~ ELSEVIER Solar Energy Materials and Solar Cells 38 (1995)

~ m Solar Energy Materials and Solar Cells ~ Solar Energy Materials and Solar Cells 38 (1995) 249-276 ELSEVIER Photochemical production of hydro...
Author: Susanna Golden
18 downloads 0 Views 2MB Size
~

m

Solar Energy Materials and Solar Cells

~

Solar Energy Materials and Solar Cells 38 (1995) 249-276

ELSEVIER

Photochemical production of hydrogen and oxygen from water: A review and state of the art Edmond Amouyal Laboratoire de Physico-Chimie des Rayonnements (CNRS, URA 75), Bat 350, Universite Paris-Sud, 91405 Orsay, Cedex, France

Abstract Photochemical hydrogen production is potentially one of the most fascinating ways for solar energy conversion and storage. Since 1977, several homogeneous, quasi-homogeneous or microheterogeneous model systems of hydrogen or oxygen generation from water, by visible-light irradiation, have been proposed and are briefly reviewed. These half photosysterns are based on different approaches: (i) multimolecular systems, (ii) systems involving a supramolecular structure of polyad type, and (iii) systems incorporated in organized and constrained or confined media. A survey of the different attempts for complete water splitting into hydrogen and oxygen is also made.

1. Introduction

The production of renewable and non-polluting fuels via the direct conversion of solar energy into chemical energy remains a fascinating challenge for the end of this century. Among various interesting reactions, the splitting of water into molecular hydrogen and molecular oxygen by visible light (reaction 1) is potentially one of the most promising ways for the photochemical conversion and storage of solar energy [1-4]. visible light

H 20

~

H2 + 1/202'

(1)

Indeed hydrogen is a valuable fuel: the free enthalpy needed in reaction 1 to produce one mole of H 2 , i.e., the energy stored per mole, is LlGg98 = 237.2 kJ . mol-I. Due to its small weight, the energy storage capacity of H2 per gram, 119000 J. g-I, is very high. It is, for example three times higher than the storage capacity of oil (40 000 J. g-I). Moreover, the water-splitting process has two other 0927-0248/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDl 0927-0248(95)00003-8

250

E. Amouyal / Solar Energy Materials and Solar Cells 38 (J995) 249-276

advantages. First, the raw material, i.e., water, is abundant and cheap. Second, the combustion of H2 in air (reverse of reaction 1) again gives water. In other words, the overall process is cyclic and non-polluting. The main disadvantage is that H2 can react explosively with 02' However, the explosive limit of H2 in air is 4.00%, less than that of butane: 1.86%. Hence, the utilization of hydrogen is no more dangerous than that of natural gas, and like natural gas hydrogen can be quite easily stored and transported. Since water does not absorb visible light, intermediates are needed to achieve the water photocleavage via a cyclic pathway (reaction 1). Numerous strategies have been described and several model systems capable of producing separately hydrogen and oxygen have been proposed. Attempts to generate simultaneously H2 and O 2 have also been reported. In this paper, I present a general survey of the principal homogeneous and microheterogeneous systems (multimolecular, supramolecular, constrained or confined systems) in which the illumination of a coloured compound acting as a photosensitizer gives rise to photoinduced redox processes. Microheterogeneous systems involving the formation of electron-hole pairs through direct excitation or photosensitization of semiconductor particles are not considered. Several reviews on these semiconductor-based systems are available [1,5,6].

2. Multimolecular systems 2.1. Ideal functions

As emphasized in the introduction, hydrogen and/or oxygen production from water by visible light requires one or several intermediates having ideally the following functions: (i) visible light absorption, (iO conversion of the excitation energy to redox energy (charges), (iii) concerted transfer of several electrons to water leading to the formation of H2 as energy-storage compound and/or to the formation of 02' Indeed, one of the main difficulties in achieving the splitting of water by means of light-induced redox processes is that hydrogen requires two electrons (reaction 2) while oxygen requires four electrons (reaction 3). 2H 20 + 2e--+ H2 + 20H-, 2H 20-+0 2 +4H++4e-,

EO (pH = 7) = -0.41 V versus NHE,

EO(pH=7) = +0.82VversusNHE.

(2) (3)

This number of charges corresponds to the most favourable thermodynamic conditions for reaction 1. In other words, this reaction is a multi-electron transfer process which requires 1.23 eV per electron transferred. Hence, photons with A < 1008 nm corresponding to a minimum energy of 1.23 eV can induce the cleavage of water.

E. Amouyal / Solar Energy Materials and Solar Cells 38 (1995) 249-276

251

2.2. General schemes for H2 and O2 production

In a first approach, photochemical systems involving several compounds were proposed. In these multimolecular systems, each function is fulfilled by one molecule, namely, (i) a photosensitizer PS able to absorb visible light to generate excited species PS • with useful redox properties (reaction 4): hu·

PS --:spS' ,

(4)

(ii) a second compound R which can be reduced or oxidized by quenching of the excited species PS' in electron transfer reactions leading to the formation of charge pairs, such as PS +, R - in the case of the oxidative quenching of PS (reaction 5):

(5) (iii) and a third compound able to collect several electrons to facilitate the exchange of two (reaction 6) or four electrons with water. This multi-electron collection and transfer can be realized by a specific redox catalyst Cat. Cat

2R-+ 2H+ ~2R + H 2 •

(6)

In such a system, the second compound R acts as an electron relay between the photosensitizer PS and the catalyst Cat mediating the electron collection. The redox potential of its reduced species R - must be less than - 0.41 V (versus NHE, pH = 7) to take part in reaction 2. In practice, difficulties arise from a fast recombination of charge pairs (reaction 7).

(7) The main problem, for these multimolecular systems and more generally for photochemical systems, is how to retard this back electron transfer reaction in order to get a charge separation of long lifetime. In the case of multimolecular systems, the back reaction should be prevented by using a fourth compound, an electron-donor D, which scavenges the oxidized photosensitizer PS + in a competitive electron transfer reaction to give the initial PS and a donor oxidation product D + (reaction 8). PS++ D

~

PS + D+,

D + ~ products.

(8) (9)

The latter rapidly decomposes irreversibly (reaction 9), and such systems have been qualified as "sacrificial". D is the only compound, apart obviously from water (H+), which is consumed. The other compounds PS, R and Cat follow catalytic cycles. Two schemes for cyclic production of hydrogen from water can be envisaged [7]. The first is called the "oxidative quenching mechanism" because it involves

252

E. Amouyal/ Solar Energy Materials and Solar Cells 38 (1995) 249-276

/'.'

D~X~-X~~

(b)

0+

Fig. 1. Schematic representation of the redox catalytic cycles in the photoreduction of water to hydrogen by visible-light irradiation of a four-component model system PS/R/D/Cat: (a) oxidative quenching mechanism, (b) reductive quenching mechanism.

oxidation of the excited photosensitizer PS' to PS + by the electron relay R (Fig. 1a). It corresponds to reaction 4 to 9. The second scheme involving reduction of the excited state photosensitizer PS' by D is called the "reductive quenching mechanism" (Fig. 1b). This primary reaction (reaction 10) yields the (10) reduced photosensitizer PS - and the oxidized donor D + which decomposes irreversibly (reaction 9). In this way, PS - can accumulate and react with an electron relay R to regenerate PS and to yield R - (reaction 11).

(11) In the presence of a suitable catalyst, R - can lead to the formation of hydrogen as in the first scheme (reaction 6). It should be remarked that PS - is a more powerful reducing species than R -. Hence, the reduction of water to H2 can be achieved directly by PS- itself in the presence of a suitable catalyst. As a consequence, this scheme involves only three components (PS, D, Cat) and the mechanism becomes simplified (Fig. 2).

E. Amouyal / Solar Energy Materials and Solar Cells 38 (1995) 249-276

253

0+

Fig. 2. Schematic representation of the redox catalytic cycles in the photoreduction of water to H 2, via reductive quenching mechanism, for a three-component model system PS/D/Cat.

Similar three-component systems for O 2 production from water have been proposed (Fig. 3). These systems require the formation, following visible-light excitation of the photosensitizer PS, of a strong oxidizing species PS +, having a redox potential EO(PS+ IPS) greater than 0.82 V (versus NHE, pH = 7). This can be achieved by using an electron-acceptor A as quencher which, once reduced to A - (reaction 12), decomposes irreversibly (reaction 13).

(12) (13)

A - - decomposition products.

The oxidized PS + can thus accumulate and lead to oxygen evolution in the presence of a suitable catalyst capable of facilitating the exchange of 4 electrons with water (reaction 14). (14)

/hV"'

~X02

A~PS"S '" 0 "2

A' Fig. 3. Schematic representation of the redox catalytic cycles in the photooxidation of water to oxygen by visible-light irradiation of a three-component model system PSI A/Cat.

254

E. Amouyal j Solar Energy Materials and Solar Cells 38 (1995) 249-276

o Fig. 4. Schematic representation of the redox catalytic cycles in the photoreduction of water to H 2 , via energy transfer (antenna effect), by visible-light irradiation of a five-component model system

PSjR.n /R/D/Cat.

Another approach (Fig. 4) consists in using the photosensitizer PS as an antenna and transferring the excitation energy to a receptor molecule Ren (reaction 15). (15) +PS + R: n. The receptor can subsequently react with the electron relay R via electron transfer (reaction 16) PS' + Ren

~

R:n+R~R:n+R-

(16)

to give a charge pair (R:n, R - for example). The reduction of water to H2 can be achieved in the presence of a sacrificial electron-donor D and a suitable catalyst (Fig. 4) as in the first scheme (Fig. la). In this five-component system PS/Ren/R/D/Cat, the energy-transfer photosensitizer PS is not involved in any redox processes as are the antenna molecules in natural photosynthesis, and the receptor Ren acts as an energy-electron relay. It should be noted that, although the multimolecular approach is the simplest manner to achieve cyclic photochemical water cleavage, its accomplishment necessitates overcoming several difficulties. Indeed, the different components of such systems must fulfil spectral, photophysical, thermodynamic and kinetic conditions. Some of them have been mentioned here. The other requirements can be found in earlier comprehensive reviews [3-6,8-10]. 2.3. First model systems for H2 production

Several sacrificial model systems of H2 production from water have been proposed since 1977. The first ones are listed in Table 1. They used acridine dyes such as acridine yellow [11] as PS. But transition metal complexes, in particular

E. Amouyal / Solar Energy Materials and Solar Cells 38 (1995) 249-276

255

Table 1 First model systems for water photoreduction to hydrogen System No. PS

1 2 3 4

Acridine Yellow [Ru(bpY)3]2+ [Ru(bpY)3]2+ [Ru(bpY)3]2 +

Reference

R

D

Cat

Eu 3+ or y3+ [Rh(bpY)3P+ MV 2 + My2+

Cysteine TEOA EDTA TEOA

Pt0 2 Shilov group 1977 [11] Lehn group 1977 [12] K 2 PtCI 6 Colloidal Pt (PYA) Orsay groups 1978 [7] Pt0 2 Gratzel group 1978 [13]

[Ru(bpY)3J2+ [7,12,13], then appeared to be remarkable photosensitizers with respect to visible light absorption, excited state properties, redox potentials and kinetic requirements [9]. The electron relay species first investigated include Eu3+ and y3+ salts [11], a transition metal complex [Rh(bpY)3P+ [12] which can transfer two electrons, and methyl viologen MYz+ the most commonly used electron relay [7,11,13]. Cysteine [11,13] and especially tertiary amines such as EDTA (ethylenediamine tetra-acetic acid) [7,11] and triethanolamine TEOA [11-13], which are rapidly decomposed when oxidized, were used as sacrificial electron-donors. Platinum compounds [7,11-13] turned out to be suitable catalysts. The first system (Table 1, system 1) has been described by Shilov et al. [11]. It consists of acridine yellow A Y as PS, cysteine as sacrificial electron-donor, salicylate complexes of Eu3+ and V3+ as R and Adams' catalyst (PtO z) as Cat. They also used EDTA, TEOA or HzS as D, MVz+ as Rand K zPtCl 6 as catalyst. The mechanism assumed was of the "reductive" type (Fig. 1b), and the quantum yield of Hz production in the case of AY /Eu 3+ /cysteine/PtO z model system was of the order of 1%. This quantum efficiency was too low so that the validity of the system was questioned at that time. Indeed, in the early 1970s, binuclear metal complexes were considered as more promising candidates for PS than organic compounds [14]. The two following systems, that of Lehn and Sauvage [12] (Table 1, system 2) and ours [7] (Table 1, system 3) used a metal complex, [Ru(bpY)3]2+, as PS. The hydrogen quantum yields 0 (1/2 Hz) were much higher (> 10%) than that of Shilov's system [11]. Consequently, it was easier to produce and characterize Hz, and even to detect the formation of Hz bubbles with the naked eye. In addition, in the case of the Orsay system (Table 1, system 3), we have clearly established the oxidative quenching mechanism for Hz production (Fig. 1a) by laser flash spectroscopy [7,15]. So it is not surprising that these two systems [7,12] which were presented at the International Conference on Photochemical Conversion and Storage of Solar Energy OPS-2) in 1978 at Cambridge [16], contributed largely to convince the most sceptic "solar experts" of the interest of the multimolecular approach. These results were rapidly reproduced by many other laboratories. It should be noted that in the Shilov system [11] and in the Lehn system [12], it was assumed that Pt particles are formed in situ through the photosensitized reduction of K zPtCI 6 , while in our system [7] it was demonstrated for the first time that colloidal metals (Pt,Au), chemically prepared and stabilized by polymers

Bipyridinium ions:

Metal complexes of Ru, Cr, Os, Ir, Pt ... :[Ru(bpY)3J2+ Metal porphyrins of Zn, Mg, Ru ... : ZnTMPyp 4 + Metal phtalocyanines of Zn, Co, Mg ...

MV 2 +

Cat

Enzymes: hydrogenase, nitrogenase

Supported metals: Pt-Ti0 2 , Rh-SrTi0 3 • Ni-Ti0 2 Metal oxides: Ru0 2 , Pt0 2 , Ir0 2 , Pd0 2 , Ti0 2 , Fe 2 0 3 Supported metal oxides: Ru0 2 + Ir0 2 /zeoJite Colloidal metal systems: Ni-Pd

Metal powders: Pt, Ru, Ni

EDTA and glycine derivatives (NPG) Colloidal metals: group VIII metals (Ir, Pt,NL),Au,Ag Amines: TEOA, TEA. Pt salts: K 2 PtCI 6 , K 2 PtCI 4

D

Metal ions: Eu 3 +, V 3 +, Cr 3 + Sulphur compounds: cysteine, thiols (mercaptoethano\), H 2 S Acridine dyes: acridine yellow, proflavine. Metal complexes of Rh, Co ... : Urea derivatives: allythiourea [Rh(bpY)3]3+, [Co(Sep)]3+ Proteins: cytochrome c3 Xanthene dyes: fluorescein, Amino acids eosin Y Cyanine dyes Carbon compounds: ascorbate, ethanol Metal ions: Eu 2 + Organic compounds: poly(pyridine-2,5-diyJ) Coenzymes: NADH, NADPH

Phenanthrolinium ions

R

PS

Table 2 Components of H 2-generating microheterogeneous systems

E. Amouyal / Solar Energy Materials and Solar Cells 38 (J995) 249-276

257

(polyvinyl alcohol PYA), can be used successfully as catalysts in a photochemical model system. Gratzel et al. [13] described at the same time a system (Table 1, system 4), similar to the Orsay system (Table 1, system 3) [7,15], but using platinum oxide Pt0 2 (Adam's catalyst) instead of colloidal platinum, and triethanolamine TEOA or cysteine instead of EDTA. With these components (Pt0 2, TEOA) [13], the H2 yields are much lower than those obtained by visible-light irradiation of the Orsay system [15,17]. 2.4. Components of water photoreduction systems

It is hardly possible to give an exhaustive list of all the multimolecular systems of H2 production from water described in the literature and based on the general schemes (Figs. 1,2,4). However, I have categorized in Table 2 the different constituents used in these systems as PS, R, D and Cat, and I shall describe only a few systems, in particular the Orsay system. 2.4.1 Photosensitizers Transition metal complexes of Ru, Cr, Os... [18,19], metalloporphyrins and metallophthalocyanines [19,20] and acridine dyes [11,21-23] are the principal classes of PS, [Ru(bpY)3]2+ being the most investigated. This complex is essentially involved in the oxidative quenching mechanism. For the reductive mechanism, one of the best candidates as PS is [Ru(bpz)3]2+ (bpz = 2,2'-bipyrazine) [24,25]. Hydrogen is produced with good yields when it is used in a photochemical system with TEOA as D, My 2+ as R and a platinum compound as catalyst [26,27]. More recently, we have shown that [Ru(bpY)2dppz]2+ (dppz = dipyrido [3,2-a:2',3'c] phenazine) can be involved as PS either in an oxidative mechanism, with EDTA as D, or in a reductive mechanism, with TEOA as quencher [28]. Interestingly, a rigid copper (I) complex [Cu(dpp)2]+ (dpp = 2,9-diphenyl-1,1O-phenanthroline) has been used as energy-transfer photosensitizer (Fig. 4) in a five-component system [29]. It should be noted that high quantum yields for H2 formation with an optimum cJ>(l/2 H 2) = 0.6 [30] have been found when aqueous solutions of a water-soluble zinc porphyrin, Zn (II) tetrakis (N-methyl-4-pyridyI) porphyrin ZnTMPyp 4 + [20,30-32], are irradiated with 550 nm light in the presence of My2+, EDTA and colloidal Pt. However, these yields decrease dramatically within an irradiation time of 4 hours [20,30]. More recently, cage complexes [33,34], cyanine dyes [35] and organic compounds absorbing visible light such as poly{pyridine-2,5-diyI) [36] have been tested as PS. 2.4.2. Electron relays Bipyridinium ions, also called viologens, are the main compounds used as electron relays R (Table 2), the most popular being methyl viologen My2+. They also provide an extended range of redox potentials [37]. We have investigated several homogeneous series of 2,2'-bipyridinium (or diquat), 4,4'-bipyridinium (or paraquat) and 1,1O-phenanthrolinium ions as mediators for water photoreduction

258

E. Amouyal / Solar Energy Materials and Solar Cells 38 (1995) 249-276

[37-39], the most efficient electron relays being MYz+ and I,Y-dimethylene-4,4'dimethyl-2,2'-bipyridinium ion [37,38]. Transition metal complexes are also interesting mediators, in particular [Rh(bpY)3]3+ which can transfer two electrons [12,40,41] and cage complexes such as [Co(sep)]3+ (sep = sepulchrate) [42,44]. [Co(sep)]3+ is, contrary to viologens, insensitive to hydrogenation, an undesired side reaction which can occur at the catalyst surface. A natural electron mediator, cytochrome c 3, which unlike My2+ is not at all toxic, has been tested in model systems in association with hydrogenase as catalyst [23,45]. It is of interest to note that the only compound used as energy-electron relay Ren in a five-component system (Fig. 4) is 9-carboxylate anthracene anion [29,46]. 2.4.3. Electron-donors

Krasna [21] thoroughly tested several classes of organic compounds as electrondonors D (Table 2), with proflavine as PS, MYz+ as R and the enzyme hydrogenase or Pt asbestos as Cat. With this model system, the most effective donors were EDTA and 1,2-diaminocyclohexane tetra acetic acid. Whitten et al. [47] found that triethylamine TEA which is not at all effective in Krasna's systems, leads to high Hz yields (0.53) in a three-component system [Ru(bpY)3]2+/TEA/PtO z, i.e., in the absence of MYz+, but in 25% wateracetonitrile mixtures. Coenzymes such as NADH and NADPH have been tested as sacrificial electron-donors [45,48]. However, the photoinduced regeneration of these natural reductants can also be achieved [48,49]. In photochemical systems involving a sacrificial electron-donor, it is interesting, from a practical point of view, to find and use electron-donors such as HzS [48] which are easily available as waste industrial products. 2.4.4. Microheterogeneous catalysts

As regards the catalyst (Table 2), and since the first report on the catalytic activity of colloidal Pt in a photochemical system of Hz production [7], we have investigated systematically colloidal metals (of groups YIII and IB essentially), metals deposited on solid supports (semiconductor, zeolite), metal and metal oxide powders [50,51]. Iridium and platinum hydrosols are extremely efficient [50,51]. Pt supported on TiO z [27,50-52] leads to similar high Hz yields. We have found that ruthenium oxides, known to be good catalysts for 0z generation from water (Section 2.6), efficiently mediate the photoreduction of water to Hz without catalyzing the undesired hydrogenation of the electron relay [53,54]. Ru0 2 and IrOz codeposited on zeolite give the highest H2 yields [50]. More recently, we have observed an improvement of catalytic activity by an alloying effect in the case of bimetallic sols of Ni-Pd [55]. It is of interest to note that the enzyme hydrogenase which contains at least Fe 4 S4 -type clusters has been used as a natural catalyst [21,23,45] but it is unstable and the H2 yields are lower than those obtained with Pt compounds [21]. 2.4.5. Homogeneous catalysts and homogeneous systems for H2 production

The great majority of catalysts described in the literature are heterogeneous. Very few homogeneous catalysts and hence homogeneous systems for H 2 produc-

E. Amouyal / Solar Energy Materials and Solar Cells 38 (1995) 249-276

Table 3 Components and quantum yields for H 2-generating homogeneous systems D Homogeneous catalyst pH PS Eu2+ [Ru(bpY)3]2+ [Ru(bpY)3]2 + ascorbate [Ru(bpY)3]2+ EDTA [Ru(bpY)3]2+ TEOA [Ru(bpY)3]2+ ascorbate [Ru(bpY)3]2+ TEOA [Ru(4,7-(CH 3)2phen)3]2+ TEOA a b

[Co(Me 6[14]diene N4 XH 2O)2]2+ [Co(Me6[14]diene N4 XH 2O)2]2+ [Rh(bpY)3P+ [Rh(bpY)3P+ [Co(bPY)nJ2 + [Co(dimethylglyoxime)2] [Co(bpY)3]2+

S S S.2

259

cfJ(H 2)

Reference

O.OS 0.0005 0.04 [41] 0.02 0.03

1979 [S6] 1979 [S6] 1979 [40] 1981 [41] 1981 [57] 1983 [S9] 1985 [58]

S S 8.7 • 0.29 8

b

In DMF /H 20 or in neat organic media (DMF, acetone, acetonitrile ... ). In SO % aqueous acetonitrile. .

tion (Table 3) have been reported (see Section 2.6 for homogeneous catalysts for O 2 production). In these systems, following visible-light excitation of the photosensitizer PS, one of the components is transformed into an unstable intermediate, for instance a metal hydride, which in turn decomposes to yield Hz and the starting component. Therefore, this component acts as homogeneous catalyst. Good candidates are inorganic compounds, such as homogeneous hydrogenation catalysts, which have a metallic site able to present different oxidation states during the catalytic cycle, and which can form an intermediate hydride, unstable in solution, to provide a pathway for H2 release. Homogeneous catalysts (Table 3) which have been first reported in 1979 are a macrocyclic cobalt (II) complex, [Co(Me 6[14]diene N4 ) (H 20)2]2+ [56], and [Rh(bpY)3]3+ [40,41]. Other CoOI) complexes have been proposed viz. [Co(bpY)3]2+ [57,58], cobaloxime [59] and other macrocyclic complexes [59]. These homogeneous systems (Table 3) consist of three components, the photosensitizers PS being essentially [Ru(bpY)3]2+ [40,41,56,57,59] or [Ru(4,7(CH 3)2 phen)3]2+ (phen = 1,1O-phenanthroline) [58], and the electron donors being EDTA [40,41], Eu(II) [56], ascorbate [56,57] or tertiary amines like TEOA [41,58,59]. As in the case of microheterogeneous systems, the mechanism of H2 production involves either a reductive [56,57,59] or an oxidative [40,41,58] quenching of PS '. These systems are efficient, especially in organic media [58,59]. In the case of the [Ru(4,7-(CH3)zphen)3]2+/TEOA/[Co(bpY)3]2+ system proposed by Sutin et al. [58], the Hz yields increase from about 0.02 in HzO to 0.29 in 50% CH 3CN-H 20 (Table 3). 2.5. The [Ru(bpY)3P + / MV 2 + / EDTA / colloidal Pt model system for from water

H2

generation

The classical system proposed in 1978 for the first time by the Orsay groups [7,15] (Table 1, system No.3) comprises [Ru(bpY)3]2+ as PS, My2+ as R, EDTA as D, and colloidal Pt as catalyst [7,15] (Fig. 5). It has been known since 1934 [60] that electron exchange between My 2+ and H2 is catalyzed by Pt. But for the first time, it has been proved that catalyst hydrosols [61] used in a model photosystem [7] mediate visible-light H2 generation from water. In previous studies [11,12] the

E. Amouya/ / Solar Energy Materials and Solar Cells 38 (J995) 249-276

260

products -

EOTA+

EOTA

Fig. 5. Schematic representation of the redox catalytic cycles in the photoreduction of water to H2 by visible-light irradiation of the Orsay model system [Ru(bpY)3j2+ /MV2+ /EDTA/colloidal Pt proposed by Moradpour et al. [7].

formation in situ of such colloids had been assumed through the reduction of Pt salts. This system produces Hz very efficiently, leads to reproducible results, and is well characterized as regards Hz formation quantum yields [50,62] and detailed mechanism [15,37]. These are the reasons why this system has been thoroughly studied by several groups [17,26,52,63-70] and is still considered as a reference for testing new PS, R, D and catalysts (Table 2), and for evaluating solar photochemical reactors on a pilot level [71]. The system has been described in many reviews and books. I just wish to recall some of its important features. When an aqueous solution containing [Ru{bpY)3F+, My2+, EDTA, and colloidal Pt is irradiated with visible light (400 nm < A < 600 nm), an important Hz evolution is observed (Fig. 5) according to the following mechanism established from laser flash spectroscopy experiments [7,15]. In deaerated solutions, the main reactions are: [ 12 +' , [Ru(bpYh 12+ hu ~ Ru(bpYh

[Ru(bpYh1 [Ru(bpYh1

z+' ko

~ [Ru(bpYh1

z+'

kq

2+

(17)

(18)

'

+ My 2 + ~ [Ru(bpYh1

3+

+ MY+-,

kb[ Ru(bpYh 12+ + MY 2+, [Ru(bpYh 13+ + MY+- ~ [Ru(bpYh ] 3+ +

kox[

EDTA~

Ru(bpYh 12+ + EDTA +,

(19) (20) (21) (22)

Net: EDTA + H+

hu Pt

~

EDTA ++ lj2H 2.

(23)

E. Amouyal / Solar Energy Materials and Solar Cells 38 (J995) 249-276

261

The proposed mechanism for EDTA degradation (reaction 24) EDTA + ~ products (mainly glyoxylic acid) ,

(24)

and for the catalytic process on metallic particles (reaction 22) have been described in detail in previous reports [15,37,51,72]. Besides the EDTA consumption (reaction 24), difficulties arise from undesirable reactions such as MY+ dimerization [72] (reaction 25) and the irreversible 2MY+'~

(MY+'h

(25)

hydrogenation of methyl viologen [15,37,72] (reaction 26) (26) The following rate constant values, as determined by flash photolysis [15] are in good agreement with published ones [65]: ko = 1.45 X 10 6

S-I

kq = 1.03 X 10 9 M- 1 S-I (for !-L kb

= 0.018 M and dried MYz+)

= 2.8 X 10 9 M- 1 S-I

k ox = 1.1 X 10 8 M- 1 S-I

We have shown [37] that kq increases with the ionic strength p. of the solution [73] with kg = 2 X 10 8 M -I S -1 extrapolated for zero p. [37]. In non-deaerated solutions, Hz formation rates and yields decrease due mainly to reaction 27 with k = 8 X 10 8 M- 1 S-I [74]. (27) This reaction leads to HzOz as a stable intermediate and to MYz+ degradation products [75,76]. 02" +H+~HOi,

(28)

2HOi ~ 0z + HzOz,

(29)

MY+'+ 02"(HOi)

~

MYO z ~ products.

(30)

The mechanism shows that Hz production depends on light intensity, pH and concentration of the four constituents of the system. An optimum quantum yield

Suggest Documents