Visible light-induced reduction of carbon dioxide sensitized by a porphyrin-rhenium dyad metal complex
5.3. Results and Discussion
Absorption and em ission properties of the dyad , ZnDMCPP-Re(bpy)(NHAc) (1)
The absorption spectra were measured for ZnDMCPP-Re(bpy)(NHAc) (1), ZnTMP(COOMe) (6), Re(bpy)(NHAc)2 (7), and ZnTMP(COOMe) (6) + Re(bp y)(NHAc)2 (7). The molar extinction coefficient of ZnDMCPP-Re(bpy)(NHAc) (1) is4.9 x 105 M- 1cm- 1 (Soret band, 428nm), 1.93104 M- 1 cm- 1 (Q band, 560nm) and that of Re(bpy)(NHAc)2(7) is 8400M- 1cm- 1 (MLC T2, 319nm, 5500 M- 1 cm- 1 (MLCT1, 365nm) in DMF. The absorption of ZnDMCPP-Re(bpy)(NHAc)(1) (solid line in Figure 1) and ZnTMP(COOMe) (6) + Re(bpy)(NHAc)2 (7) (dotted line in Fi gure 1) were nearl ythe same, which shows that the zinc-porphyrin and rhenium-complex do not have astrong el ectronic interaction each other.
115
Figure 1 Absorption spectrum of dyad ZnDMCPP-Re(bpy)NHAc (1) compared with those of ZnTPP(COOMe) (6), Re(bp y)(NHAc)2(7), and (6) + (7). All spectra were measured i n DMF in concentration of 2 x 10- 6 M.
The electrochemical properties of the ZnTMP(COOMe) (6), Re(bpy)(NHAc)2 (7), and ZnDMCPP-Re(bp y)(NHAc) (1) were examined by c yclic voltammetry. The electrochemical behavior of the compounds is summarized in Table 1. The potentials for oxidation and reduction were nearl y the same, comparing the ori ginal mol ecules and the dyad, again indicating an egligible interaction between the zinc-porph yrin and rhenium-complex in the ground state.
Table 1. Potentials for oxidation and reduction of ZnTMP(COOMe) (6), Re(bpy)(Ac)2(7) and ZnDMCPP-Re(bpy)(NHAc) (1).
Eo x/ V Er e d/ V
C o mp o u n d Z n P - o x ( 1 ) Z n P - o x ( 2 ) R e - o x ( 1 ) Z n P - r e d ( 1 ) R e - r e d ( 1 )
Z n T M P ( C O O M e ) (6) 0 . 5 8 0 . 8 8 - 1 . 7 4
R e ( b p y ) ( A c )2(7) 1 . 0 6 - 1 . 7 6
Z n D M C P P - R e ( b p y ) ( N H A
c ) (1) 0 . 4 0 0 . 6 5 1 . 0 4 - 1 . 7
1
Th e s e va lu e s we re ob t a ine d fo r the c o mp o un d s (5x10- 4M) in CH2C l2 co n ta in in g 0 .1 M B u4NB F4 e le c tro l yte with a swee p ra te o f 1 00 mV s- 1 with a P FC E wo rk in g e le c trod e an d A g /A gN O3 re fe ren c e .
The energy l evel of the charge -separated state ZnDMCPP˙+-Re(bp y˙ˉ)(NHAc) was estimated to be 1.99 eV. 3 8 Considering the d yad S1 state (2.04 eV, based on the fluorescence wavelength), the electron transfer, from the excited zinc -porphyrin
116
moiet y to the rhenium -complex one within the molecule, should proceed either via S1(Q-band) or S2(Soret-band) excitation.2 0 Actuall y, the relati ve fluoresc ence int ensit y of the d yad (1) by S1
excitation is weaker by 38% compared with that of the parent ZnTMP(COOMe), strongl y suggesting that the deca y processes from the S1 state involve the intramolecular electron transfer t o form the charge-separatedstate.
Photochem ical reduction of CO2 sensitized by ZnDMCPP-Re(bpy)(NHAc) (1) with triethylam ine
Before studying the photoelectrochemical CO2 reduction on the NiO semiconductor, we examined CO2 reduction using the ZnDMCPP-Re(bpy)(NHAc)(1) d yad as a sensitizer and triethylamine (TEA), which is frequentl y used in CO2 reduction schemes as an electron donor. Zn DMCPP-Re(bpy)(NHAc)(1) (10.3M) i n DMF/TEA (4:1, v/v) was irradiated at 43 0 nm under CO2-saturated condition s.
Upon visible light irradiation, CO was detected as the sole product and hydrogen was not formed at all. The time course of CO production during the visible light irradiation is shown in Figure 2 .
Figure 2 Photoreduction of CO2 to CO sensitized by t he dyad (1) in CO2 saturated DMF/TEA (4/1 v/v) .([1] = 1.03 x 10- 6M, excitation wavelength = 430 nm with 8.85 mW.)
The quantum yield for CO production at the initial stage was estimated to be 0.12% from the slope of the plots in Figure 2 (time = 30 -240 min).
The turnover number (TNC O) was 14 after light irradiation for 30 h . These results indicate that the dyad (1) actuall y induces CO2 reduction upon visible light irradiation, although the reactivit y is relativel y modest, as easil y anticipated b y our previous re port on a similar d yad , which had no anchoring substituent on the meso-phen yl group of the zinc-porph yrin.2 0 In the present case of the dyad(1), we can simpl y understand the react ion pathways with dual modes of Route (I) through the upper excited S2 level and and (II) via the S1 level shown in Scheme.
117
Scheme 3 Dual modes of the photoreduction of CO2 sensitized by the dyad (1) in DMF.
Photochem ical behavior of the dyad (1) on NiO as a promising candidate for a non-sacrificial electron donor
One of the promising options to improve the modest photoreactivit y of the dyad (1), which utilizes a sacrificial electron donor, and to develop a more realistic approach for artificial photosynthesis, would be to choose a more effici ent electron -donating s ystem, which would lead to the possibilit y of a non-sacrificial one. To this end, we here sought to construct a CO2 photochemical reduction system sensitized by the dyad (1) on the NiO p-type semiconductorwithout any sacrifi cial electron donor, which would be a half -reaction that would be employed at the reduction terminal end of a full artificial photosynthetic system to be constructed. Nickel (II) oxide is a well known p-t ype sem iconductor,
3 9 - 4 3 which has its valence band at around +0.1 V vs . standard hydrogen
electrode(SHE) and is expected to donate an electron to the ex cited state and/or radical cation of Zn -DMCPP in the dyad (1) (Eo x =+0.4 V vs. SHE).The fluorescence behavior of ZnTCPP (5) and t he dyad (1) adsorbed on NiO were thus examined to check whether or not the hole injection into NiO from ZnTCPP (5) and the dyad (1) take place.
ZnTCPP (5) was adsorbed on the NiO nanoparticle/FTO -electrode (see Experimental) and the fluorescence from 5 upon picosecond laser excitation was monitored b y a streak camera (see Experimental). As shown in Figure 3, the fluorescence lifetime of 5 adsorbed on NiO is greatl y diminished , down to 30 ps , compared with 1.9 ns for 5 in DMF solution, strongl y suggesting that an efficient electron transfer to the excited 5(98.4% efficiency, (1/30ps)/[(1/30ps)+(1/1.9ns)]=0.984), that is, an effici ent hole injection into NiO from the excited 5 to form the radical anion of 5, is induced upon laser excitation of 5 on NiO.
118
Figure 3 Fluorescence deca y profile of a) ZnTCPP (5) in H2O, on NiO, and b) the dya d ZnDCPP -Re(bp y)NHAc (1) in DMF and on NiO.
(Excited at = 400 nm with an Nd3 +YAG laser-pumped OPG system, [5]
= 2 x 10- 6M in H2O, [1] = 2 2 x10- 6 M in DMF, see Experimental for the preparation of the dye samples on NiO.)
Figure 4 Transient absorption spectra of a) ZnTCPP (5) in de-aerated H2O, b) on NiO under nitrogen atmosphere observed at 10 ns delay time after the laser pulse ( = 400 nm) from an excimer laser-pumped dye laser. (Lambda Physik , Compex 102 + Lumonix, HD-300; FWHM 10 ns, 1 Hz)). ([5] = 2 x 10- 6M in H2O, [1] = 2 2 x10- 6 M in DMF, see Experimental for the preparation of the dye samples on NiO.)
119
Actuall y, nanosecond laser flash photol ysis of ZnTCPP/NiO exhibits a formation of the radical anion with a t ypi cal transient spect rum (sharp band with m a x = 450 nm, Fig. 4a)) , which is evidentl y different in its
m a x and shape with those of the triplet state of ZnTCPP (3ZnTCPP*) in
aqueous solution (tailed broad band withm a x = 460 nm, Fig. 4b)).
The radical anion of ZnTCPP on NiO deca yed with single-exponential behavior(lifetime 180 ns ), which corresponds to a charge recombination between the radical anion of the adsorbed ZnTCPP and the hole on the surface of NiO.(Figure 5) In the case of the dyad (1), the fluorescence of 1is also efficient l y quenched (lifetime 29 ps on NiO) as compared with that in DMF solution (lifetime 1.6 ns), the hole injection
efficienc y can be calculated to be 98.2%
((1/29ps)/[(1/29ps)+(1/1.6ns)]=0.982 ).
Figure 5 Deca y profi le of the radical anion of ZnTCPP (5) on NiO under nitrogen atmosphere observed at 10 ns delay time after the l aser pulse ( = 400 nm) from an excimer laser-pumped dye laser (Lambda Physik , Compex 102 + Lumonix, HD -300; FWHM 10 ns, 1 Hz)). (See Experimental for the preparation of the dye sample on NiO.)
These results indicate that visible light irradiation of the dyad (1) on NiO would induce an effici ent formation of the one-electron-reduced form of the rhenium-complex moiet y ei ther through 1) a fast electron transfer from the valence band of NiO to the excited zinc-porphyrin moiet y of 1 in its S1 state followed by a charge mi gration to the rhenium-complex one within the dyad, or 2) through an electron transfer from the valence band of NiO to neutralizethe hole (radical cation of zinc-porphyrin within the charge separated species, [ZnDMCPP]+ .-[Re(bpy)(NHAc)]- ., which would be formed through the S2 excited state wit hin 0.5 ps.2 0 Either route for the electron transfer from the valence band of NiO eventually lead s to the formation of the one-electron-reduced form of rhenium complex moiet y within the dyad, which will induce the reduction of CO2. (Scheme 3)
120
Photoelectrochem ical reduction of CO2 by the dyad on NiO
As discussed above, upon visible light irradiation of the adsorbed zinc-porph yrin on NiO nanoparticl es , an efficient one -electron reduction of the porphyrin speci es was confirmed. We thus tried to develop the system to include a photoelectrochemical reduction of CO2
on the catal ytic site of the rhenium-complex within the dyad (1), which is expected to act as a half-terminal end for the reduction side of an artificial photosynt hetic syst em without an y sacrificial reagent. An electrochemical s ystem, composed of three ele ctrodes, with zinc-porh yrin/NiO/FTO as a working electrode (see Experimental), platinum wire as a counter electrode, and Ag/AgCl as a reference electrodein DMF in the presence of tetrabutyl ammonium hexafluorophsophate (0.1 M) as an electrol yteunder CO2-saturated conditions was set up , and visible light (=430 nm, 6mW)irradiation of the zinc-porph yrin on the working el ectrode was carried out from the outside window, as depicted in Figure 6.
Figure 6 Schematic illustration of photo-electrochemical setup .
When the dyad (1) was adsorbed on NiO nanoparticles on the FTO electrode (dyad -onl y s ystem), visible light irradiation induced a constant cathodic photocurrent of around 20 A/cm2, as shown in Figure 7.
121
Figure 7 Cathodic photocurrent on visible light irradiation of the dyad (1) adsorbed on Ni O/FTO . (An electrochemical system, composed of three electrodes , wi th zinc-porh yrin/NiO/FTO as a working electrode (see Experimental), platinum wire as a counter electrode, and Ag/AgNO3 as a reference electrode in DMF in the presence of tetrabut ylammonium hexafluorophsophate (0.1 M) as an el ectrol ytein CO2-saturated DMF, visible light (=430 nm, 6mW) irradiation of the zinc-porph yrin on the working electrode was carried out from the outside window.)
The IPCE (incident photon to current efficienc y) was estimated to be 0.91%, and the APCE (a bsorbed photon to current efficiency) , which corresponds to the quantum yield of photocurrent , was 2.3%. The photocurrent for the dyad -onl y system graduall y decreased duringprolonged light irradiation for ca. 50 h to zero current . After the prolonged visible light irradiation of the system, however, GC -anal ysis of a gas sample from the inside the cel l showed an actual formation of CO (0.93mol) wi th an integrated photocurrent of 3.0 coulomb.
Hydrogen was not detected at all , as in the case of Re(bp y)(NHAc)2 in
DMF.2 0 The Faradaic efficienc y for the CO formation was thus
calculated to be 6.2%. The turnover number (TNC O) was 10 after the light irradiation for 50 h . The NiO/FTO electrode did not show detectable wei ght loss after the photoreaction and was able to be repeatedl y used b y desorption and re -adsorption of the dyad. This is the first example of the photochemical reduction of C O2 upon visible light irradiation of a sensitizer with catal yst adsorbed on p-t ype semiconductor NiO, while photoelectrochemical reduction of CO2 on p-t ype semiconduct ors with light irradiation of the sem iconductor themselves have been reported.2 2 - 3 0 However, the rather m odest APCE (2.3%) and the Faradaic efficienc y (6.2%) in the present work were
122
contrar y to expectation from the efficient fluorescence quenching behavior described above. There should be factors controlling the apparent photoelectroche mical reactivit y in the total process , composed of 1) the injection of electron s from NiO to the exciteddyad (1) competing with the charge recombination with the hole on NiO, 2) charge migration from the zinc-porph yrin moiety to the rhenium-complex one, and 3) CO2 reduction processes , including electrophilic attack of the CO2 to the metal center of rhenium. To explore these factors and obtaindeeper insight into the processes, we further examined another system , composed of ZnTCPP (5) adsorbed on NiO/FTO plus Re(bpy)(NHAc)2 (7) i n DMF solution, where 7 is dissolvedin the bulk DMF solution and can receive an electron from the photogenerated one -electron reduced 5 adsorbed on NiO. Very interestingl y, a larger constant cathodic photocurre nt,ca.110 A/ cm2, was observed in th i s case.(Figure 8) The photocurrent, however, also graduall y decreased to nearl y zero level after 30 h of light i rradiation . Formation of CO was also observed and hydrogen was not detected.
The IPCE (5%), APCE (12.5%) and Faradaic efficienc y (13%) were much higher than those for the dyad -onl y s ystem, while the turnover number (TNC O) was very modest (1.5). The Faradaic efficiency (13%)is comparable with the quantum yi eld of photoreduction of CO2 by the parent Re(bpy)(NHAc)2(7) (ca. 10%),2 0 where the observed quantum yi eld correspond ed to the efficiency of the reaction after the one-electron reduced form of Re(bp y)(NHAc)2 (7) was generated by light irradiation, because the photochemical one -electron reduction process by TEA is nearl y quantitative. This means that once the one-electron-reduced form of 7 is generated in solution by receiving an electron from the photogenerated one -electron-reduced form of 5 on NiO, the reduction of CO2 proceeds i n a similar manner with 7 in solution.
123
Figure 8 Cathodic photocurrent upon visible light irradiation of the system with ZnTCPP (5) adsorbed on NiO/FTO plus Re(bpy)(NHAc)2
(7) in DMF solution. (An electrochemical system, composed of three electrodes, with zinc-porh yrin/NiO/FTO as a working electrode (see Experimental), platinum wire as a counter electrode, and Ag/AgNO3 as a reference electrode in DMF in the presence of tetrabut yl ammonium hexafluorophsophate (0.1 M) as an electrol yte in CO2-saturated DMF, visible light (=430 nm, 6mW) irradiation of the zinc-porphyrin on the working electrode was carried out from t he outside window.)
In the case of the dyad 1 on NiO (dyad -onl y s ystem), the Faradaic efficienc y (6.2%) is also comparabl e to the quantum yield of photoreaction for 7. This should be quite interesting for the viewpoint of the reaction mechanism of CO2 reduction by the rhenium-complex.
In spite of th e extensive studies on the reaction mechanism,6 , 3 1 - 3 6 the details are still not so clear, especiall y on how the second electron is provided, from which species and to which intermediate. Clearl y, the intermediate complex in the photoreduction of C O2 b y the rhenium-complex in solution would accept the second electron from another one-electron-reduced rhenium-complex in solution by a diffusional collision process, while the dyad 1 adsorbed on NiO is fixed on the surface and is thus believed to undergo no such collisional process. The dimerization mechanism,6 b , 3 5 thus, does not take place in this case. The present experiment on the dyad 1 on NiO clearl y demonstrates that the second electron can onl y come from beneath the adsorbed molecule, from the semicon ductor NiO through the zinc-porph yrin moiet y within the d yad. This finding is very cruci al from the viewpoint of the reaction mechanism and designing a CO2
photoreduction syst em with a non-sacrificial electron donor.
Im proving the reactivity by introduc ing a concept of
“electron-harvesting” sensitized by surrounding m olecules
The current d yad -onl y s ystem on NiO would surel y serve as a promising candidate for the reductive terminal end of an artificial photosynthetics ystem , even though the net reactivit y of the current half system for CO2 reduction is rather modest. The higher reactivit y of the system with ZnTCPP(5)/NiO plus 7, however, ma y have some suggestive features for improving the reactivit y of the d ya d system.A deeper insi ght into the reactivit y wa s examined as follows. The net two-electron transfer from NiO to the terminal rhenium-complex, which corresponds to the cathodic photocurrent, is considered to be composed of 1) the charge separation to form [ ZnDMCPP]+ .-[Re(bp y)(NHAc)]- . followed by an elect ron transfer from Ni O resulting in the formation of ZnDMCPP-[Re(bpy)(NHAc)]- .or 2) the first electron transfer from the valence band of Ni O to the excited zinc -porph yrin moiet y of 1 to form [ZnDMCPP]- .-Re(bpy)(NHAc) followed by the ele ctron migration within the diad (1) leading to ZnDMCPP-[Re(bpy)(NHAc)]- ., 3)the
124
electrophilic attack to the one -electron-reduced rhenium-complex by CO2 to form a CO2-incorporated intermediate -complex, 4) the second electron injection from NiO to the excite d zinc-porph yrin moiet y of 1, and 5) the second electron migration to the CO2-incorporated intermediate-complex to lead to the reduction of CO2.(Scheme 4 )
Scheme 4 Elecron flow in the dyad (1) on NiO composed of five
processes.1) the charge separation to form
[ZnDMCPP]+ .-[Re(bpy)(NHAc)]- . followed b y an electron transfer from NiO resulting in the formation of ZnDMCPP-[Re(bpy)(NHAc)]- . or 2) the first electron transfer from the valence band of NiO to the excited
Zn-porph yrin moiet y of 1 to form
[ZnDMCPP]- .-Re(bpy)(NHAc) followed by the electron migration within the diad (1) leading to ZnDMCPP-[Re(bpy)(NHAc)]- . (Top column and the second one), 3) electrophilic attack to the one -electron reduced rhenium -complex by CO2 to form CO2 incorporated intermediate-complex (The third column, right), 4) the second electron injection from NiO to the excited zinc -porphyrin moiety of 1(The fourth column, right , left arrow), and 5) the second electron migration to the CO2 incorporated intermediate -complex to lead to the reduction of CO2(The fourth column, right , ri ght arrow ).
Each electron transfer (1) -5)) would proceed subsequentl y step by step, that is, even when the second electron is injected to the excited
125
zinc-porph yrin moiet y from NiO (4)), the electron would not be transferred to the terminal rhenium-complex, when proce ss 3 has not been completed, to form the doubl y one -electron-reduced form such as [ZnDMCPP]- .-[Re(bpy)(NHAc)]- .. The radical anion of zinc -porphyrin would be obliged to undergo a back-el ectron transfer to the hole of NiO within 180 ns, as observed above in the decay, without migrating to the terminal rhenium-complex , to lead to the low net electron transfer efficienc y. Detailed consideration thus leads to the conclusion that the injection of an electron (process es 1, 2 and 4) from NiO and the reaction with CO2 at the terminal rhenium-complex (process 3) require a sort of “harmonization” to be well balanced with each other. The flux of electron injection and its timing should be harmonized with the reaction with CO2 at the terminal rhenium-complex. On the other hand, in the s yst em of 5/NiO plus 7 in solution, the electron bal ance ma y be more feasibl y harm onized by capturing the electron of [ZnTCPP (5)]- . on NiO randoml y by the rhenium-com plex, 7, in solution through a diffusional collision. In th at case, the resultant one -elect ron-reduced form of 7 can temporaril yremain in solution and would serve also as the reservoir for the second electron forthe reduction of CO2. When the one-electron-reduced 7 in solution reacts with CO2 to form the CO2-incorporated intermediate, another one -electron-reduced 7 alread y present in solution could provide the second electron at a reaction rate depending on the rate constant and the concentration of the one-electron-reduced form of 7 itself in solution. The lower APCE in the case of the dyad–onl y s ystem (1) than that for the second case, i.e., ZnTCPP(5)/NiO plus 7 in solution, strongl y suggests that the timing and the frequenc y of exciting the zinc-porphyrin moiet y in the dyad (1) are not well harmonized with the reaction of the rhenium-complex with CO2 at the terminal of the dyad, that is, the excitation frequenc y mi ght be larger compared with the latter reaction, which will fail in the second electron migration but can induce its recombination with the hole on NiO within 180 ns. Accordi ng to the photon -flux -densit y consideration [2]for the current system with =4.9 x 105 M- 1cm- 1 under the light intensit y of 6mW at 420 nm, the average frequenc y of exciting the zinc-porphyrin moiet y can be estimated to be 2.0 s- 1 when the chromophore is fixed on the elect rode. The reaction of the one-electron-reduced form of the rhenium-complex with CO2 would be much slower than the excitation rate. Actuall y, the lifet ime of the one-electron-reduced form of some rhenium-complex under CO2
saturated condition s was reported to be long in the order of several tens of seconds.8
These considerations strongl y prompted us to examine thecontrol of the frequenc y of excitation , which corresponds to the frequenc y of the second electron appearing on the zinc-porphyrin moiet y. One of t he options to control the frequenc y should be changing the incident light intensit y. From the viewpoint of artificial photosynthesis, however, sunl ight radiation should be full y utilized.
126
Controlling the inci dent light intensit y, lowering it in this case, has thus not appeared attractive, so that other options should be more preferable to improve the reactivit y. To control the timing and frequenc y of excitation is not straight forward, such that we sought to consider another approach to providingan excess amount of the second electron in the surrounding area of the dyad. When proper electron donors are available in the neighborhood of the dyad, t hey could provide their electrons to the zinc-porphyrin moiet y in the dyad (1) when the terminal rhenium-complex is prepared to accept the second electron by forming the CO2-incorporated rhenium-complex through the reaction of the first one -electron-reduced rhenium-complex with CO2. The approachma y be termed “sensitization through electron harvesting.” With this idea, we examine the sensitization by the same chrom ophore, ZnTCPP(5) with the dyad. ZnTCPP(5) and the dyad (1) were co-adsorbed on NiO in 1:1, 9:1, and 24:1 ratios without changing the total concentration (co -adsorption system) , and visible light irradiation ( = 420 nm) was carried out in CO2-saturated DMF. Since the extinction coefficients at the excitation wavelength ( = 420 nm) of ZnTCPP (5) and the dyad (1) are nearl y the same, the two different NiO films should have the same light absorption , but the dyad has onl y 1/2, 1/10 and 1/25 of the absorption among the total number of photons absorbed in each co -adsorption system. The APCE was thus calcul ated on the basis of the light absorption by the dyad itself. Very interesting results were obtained. The APCE was very much enhanced in the co-adsorption system (APCE = 14% for the 1:1 system, 66.5% for the 9:1 system, and 48.3% for the 24:1 system) as compared with the dyad-onl y system (APCE = 2.25%). Among the co -adsorption conditions, prolonged visible light irradiation was carried out in the case of 24:1 co -adsorption, as shown in Table 2.
Table 2 Photoelectrochemical reduction of CO2 sensitized by the dyad (1) on NiO
S ys t e m P o r p h yr i n Ad s o r b e d o n
N i O / mo l
P h o to - c u r r e n ta/A
I P C E AP C E / %
I n t e g r a t e d n u m b e r o f e l e c t r o n / c o u l o mb
C O /m o l
T NC
Ob F a r ad a i c y i e l d /
%
R v a l u ec D ya d -
o n l y s ys t e m
D ya d (1)
9 . 0 x 1 0- 8 2 0 0 . 9 1
2 . 2 5 3 . 0 0 . 9
3 1 0 6 . 2 1 . 0 C o -
a d s o r p t i o n s ys t e m
D ya d (1) 3 . 6 x 1 0- 9 Z n T C P P (5) 8 . 6 4 x 1 0- 8
1 7 0 . 7 7
4 8 . 3 6 . 0 0 . 4
6 1 2 2 1 . 5 3 5 . 3
a) Cathodic photocurrent at the initial stage within 5min.
b) Turnover number of CO formation based on the dyad (1) concentration used.
c) R value relative to the dyad -onl y s ystem .
127
The cathodic photocurrent s in the initial stage for several tens of minutes irradiation were ver y similar among the d yad -onl y system (20
A/cm2, IPCE = 0.91%) and the co -adsorption system (17 A/cm2, IPCE = 0.77%), while the APCE for th e dyad-onl y s ystem (2.25%) was very much enhanced , by a factor of 21.5 in the co -adsorpt ion system (48.3%). The photocurrent for the co -adsorption system exhibited a ratherstable profile; even during prolonged light irradiation( 115 h) onl y a 50% decrease was observed. This is quite in contrast to the result for the dyad-onl y s ystem, in which the photocurrent nearl y disappeared after 50 h li ght irradiation, as described above. The turnover number of CO formation for the co -adsorption system (TNC O) was also very much enhanced, b y a factor of ten , reaching 120, and could be much higher, if the light irradiation were prolonged further. The apparent Faradaic efficienc y based on the amount of CO formation (0.46 m ol) and the integrated photocurrent ( 6.0 C) was 1.53%. The curious point on how much enhancement is attained by the co -adsorption syst em can be summarized by com paring the relative reactivit y from the viewpoint of 1) APCE at the initial stage, 2) TNC O, and 3) the ratio of the apparent quantum yield (R) that was calculated by R = (APCE x Faradaic efficienc y)c o - a d s o r p t i o n/(APCE x Faradaic efficienc y)d y a d - o n l y, as tabulated in Table 2. The R value should represent the relative reactivity between the two systems(d yad -onl y and co-adsorption) when th e photocurrents exhibit constant values in both cases during the light irradiation , because the term (APCE x Faradaic efficienc y) corresponds to the quantum yi eld in each case. However, such an assumption is not valid here: the photocurrent decreased in th e case of dyad -onl y s ystem, while the co-adsorption system showed rather good stabilit y. Even in the present case, the R value can be a measure of the relative reactivit y.
In all three aspects described above, the co -adsorption syst em showed a remarkable enhancement compared with the dyad -onl y s ystem. The enhancement factor of the R value (ca. 5.3) indicates that at least 5 to 6 sensitizer molecules (ZnTCPP (5)) surrounding the dyad (1) on NiO (the first coordination region) are participating in the photoreduction of CO2 in the terminal end of the dyad b y t ransferring their electrons that were generated b y the electron transfer from NiO to the excited ZnTCPP (5). The mean distance between the adsorbed ZnT CPP (5) and the dyad (1) can be estimated from the numberof adsorbed molecules and the surface area of NiO. The mean occupation areas in these case s are ca. 0.5 nm2 in both cases for the dyad -onl y s ystem and the co-adsorption one, which means that the intermolecul ar distance between the molecules on NiO is ca. 0.7 nm. Electron migration by hopping can proceed over such a distance . The results thus can be rationalized as follows. The sensitizer, ZnTCPP (5), co-adsorbed with the dyad (1) surel y serve s as a sensitizer through the subsequent processes of 1) light absorption to form the radical anion of 5 (light harvesting) b y electron transfer from NiO, 2) electron migration by hopping over the neighboring ZnTCPP molecules, competing with the
128
charge recombination with the hole on NiO, 3) electron transfer to the zinc-porph yrin moiet y of the dyad (1) (electron harvesting). The sensitizer ZnTCPP served as an electron reservoir for the dyad (1) as well as the light -harvesting agent. The enhancement effect mi ght be termed “sensitization through li ght harvesting coupled with electron harvesting.” Even t hough the experiments are yet preliminary, the se studies suggest a promising example of the design of a CO2
photoreduction s yst em as the reduction terminal end o f an artificial photosynthetic syst em. We have been developing the two -electron activation of water induced by one photon excitation.4 4 - 5 0 Coupling the CO2 reduction system with the oxidative activation of water both as the terminal ends would be most attractive for realizing an artificial photosynthesis by visible light. Further studies to optimize the conditions and to clarify the detail s of t he sensitizat ion mechanism are now in progress.