3950
J. Phys. Chem. A 2010, 114, 3950–3955
Potential of Nano-ZnS as Electrocatalyst† Michael Bredol* and Michał Kaczmarek Department of Chemical Engineering, Fachhochschule Mu¨nster - UniVersity of Applied Sciences, Steinfurt, Germany ReceiVed: July 31, 2009; ReVised Manuscript ReceiVed: September 12, 2009
ZnS is a versatile wide-gap semiconductor that shows remarkable chemical stability against oxidation and hydrolysis. These properties are retained when the particle size steps down to 10 nm and below, and thus ZnS nanoparticles are interesting entities for optical and catalytic functions, where they might be exposed to quite aggressive environments. Moreover, ZnS may be doped by a host of elements, so that the electrical and optical properties can be tuned over an extended range. Neither Zn nor S are in short supply, and ZnS as compound or in view of the constituting elements is not toxic. Therefore, functional materials based on ZnS will be ideal for “green chemistry” applications, since they are suited also for large scale manufacture and will pose no additional problems when released into the environment. The article describes first results for ZnS-nanoparticles acting as electrocatalysts, a field that has not been covered so far by ZnS-derived materials. The catalytic action is evaluated for the decomposition of ethanol, a potentially abundant fuel for mobile electricity generation, since it can be fabricated by fermentation from a broad range of organic materials. Introduction Semiconductors are defined by the width of their energy gap and their associated conductivity properties. Apart from the classical use in solid state electronics, they find application, e.g., in catalysis (since they can store and shuttle electrons or holes in redox systems) or photovoltaics (because they may either generate electron/hole pairs or accept electrons or holes during the process of charge separation). Regarding “Green Chemistry”, such functions are central, since they promise smart ways for the fabrication of essential chemicals, tailored electrocatalysts for batteries and fuel cells, or cheap and mobile electricity generation. However, up to now catalytic functions in electrochemistry traditionally have been mostly the field of precious metal nanoparticles, typically from the platinum group. These particles may have high catalytic activity but suffer from several problems, which often prevent their actual use. For instance, most precious metal particle catalysts are poisoned either by CO or H2 and thus can only work with highly purified source materials in a reliable fashion. Energy sources like raw biogenic methane can thus not be employed in fuel cells based on such electrocatalysts. Electrocatalytic semiconductors, especially those based on chalcogenides would circumvent such problems, due to the lack of specific affinity to CO and the abundant presence of chalcogenide functions in the catalyst itself. However, only semiconductors with a wide band gap are suited in this context, since they offer the necessary stability against redox activity that is needed for catalytic operation over many cycles. Typical examples of this kind are ZnO, TiO2, and ZnS, as demonstrated by the fact that they are or have been used as pigments or UV filters. Their photocatalytic activity under diverse environmental conditions is well described, but there is only very spase information about their activity in electrocatalytic applications. On the other hand, heterogeneous catalytic particles need to be † Part of the special issue “Green Chemistry in Energy Production Symposium”. * Corresponding author. E-mail:
[email protected].
Figure 1. Schematic representation of ZnS nanoparticles in mesoporous carbon in contact with ethanol-containing electrolyte.
nanodispersed in most cases to present a surface large enough for high conversion rates, and fortunately the preparation and immobilization of semiconductor nanoparticles has made large progress, so that quite a broad range of procedures is now available from the literature. In the presence of nano-ZnS, reports claimed the reduction of CO2 under irradiation with light,1 down to carbonic acids.2 The mechanism was explained by the high reductive potential of electrons in ZnS after photoexcitation (-1.75 V relative to a standard hydrogen electrode SHE3). Accordingly, holes would have a (oxidative) potential of about +1.85 V relative to SHE. Of course, electrons and holes injected by electrochemical processes will have similar properties. On the other hand, very efficient photooxidation was observed in the presence of nanoZnS agglomerates4,5 together with approaches to mass production of suited mesoporous aggregates.6 Similar observations were made in trials to generate prebiotic organic molecules on ZnS surfaces.7 These examples show that ZnS indeed is catalytically active in redox reactions with organic components, and thus its
10.1021/jp907369f 2010 American Chemical Society Published on Web 10/09/2009
Nano-ZnS as Electrocatalyst
J. Phys. Chem. A, Vol. 114, No. 11, 2010 3951
TABLE 1: Electrochemistry of Ethanol Oxidation, Formulated as Half-Reaction in Acidic Solution, Together with Associated Standard Reduction Potentials process
∆RGL/(kJ mol-1)
EL/V
C2H5OH f CH3CHO + 2H+ + 2eCH3CHO + H2O f CH3COOH + 2H+ + 2eCH3COOH + 2H2O f 2CO2 + 8H+ + 8eC2H5OH + 3H2O f 2CO2 + 12H+ + 12e-
46.66 -24.65 75.44 97.46
0.242 -0.128 0.098 0.084
TABLE 2: Free Enthalpies of Combustion in O2 for Ethanol and Its Intermediates in Electrochemical Oxidation process
∆cGL/(kJ mol-1)
C2H5OH + 0.5O2 f CH3CHO + H2O CH3CHO + 0.5O2 f CH3COOH CH3COOH + 2O2 f 2CO2 +2H2O C2H5OH + 3O2 f 2CO2 + 3H2O
-190.5 -268.3 -849.8 -1308.6
potential as electrocatalyst on the surface of electrodes was the subject of the present study. In this work, we will explore the usability of ZnS nanoparticles as electrocatalysts for direct conversion of ethanol in fuel cells. Such catalysts are needed urgently, since ethanol can be made from a broad variety of organic materials by fermentation and, at present, a lot of effort is invested into processes that harvest complete plants (including their cellulose fraction), converting them to bioethanol. Available electrocatalytic systems on the basis of precious metals do not work well here, since they are quickly poisoned by decomposition products or biogenic impurities. The best systems in this respect are complicated, often ternary structures.8-10 Problems with the electrocatalysts also hamper the introduction of direct methanol fuel cells and prevent them from delivering electricity with the efficiency possible from their thermodynamic potential. As a consequence, fuels often have to be re-formed on board to hydrogen to feed a conventional fuel cell, which adds losses and makes the systems much more complicated than necessary. Problems like this will also occur with higher alcohols like butanol, which would be preferred over ethanol because of their higher energy density. Electrocatalytic mineralization of ethanol with O2 to H2O and CO2 is a very demanding task; it includes the transfer of 12 electrons per molecule of ethanol and proceeds via two quite stable intermediates, namely acetaldehyde and acetic acid (Table 1). In a realistic setting (pH approximately 2, diluted alcoholic fuel), the electrode potential will differ from the standard potential by some millivolts, but the general thermodynamic relations will be unchanged. For full decomposition, the intermediate species must not diffuse away from the electrocatalytically active surface but rather be kept in close contact. Thus, a suitable catalyst support needs to be mesoporous; the pores then need to accommodate the catalytic particles as well as all intermediates. For efficient use of the energy contained in ethanol, it will definitely not be sufficient to exploit only the first two steps of oxidation to acetic acid, since they carry only 35% of the total energy. Table 2 shows this from the free enthalpies of combustion associated with the individual steps; acetic acid thus needs to be decomposed as well as ethanol and acetaldehyde. Although ZnS as such can be prepared as a mesoporous material, it will not be suited as electrocatalyst in its pure form, due to its poor electrical conductivity. In Figure 1, a suitable composite structure is sketched schematically, together with the species to be transported and those to be kept inside the pores during operation. The support of course needs to be electrically
conductive, but chemically inert; thus carbons will be most useful here. The preparation of mesoporous carbons with well designed pore structures has made large progress in the past years; meanwhile there are many potentially suitable ways to generate carbon supports with well-defined porosity, e.g., by templating through lyotropic mesophases and so-called “nanocasting”.11 However, most of these carbons are amorphous in nature, whereas graphitic carbon would be desired because of its high electrical conductivity,12 which unfortunately only forms at very high temperature. But also in this respect, progress has been made, and there are reports about graphitic carbon with high surface area.13 For demonstration purposes, in the experiments described in this article the support for ZnS nanoparticles was either conductive glass or commercial carbon paste, partially mixed with activated carbon. Of course, such electrodes will not contain pores with well-defined size and order, and thus future work will transfer the results from such ill-defined supports to mesoporous carbons with specially designed mesostructures. Competing with ethanol oxidation, ZnS particles may be degraded by oxidation or reduction. Typical processes can be associated with an electrochemical standard potential as well (Table 3). However, the standard potentials of all oxidative processes possible are higher than for ethanol oxidation, whereas reduction only occurs at very negative potentials. Therefore, ZnS should always be stable in the presence of ethanol against oxidation or reduction. This means that ZnS-based electrocatalysts will need to operate always in the presence of fuel, never without, to protect them against decomposition. Moreover, the redox states of ZnS at standard potentials between oxygen reduction and ethanol oxidation may provide intermediate states for electron transfer, as desired in redox catalaysts. Another potential degradation mechanism for ZnS (nano-) particles is dissolution in acid, since useful electrolytes and membranes in a fuel cell typically will be acidic. The relevant thermodynamical data are tabulated in Table 4. According to these data, the activity of Zn2+ in equilibrium at pH ) 2 is smaller than 3 × 10-6; therefore, ZnS can be considered to be stable against hydrolysis down to pH values of 2. A more general problem in this context is the necessary ligands, protecting the nanoparticles against agglomeration and providing the necessary functionalities for interfacing with the matrix. In our case, the environment of the ZnS particles will be acidic; therefore, the particles should be hydrophilic to guarantee close contact with the surrounding medium. To facilitate charge transfer under these conditions, a ligand as small as possible was chosen, namely cysteamine (2-aminoethanethiol), which adsorbs readily on Zn-rich particle surfaces and generates nanopowders with excellent redispersibility in acidic solution (due to the basic amino group). The ligand sphere thus will be extremely thin, hopefully not limiting the charge transfer from solution to solid. The general ability of stabilized ZnS nanoparticles to exchange charge even in the presence of a large
3952
J. Phys. Chem. A, Vol. 114, No. 11, 2010
Bredol and Kaczmarek
TABLE 3: Thermodynamics of ZnS Oxidation and Reduction, Formulated as Half-Reacion in Acidic Solution, Together with Associated Standard Reduction Potentials process
∆RGL/(kJ mol-1)
EL/V
H2O + ZnS f ZnO + S + 2H+ + 2e3H2O + ZnS f ZnO + SO2 + 6H+ + 6e4H2O + ZnS f Zn2+ + SO42- + 8H+ + 8eZnS + 2H+ + 2e- f Zn + H2S
120.12 294.19 258.22 173.46
0.62 0.51 0.33 -0.90
TABLE 4: Thermodynamics of ZnS Hydrolysis in Acidic Solution process +
ZnS + 2H f Zn + H2S ZnS + H+ f Zn2+ + HS2+
∆RGL/(kJ mol-1)
K298
26.40 66.31
2.36 × 10-5 2.38 × 10-12
ligand sphere has been demonstrated recently by our lab in the context of photovoltaic activity,14 although at low current density. From the literature, it is well-known that not only the band gap energy but also the absolute positions of valence band and conduction band with respect to vacuum energy are a function of the particle size.15 On the other hand, the ZnS particles employed need to be nanosized anyhow, since they need to be accommodated in carbon support and need to expose a catalytic surface as large as possible to the liquid phase. If necessary, adjustment of levels will therefore be possible not so much by size control, but rather by alloying the nanoparticles with ZnSe or other dopants. Energy shifts to be expected by such measures are documented in the literature,16 also in view of electrochemical alignment. The ZnS particles used in this study were doped with manganese, since luminescence from Mn2+ ions allows us to check for the (persistent) presence of ZnS nanoparticles. The spectral position of Mn2+ emission is not affected much by the size of the hosting particle, in contrast to the excitation spectra, due to the inset of (weak) quantum confinement. Therefore, the size of the catalytic particles can be estimated from excitation spectra,17 also in powders, deposited layers or after usage. Experimental Section ZnS nanoparticles were made in a conventional manner taken from the literature (arrested precipitation). In a typical procedure, zinc acetate (Sigma-Aldrich), manganese acetate (Merck), and cysteamine (Merck) were dissolved in deionized water in molar ratios of 1:0.19:0.87, with Zn concentrations of approximately 0.8 mol/L. Na2S solution (Riedel-de-Haen) was injected under stirring into this solution (in under-stoichiometric total amount with a Zn/S ratio of ca. 1.8:1). The resulting turbid dispersion was refluxed for 3 h; the nanoparticles then could be precipitated from the clear dispersion by addition of ethanol and were isolated by centrifugation and drying. Particle sizes were measured either directly after synthesis or after redispersion of powder in water, by dynamic light scattering (DLS) in a Malvern High Performance particle sizer. Optical spectra were recorded on a Shimadzu fluorescence spectrometer at room temperature; powders were dispersed in water and measured in cuvettes. Electrodes without carbon support were prepared from a dispersion of ZnS nanoparticles in ethanol (0.3 g in 2 mL ethanol, ultrasonicated), flooding an ITO substrate (15 cm2, previously cleaned with ethanol) with this dispersion, and subsequent drying in air. To improve adhesion, some electrodes were baked at 170 °C in air. Alternatively, carbon-supported electrodes were prepared from copper plates of the same size (previously cleaned in
Figure 2. Schematic of experimental fuel cell using a FeCl3/FeCl2 counter electrode.
hydrochloric acid and ethanol). Activated carbon powder (Merck, typical particle size 20 µm) was dried in vacuum at 200 °C for 2 h. A mixture of 0.32 g of dry carbon powder and ZnS nanoparticles (0.10 g in water, 1 mL) was prepared and dried. The resulting ZnS-loaded carbon powder was mixed with 1.0 g of printable carbon paste (DuPont, diluted with 1 mL of tetrahydrofuran) and applied to the copper plate by tape-casting. The final step of the procedure was drying and baking at 170 °C in air. As a control, this procedure has been used as well on ITO-coated glass (cleaned with EtOH) instead of copper, leading essentially to the same results. All electrochemical measurements were made at room temperature, using a Zahner IM6 electrochemical workstation. Cyclovoltammograms were recorded with a platinum sheet as counter electrode; voltage sweeps were performed with 50 mV/ s. Electrolyte in contact with the experimental electrode in all cases was 1m KNO3 at a pH of 2 to 3. Experimental fuel cells were constructed using a Meinsberger Korrosionsmesszelle, using electrodes as described above for oxidation of ethanol and Fe3+/Fe2+/Pt electrodes as oxidative counterelectrodes; with equal activities of Fe3+ and Fe2+ (0.2 m of the chlorides) the potential is equal to the standard potential of 0.77 V (O2 electrodes in acidic solution would have a standard potential of 1.23 V). The half-cells were connected by a small glass membrane (diameter 1 mm) and operated in short-circuit mode by setting the external potential difference of the cell to zero by the Zahner IM6 in potentiostatic mode (for a sketch, see Figure 2). Open-circuit voltages were determined in galvanostatic mode with the current set to zero. The whole electrochemical system was deaerated and bubbled with N2 during operation; the constant N2 stream continuously stirred the electrolyte. To check electrocatalytic activity, the short circuit current was recorded over time, and ethanol or other organics were added to the electrolyte at fixed time intervals to examine the effect on the current.
Nano-ZnS as Electrocatalyst
Figure 3. Particle diameter of ZnS:Mn dispersed in water; measured by DLS and plotted as volume statistics.
Figure 4. Excitation (monitor: 590 nm) and emission spectra (excited at 300 nm) of ZnS:Mn nanoparticles dispersed in water.
J. Phys. Chem. A, Vol. 114, No. 11, 2010 3953
Figure 5. Cyclovoltammogram of ITO electrodes; blank as well as covered with nano-ZnS.
Figure 6. Cyclovoltammogram of ITO electrodes in the presence of 10% of ethanol or acetic acid, respectively. Electrolyte: KNO3.
Results and Discussion Diameters of redispersed particles (in slightly acidic water) were determined by DLS to be on the order of 4-5 nm (Figure 3). Additionally, this is visible from the position of the excitation spectrum of the ZnS:Mn, with its peak being close to 300 nm.17 Figure 4 shows excitation (monitor: 590 nm) and emission (excitation: 300 nm) spectra of the ZnS:Mn nanopowder dispersed in water. To prevent problems with metallic redox chemistry, glass plates covered with indium tin oxide (ITO) were used as the electrode base in those cases where no protecting carbon layer was present. ZnS nanoparticles were deposited from ethanolic dispersion on this base. In a series of cyclovoltammetric measurements, the stability of ITO as well as of ZnS-covered ITO-glass was checked in pure KNO3 electrolyte. Figure 5 shows in an extended scan that there is a window of chemical stability between approximately -0.8 and +1.7 V, proving that ZnS nanoparticles are stable enough under the conditions necessary in an ethanol-burning fuel cell; below -0.8 V ITO and ZnS decompose. Potentials larger than 2 V and lower than -2 V lead to decomposition of water. Adding 10% of ethanol or acetic acid to the KNO3 electrolyte does not lead to any electrochemically induced reaction within the stability region (Figure 6 shows scans from -1 to +1 V). Only when adding ethanol or acetic acid to the electrolyte in contact with a nano-ZnS-covered electrode oxidation peaks are observedsfor ethanol-containing as well as for acetic acidcontaining systems (Figure 7).
Figure 7. Cyclovoltammogram of nano-Zns covered ITO electrodes in the presence of 10% of ethanol or acetic acid, respectively. Electrolyte: KNO3.
After we demonstrated the stability of ZnS as well as its electrocatalytic activity on electrodes covered with ZnS nanopowder, but without any further support, experimental ethanolcontaining electrochemical cells were tested using electrodes with ZnS nanopowders immobilized in carbon, always compared to blanks (pure ITO-covered glass plates, carbon electrodes without ZnS). These experiments were performed by setting the potentiostat controlling the cell to 0 V (short-circuit conditions), thus recording maximal currents flowing through the electrodes, as driven by the internal electrochemistry. Adding ethanol to the electrolyte in contact with a nano-ZnS-covered
3954
J. Phys. Chem. A, Vol. 114, No. 11, 2010
Figure 8. Short-circuit current in a fuel cell with blank ITO electrode, blank carbon electrode, and carbon/ZnS electrode after ethanol addition.
Bredol and Kaczmarek at regular time intervals into the carbon/ZnS-electrode compartment. Figure 9 shows the results of such measurements; as with EtOH, the carbon blanks show sharp peaks (adsorption), but in the presence of ZnS:Mn, continuous decomposition occurred. Since acetaldehyde and acetic acid can be oxidized as well as reduced, negative currents can occur during the adsorption phase. Comparing Figure 9 and Figure 8, it can be seen that acetaldehyde and acetic acid are active as well as ethanol on the electrodes tested, but with much smaller currents. Therefore, in the present state conversion of ethanol will be mainly restricted to oxidation to acetaldehyde, but optimization of the electrode structure might lead to configurations that allow for more complete consumption of ethanol. A key to this end will be the use of structured and highly conductive carbons, since they have the potential to immobilize the intermediates. The electrodes tested so far contain ZnS nanoparticles only in small concentrations, embedded in amorphous, unstructured carbon; such systems can offer only limited capability for encapsulation. Another open question is the number of cycles the ZnS nanoparticles can survive; for the moment the best electrodes prepared have been reused after intermediate cleaning with water and storage in ambient atmosphere for more than ten experimental runs. Conclusion and Outlook
Figure 9. Short-circuit current in a fuel cell with blank carbon electrode and carbon/ZnS electrode upon injection of acetaldehyde or acetic acid.
electrode under these circumstances leads to current peaks, which are clearly correlated to the injection times, pointing to the catalytic action of nano-ZnS (Figure 8). Before ethanol addition, the current observed in the presence of ZnS points to oxidation of ZnS; in the presence of ethanol, ZnS is protected (see Table 3), and already oxidized species may be reduced again. This is visible also from the open circuit potential of about 250 mV in the presence of ethanol; this value is too high to be explained by ZnS oxidation alone (see Table 3), but considerably smaller than expected for ethanol oxidation (Table 1). This overpotential will be due to the low conductivity of the connecting membrane as well as to still exisiting catalytic barriers. Repeated ethanol addition increases the flowing current stepwise in the case of ZnS-containing electrodes. From Figure 8, it can be seen that a pure carbon electrode also leads to a sharp current peak (adsorption of ethanol), but not to continuous decomposition. With a ITO blank electrode, these effects are even smaller. The current in the test cell is very small; this is due to the small diameter of the membrane between the ethanol electrode and the Fe3+/Fe2+ electrode. Therefore, future work will have to concentrate on the increase of the current density of the electrodes, and of course on the demonstration of operation with an air or oxygen electrode. To find out whether the carbon/ZnS electrodes are electrocatalytically active also against the intermediates, similar runs have been made with acetaldehyde or acetic acid being injected
Electrocatalytic action of nanoparticulate ZnS in a carbon support has been demonstrated in an aqueous KNO3 electrolyte containing ethanol. To which extent direct conversion to CO2 is possible has to be investigated in future work. With improved cell design and electrode structure, utilization in fuel cells burning ethanol directly is a promising outlook. Whether also higher alcohols like butanol can be consumed, remains open for future work as well. Further improvements are expected from a more rigorous control of the electrode microstructure; mesoporous carbons, and ZnS particles with specific morpholopgy will have to be combined for such work. Acknowledgment. This work was partially funded by the state of North-Rhine-Westphalia in Germany in the framework of “competence platforms”. M.K. is supported by Fachhochschule Mu¨nster with a stipend for the development of young researchers. References and Notes (1) Fujiwara, H.; Hosokawa, H.; Murakoshi, K.; Wada, Y.; Yanagida, S. Langmuir 1998, 14, 5154–5159. (2) Kanemoto, M.; Shiragami, T.; Pac, C.; Yanagida, S. J. Phys. Chem. 1992, 96, 3521–6. (3) Kanemoto, M.; Hosokawa, H.; Wada, Y.; Murakoshi, K.; Yanagida, S.; Sakata, T.; Mitsuru Ishikawa, H. M.; Kobayashi, H. J. Chem. Soc., Faraday Trans. 1996, 92, 2401–2411. (4) Hu, J.-S.; Ren, L.-L.; Guo, Y.-G.; Liang, H.-P.; Cao, A.-M.; Wan, L.-J.; Bai, C.-L. Angew. Chem., Int. Ed. 2005, 44, 1269–1273. (5) Mu¨ller, B. R.; Majoni, S.; Meissner, D.; Meming, R. J. Photochem. Photobiol., A 2002, 151, 253–265. (6) Yang, J.; Peng, J.; Zou, R.; Peng, F.; Wang, H.; Yu, H.; Lee, J. Nanotechnology 2008, 19, 255603. (7) Zhang, X. V.; Ellery, S. P.; Friend, C. M.; Holland, H. D.; Michel, F. M.; Schoonen, M. A. A.; Martin, S. T. J. Photochem. Photobiol., A 2007, 185, 301–311. (8) Kowal, A.; Li, M.; Shao, M.; Sasaki, K.; Vukmirovic, M.; Zhang, J.; Marinkovic, N.; Liu, P.; Frenkel, A.; Adzic, R. Nat. Mater. 2009, 8, 325. (9) Mann, J.; Yao, N.; Bocarsly, A. B. Langmuir 2006, 22, 10432– 10436. (10) Chan, K.-Y.; Ding, J.; Ren, J.; Cheng, S.; Tsang, K. Y. J. Mater. Chem. 2004, 14, 505–516. (11) Lu, A.-H.; Li, W.-C.; Schmidt, W.; Kiefer, W.; Schu¨th, F. Carbon 2004, 42, 2939–2948. (12) McCreery, R. L. Chem. ReV. 2008, 108, 2646–2687.
Nano-ZnS as Electrocatalyst (13) Wang, J. N.; Zhao, Y. Z.; Niu, J. J. J. Mater. Chem. 2007, 17, 2251–2256. (14) Bredol, M.; Matras, K.; Szatkowski, A.; Sanetra, J.; Prodi-Schwab, A. Sol. Energy Mater. Sol. Cells 2009, 93, 662–666. (15) Hikmet, R.; Talapin, D.; Weller, H. J. Appl. Phys. 2003, 93, 3509– 3514.
J. Phys. Chem. A, Vol. 114, No. 11, 2010 3955 (16) Van de Walle, C. G.; Neugebauer, J. nature 2003, 423, 626628. (17) Yang, P.; Szatkowski, A.; Bredol, M. J. Sol-Gel Sci. Technol. 2009, 51, 306.
JP907369F
