixture solution of silver nitrate and ammonia). Sample

Phenomenon

HCOOH

The precipitate immediately appears at room temperature A small amount of precipitate appears after the solution is heated A small amount of precipitate appears after the solution is heated

CH3OH Vaporized solution

process. The gas was introduced into the solution of Ca(OH)2, then precipitate formed. So the reaction is expected to carry out according to the following equation. For this reason, the methanol conversion rate can be calculated through the carbon content change in the solution. CH3 OH H2 O 6PMo12 O340

catalyst

! 6PMo12 O440 CO2 6H

Figure 6 shows the methanol conversion when POM and methanol concentrations are 0.1 and 0.05 mol L–1, respectively over the different catalysts at 25 °C. The experiment results showed that the methanol conversion over Pt/Al2O3 catalyst is significantly higher than that over Ru/Al2O3. With the increase of reaction time, the difference of methanol conversion between two catalysts increases gradually. After reacted for 60 min, methanol conversion over Pt/Al2O3 catalyst is 22%, still higher than that over Ru/Al2O3, which is con-

60


Fig. 7 The influences of temperature on the conversion of methanol for different reaction time.

sistent with the results of H2 chemisorptions on two kinds of catalysts. The effect of reaction temperature on methanol conversion for the reaction over Pt/Al2O3 catalyst is showed in Figure 7. The methanol concentration is 0.05 mol L–1. As shown in the figure, at the same reaction temperature, the methanol conversion increases with increasing of the reaction time, and the growth rate gradually increases as reaction temperature increases. For the same reaction time, the methanol conversion rapidly grows as temperature increases, which indicates that reaction of POM and methanol is mainly controlled by kinetics. Figure 8 shows the relation of methanol conversion and the methanol concentration at the reaction temperature of 80 °C with the reaction time of 20 min. At the same reaction temperature and for the same reaction time, the methanol conversion slowly increases with increasing its concentration. 3.3 The Electrochemical Characteristic of Reduced POM on Au Electrode Power was generated on fuel cell anode using reduced POM reported by Dumesic, which shown that the reduced POM could be regenerated through electrochemical oxidation


FUEL CELLS 14, 2014, No. 1, 56–63


Fig. 8 The effect of concentrations on methanol conversion.

Fig. 9 The cyclic voltammogram curves of H2SO4, CH3OH, and reduced POM (H2SO4 (0.5 mol L–1), CH3OH (0.05 mol L–1)/H2SO4 (0.5 mol L–1)) and reduced POM (0.05 mol L–1)/H2SO4 (0.5 mol L–1)) on Au electrode.

FUEL CELLS 14, 2014, No. 1, 56–63


The cyclic voltammogram curves shapes of CH3OH (0.05 mol L–1)/H2SO4 (0.5 mol L–1) and H2SO4 (0.5 mol L–1) are nearly same. There were not any oxidation peaks of methanol. In fact the gold electrode has no electrochemical catalytic activity for methanol, which is consistent with references [13]. Usually, the expensive metal of Pt is used as catalyst for the oxidation of methanol. Cyclic voltammogram curve of reduced POM shows three pairs of redox reaction peaks near 0.085, 0.133, and 0.309 V. It illustrates that the redox process of POM on Au electrode is a reversible process with multi-electron transfer. The electrochemical reaction equation for three pairs of redox peaks can be expressed as following [14]: PMo12 6 O40 3 2e , PMo10 6 Mo2 5 O40 5 PMo10 6 Mo2 5 O40 5 2e , PMo8 6 Mo4 5 O40 7 PMo8 6 Mo4 5 O40 7 2e , PMo6 6 Mo6 5 O40 9 Comparing the electrochemical redox reaction process of methanol with POM, the redox peak areas and current densities of POM were significantly higher than those of methanol, which indicates that POM has better energy-storage capability with the existence of Au electrode. The cyclic voltammogram curves of reduced POM (0.05 mol L–1)/H2SO4 (0.5 mol L–1) at 25, 40, 60, and 80 °C, respectively are given in Figure 10. Though, at high temperature of 80 °C, there is some fluctuation in the curve, but change is not enough to influence the result because the variation is far less than that in compared redox peaks. When the solution temperature increases at the low temperature (