Parallel pathways of ethanol oxidation: The effect of ethanol concentration

Journal of

Electroanalytical Chemistry Journal of Electroanalytical Chemistry 578 (2005) 315–321 www.elsevier.com/locate/jelechem

Parallel pathways of ethanol oxidation: The effect of ethanol concentration G.A. Camara, T. Iwasita

*

Instituto de Quı´mica de Sa˜o Carlos/USP, C.P. 780, 13560-970 Sa˜o Carlos, SP, Brazil Received 6 October 2004; received in revised form 18 January 2005; accepted 19 January 2005 Available online 16 February 2005

Abstract In this work, we investigate the effects of ethanol concentration on the yields of CO2, acetic acid and acetaldehyde as electrooxidation products. FTIR spectra show that the main oxidation product at low ethanol concentrations is acetic acid, CO2 being produced to a minor extent. For concentrated ethanol solutions, the pathway producing acetaldehyde becomes dominant. The CO2 production passes through a maximum at 0.025 M C2H5OH and then decreases. The intensity of the IR absorption band for COad, which is one of the adsorbed intermediates, is practically independent of ethanol concentration.
2005 Elsevier B.V. All rights reserved. Keywords: Ethanol electrooxidation; FTIR; Ethanol concentration; Electrocatalysis

1. Introduction The electrochemical oxidation of ethanol on platinum electrodes has been a subject of permanent interest in the last twenty years [1–7]. Ethanol, being a renewable fuel, would be a candidate for fuel cell applications. Such interest is justified by the energy content of ethanol, which corresponds to 12 e
per molecule for total ethanol oxidation. However, the total conversion of ethanol to CO2 is the central challenge in the electrocatalysis of this alcohol [1,3,5]. Several spectroscopic and electrochemical techniques have been applied to elucidate the mechanism of ethanol electrooxidation on Pt-based materials [2,3,8–10]. Ethanol undergoes parallel reactions, producing acetaldehyde, acetic acid and CO2 as shown in Fig. 1, where the respective number of electrons for the different path* Corresponding author. Tel.: +55 16 33739950; fax: +55 16 33739952. E-mail addresses: [email protected] (G.A. Camara), iwasita@ iqsc.usp.br (T. Iwasita).

0022-0728/$ - see front matter
2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2005.01.013

ways are indicated. It was shown in recent studies using Pt, Pt–Sn and Pt–Re as catalysts [11–13] that ethanol oxidation pathways are sensitive to the catalysts nature and bifunctional mechanism and ligand effect have been suggested as responsible for the gain in the catalytic activity. On the basis of FTIR spectra, the yield of products in 0.1 M ethanol solution has been reported in early papers by Weaver and co-workers [2,3]. The polluting nature of acetaldehyde and the ‘‘end-product’’ character of acetic acid are disgusting factors in this system and, therefore, the search for catalysts to produce selectively oxidation to CO2 is a major goal of research activities on ethanol today. However, to our knowledge, no quantitative and systematic investigation of the effect of ethanol concentration on the parallel reactions was reported until now, although it was early recognized that the relative amount of products depends on ethanol concentration [9,14]. This issue is the main concern of the present work. We present here a systematic study on the concentration-dependence of the yield of products on polycrystalline Pt, as measured from in situ FTIR spectra. The

316

G.A. Camara, T. Iwasita / Journal of Electroanalytical Chemistry 578 (2005) 315–321

CO2 -

12e

-

10e 2e C2H5OH

dow material has the advantage of being transparent to IR radiation down to ca. 800 cm
1. Spectra were computed from the average of 30 interferograms. The spectral resolution was set to 8 cm
1.

-

CH3CHO -

3. Results and discussion

2e -

3.1. Potentiodynamic data

Fig. 1. Schematic representation of the parallel pathways for ethanol oxidation.

study is complemented with corresponding data on the total current for ethanol oxidation at a constant potential of 0.5 V.

2. Experimental The working electrode was a smooth polycrystalline Pt disk with a geometric surface area of 0.78 cm2. The real surface area of this electrode was determined by integration of the charge in the hydrogen region of the cyclic voltammogram. This area was used in the calculation of the current densities and the amounts of products measured along this work. The counter electrode was a platinum sheet with 1.2 cm2. All potentials were measured against a reversible hydrogen electrode in the same electrolyte. Before each experiment, the Pt disk was flame annealed for 1 min and cooled in an argon atmosphere. The electrode was immediately transferred to the electrochemical cell containing 0.1 M HClO4. The state of the electrode surface was checked by cyclic voltammetry in a meniscus configuration, between 0.05 and 1.4 V with a scan rate of 50 mV s
1. Then, the potential was kept at 0.05 V and ethanol (J. T. Baker) was admitted in the cell to reach the given concentration. All the experiments were performed at room temperature (25 ± 1.0 C). The solutions were prepared with Milli-Q water (18.2 MX cm) and deaerated with N2 (4.6) or Argon (4.8). FTIR spectra were measured using ethanol at different concentrations in the range between 10
2 and 1.0 M using 0.1 M HClO4 as supporting electrolyte (Alfa Aesar) p.a. The cell used for these experiments is described in detail elsewhere [15]. Reflectance spectra were calculated as the ratio (R/Ro) where R represents a spectrum at the sample potential and Ro the spectrum collected at 0.05 V. Positive and negative bands represent, respectively, the consumption and production of substances at the sample potential. The electrochemical IR cell was fitted with a planar ZnSe window. This win-

Fig. 2 shows cyclic voltammograms recorded at 50 mV s
1 for two ethanol concentrations. Two oxidation peaks, centered at ca. 0.8 and 1.3 V, are observed during the positive-going scan and a reactivation peak, which becomes more prominent in the 1.0 M solution, is observed on the negative-going scan. For all concentrations studied in this work, in situ FTIR spectra were measured at different potentials. As an example, the series of spectra for the 1.0 M solution is shown in Fig. 3. First, a reference spectrum Ro was measured at 0.05 V and then sample spectra were collected after applying successive potential steps of 50 mV up to a potential of 1.0 V. The most important features related to ethanol oxidation products are listed in Table 1. It is worth noting that the band for CO2 (2343 cm
1) is a relatively weak feature in the spectra, even at the highest potentials. Taking into account that CO2 has the highest absorptivity from all soluble products (see below), the weak band intensity indicates a low level of CO2 formation. This was observed along the series of concentrations studied in the present work. The band observed at 933 cm
1 is the only acetaldehyde feature not superimposed on other bands and will, therefore, be used for quantitative purposes [2,3]. The

5

-2

CH3COOH

Current density / mA cm

4e

4

3

2

1

0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Potential / V vs. RHE

Fig. 2. Cyclic voltammograms at 50 mV s
1 for the electrooxidation of: 0.1 M (dashed line) and 1 M of ethanol (full line) on smooth polycrystalline Pt. Sweeps are from 0.05 to 1.45 V vs. RHE. T = 25 C.

G.A. Camara, T. Iwasita / Journal of Electroanalytical Chemistry 578 (2005) 315–321 40

0.10 0.35 0.40 0.50

317

(a)

30 20

0.60 0.70

10

0.80

0.5 M 1.0 M 0.1 M 0.01 M

0 0

500

1000

1500

Time / s

0.90

10

1357

2050

(b) 8

1.00 933

6

2343 1280

4

R/Ro = 1%

2 1720

2500

2000

0 1500

1000 -1

Wavenumber/ cm

Fig. 3. In situ FTIR spectra for a polished Pt electrode at different potentials as indicated. 1.0 M C2H5OH + 0.1 M HClO4. Reference spectra taken at 0.05 V. Sample spectra: 100 interferometer scans measured after applying successive potential steps each of 50 mV.

0.0

0.3 0.6 0.9 EtOH concentration / M

1.2

Fig. 4. (a) Chronoamperometric curves for polycrystalline Pt after application of a potential step from 0.05 to 0.5 V in 1.0 M C2H5OH + 0.1 M HClO4. (b) Current density (referred to real surface area) for ethanol oxidation taken from the curves in (a), after 20 min of polarization, plotted as a function of the concentration of ethanol. T = 25 C.

same is valid for the band at 1280 cm
1, corresponding to acetic acid. 3.2. Quasi-stationary current data: Dependence on ethanol concentration The concentration dependence of the current for ethanol oxidation was measured at a constant potential of 0.5 V. Chronoamperometric curves, obtained after applying a potential step from 0.05 to 0.50 V, are shown in Fig. 4(a). The current densities refer to the real surface area as determined from the charge in the H-adsorption region of the voltammogram. The initial current decay can be addressed to a great extent to the increasing surface coverage with partially oxidized intermediates [8]. Most of the currents present a quasi-sta-

Table 1 Band assignment for ethanol oxidation products Frequency (cm
1)

Species

References

2343 2050 1720 1280 1357 933

CO2 (assym. str.) COL (C–O str.) –CO str. (carbonyl group) Acetic acid (–COOH group) Acetaldehyde (CH3 sym. bend) Acetaldehyde (C–C–O assym. str.)

[16] [18] [16] [1,16] [17] [3]

tionary behavior after 20 min of polarization; the current density for this time is plotted against the ethanol concentration in Fig. 4(b). From repetitive measurements, we estimate that current values have a maximum variance of ca. ±10%. In Fig. 4(b) we observe an increase of current with ethanol concentration, showing up a maximum at ca. 0.5 M. The reason for this behavior can be understood on the basis of the dependence of the yields of products on ethanol concentration, which were calculated from FTIR measurements as shown below. 3.3. Quantification of soluble products In situ FTIR spectra were obtained in solutions of different ethanol concentration at a constant potential of 0.5 V. Some selected spectra obtained after 10 min of polarization at 0.5 V are shown in Fig. 5(a) and (b) shows, enlarged, the band for linearly bonded CO. Interestingly, the latter exhibits a constant intensity, independently of ethanol concentration. At a first glance, adsorbed COL (2056 cm
1), CO2 (2343 cm
1) and acetic acid (1280 cm
1) are present for all concentrations. Acetaldehyde (933 cm
1) was not detected at the lowest concentrations (0.01–0.05 M).

318

G.A. Camara, T. Iwasita / Journal of Electroanalytical Chemistry 578 (2005) 315–321

0.01 M

0.05 M 0.025 M

0.10 M

0.05 M

0.25 M 0.10 M

1861

1280

0.50 M

0.50 M 2050

1M

1.0 M

2343

2050

1354

933

R/Ro = 1 % R:Ro = 0.3 %

2500

2000 1500 -1 Wavenumber/ cm

1000

2200

2000

1800

Wavenumber / cm

-1

(b)

(a)

Fig. 5. (a) In situ FTIR spectra taken after 10 min of polarization at 0.5 V vs. RHE. Polycrystalline Pt in solutions of different ethanol concentrations as indicated. Base electrolyte 0.1 M HClO4. Background spectrum collected at 0.05 V. Room temperature (25 C). (b) Enlarged view of the band for linearly bonded CO.

Q ¼ Ai =eeff :

ð1Þ

40

Ald. -2

30

Q / nmol cm

The individual yields of soluble products were calculated after evaluation of the respective integrated band intensities. Values of the effective absorption coefficient, eeff, were taken from the work of Weaver and co-workers [2,3]; these were: 3.5 · 104, 5.8 · 103 and 3
1
2 2.2 · 10 M cm for CO2, acetic acid and acetaldehyde, respectively. The amount (Q/mol cm
2) of a given species inside the thin layer cavity follows the relationship

20

HAc 10

CO2x10 0

0.0

Values of Q were evaluated from time-resolved spectra taken after applying a potential step from 0.05 to 0.5 V. The values obtained after 10 min of polarization are plotted in Fig. 6 as a function of ethanol concentration. In order to check any time-dependent effect, either inherent to the electrochemical process or due to the possible diffusion off the thin layer, the evaluation was made also in spectra taken at shorter times. The behavior observed for all times was similar to that shown in Fig. 6. The yields of CO2 are very low for all concentrations; observe that Q values for CO2 in Fig. 6 are 10-fold magnified. After a sharp initial increase at low concentra-

0.2

0.4

0.6

0.8

1.0

Concentration / M

Fig. 6. Amounts (Q) of CO2, acetic acid and acetaldehyde after 10 min of polarization at 0.5 V for different concentrations of ethanol in 0.1 M HClO4. CO2 (d), acetic acid (s) and acetaldehyde (m), respectively.

tions, the CO2 production goes through a maximum for 0.025 M C2H5OH and then slowly decays as the ethanol concentration increases. A comparison with Fig. 5(b) shows that the changes in QCO2 are not accompanied by corresponding changes in the band intensity for adsorbed CO, i.e., the surface coverage with CO seems to be somehow buffered in this system.

G.A. Camara, T. Iwasita / Journal of Electroanalytical Chemistry 578 (2005) 315–321

For concentrations of ethanol below 0.1 M, acetic acid is the major oxidation product (six times higher than CO2). Also, acetic acid exhibits a maximum yield for ethanol concentrations between 0.1 and 0.4 M and then decays. First detectable signal for acetaldehyde was observed at 0.1 M C2H5OH. There is an intrinsic difficulty in detecting acetaldehyde, due to its low absorptivity; but considering the rapid growth of the amount of acetic acid at low ethanol concentrations, it could be possible that the absence of acetaldehyde under these conditions may be caused by its further conversion to acetic acid. But, in fact, this is just a hypothesis. Although it is well known that oxidation of acetaldehyde yields mainly acetic acid [17], it is not possible to establish by simple inspection of Fig. 6 whether the pathway forming acetic acid requires a previous formation of acetaldehyde. If this were so, then we could state that from 0.1 M onwards only a small, constant, fraction of acetaldehyde is converted to acetic acid. An alternative pathway is thinkable, where some intermediate of adsorbed ethanol forms acetic acid directly. This possibility is sketched in Fig. 1. Above 0.1 M C2H5OH, acetaldehyde is the only product increasing with increasing ethanol concentration, up to an almost constant maximum at 0.5 M C2H5OH. 3.4. Adsorbed ethanol and oxidation mechanism A remarkable issue in the hitherto presented results is the fact that both products requiring additional oxygen for the oxidation, namely CO2 and CH3COOH, exhibit a maximum for the dependence of Q on the reactant concentration. This behavior can be rationalized via a Langmuir–Hinshelwood (LH) mechanism, which is well established for the oxidation of adsorbed CO: COad þ OHad ! CO2 þ Hþ þ e


319

being strongly bound to the surface. As discussed earlier, all these intermediates can be only oxidized to CO2 and none of them could be identified as the adsorbed intermediate forming acetaldehyde or acetic acid [8]. As an interesting fact in the foregoing discussion, some stripping experiments for adsorbed ethanol are shown in Fig. 7. For these experiments, ethanol was admitted in a cell of small volume (ca. 2 ml) at a constant potential of 0.35 V. After 10 min of adsorption, the solution was exchanged by pure supporting electrolyte (0.1 M HClO4) and a CV was started towards positive potentials. As a control for the absence of bulk ethanol during the stripping of adsorbates, the second scan always exhibited the same current as the base electrolyte. According to Fig. 7, strongly adsorbed species are oxidized in two potential regions, as already shown in an earlier paper [8]: a well-defined peak at ca. 0.65 V due to adsorbed CO and a broad oxidation peak overlapped on the Pt oxide region, due to H-containing adsorbates. Unfortunately, no data were obtained with higher ethanol concentrations due to the increasing difficulty to eliminate residual ethanol from the bulk; but, for the three ethanol concentrations shown in the figure (from 5 · 10
4 to 10
2 M) there are no pronounced differences in the charge for stripping the adsorbates. If this result is extrapolated to higher ethanol concentrations, we can consider that the surface is saturated with strongly adsorbed species for all bulk concentrations used here. Recall that the coverage with COad is practically constant as shown in Fig. 5(b). Returning to our interpretation of the concentration dependence of the pathway forming CO2 and acetic acid, we now state that not only strongly adsorbed species but

ð2Þ

For acetic acid, a corresponding step should involve adsorbed OH and some adsorbed intermediate, not yet identified. In a LH mechanism, the rate of reaction is expected to depend on the coverage with both adsorbed reactants. Thus, a reaction rate passing through a maximum for a given reactant concentration would indicate that the reaction partners (adsorbed organic species and adsorbed water) compete for active surface sites. For the organic species, the coverage could depend on the bulk concentration of the reactant. But the present spectroscopic data (Fig. 5(b)) indicate a constant band intensity for adsorbed CO, which is the intermediate for CO2 formation, at least at the potential1 of 0.5 V. Obviously, the present case involves a surface covered with a number of other adsorbed species from ethanol, some of them 1 At higher potentials, namely in the Pt oxide region, other strongly absorbed intermediates of ethanol undergoes oxidation to CO2 [8].

50

0

-50

HClO4 0.1 M -4

5 x 10

-3

1 x 10

-100

-2

1 x 10

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Potential / V vs. RHE Fig. 7. Stripping CV for strongly adsorbed species formed after ethanol interaction with polycrystalline Pt at 0.35 V. Sweep rate 20 mV s
1; electrolyte 0.1 M HClO4. (–––) Clean Pt surface; ethanol concentrations: 5 · 10
4 (- - -); 10
3 (  ); 10
2 M ( - ).

320

G.A. Camara, T. Iwasita / Journal of Electroanalytical Chemistry 578 (2005) 315–321

also weakly bonded intermediates are present at the Pt surface. Whatever its nature, also weakly adsorbed species contribute to blocking surface impeding OH formation and inhibiting those pathways that require additional oxygen atoms for oxidation. The behavior of Q for acetaldehyde (Fig. 6) indicates that the corresponding pathway is less site demanding. A similar behavior was observed recently for the production of formaldehyde during methanol oxidation [19] and was interpreted in terms of a reverse Eley–Rideal (ER) mechanism [20]. Such a mechanism could also rationalize the results for acetaldehyde obtained here. 3.5. The charge involved in the different pathways Since appreciable differences of charge are involved in the formation of the soluble products, further information on the pathways of ethanol oxidation can be envisaged via the calculation of the total charges on the basis of 12, 4 and 2 e
per molecule for CO2, acetic acid and acetaldehyde, respectively. These values were added to get the total charge plotted in Fig. 8(a) and the individual charge yield of each product plotted in Fig. 8(b). The curve for the total charge (Fig. 8(a)) is in good agreement with the current-concentration dependence of Fig. 4 as the charge passes through a maximum at

Total charge / nF cm-2

150

100

50

(a)

100 0.0

0.2

0.4

0.6

80

0.8

1.0

Ald.

Charge %

HAc 60 40

(b) 20

CO2

0 0.0

0.2

0.4

0.6

0.8

1.0

EtOH conc./ M

Fig. 8. (a) Calculated total charge based in the yields of soluble products of Fig. 6 as a function of ethanol concentration. (b) Percentual contribution of each product to the total charge.

about 0.4 M C2H5OH. With respect to the pathway contributions (Fig. 8(b)), it can be observed that for the lowest concentration studied here, 0.01 M C2H5OH, CO2 and acetic acid are, respectively, responsible for 31% and 69% of the total charge produced. As acetaldehyde begins to be produced (when C2H5OH = 0.1 M), the contribution of CO2 and acetic acid clearly decays. Above 0.4 M, most of the reaction occurs via acetaldehyde, which is responsible for 86% of the total charge for 1.0 M C2H5OH. The search for catalysts to produce selectively CO2 is a major goal of research activities on ethanol today. The results discussed hitherto show the power of infrared spectroscopy to monitor, in situ, the parallel oxidation pathways of ethanol. Moreover, IR data allow to reproducing, qualitatively, the electrochemical results. This approach is useful for the analysis of new catalysts for ethanol oxidation and, in general, to study processes involving parallel reaction paths. A similar work on a PtRu catalyst is in preparation and will be published in due course [21].

4. Concluding remarks  On polycrystalline platinum electrodes in 0.1 M HClO4, the current for ethanol oxidation increases with the reactant concentration up to a maximum of ca 8 lA cm
2 (referred to real surface area) for a C2H5OH concentration of about 0.5 M. Higher C2H5OH concentrations produce lower current densities.  The contributions of the different pathways, producing CO2, acetic acid and acetaldehyde, are strongly dependent on ethanol concentration.  Higher yields for acetic acid and, to a much lower extent, for CO2 are obtained only at low concentrations of ethanol (below 0.1 M), though the current density for these conditions is much too low (ca. 20% of the maximum).  Above 0.1 M, the pathways forming CO2 and acetic acid undergo a significant inhibition, probably due to the limited availability of free sites to adsorb water, which is the O-donor in the respective oxidation process (LH mechanism).  Acetaldehyde was not detected at ethanol concentrations below 0.1 M. But for higher ethanol concentration this product is greatly favored, being the main product for ethanol concentrations higher than 0.2 M. Due to the lowering of the charge per ethanol molecule, this pathway causes a large fall of efficiency.  The in situ infrared method shows up as an efficient method to study complex mechanism involving parallel pathways.

G.A. Camara, T. Iwasita / Journal of Electroanalytical Chemistry 578 (2005) 315–321

Acknowledgments The authors thank FAPESP and CAPES for financial support.

References [1] X.H. Xia, H.-D. Liess, T. Iwasita, J. Electroanal. Chem. 437 (1997) 233. [2] L.-W.H. Leung, S.-C. Chang, M.J. Weaver, J. Electroanal. Chem. 266 (1989) 317. [3] P. Gao, S.-C. Chang, Z. Zhou, M.J. Weaver, J. Electroanal. Chem. 272 (1989) 161. [4] L.-W.H. Leung, A. Wieckowski, M.J. Weaver, J. Phys. Chem. 92 (1988) 6895. [5] V.M. Schmidt, R. Ianniello, E. Pastor, S.R. Gonza´lez, J. Phys. Chem. 100 (1996) 17901. [6] J.F.E. Gootzen, W. Visscher, J.A.R. van Veen, Langmuir 12 (1996) 5076. [7] G.A. Camara, R.B. de Lima, T. Iwasita, Electrochem. Commun. 6 (2004) 812. [8] T. Iwasita, E. Pastor, Electrochim. Acta 39 (1994) 531.

321

[9] J.P.I. de Souza, J.B. Rabelo, I.R. de Moraes, F.C. Nart, J. Electroanal. Chem. 420 (1997) 17. [10] J.P.I. de Souza, S.L. Queiroz, K. Bergamaski, E.R. Gonzalez, F.C. Nart, J. Phys. Chem. B 106 (2002) 9825. [11] F. Vigier, C. Contanceau, A. Perrard, E.M. Belgsir, C. Lamy, J. Appl. Electrochem. 34 (2004) 439. [12] C. Lamy, S. Rousseau, E.M. Belgsir, C. Contanceau, J.M. Leger, Electrochim. Acta 22–23 (2004) 3901. [13] F. Vigier, C. Contanceau, F. Hahn, E.M. Belgsir, C. Lamy, J. Electroanal. Chem. 563 (2004) 81. [14] T. Iwasita, B. Rasch, E. Cattaneo, W. Vielstich, Electrochim. Acta 34 (1989) 1073. [15] T. Iwasita, F.C. Nart, in: H. Gerischer, Ch.W. Tobias (Eds.), Advances in Electrochemical Science and Engineering, vol. 4, VCH, Weinheim, 1995. [16] G. Socrates, Infrared Characteristic Group Frequencies, Wiley, New York, 1966. [17] B. Rasch, T. Iwasita, Electrochim. Acta 35 (1990) 989. [18] A. Bewick, K. Kunimatsu, S. Pons, Electrochim. Acta 25 (1980) 465. [19] E.A. Batista, G.R.P. Malpass, A.J. Motheo, T. Iwasita, J. Electroanal. Chem. 571 (2004) 273. [20] R.I. Masel, Principles of Adsorption and Reaction on Solid Surfaces, Wiley, New York, 1996. [21] G.A. Camara, R.B. de Lima, T. Iwasita, in preparation.