Cover Page
The handle http://hdl.handle.net/1887/36115 holds various files of this Leiden University dissertation
Author: Limburg, Bart
Title: Photocatalytic redox reactions at the surface of liposomes Issue Date: 2015-11-12
Chapter 7
7
Stabilization of a ruthenium photosensitizer for photocatalytic water oxidation by binding to a
liposome bilayer
A ruthenium(II) trisbipyridine photosensitizer was anchored to the lipid bilayer of liposomes together with a Ru-, Co-, or Ir-based water-oxidation catalysts in order to study the influence of liposomes on photocatalytic water oxidation. Membrane anchoring caused a large shift in the quantum yield of oxidative quenching of the photosensitizer excited state, which decreased from 180% in homogeneous solution to 8% at the surface of liposomes. In the latter system the electron-transfer rate between the photosensitizer and the water-oxidation catalyst is increased relative to the oxidative quenching rate. Consequently, the rate-limiting step in photocatalytic water oxidation is the oxidative quenching at the surface of liposomes, whereas it is the reduction of the oxidized photosensitizer by the catalyst in homogeneous solution. Overall, compared to a homogeneous solution a lower oxygen-production rate was observed when photocatalytic water oxidation occurred at liposomes. As a result, the stability of the liposomal system is increased compared to that of the homogeneous system because degradation of the photooxidized photosensitizer is lowered by fast electron transfer to the catalyst.
Finally, the liposomal system was found to be more tolerant to changes in light intensity, which are expected for sunlight-driven solar fuel production devices.
This chapter is to be submitted as a full paper: B. Limburg, J. Wermink, S.S. van Nielen, R. Kortlever, M.T.M. Koper, E. Bouwman, S. Bonnet, manuscript in preparation.
Chapter 7
7.1 Introduction
To fulfill the future energy needs of our growing society, alternative sources of energy are needed. One of the most promising alternatives to the burning of fossil fuels is utilizing the energy of the sun, which can be stored by using photocatalytic water splitting. In this reaction water is converted into its parent elements, dihydrogen and dioxygen, under the influence of visible light as energy source. The reaction requires multiple steps to be catalyzed. First, light must be absorbed and a charge-separated state must be created. As photoinduced redox reactions comprise inherently single electron-transfer processes, catalysis is needed to temporarily store the electrons or holes required for proton reduction or water oxidation, respectively.
Thanks to the research efforts of the chemical community in artificial photosynthesis, many compounds have been reported that fulfill these challenging catalytic tasks. Especially the catalyst that is able to convert water into dioxygen has been the subject of elaborate studies leading to catalysts that are very active and stable under chemocatalytic conditions using the sacrificial oxidant cerium(IV) ammonium nitrate.1-4 In contrast, the molecule that is responsible for the absorption of light and the subsequent first essential reaction leading to a charge-separated state, i.e., the photosensitizer (PS), has been much less optimized. The most-studied photosensitizers are [Ru(bpy)3]2+ and its derivatives. Unfortunately, the oxidized PS, [Ru(bpy)3]3+, is unstable under operating conditions,5-7 leading to only moderate photocatalytic activity for water oxidation catalysis (see Chapter 6). Furthermore, many intermediates are produced during the photocatalytic oxidation of water that can react with each other. To prevent these charge-recombination reactions, electron transfer between the water oxidation catalyst and the photosensitizer should be fast and unidirectional. Unidirectional processes in which intermediates are separated in space are, however, impossible to perform in homogeneous solution, in which most of the water-oxidation catalysts were tested.
Therefore, we undertook the anchoring of known photosensitizers and water- oxidation catalysts to the surface of liposomes. Such anchoring has indeed been proven useful for various photocatalytic redox reactions related to energy conversion.8-10 By anchoring the components of a photocatalytic system to a bilayer membrane, electron transfer might be made unidirectional, and because the compounds are close to each other at the nanometer scale it is likely that electron-transfer processes between the different components will occur more rapidly than in homogeneous solution. We report the effects of such membrane
anchoring on the kinetics and mechanism of photocatalytic water oxidation, and interpret the increased stability observed for membrane-anchored water- oxidation systems compared to homogeneous conditions.
7.2 Results
7.2.1 Synthesis
The [Ru(bpy)3]2+ photosensitizer, as well as three different water-oxidation catalysts based on different metal ions (i.e. Ru2+, Co2+ and Ir3+), were functionalized with alkyl tails to allow for anchoring to the lipid bilayer of liposomes. The structures of the complexes used in this study are depicted in Scheme 7.1. All four compounds were synthesized as described in the experimental part and characterized by e.g. NMR, mass spectrometry, UV-vis spectroscopy, and elemental analysis. In the photosensitizer [1]2+ one of the bipyridine ligands was covalently linked to two alkyl tails via amide linkers. We found that one of the (unidentified) side-products of the synthesis of the photosensitizer [1](PF6)2 (see Scheme 7.1) quenched the photocatalytic activity of the water-oxidation system, even when present in quantities not detectable by NMR spectroscopy. Thus extensive purification by column chromatography was required to obtain compound [1](PF6)2 in a pure form. The neutral compounds [2], [3] and [4] (Scheme 7.1) were based on the three known water oxidation catalysts [Ru(bda)(pic)2] (H2bda = 2,2’-bipyridine-6,6’-dicarboxylic acid, pic = 4- methylpyridine),1 [Co(salphen)]11 (H2salphen = N,N’-bis(salicylaldehyde)-1,2- phenylenediamine) and [Ir(Cp*)(dmiz)Cl2] (dmiz = 1,3-dimethylimidazol-2- ylidene),2 respectively. Also for these catalysts the alkyl tails were covalently attached via an amide linker on the pyridine, salphen, or carbene ligand,
Scheme 7.1. Molecular structures of alkyl-tail functionalized photosensitizer [1]2+, and water
Chapter 7
respectively. The preparation of compound [2] was difficult due to its sensitivity to dioxygen. For this compound it was necessary to perform column chromatography under argon using deaerated solvents to obtain a brown RuII compound that is characterized by a diamagnetic NMR spectrum. When these precautions were not taken, the sample turned green, indicating the formation of RuIII species.
For the preparation of liposomes, the photosensitizer [1](PF6)2 and catalyst [2], [3], or [4], were mixed in various ratios (see Table 7.1) from chloroform stock solutions with the lipids 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) or sodium 1,2-dimyristoyl-sn-glycero-3-phosphoglycerol (NaDMPG). A small amount of sodium 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine-N- [methoxy(polyethylene glycol)-2000] (NaDSPE-PEG2K) was also introduced in a 1:100 ratio relative to DMPC/NaDMPG to stabilize the liposome suspension in water and avoid aggregation. The solution of lipids was evaporated and dried to form a lipid film that was hydrated with a phosphate buffer (10 mM, I = 50 mM, pH = 7.0). The suspension was then sonicated for 30 min at 323 K and left to equilibrate overnight. According to DLS, all prepared liposomes had an average diameter of 100 ± 15 nm, and were characterized by PDIs smaller than 0.2. The liposomes were further diluted with a phosphate buffer containing Na2S2O8 so that the final concentration of [1]2+ was 50 µM. Table 7.1 summarizes the composition of all liposome samples used in this study.
7.2.2 Dioxygen evolution on liposomes
To measure the photocatalytic generation of O2, the liposome-containing solution was initially deaerated by gentle helium bubbling. The sample was stirred constantly, and visible light irradiation was performed using either a blue LED or white light from a 1 kW Xenon lamp. The evolution of O2 was measured in the gaseous phase by GC and in solution by a NEOFOX fluorescence probe. The experimental setup for photocatalytic water oxidation is described in Appendix E.
Upon irradiation the samples LC12 (Liposome sample of DMPC, containing compounds [1]2+ and [2]) and LC13 evolved dioxygen, whereas the samples LC14, LC1 and LG12 did not show any O2 evolution. Figure 7.1 shows the evolution of O2 as a function of irradiation time for LC12, LC13, and for the homogeneous system H described in Chapter 6 for comparison. The turnover number (TON) was calculated by dividing the maximum amount of O2 produced as measured by GC in the headspace by the number of moles of catalyst present.
The photocatalytic turn-over number (PTON) was calculated by dividing the maximum amount of O2 produced by the number of moles of photosensitizer
introduced in the system, and by multiplying by a factor 4. This factor 4 accounts for the fact that 4 molecules of photosensitizer must turnover once (or one molecule must turnover 4 times) for the formation of each molecule of O2. With the exception of the experiments described in Section 7.2.7, the TON and PTON are proportional due to the fact that the relative proportions of the photosensitizer and water-oxidation catalyst was kept constant (10:1). In such conditions, both the TON and the PTON provide an equivalent quantification of the stability of a photocatalytic system, and this independently of the species that decomposes the fastest. However, because we assume that the decomposing species in our experiments is the photosensitizer (see Chapter 6 and Section 7.2.4), we will use the PTON as a measure of stability of the photocatalytic system. Likewise, the maximum turnover frequency (TOF) and maximum photocatalytic turn-over frequency (PTOF) were calculated from the initial slope of the evolution of dissolved oxygen, as measured by the NEOFOX probe. As we expect that the catalyst does not limit the rate of the evolution of O2, we will use the PTOF as a measure of activity of the photocatalytic system.
Most importantly, both liposomal samples LC12 and LC13 evolve dioxygen at a lower rate than the homogeneous sample H (see also Chapter 6), but they continue to do so for a longer period of time (Figure 7.1). The liposomal samples thus show a higher stability, i.e., they produce more O2 than the homogeneous control. The fact that compound [4] does not catalyze the evolution of O2 in the liposomal sample is surprising and as of yet unexplained, as it was shown in Chapter 6 that [Ir(Cp*)(dmiz)(OH)2] does function as water-oxidation catalyst in
Table 7.1 Compositions of liposome samples.
Samplea [lipid] (mM) [catalyst] (µM) [Na2S2O8] (mM) PSe LC12 DMPC (0.125 – 1)b [2] (0.94 – 7.5)c 1 – 33d [1](PF6)2
LC13 DMPC (0.19) [3] (5) 2.5 [1](PF6)2
LC14 DMPC (0.19) [4] (5) 2.5 [1](PF6)2
LG12 NaDMPG (0.19) [2] (5) 2.5 [1](PF6)2
LC1 DMPC (0.125 – 1) – 1 – 33d [1](PF6)2
LG1 NaDMPG (0.19) – 2.5 [1](PF6)2
He – [Ru(bda)(isoq)2] (5) 1 – 33d [Ru(bpy)3]Cl2 a Samples were prepared in a phosphate buffer (10 mM, I = 50 mM, pH = 7.0). Indicates concentrations are bulk concentrations. b If the bulk concentration of DMPC is not specified in the figure or text, then a standard 0.19 mM was used. c If the bulk concentration of catalyst is not specified in the figure or text, then a standard 5 µM was used. d If the bulk concentration of Na2S2O8
is not specified in the figure or text, then a standard 2.5 mM was used. e The photosensitizer was present in 50 µM bulk concentration.
Chapter 7
homogeneous photocatalytic conditions. This first surprising result demonstrates that not every catalyst that is active in homogeneous solution can be used for supramolecular photocatalysis, and that more research is necessary to be able to employ Ir-based catalysts on the surface of liposomes. On the other hand, liposomes functionalized with catalyst [2] but based on negatively charged lipids, i.e., NaDMPG (LG12) are also inactive for photocatalytic water oxidation. Such inactivity may be due to electrostatic repulsion of the peroxodisulfate electron acceptor from the photoactive lipid bilayer (see Section 7.2.5). Overall, these observations demonstrate that liposomal systems bring additional complexity to the photocatalytic water-oxidation reaction, compared to homogeneous media.
7.2.3 Varying the light intensity
The anchoring of the photosensitizer and of a water-oxidation catalyst to the surface of DMPC liposomes was found to lead to an increase in stability of the
Figure 7.1. Evolution of dioxygen vs. irradiation time for photocatalytic liposomes irradiated in presence of Na2S2O8. The produced O2 was measured in solution by a fluorescence probe (solid curves) and in the headspace by GC (data points). Before the headspace was injected into the GC, the dissolved oxygen was equilibrated with the headspace by bubbling the headspace through the solution (see Appendix E). The blue data represent sample LC12, the green data represent sample LC13 and the red data represent the homogeneous control H. Conditions: DMPC (0.19 mM), NaDSPE-PEG2K (1.9 µM), catalyst (5 µM), [1](PF6)2 (50 µM), Na2S2O8 (2.5 mM) in 5 mL phosphate buffer (10 mM, I = 50 mM, pH = 7.0), λirr = 450 nm, θ0 = 30 nmol·s–1, T = 298 K. The data points for the GC measurement decrease after a maximum due to consumption of the headspace, together with increasing error in measurement due to air leaking, see Appendix E. The dips in the curves of the emission probe, and the decrease after a maximum, are caused by the equilibration procedure, see Appendix E. The corresponding PTON was calculated from the measured amount of O2 present by dividing by the number of moles of [1]2+ and multiplying by 4 (4 electrons are removed from water in order to produce O2).
system compared to homogeneous solution.10 To find the reason for such increased stability the dependence of the maximum PTOF and PTON of liposome samples LC12 and LC13 on the light intensity was investigated. Due to the use of different light sources (blue vs. white light), and due to the different overlap of the irradiation spectrum and absorbance spectrum of the photosensitizer, the light intensity was expressed as the number of excited states produced per second at t = 0, θ0 (see Appendix F). Up until very high values of θ0, i.e., beyond a value corresponding to the solar constant of 136 mW·cm–2 (θ0 > 0.21 µmol·s–1), the initial PTOF increased almost linearly as a function of θ0 (Figure 7.2a), indicating that the photocatalytic water-oxidation process is limited by the amount of incoming photons, similar to homogeneous samples (see Chapter 6).
This phenomenon was observed for both samples LC12 and LC13, which both showed a maximum initial PTOF of 0.15 s–1 at θ0 = 0.49 µmol·s–1. As a comparison, under homogeneous conditions sample H showed a slightly higher maximum PTOF of 0.2 s–1 (see Chapter 6).
Remarkably, the stability (PTON) did not vary significantly with the light intensity (Figure 7.2b). This is in contrast to what was observed with the homogeneous sample H, for which the PTON (and thus also the TON) decreased steadily as the light intensity increased (see Chapter 6). By definition, when θ0
increases the concentration of excited state *[1]2+ will increase. In sample H decomposition of the photocatalytic system is mostly due to the instability of the oxidized photosensitizer [Ru(bpy)3]3+ in homogeneous aqueous solution. An increase in θ0 thus leads to a higher concentration of [Ru(bpy)3]3+, which leads to faster decomposition. For liposome systems, it is probable that an increase in light intensity does not lead to a higher concentration of [1]3+, thus not increasing the rate of decomposition. This in turn explains the higher stability of the photocatalytic system LC12 relative to H especially at high values of θ0. At a light intensity that equals that of the sun, liposomal system LC12 showed a PTON of 21, whereas the homogeneous system H stopped evolving dioxygen after a PTON of 13. Additionally, liposomal system LC13 showed a PTON of 18. At lower light intensities, the PTONs of the homogeneous and liposomal systems were more comparable, ranging between 14 and 16 for both water-oxidation catalysts [Ru(bda)(pic)2] and [2]. It is therefore important to study systems at a light intensity comparable to that of the sun. In summary, the liposomal systems are more tolerant to changes in light intensities, which is important as systems driven by sunlight are expected to see changing weather conditions.
Chapter 7
Figure 7.2. Dependence of the PTOF (a) and PTON (b) of photocatalytic water-oxidation at the surface of liposomes on the amount of photons absorbed by [1]2+ per second at t = 0. Light grey triangles: sample LC12; dark grey diamonds: LC13. Conditions: DMPC liposomes (0.19 mM containing 1.9 µM NaDSPE-PEG2K, size 100 nm), catalyst (5 µM), photosensitizer [1](PF6)2 (50 µM), electron acceptor Na2S2O8 (2.5 mM), in 5 mL phosphate buffer (10 mM, I = 50 mM, pH = 7.0) , T = 298 K. Irradiation was performed using either a blue LED (filled data points) or a Xenon lamp fitted with UV and IR filters (open data points).
7.2.4 Monitoring oxidative quenching of the photosensitizer
To investigate why the PTON of liposome sample LC12 does not significantly change with light intensity a UV-vis spectroscopy study was performed. First, the extinction coefficient of [1]2+ was determined in acetonitrile (ε450 = 12.3 · 10Z M :· cm :). Afterwards, the extinction coefficient of [1]3+ was determined by spectro-electrochemistry: an acetonitrile solution of [1](PF6)2 was prepared and electrochemically oxidized at a carbon sponge bulk electrode at a potential E
= +1.7 V vs. Ag|AgCl. The evolution of the UV-vis spectrum of the solution during oxidation is depicted in Figure 7.3a. During oxidation the MLCT-band of [1]2+ at 450 nm was gradually replaced by a new band at 320 nm ascribed to [1]3+ (ε320 = 33.0 · 103 M–1·cm–1 in acetonitrile).
In a second step, a liposome sample without catalyst (LC1, see Table 7.1) was prepared, the electron acceptor Na2S2O8 was added, and the sample was deaerated by argon bubbling. Upon irradiation with blue light (λirr = 450 nm, I = 8 mW·cm–2, θ0 = 33 nmol·s–1), a similar change in the UV-vis spectrum of the solution was observed (Figure 7.3b), i.e., a new band appeared at 320 nm concomitantly to the disappearance of the MLCT band of [1]2+ at 450 nm. The evolution of the spectrum of sample LC1 under irradiation slightly differed from the one obtained by electrochemical oxidation (Figure 7.3a), which is ascribed to
the different conditions of the two experiments, i.e., a homogeneous acetonitrile solution vs. DMPC liposomes in an aqueous solution. However, the similar spectroscopic changes in the electrochemical and photochemical experiments are interpreted as a consequence of the oxidation of [1]2+ to [1]3+. Thus, in the following analysis the extinction coefficient of [1]3+ determined by spectroelectrochemistry in acetonitrile is used as an approximation of that of [1]3+ in liposomes.
Figure 7.3. a) Difference UV-vis spectra during electrochemical oxidation of an acetonitrile solution of [1](PF6)2. Conditions: tetrabutylammonium hexafluorophosphate (0.1 M), 0.1 mM [1](PF6)2 in acetonitrile, glassy carbon sponge working and counter electrodes, and Ag|AgCl reference electrode. Electrolysis was performed at E = +1.7 V and the working electrode was isolated from the counter electrode by a nafion membrane. b) Difference UV-vis spectra during photochemical oxidation of sample LC1 irradiated with a blue LED (λirr = 450 nm, I = 8 mW·cm–2, θ0 = 33 nmol·s–1) in the presence of Na2S2O8 (2.5 mM). The spikes in the absorbance data are caused by gas bubbles.
Conditions: phosphate buffer (3 mL, 10 mM, I = 50 mM, pH = 7.0), T = 298 K. One spectrum per 6 seconds.
From the extinction coefficients at 450 nm and 320 nm the absolute concentrations of [1]2+ and [1]3+ during photocatalytic experiments can be estimated, as well as the number of mol of [1]3+ , PHTµ?, and the sum PHT¶·¶ of the number of mol of [1]2+ and [1]3+. A plot of PHTµ? and PHT¶·¶ as a function of irradiation time can then be made for each irradiation experiment, assuming that the absorbance due to the catalyst is negligible at the concentration and wavelengths used (Figure 7.4). When sample LC1 was irradiated with blue light in presence of Na2S2O8, i.e., in the absence of catalyst, the oxidized photosensitizer [1]3+ was obtained in much lower quantities than in the corresponding
Chapter 7
homogeneous system containing [Ru(bpy)3]Cl2 instead of [1](PF6)2. The initial quantum yield for the formation of [1]3+ (Ψ[1]3+) was calculated to be 8% at the liposome surface, compared to 180% for the formation of [Ru(bpy)3]3+ in homogeneous conditions (data taken from Chapter 6). Due to the low signal-to- noise ratio in the liposomal conditions, and due to the approximation made for the extinction coefficients, the error on this value of the quantum yield Ψ[1]3+ is rather large. However, the rate of oxidative quenching of the excited state of the photosensitizer obtained for the liposomal system, and thus Ψ[1]3+, is at least one order of magnitude lower than that found for the homogeneous system in Chapter 6. Concurrently, PHT¶·¶ decreases slower in the liposome system, i.e., the photosensitizer decomposes less fast. Both facts are linked, as a lowered oxidative quenching rate leads to lower concentrations of [1]3+ at the lipid bilayer, which in turn increases the stability of the photosensitizer.
7.2.5 Why is oxidative quenching slow on the membrane?
The lower oxidative quenching rate of *[1]2+ by S2O82– at the surface of the DMPC membrane may be explained by two different effects. First, the electronic properties of [1]2+ are significantly different from those of [Ru(bpy)3]2+. The
Figure 7.4. Plot of the number of moles of ruthenium photosensitizers as a function of irradiation time, showing the fate of the photosensitizers during photochemical oxidation by S2O82–. Black traces: liposome sample LC1, PHTtot is the sum of the number of moles of [1]2+ and [1]3+ ( PHTS? and
PHTµ?, respectively) in the sample. Grey traces: homogeneous sample containing [Ru(bpy)3]Cl2 (50 µM). P¹º"»¼€µT¶·¶is the sum of the number of moles of [Ru(bpy)3]2+ and [Ru(bpy)3]3+ in the sample.
P¹º"»¼€µTµ? is the number of moles of [Ru(bpy)3]3+. Data for the homogeneous system was taken from Chapter 6. Conditions: [Na2S2O8] = 2.5 mM, phosphate buffer (10 mM, I = 50 mM, pH = 7.0), T = 298 K, irradiation with the blue LED (I = 8 mW·cm–2, θ0 = 33 nmol·s–1).
presence of two electron-withdrawing amide substituents on one bipyridine ligand increases the redox potential of the RuIII/RuII couple by 80 mV, to an E½ of 1.34 V vs. NHE1 (see Figure D.1, Appendix D). In addition, the phosphorescence emission maximum is redshifted for [1]2+ (653 nm) compared to [Ru(bpy)3]2+
(597 nm, see Figure D.1, Appendix D). Due to the combination of these two effects, the redox potential of the [1]3+ / *[1]2+ couple (E½ ≈ –0.6 V vs. NHE) is approximately 0.2 V more positive than that of the [Ru(bpy)3]3+ / *[Ru(bpy)3]2+
couple (–0.82 V vs. NHE). The thermodynamic driving force for oxidative quenching is thus lower for [1]2+ than for [Ru(bpy)3]2+. The second effect originates from the destabilization of charges in a lipophilic environment. The lipophilic and positively charged environment created by the high local concentration of [1]2+ in a bilayer made of neutral lipids destabilizes the oxidized state [1]3+ and thus decreases the thermodynamic driving of oxidative quenching force further.
To test which of these hypotheses best explains the lower oxidative quenching rate of *[1]2+ by S2O82–, the two photosensitizers [Ru(bpy)3]Cl2 and [1](PF6)2 were tested in the same medium, i.e., a 1:1 mixture of acetonitrile and water containing Na2S2O8 (2.5 mM). After deaeration, the mixture was irradiated with a blue LED (θ0 = 33 nmol·s–1) and the evolution of the UV-vis spectra was monitored. Under these conditions, the oxidized photosensitizer does not decompose on a timescale competing with oxidative quenching. Consequently, both *[Ru(bpy)3]2+ and *[1]2+
are fully oxidatively quenched by S2O82– with a quantum yield of 40% and 10%, respectively, that remains constant during irradiation. The differences in oxidation potential and 3MLCT state energy thus have already a significant impact on the oxidative quenching rate of *[Ru(bpy)3]2+ and *[1]2+ by S2O82–. However, the environment clearly plays a role as well, as in homogeneous phosphate buffer
*[Ru(bpy)3]2+ is oxidatively quenched with a much higher quantum yield (180%) compared to the 1:1 acetonitrile:water solution (40%).
The formation of [1]3+ on the NaDMPG liposome sample LG1 was also monitored as another demonstration of the importance of the environment in which the photosensitizer is located for the rates of electron transfer . In Section 7.2.2 the inactivity of NaDMPG liposomal system LG12 for water oxidation was already described. More specifically, according to UV-vis spectroscopy the formation of [1]3+ does not occur upon blue-light irradiation of LG1 in the presence of S2O82–. Thus, on NaDMPG membranes the oxidative quenching
1 As reference, Ag|AgCl calibrated with ferrocene (Fc) was used. Potentials were
Chapter 7
reaction is completely inhibited. This effect is likely due to the electrostatic repulsion between S2O82– and the membrane, which prevents oxidative quenching to occur with significant rates. In summary, the decrease in oxidative- quenching quantum yield at the surface of liposomes is a combination of electronic changes in the photosensitizer, and of the environmental changes due to the anchoring at the surface of a lipid bilayer.
7.2.6 Varying the concentration of Na2S2O8
The results discussed in Section 7.2.4 and 7.2.5 thus indicate that oxidative quenching might be limiting the quantum yield of O2 production at the surface of DMPC liposomes. To verify this hypothesis, the O2 evolution upon blue-light irradiation of liposomes LC12 was monitored at different concentrations of Na2S2O8 (Figure 7.5a). Over the range of concentrations tested (1 – 33 mM), neither the stability (as indicated by the PTON), nor the rate of O2 evolution (expressed as the PTOF), changed significantly. In contrast, the stability and activity of homogeneous sample H is greatly influenced by a change in the concentration of S2O82–. Especially the PTON of sample H decreases as the concentration of Na2S2O8 increases (see Chapter 6). If oxidative quenching is rate- limiting the production of O2 at the surface of liposomes, an increase in PTOF would be expected with increasing concentration of S2O82– due to an increased rate of oxidative quenching. To explain these seemingly contradictory results, the oxidative quenching of *[1]2+ was monitored for different concentrations of Na2S2O8 using UV-vis spectroscopy, as described above. For liposomal sample LC1, increasing the concentration of S2O82– (2.5 – 33 mM) did not lead to an increase in the quantum yield of oxidative quenching. The evolution of the number of moles of [1]2+ ( PHTS?), [1]3+ ( PHTµ?), and their sum PHT¶·¶ as a function of irradiation time is depicted in Figure 7.5b. Surprisingly, the traces overlap within experimental error, showing that the oxidative quenching of *[1]2+ at the surface of DMPC liposomes is zeroth order in S2O82–. We ascribe this effect to the adsorption of S2O82– ions to the positively charged bilayer membrane (see Chapter 3). If S2O82– adsorbs strongly to the bilayer membrane, its effective local concentration will be high and increasing the bulk concentration of S2O82– may not lead to a further increase in the local concentration. The liposomal system is thus much more tolerant to changes in the bulk concentration of S2O82–, and is much more stable at higher concentrations of S2O82–, than in homogeneous conditions: the maximum PTON reached by sample LC12 at a Na2S2O8
concentration of 33 mM was 15, while the maximum PTON of sample H was 3.
Furthermore, the observed zeroed order in the concentration of S2O82– for the rate of oxidative quenching is not contradictory to the hypothesis that oxidative quenching is rate-limiting the production of O2 at the surface of liposomes.
Figure 7.5. a) PTON (triangles, left axis) and initial PTOF (squares, right axis) of photocatalytic water oxidation of liposome sample LC12 as a function of the concentration of Na2S2O8. Bulk concentrations: DMPC (0.19 mM), NaDSPE-PEG2K (1.9 µM), [2] (5 µM), [1](PF6)2 (50 µM) in 5 mL phosphate buffer (10 mM, I = 50 mM, pH = 7.0). Irradiation: λirr = 450 nm, I = 8 mW·cm–2, θ0 = 30 nmol·s–1. b) Evolution of the number of mol of [1]2+, [1]3+, and of their sum ( PHTS?, PHTµ?, and PHT¶·¶, respectively), as a function of irradiation time upon blue light irradiation of sample LC1 for different concentrations of Na2S2O8. From light grey to dark lines: 2.5 mM, 10 mM, 33 mM Na2S2O8. Bulk concentrations: DMPC (0.19 mM), NaDSPE-PEG2K (1.9 µM), [1](PF6)2 (50 µM) in 3 mL phosphate buffer (10 mM, I = 50 mM, pH = 7.0). Irradiation: λirr = 450 nm, I = 8 mW·cm–2, θ0 = 33 nmol·s–1, T = 298 K.
7.2.7 Effect of the local concentration of catalyst [2]
In order to prove that oxidative quenching is the rate-limiting step in the photocatalytic oxidation of water at the surface of liposomes, the local concentration of catalyst [2] was changed. For liposomal systems in which the photoactive components are bound to the bilayer membrane, the local concentrations are likely to have a great impact on the rate of electron-transfer reactions. However, there are two ways to change the local concentration of a compound in a bilayer membrane. First, the bulk concentration of the lipid can be lowered (method A), which will increase the local concentrations of all liposome- bound species, while the total (bulk) concentration of these species stays the same. Second, the bulk concentration of one of the photoactive species can be changed while keeping the lipid bulk concentration constant (method B). In this case both the bulk and the local concentration of this component change, but the
Chapter 7
local concentrations of the other species remain the same. Method A and method B are thus inherently different.
To study the effect of the local concentration of water-oxidation catalyst [2]
we first varied the bulk concentration of DMPC for sample LC12, while keeping the bulk concentration of [2] and [1]2+ constant at 5 µM and 50 µM, respectively (Method A). Increasing the concentration of DMPC thus decreases the local concentration of the catalyst and of the photosensitizer. Figure 7.6a shows how the maximum PTOF and PTON of photocatalytic water oxidation for liposome sample LC12 changed with the concentration of DMPC. At very low DMPC concentrations the initial PTOF and PTON increase upon increasing DMPC concentration, to reach a maximum initial PTOF and PTON at a DMPC concentration of 0.19 mM. Above this value, the initial PTOF and PTON both decreased. In Chapter 6 we described that in homogeneous solution the rate- limiting process of photocatalytic water oxidation is the electron transfer from the water-oxidation catalyst to the oxidized photosensitizer. Thus, it is desirable to keep [1]2+ and [2] in a high local concentration in the membrane, to increase the rate of the reduction of [1]3+ by [2]. By decreasing the concentration of DMPC from 1 to 0.19 mM, the rate of O2 production is indeed increased as a consequence of the increased local concentration of [1]2+ and [2]. However, decreasing the DMPC concentration below this threshold leads to an inverse effect, i.e., the PTOF and PTON begin to decrease. We ascribe this effect to the higher activation energy required to transfer an electron from [1]2+ on highly functionalized liposomes to S2O82– in the solution: as the local concentration of [1]2+ increases, the environment becomes more positively charged and thus the oxidation of [1]2+ to [1]3+ will become thermodynamically less favorable. This effect was experimentally visualized by monitoring the oxidative quenching of
*[1]2+ in liposome samples LC1 with two different DMPC bulk concentrations (0.13 and 0.19 mM, see Figure 7.6b). When the concentration of DMPC was lowered from 0.19 mM to 0.13 mM, [1]3+ was formed twice slower in the absence of catalyst, and less of it was obtained. Thus, the “volcano plot” shown in Figure 7.6a is due to the combination of the two competing effects described above, with an optimal DMPC bulk concentration of 0.19 mM. At this concentration, the local concentration of [1]2+ and [2] are high enough to ensure that the electron- transfer rate between [1]3+ and [2] is fast and compete with the decomposition of [1]3+. Meanwhile, the local concentration of [1]2+ is not too high, which keeps the positive charge density of the membrane low enough to allow for oxidative quenching to occur with sufficient thermodynamic driving force.
Figure 7.6. a) Dependence of the PTON (triangles, left axis) and initial PTOF (squares, right axis) of photochemical water oxidation at liposomes on the bulk concentration of DMPC in sample LC12.
Conditions: [2] (5 µM), [1](PF6)2 (50 µM), Na2S2O8 (2.5 mM) in 5 mL phosphate buffer (10 mM, I = 50 mM, pH = 7.0), blue light (λirr = 450 nm, I = 8 mW·cm–2, θ0 = 30 nmol·s–1). The error bars represent the standard deviation for at least two duplicate experiments. b) Evolution of the number of moles of [1]2+ ( PHTS?), [1]3+ ( PHTµ?), and of their sum ( PHTtot) as a function of irradiation time of sample LC1 for different bulk concentrations of DMPC. Grey line: [DMPC] = 0.13 mM, black line:
[DMPC] = 0.19 mM. Conditions: Na2S2O8 (2.5 mM) [1](PF6)2 (50 µM) in 3 mL phosphate buffer (10 mM, I = 50 mM, pH = 7.0), blue light (λirr = 450 nm, I = 8 mW·cm–2, θ0 = 33 nmol·s–1), T = 298 K.
The side-effect of lowering the bulk concentration of DMPC is that besides the local concentration of [2], also the local concentration of [1]2+ is changed, which leads to an increasingly positively charged liposomal membrane and thus to a higher energy barrier for oxidative quenching. To study the effect of changing the catalyst concentration without changing the bulk and local concentration of [1]2+
(which were kept at 50 µM and 27 mol%, respectively), only the bulk concentration of [2] was varied while keeping the bulk concentration of DMPC and [1](PF6)2 constant (method B). The two experiments remain comparable by the PTOF and PTON because the bulk concentration of [1]2+ does not change.
Thus, the PTOF and PTON were plotted as a function of the local concentration of [2] (in mol%) in the membrane (Figure 7.7a and Figure 7.7b). At a low local concentration of [2], the same trend was found as with method A: the PTOF and PTON decreased with the local concentration of [2]. This result confirms that the local concentration of [2] indeed dictates the behavior of the system and that the local concentration of [1]2+ does not play a significant role. At the highest local concentration of [2] (4 mol%), however, both methods led to significantly different results (Figure 7.7): the higher local concentrations of [1]2+ obtained by
Chapter 7
method A (40 mol%) compared to B (27 mol%) decreased the PTOF and PTON.
Thus, for an optimized system it is beneficial to have a local concentration of [1]2+
that is smaller than 27 mol%, and a local concentration of [2] that is higher than 2.7 mol%. Furthermore, because the trend between method A and B is the same below a local concentration of [2] of 2.7 mol%, it can be excluded that the catalyst is the decomposing species, or that the rate at which the catalyst turns over is limiting the over O2-production rate, which justifies our use of PTOF and PTON to describe the system instead of TON and TOF.
Figure 7.7. Dependence of the PTOF (a) and PTON (b) of photocatalytic water oxidation in sample LC12 irradiated with blue light on the local concentration of [2] in mol%, relative to DMPC, for two methods (see text). Method A: [DMPC] varies, [1]2+/[2] = 10 (grey diamonds). Method B:
concentration of [2] varies, [1]2+/DMPC = 0.27 (black squares). Bulk concentrations: [1](PF6)2 (50 µM), Na2S2O8 (2.5 mM) in 5 mL phosphate buffer (10 mM, I = 50 mM, pH = 7.0). Irradiation: λirr = 450 nm, I = 8 mW·cm–2, θ0 = 30 nmol·s–1, T = 298 K.
7.2.8 Oxidative quenching in the presence of [2]
The complicated effects of local concentrations of [2] in the membrane on the rate of the water-oxidation reaction and on the overall stability of the photocatalytic system were also studied by monitoring the fate of the photosensitizer by UV-vis spectroscopy using method B. For liposome sample LC12 containing 2.7 mol% of [2], the formation of [1]3+ was not observed (Figure 7.8a). Instead, only a steady decrease in the number of moles of [1]2+, and thus of
PHT¶·¶, was observed. Apparently, due to efficient electron transfer from [2] to [1]3+, followed by the oxidation of water, in such conditions the concentration of [1]3+ is kept below our detection limits. However, at decreasing concentrations of [2] the formation of [1]3+ was observed more and more (Figure 7.8a), indicating a
less efficient electron transfer as indeed the rate of reduction of [1]3+ depends on the local concentration of [2]. A higher concentration of [1]3+ (facilitated by a lower concentration of [2]) furthermore led to faster decomposition as indicated by a faster decrease in PHT¶·¶ in Figure 7.8b and a lower PTON in Figure 7.7b.
However, even when the formation of [1]3+ cannot be observed in our experimental setup, i.e. at a local concentration of [2] of 2.7 mol%, the photosensitizer decomposes over time as indicated by a decrease in PHT¶·¶ in Figure 7.8b. This indicates that decomposition of [1]3+ is not the only decomposition pathway of the photosensitizer. The nature of this additional decomposition pathway is however not subject of the work presented here. In conclusion, at a local concentration in [2] of 2.7 mol% the electron transfer from catalyst [2] to S2O82– via the photosensitizer occurs without visible formation of [1]3+, which means that the photocatalytic oxidation of water under these conditions is limited mainly by the rate of oxidative quenching of *[1]2+ by S2O82–.
7.3 Discussion
By anchoring the photosensitizer [1]2+ and a molecular water-oxidation catalyst to the membrane of a liposome the kinetics of electron-transfer reactions are changed significantly compared to those in homogeneous solution. These
Figure 7.8. Evolution of the amount of [1]3+ ( PHTµ?) (a) and of the sum of the amounts of [1]2+ and [1]3+ ( PHTtot) (b) for different concentrations of catalyst [2] in the membrane of liposome sample LC12. From grey to black: 0 mol%, 0.5 mol%, 1.3 mol%, 2 mol%, 2.7 mol%. Conditions: bulk concentrations [1](PF6)2 (50 µM), Na2S2O8 (2.5 mM), DMPC (0.19 mM), NaDSPE-PEG2K (1.9 µM) in 3 mL phosphate buffer (10 mM, I = 50 mM, pH = 7.0), λirr = 450 nm, I = 8 mW·cm–2, θ0 = 33 nmol·s–1, T = 298 K.
Chapter 7
changes can be attributed to two effects, as depicted in Figure 7.9. The first effect is the electronic influence of the amide substituents of one of the bipyridine ligands on the photosensitizer. The energy of the triplet metal-to-ligand charge- transfer (3MLCT) excited state decreases, and that of the charge-separated state (CSS: [1]3+ + SO42– + SO4•–) increases. The second effect is a destabilization of [1]3+
in the membrane, effectively increasing the energy of the CSS, and thus decreasing the driving force of oxidative quenching further. The combination of these effects is depicted in Figure 7.9 by an increase in energy of the CSS. As depicted in the figure, this leads to an overall increased activation energy (Eact), i.e., a decreased electron-transfer rate from *[1]2+ at the liposomes surface to to S2O82– (red curves), compared to that from *[Ru(bpy)3]2+ in homogeneous solution (black curves).
Thus, for liposome sample LC12 the electron transfer from the water- oxidation catalyst to the photosensitizer (see r2 in Scheme 6.1, Chapter 6) is faster than the oxidative quenching of *[1]2+ (r1 in Scheme 6.1, Chapter 6), which
Figure 7.9. Schematic representation of the changes occurring on the oxidative quenching of *PS by S2O82– when going from the homogeneous system H (black) to the system on liposomes LC12 (red).
For the purpose of clarity, the ground state of PS=[1]2+ (sample LC12) is set to the same potential energy as the ground state of PS=[Ru(bpy)3]2+ (sample H). The singlet metal-to-ligand charge- transfer (1MLCT) excited state is omitted for clarity, and thus the depicted absorption of a photon (hν) is a combination of photon absorption and intersystem crossing from 1MLCT to 3MLCT. GS = ground state, CSS = charge separated state = PS+ + SO42– + SO4•–. The horizontal axis describes the reaction coordinate of the electron transfer reaction.
means that the rate-limiting step of the overall photochemical water oxidation has changed upon anchoring both photoactive molecules on the lipid bilayer. In homogeneous solution, the rate-determining step of the photoreaction is the electron transfer from the water-oxidation catalyst to the oxidized photosensitizer [Ru(bpy)3]3+ (r2 in Scheme 6.1, Chapter 6), whereas at liposomes (LC12) it is the oxidative quenching of the excited photosensitizer *[1]2+ by S2O82–
, provided that the local concentration of [2] in the lipid membrane is high enough (> 2.7 mol%). Such a decrease in reactivity between a photosensitizer anchored to the membrane and an electron acceptor in the surrounding aqueous solution has also been described in Chapter 4 twice and thus appears to be a rather common phenomenon.
This shift of the rate-limiting step with liposomes has consequences for the stability of the photochemical system as well. As all molecules of [1]3+ that are formed photochemically are reduced very quickly by the catalyst, increasing the light intensity to that of the sun does not significantly change the stability of the system (Figure 7.2b). Instead under such conditions another pathway must lead to the decomposition of the photosensitizer. The origin of this decomposition pathway is, however, subject of another discussion. In comparison, under homogeneous conditions a higher light intensity leads to a build-up in the concentration of the unstable species [Ru(bpy)3]3+, which decreases the overall stability of the system. In addition to being more tolerant to variations of the light intensity the liposomal system is more tolerant to changes in the concentration of Na2S2O8. In recent reports about photocatalytic water oxidation, a wide variety of concentrations of Na2S2O8 have been employed (0.1 mM,12 2.5 mM,10 5 mM,11,13 10 mM,14 35 mM,15 or even 50 mM5), which makes comparison difficult as the rates of O2 production and maximum amount of O2 produced in homogeneous solution are highly dependent on the concentration of Na2S2O8. Overall, supramolecular systems such as LC12 liposomes may allow for better comparison of catalysts or photosensitizers, as such liposomal systems are more tolerant to changes in experimental conditions than homogeneous systems.
7.4 Conclusion
The kinetics and stability of photocatalytic water oxidation at the surface of neutral liposomes was investigated for three catalysts based on Ru, Co, and Ir.
The liposomal system containing the Ir-complex [4] did not perform any photocatalytic water oxidation, while catalysts [2] and [3] were photocatalytically active with comparable rates and stabilities. For the liposomal
Chapter 7
system LC12 the rate-limiting step appears to be the oxidative quenching of
*[1]2+ by S2O82–, in contrast to what was observed in homogeneous solution where the electron transfer between the oxidized photosensitizer and the water- oxidation catalyst is rate-limiting (Chapter 6). This finding explains the relatively slow kinetics of water oxidation at the surface of liposomes compared to that in homogeneous solution, but also the increased stability of the liposomal system.
Photocatalytically active liposomes are more tolerant than the homogeneous system to changes in experimental conditions such as light intensity or Na2S2O8
concentration. Thus, liposomes are a promising scaffold for water splitting, provided that the inefficient oxidative quenching of *[1]2+ can be overcome.
7.5 Experimental
7.5.1 General
UV-vis absorbance spectra were recorded on a Varian Cary50 spectrophotometer. All reagents were obtained from Sigma-Aldrich and used as received. N-dodecylnicotinamide was prepared according to a literature procedure.16 Gas chromatography was performed on a Shimadzu GC2010 fitted with a Supelco Carboxen 1010 molecular sieve column, and a Restek Hydroguard fused silica precolumn. See Appendix E for an overview of the GC setup. Both columns were kept at 308 K. Helium was used as the carrier gas, and compounds were detected using a TCD detector operated at 80 mA. O2 was calibrated by the external reference method by injection of known amounts of O2 into the system.
Dissolved oxygen was measured in solution with the Ocean Optics NeoFox system fitted with a Foxy-AR probe. The probe was recalibrated at 298 K in air and in helium prior to each experiment. As irradiation source, an OSRAM Opto Semiconductors LD W5SM LED (λirr = 450 nm, Δλ½ = 25 nm), or a LOT Oriel Xenon lamp fitted with UV and IR filters connected via an optical fiber, was used.
The spectrum of the light coming out of the optical fiber can be found in the Appendix F. A 10 mM phosphate buffer (I = 50 mM) was prepared by dissolving a mixture of KH2PO4 (626 mg, 4.6 mmol), K2HPO4·3 H2O (1.186 g, 5.2 mmol), and K2SO4 (1.70 g, 9.74 mmol) in Milli-Q water (1 L), to reach a final pH of 7.04 at 23
°C. K3[Fe(C2O4)3]·3 H2O was prepared following a literature procedure17 and used within one week. Photocatalytic water-oxidation experiments were performed as explained in Chapter 6 and Appendix E.
7.5.2 Syntheses
N-(2-(1H-imidazol-4-yl)ethyl)dodecanamide, compound 5. Histamine·2 HCl (1.84 g, 16.6 mmol) and Et3N (4.2 mL, 3.1 g, 30 mmol) were dissolved in DMF (30 mL) and dodecanoyl chloride (2.3 mL, 2.2 g, 10 mmol) was added dropwise to the solution. The solution was heated to 323 K and stirred for 20 hours. Afterwards, the solution was cooled to room temperature and H2O (30 mL) was added. A white precipitate formed, which was filtered and washed with NaHCO3 (sat. aq.), H2O and Et2O. The white solid was dried in vacuo to yield 5 (1.97 g, 6.7 mmol, 67%). 1H NMR (300 MHz, CDCl3) δ 7.55 (s, 1H, 1), 6.80 (s, 1H, 2), 6.34 (s, 1H, 6), 3.52 (q, J = 6.4 Hz, 2H, 5), 2.80 (t, J = 6.5 Hz, 2H, β), 2.15 (t, J = 6.5 Hz, 2H, 4), 1.60 (p, J = 7.0 Hz, 2H, γ), 1.26 (s, 16H, δ), 0.88 (t, J = 6.7 Hz, 3H, ω). 13C NMR (75 MHz, CDCl3) δ 135.80 (3), 134.77 (1), 116.44 (2), 39.33 (5), 36.89 (4), [31.92, 29.62, 29.51, 29.36, 29.32, 27.00, 25.79, 22.67] (β+γ+δ) , 14.04 (ω). HRMS [C17H31N3O+H]+: 294.25394 found, 294.25399 calculated.
N-(2-(N,N’-dimethyl-1H-imidazolium-4-yl)ethyl)dodecanamide iodide, compound 6I. Compound 5 (293 mg, 1 mmol), K2CO3 (691 mg, 5 mmol) and MeI (0.3 mL, 0.68 g, 4.8 mmol) were suspended in DMF (50 mL). The mixture was heated and refluxed for one hour. The solution was evaporated to approximately 1 mL under reduced pressure and afterwards CHCl3 (30 mL) was added, leading to the precipitation of excess K2CO3, which was filtered off. The filtrate was evaporated to dryness to yield 6I (337 mg, 0.75 mmol, 75%). 1H NMR (300 MHz, CDCl3) δ 9.21 (s, 1H, 1), 7.25 (s, 1H, 6), 7.23 (s, 1H, 2), 3.96 (s, 3H, 8), 3.94 (s, 3H, 7), 3.57 (q, J = 6.1 Hz, 2H, 5), 2.97 (t, J = 6.0 Hz, 2H, β), 2.25 (t, J = 7.2 Hz, 2H, 4), 1.60 (dt, J = 14.1, 7.7 Hz, 2H, γ), 1.25 (s, 16H, δ), 0.88 (t, J = 6.7 Hz, 3H, ω). 13C NMR (75 MHz, CDCl3) δ 174.64 (α), 136.56 (1), 134.11 (3) 121.36 (2), 36.87 (8), 36.84 (5), 36.69 (4), 34.55 (7), [32.05, 29.80, 29.78, 29.71, 29.62, 29.51, 29.49, 25.93,
Chapter 7
24.01, 22.83] (β+γ+δ), 14.28 (ω). HRMS [C19H36N3O+H]+: 322.28519 found, 322.28529 calculated.
[Ir(Cp*)(6)Cl2], compound [4]. 6I (225 mg, 0.5 mmol) was dissolved in dry, deaerated DCM (30 mL). To this solution was added Ag2O (70 mg, 0.30 mmol) and the mixture was stirred for 4 h under argon. The mixture was filtered through celite and [Ir(Cp*)Cl2]2 (200 mg, 0.25 mmol) was added. The mixture was stirred overnight under argon and afterwards filtered through celite three times.
The solution was evaporated and purified by column chromatography (SiO2, DCM/MeOH 92/8, Rf = 0.5). The orange band was collected and further purified by size-exclusion chromatography to remove traces of [Ir(Cp*)Cl2]2 (Sephadex LH-20, MeOH). The orange band was collected and evaporated to yield [4] (148 mg, 0.21 mmol, 41%). 1H NMR (300 MHz, CDCl3) δ 6.76 (s, 1H, 2), 6.60 (t, J = 5.9 Hz, 1H, 6), 3.86 (s, 3H, 8), 3.81 (s, 3H, 7), 3.27 (qt, J = 7.8, 4.0 Hz, 2H, 5), 2.64 (dtd, J
= 30.3, 15.3, 7.0 Hz, 2H, 4), 2.25 – 2.10 (m, 2H, β), 1.62 (s, 15H, m), 1.25 (s, 16H, δ), 1.01 – 0.74 (m, 3H, ω).13C NMR (75 MHz, CDCl3) δ 174.02 (α), 156.56 (1), 132.81 (3), 120.59 (2), 88.71 (q), 38.44 (8), 37.17 (5), 36.59 (β), 35.79 (7), [31.99, 29.75, 29.71, 29.64, 29.55, 29.45, 25.87, 22.77] (γ+δ), 24.89 (4), 14.23 (ω), 9.28 (m). Elemental analysis calcd (%) for C29H50Cl2IrN3O: C 48.39, H 7.00, N 5.84;
found: C 47.95, H 7.02, N 5.81. HRMS [C29H50Cl2IrN3O–Cl]+: 684.32599 found, 684.326592 calculated.
2,2-bipyridine-6,6’-dicarboxylatobis-N-dodecylisonicotinamide
ruthenium(II), compound [2]. The procedure was adapted from Duan et al.1 During the synthesis, contact with O2 was avoided as much as possible. All solvents were deaerated by bubbling with Ar before use. A mixture of 2,2’- bipyridine-6,6’-dicarboxylic acid (61 mg, 0.25 mmol), [Ru(dmso)4Cl2] (0.12 g, 0.25 mmol) and Et3N (0.20 mL, 1.4 mmol) in MeOH (5 mL) was refluxed for 2 hours under argon. The color of this solution changed from orange to purple.
Then, N-dodecylisonicotinamide (0.71 g, 2.4 mmol) in MeOH (8 mL) was added.
The reflux was continued for an additional 20 h, and the mixture was evaporated to dryness under reduced pressure. The red-brown residue was dissolved in DCM (30 mL). H2O (30 mL) was added and the mixture was extracted with DCM (2 x 25 mL). The combined organic layers were dried over MgSO4, filtered, and the filtrate was concentrated under reduced pressure. The crude product was purified by column chromatography under argon (SiO2, deaerated DCM/MeOH 94/6 to 88/12). After evaporation of the solvents, the pure product was obtained as brown-red powder (0.13 g, 0.14 mol, 57%). 1H NMR (400 MHz, CDCl3) δ 8.39 (d, J = 8.1 Hz, 2H, B5), 8.28 (t, J = 5.8 Hz, 2H, A6), 8.08 (d, J = 7.7 Hz, 2H, B3), 7.92–
7.81 (m, 6H, B4 + A3), 7.48 (d, 4H, A2), 3.36 (q, J = 1.7 Hz, 4H, α), 1.55 (t, J = 7.3 Hz, 4H, β), 1.25 (m, 36H, γ), 0.87 ppm (t, J = 6.7 Hz, 6H, ω). 13C NMR (100 MHz, CDCl3/MeOD 9/1) 153.19 (A2), 132.61(B4), 126.62 (B3), 124.46 (B5), 122.48 (A3), 40.51 (α), [32.26, 29.71, , 27.18, 22.78] (γ), 29.41 (β) 14.05 ppm (ω). Not all quaternary carbons could be found. ESI MS m/z (calc): 924.5 (924.1, [M–e]+), 947.4 (947.4, [M+Na]+). Elemental analysis calcd (%) for C48H66N6O6Ru: C 62.38,H
