Controlling the Charge Transfer Mechanism and Efficiency by Means of Different C70 Regioisomeric Adducts

Here, differences stemming from two newly synthesized electron donor–acceptor conjugates, α‐ or β‐regioisomeric adducts featuring orthogonal arrangements of an electron‐donating free‐base porphyrin (H2P) and electron‐accepting C70, are reported. Key to a full‐fledged investigation in terms of both experiments and theory is the use of a rigid linker to separate the electron acceptor and donor. Both conjugates are experimentally investigated by means of femtosecond/nanosecond transient absorption measurements, which is further supported by radiation chemical studies based on pulse radiolysis. Significant regioisomeric differences are seen in the charge recombination kinetics, in general, which are as large as 1.4‐fold, and in the relative distributions of the charge recombination pathways, in particular, which reach nearly 1.5‐fold. Clearly, α‐regioisomers of C70 adducts are much better suited for the construction of future energy conversion systems than β‐regioisomers; a fact that relates to the superior electron delocalization within the carbon caps of C70 made out of “corannulenoid” fragments relative to the equator region of fullerenes featuring extended “phenanthrenoid” rings.

Here, differences stemming from two newly synthesized electron donor-acceptor conjugates, αor β-regioisomeric adducts featuring orthogonal arrangements of an electron-donating free-base porphyrin (H 2 P) and electron-accepting C 70 , are reported. Key to a full-fledged investigation in terms of both experiments and theory is the use of a rigid linker to separate the electron acceptor and donor. Both conjugates are experimentally investigated by means of femtosecond/ nanosecond transient absorption measurements, which is further supported by radiation chemical studies based on pulse radiolysis. Significant regioisomeric differences are seen in the charge recombination kinetics, in general, which are as large as 1.4-fold, and in the relative distributions of the charge recombination pathways, in particular, which reach nearly 1.5-fold. Clearly, α-regioisomers of C 70 adducts are much better suited for the construction of future energy conversion systems than β-regioisomers; a fact that relates to the superior electron delocalization within the carbon caps of C 70 made out of "corannulenoid" fragments relative to the equator region of fullerenes featuring extended "phenanthrenoid" rings.
regioisomers, as well as higher costs. [32,35] The two major regioisomers of C 70 monoadducts are its αand β-regioisomers, which disclose, however, no notable differences in terms of their reduction. As a matter of fact, larger fullerenes have attracted much less attention, even though they are considered promising structures for the conversion of photochemical energy into storable energy forms. [36][37][38] Therefore, studies elucidating the mechanistic details of photoinduced electron transfer reactions in covalent electron donor-acceptor systems featuring fullerenes, which are larger than C 60 , are much needed.
By virtue of their absorption cross-section in the visible part of the optical spectrum, their synthetic availability, and their photochemical stability, porphyrins have emerged as the perfect complement to the electron accepting fullerenes. [39][40][41][42][43] Porphyrins are, however, not only suitable light harvesters, but also good electron donors. [2,14] Past work in the field has, nevertheless, been almost exclusively conducted with C 60 -based electron donor-acceptor systems. Considering the limited publications with respect to the buckminsterfullerene, covalently linked electron donor-acceptor architectures based on C 70 as the electron acceptor, [30,44,45] on one hand, and porphyrins as the light harvester/electron donor, on the other hand, we report here on two different electron donor-acceptor conjugates featuring a meso-tetraphenyl porphyrin covalently linked to C 70 . A full-fledged photochemical, radiation chemical, and quantum chemical investigation provides sound insights into to the nature of the photoinduced charge separation and recombination.

Synthesis and Characterization
An addition of the N-arylaziridine-2,3-dicarboxylates to fullerene C 70 , described previously, [35] was used for the construction of C 70 conjugates with meso-tetraphenylporphyrin (TPP). However, an attempted addition of porphyrinic aziridine 1 with carboxyethyl moieties (Figure 1a), in contrast to previously report for C 60 , [46] affords totally insoluble products, which led us to introduce solubilizing groups. For this reason, porphyrinic aziridine 2, bearing long and solubilizing alkyl chains, was obtained in the same manner with analogous aziridine 1. Indeed, the introduction of the solubilized carboxy-n-octyl moieties to the porphyrinic aziridine 2 allows us to obtain the desired adduct, soluble enough for purification and characterization.
According to the literature, the addition of such aziridines to fullerenes processes as a sequence of concerted stages of aziridine ring opening and 1,3-dipolar cycloaddition with extremely high stereospecificity, while the addition of cis-aziridine affords only trans-adducts. [35,47] In principle, monoaddition to C 70 may afford four different position regioisomers, so-called siteisomers. Practically, the reaction mixtures consist of two isomers, namely, a major "α" [1, 9] adduct and a minor "β" [7,8] adduct. Both of these isomers are distinguishable by their characteristic patterns in the NMR and UV-vis spectra.
The addition of 2 to C 70 under well-established conditions (odichlorobenzene, 100 C) affords a reaction mixture. It contains, according to the 1 H NMR spectrum, DL1 and DL2 in 1:2.72 ratio.
They are primarily assigned as βand α-isomers, respectively, on the basis of the chemical shifts known for adducts of the same type. [35] Accordingly, the site selectivity is in sound agreement with previous reports and is due to π-π interactions between C 70 and the porphyrin. Seemingly, the isomeric ratio is lower than found recently. High-performance liquid chromatography (HPLC) separation affords the two isomeric DL1 and DL2 in 28% and 24% yields, respectively (Figure 1b). 1 H NMR spectra of the products reveal the fingerprints of the individual isomers. [35] These are singlets of pyrrolidine protons at 5.66 and 5.95 ppm for DL1 and 5.82 and 6.15 ppm for DL2, from which we conclude that they are the cis-β and cis-α isomers, respectively. Their absorption spectra, in which only the DL1 isomer shows an absorption maximum typical for cis-β isomers, further support our assignment.

Density Functional Theory (DFT) Results
Electronically excited states of the C 70 -porphyrin electron donoracceptor systems were studied by means of quantum chemical calculations on the model compounds DL1 0 and DL2 0 , related to DL1 and DL2. First of all, the geometries of the ground states were optimized using DFT at the B3LYP/6-31G(d) level of theory. For both isomers, two energy minima corresponding to two different conformers were found. These are, on one hand, the "extended" (e) conformer, in which the porphyrin plane is oriented perpendicular to the tangent plane of the fullerene, and, on the other hand, the "bent" (b) conformer with a significant deviation from perpendicularity ( Figure S38, Supporting Information). For DL1 0 and DL2 0 , the extended conformer is %5 kcal mol À1 more stable than the bent conformer, and the conformers of DL2 0 are %3 kcal mol À1 more stable than those of DL1 0 . In all cases, the HOMOs are localized on the porphyrin, while the LUMOs are localized on C 70 . The HOMO-LUMO gaps are appreciably larger (0.14 eV) for the bent conformers than for the extended conformers, while only subtle differences (0.06 eV) evolve between the αand β-isomers (Table S6, Supporting  Information).
Next, the electronically excited states of the most stable conformers for DL2 0 and DL2 0 e were computed in four different solvents: toluene, anisole, tetrahydrofuran (THF), and benzonitrile. The first electronically excited state correlates with an excitation of the HOMO on the porphyrin to the LUMO on C 70 ( Figure S39, Supporting Information) and, represents, in turn, a charge-separated state. In addition, we derived the relative rate constants for charge recombination, k CR , in the four different solvents with the aid of the Marcus theory. A detailed description of the calculations is provided in the Supporting Information. Notably, the quantitative match with the experimental values was poor. A reasonable rationale is based on the time dependency of the solvent relaxation, especially at very short timescales, which is comparable to the dielectric relaxation time. This is unaccounted for by the polarizable continuum model (PCM) model. Still, the correct trend of k CR was found k benzonitrile % k THF < k anisole << k toluene (1)

Radiation Chemistry
To establish the transient absorption spectra of the one-electron reduced form of the C 70 -derivatives, pulse radiolytic experiments were performed under reducing conditions with the respective C 70 -R references: C 70 -H, C 70 -Br, and C 70 -CH 3 ( Figure 2). The reducing conditions were obtained by irradiating solutions of the C 70 -R references in N 2 -saturated mixtures of toluene, acetone, and isopropanol (8:1:1 v/v/v) with short electron pulses.
These conditions led to the formation of the (CH 3 ) 2 C • (OH) radicals, [48,49] which are known as strong reducing agents (À1.2 V vs saturated calomel electrode/À1.1 V vs normal hydrogen electrode), [49][50][51] and, which are capable of reducing fullerenes. [52] As such, (CH 3 ) 2 C • (OH) radicals are expected to reduce C 70 -H, C 70 -Br, and C 70 -CH 3 to their corresponding one-electron reduced forms. The main reaction pathway includes Upon pulse radiolysis of the C 70 -R references, initially after the electron pulse, the transient absorptions of (CH 3 ) 2 C • (OH) with its featureless band in the UV region of the optical spectrum are discernible. It decays and gives rise to a new set of transient absorptions throughout the UV, visible, and NIR parts of  www.advancedsciencenews.com www.small-structures.com the optical spectrum- Figure 3 for C 70 -H, Figure S8, Supporting Information, for C 70 -Br, and Figure S9, Supporting Information, for C 70 -CH 3 . For example, the transient absorption spectrum of (C 70 -H) •À gives rise to maxima at around 370, 420, and 580 nm, followed by a broad transient absorption throughout the visible and NIR range of the optical spectrum with minor maxima around 740, 810, and 1300 nm. All of the aforementioned is flanked by distinct minima around 400 and 460 nm, where C 70 -H exhibits ground state absorptions. A closer look at the time evolution of the transient absorption spectra discloses that the initially formed transient absorption of (CH 3 ) 2 • COH decays and its decay goes hand in hand with the rise of the spectroscopic signature of the one-electron reduced form of the C 70  • COH in toluene, acetone, and isopropanol (8:1:1 v/v/v) mixtures were determined via analyses of the pseudo-first-order rate constants plotted versus the concentrations of the C 70 -R references, as shown in Figure 3c. To this end, rate constants of 6.4 Â 10 9 M À1 s À1 (C 70 -H), 1.9 Â 10 9 M À1 s À1 (C 70 -Br), and 1.1 Â 10 9 M À1 s À1 (C 70 -CH 3 ) were derived from the linear relationships. These values are quite similar to those previously obtained for C 60 . [50] A closer inspection of the time evolution reveals that the one-electron reduced form of the C 70 -R references further reacts on timescales of hundreds of microseconds. Feasible rationales are dimerization reactions, with an unreacted C 70 -R to afford (C 70 -R) 2 •À and/or with another (C 70 -R) •À to form (C 70 -R) 2 2À . Such reactions are not unexpected, because they are reported in the literature for C 60 . [50]

Photophysical Study
First, photophysical studies, namely, steady-state absorption and fluorescence spectroscopy measurements, were performed with the C 70 -R references in toluene. As a result, extinction coefficients of 2.3 Â 10 4 M À1 cm À1 and fluorescence quantum yields of 0.0002 were obtained for C 70 -H (see also Table S3, Supporting Information), which are in sound agreement with previously published values. [53] In addition to the absorption spectra of the C 70 -R references (see Figure S10, Supporting Information), photophysical studies were performed with the corresponding H 2 P (TPP) reference and the electron donoracceptor conjugates DL1 and DL2. In the case of H 2 P, absorption spectra in various solvents with different polarities and viscosities, that is, toluene, anisole, THF, and benzonitrile, were investigated.
Absorption spectra exhibit in all cases Soret-band absorptions and a set of Q-band absorptions, which are typical features of meso-substituted free-base porphyrin derivatives. [54] In all solvents, the Soret-band absorptions are discernible in the range from 420 to 430 nm, followed by four Q-band absorptions in the range between 513 and 650 nm. In the case of DL1 and DL2, an additional band at around 360 nm is observed and it is attributed to the presence of C 70 (see Figure 4).  Interestingly, the extinction coefficients of DL1 and DL2 are significantly lower compared to the H 2 P reference, which is in line with previously reported porphyrin-fullerene electron donoracceptor systems and prompts to electronic communication between the photo-and redox-active constituents. [55] Insights into the electronic interactions between the electron donor and the acceptor came from fluorescence studies. The H 2 P reference exhibits strong emission, in the range between 620 and 760 nm in anisole upon 430 nm photoexcitation into the Soret-band absorption, with maxima at 651 and 718 nm. The fluorescence quantum yield and the lifetime of H 2 P in anisole were found to be 0.116 and 10.74 ns, respectively. Despite the fact that the fluorescence spectra of both electron donor-acceptor systems were identical to those of the H 2 P reference, the fluorescence quantum yields as well as the corresponding lifetimes were significantly quenched. In particular, fluorescence quantum yields of DL1 and DL2 in anisole were found to be 0.0015 and 0.0016, respectively. Relative to the H 2 P reference, these quantum yields correlate to a quenching of %99%. In toluene, THF, and benzonitrile the fluorescence quantum yields were calculated to be 0.0008, 0.0011, and 0.0013, respectively, for DL1 and 0.0015, 0.0007, and 0.0009, respectively, for DL2. Such a significant quenching suggests an additional deactivation channel, electron or energy transfer starting from the first singlet excited state of the electron-donating H 2 P.
Next, to reveal the nature of the fluorescence quenching, femtosecond and nanosecond transient absorption spectroscopy measurements were performed photoexciting the electron donor-acceptor compounds DL1 and DL2 as well as H 2 P at 430 nm and the C 70 -R reference compounds at 387 nm. Initially, toluene solutions of the C 70 -R references, that is, C 70 -H, C 70 -Br, and C 70 -CH 3 were prepared. H 2 P and the electron donor-acceptor compounds DL1 and DL2 were investigated in toluene, anisole, THF, and benzonitrile.
Exemplarily, for C 70 -H upon photoexcitation at 387 nm, the transient absorption changes were analyzed using the Glotaran [56] program with a "two species" kinetic model. After direct photoexcitation, transient maxima at 500 to 800 nm as well as in the NIR region evolve as the first species, that is, the first singlet excited state of C 70 -H. Its lifetime is with 727 AE 3 ps, in perfect agreement with previously reported lifetimes for C 70 derivatives. [53,57] The singlet excited state decay goes hand in hand with the growth of the second species, for which transient absorption bands at around 590, 619, and 690 nm as well as at in the NIR region at around 1050 nm were observed. Accordingly, the second species is the triplet excited state ( Figure S18, Supporting Information), with a lifetime of 38 AE 1 μs. [58,59] Photoexciting the H 2 P reference at 430 nm, that is, exciting into the Soret-band absorption, several transient absorption maxima in the visible region up to 700 nm and a strong band in the NIR around 1080 nm are seen. These observed spectral features are related to the singlet excited state of the free-base porphyrin, and are not fully decaying within the time range of our femtosecond transient absorption setup (6 ns). Corresponding nanosecond transient absorption measurements were conducted to follow the complete decay of the first excited singlet state of H 2 P and show the known intersystem crossing into the corresponding triplet manifold. [54] As a matter of fact, lifetimes of the first excited singlet and triplet state of 11.2 AE 0.1 ns and 18.9 AE 0.3 μs, respectively, were derived in anisole (see also Figure S14-17, Supporting Information, and Table 1 for the spectra and values in other solvents).
Finally, transient absorption spectroscopy measurements based on femtosecond and nanosecond photolysis at 430 nm were performed with the electron donor-acceptor compounds DL1 and DL2 in toluene, anisole, THF, and benzonitrile. Taking a look at the transient absorption spectra, several maxima and minima are observed in the visible and NIR regions. Due to this more complex picture, a three-species model based on target analysis (see Figure S37, Supporting Information) using Glotaran was conducted giving the following results. The transients directly formed after photoexcitation (see Figure 5 and Figure S24-36, Supporting Information) match well the transient absorption spectra of the H 2 P reference first singlet excited state and are thus dedicated to the first singlet excited state located at the free-base porphyrin in DL1 and DL2. The transient absorption of the first singlet excited state decays rapidly, for example, with lifetimes of 49 AE 1 ps for DL1 and 65 AE 2 ps for  DL2 in anisole (see Figure 5 and Table 1), and gives rise to a new set of spectroscopic features. More precisely, a diagnostic and broad transient absorption band in the NIR region between 1100 and 1400 nm with a maximum at 1300 nm was seen for DL1 and DL2, which is in sound agreement with the radiation chemical reduction of all C 70 -R references (see Figure 3a and Figure S8a and S9a, Supporting Information) forming C 70 •À .
Furthermore, the transient absorption maxima in the visible region at around 450 and 535 to 680 nm are in good agreement with the corresponding one-electron oxidized form of the H 2 P reference (H 2 P •þ ). [60] Accordingly, we corroborated a photoinduced electron transfer from H 2 P to C 70 to afford a chargeseparated state as the second species (red line). Afterward, the charge-separated state decays, with a lifetime of 559 AE 11 ps, partially back to the ground state as well as partially to the corresponding triplet excited state of H 2 P; a finding that is in sound agreement with previous findings on C 60 . [61] Note that no evidence for the triplet excited state of C 70 was gathered  Corresponding singlet and triplet excited state lifetimes were obtained from global analyses by Glotaran. b) Corresponding singlet excited state, charge recombination, and triplet excited state lifetimes were obtained from target analyses by Glotaran using the model shown in Figure S37, Supporting Information. regardless of a parallel or a sequential model. The latter matches the previously obtained triplet excited state of the H 2 P reference (for further details, see also Table S4, Supporting Information). When analyzing the scaling factors from the analysis of the transient absorption matrices, the relative distribution of the corresponding charge recombination pathways changes with the solvent polarity. As a general trend, with increasing solvent polarity the likelihood of the direct decay of the charge-separated state back into the ground state increases (for details, see also Figure S37 and Table S4 and S5, Supporting Information).
Considering that 430 nm photoexcitation of DL1 and DL2 also excites C 70 , its triplet excited state features are discernable in the NIR region between 870 and 1050 nm. Please note that this pathway is however only contributing less than 5% to the corresponding spectra.
In principle, DL2 is subject to the same photoinduced electron transfer mechanism as DL1. However, DL2 reveals significant longer charge separation lifetimes, especially in viscous anisole and benzonitrile, where the values are 65 AE 2 and 14 AE 1 ps for DL2 relative to 49 AE 1 and 10 AE 1 ps for DL1. In addition, the scaling factors and, in turn, the relative distribution of the corresponding pathways change. F 3 , that is, the charge recombination to the ground state, and f 4 , that is, the charge recombination to the triplet state of the porphyrin, differ strongly in anisole and THF. As a matter of fact, f 4 is doubled (see also Figure S37 and Table S4 and S5, Supporting Information) and indicates that the different C 70 adduct isomers influence the corresponding pathways significantly. Such differences are related to regioisomeric effects and are documented in similar C 60 -based systems, for example, impacting the performance in bulk heterojunction solar cells consisting of different C 60 adducts. [62,63] However, such performance effects in C 70 -based devices have been only discussed in terms of the different C 70 isomeric adducts affecting the packing on films and altering the performance by this effect, but not by the impact of the different electron transfer rates. [64] Potentially, the implementation of electron-donating porphyrins or phthalocyanines into C 70 -based solar cells should show differences in the power conversion efficiency performance not only due to structural or packing effects, but also due to the distribution of the corresponding reaction pathways from the chargeseparated state. Using different fullerene regioisomeric adducts is an interesting direction toward the control of the distribution of states in advanced electron donor-acceptor systems.

Conclusion
Here, we report on the synthesis of two new electron donoracceptor systems containing different isomeric adducts of C 70 , that is, the αand β-regioisomers. The rigid nature of the linker enforces a scenario, in which the porphyrin is locked into "edge to face" configuration relative to the fullerene and, in turn, affords an ideal platform to probe the regioisomeric dependence of electron transfer. Of great relevance is the fact that we documented experimentally the existence of an electron transfer from the electron-donating porphyrins in their excited states to both regioisomeric adducts of the electron-accepting C 70 ( Figure 6). From our full-fledged investigations we conclude sizeable differences not only in the charge recombination kinetics for the αand β-regioisomers, but also in the relative distributions of their charge recombination pathways. Moreover, we demonstrate full control over the charge transfer mechanism by, for example, tuning solvent parameters such as polarity and viscosity. The α-regioisomeric adducts of C 70 are superior, a fact that relates to the unique electron delocalization within the carbon caps of fullerenes. As such, placing the electron acceptor in close proximity to the C 70 caps with their "corannulenoid" fragments rather than the equatorial regions made out of "phenanthrenoid" rings is beneficial in terms of suppressing the charge recombination. These findings constitute an intriguing possibility for future bioinspired and nature-mimicking electron donor-acceptor systems based on fullerenes for solar energy conversion.

Experimental Section
General Methods: NMR spectra were recorded on a Bruker Avance 400 MHz spectrometer. Fourier-transform infrared spectroscopy (FTIR) spectra were obtained on the Shimadzu IRAffinity-1 in KBr pellets. High resolution mass spectrometry (HRMS) spectra were recorded on a Bruker maXis spectrometer with ESIþ ionization. HPLC was performed on a Shimadzu Prominence LC20 chromatograph with a diode array detector on a Nacalai Cosmosil PBr column in toluene:dichloromethane (DCM) as an eluant.
The residue was dissolved in dry ether (3 mL), five drops of BF 3 *Et 2 O was added. Then a solution of octyl diazoacetate (18 mg, 91 μmol) in ether (2 mL) was added dropwise. The reaction mixture was stirred for 2 h at RT, 15 drops of Et 3 N was added, and then the reaction mixture was poured in water and extracted with DCM. The organic layer was separated and dried over Na 2 SO 4 . The desiccant was filtered off, the solvent was removed in vacuo, and the product was purified by preparative TLC (SiO 2 , PE-DCM 2:1 þ 0.1% Et 3 N). As a result, pure porphyrin 2 (27 mg, 28 μmol, 35%) was obtained.   (50 mg, 60 μmol) were heated in 1,2-dichlorobenzene (4 mL) at 100 C for 44 h. The solvent was removed in vacuo, and the residue was fractioned by flash chromatography (SiO 2 , toluene-EtOAc) and purified by HPLC on a PBr column with DCM-toluene as an eluant. As a result, dyads DL1 (14 mg, 7.6 μmol, 28%) and DL2 (12 mg, 6.5 μmol, 24%) were obtained.  13 C NMR could not be recorded due to low solubility of the product in organic solvents.
Quantum Chemical Calculations: All quantum chemical calculations were performed using the Gaussian 16. [69][70][71][72] Geometry optimization was performed by DFT at B3LYP/6-31G(d) level of theory. The electronically excited states were studied using time-dependent density functional theory (TD DFT) at B3LYP 6-31G(d) level in four different solvents, toluene, anisole, THF, and benzonitrile, with the PCM of solvents. The electronic configuration of each state was analyzed using the CI expansion coefficients.
UV-Vis Absorption Spectroscopy: Steady-state absorption spectra were recorded using a Perkin Elmer Lambda 2 UV-vis two-beam spectrophotometer with a slit width of 3 nm and a scan rate of 480 nm min À1 . A quartz glass cuvette of 10 Â 10 mm was used.
Emission Spectroscopy: Steady-state emission was performed using a Horiba Jobin Yvon FluoroMax-3 spectrometer using a slit width of 3 nm for excitation and emission and an integration time of 0.1 s. A quartz glass cuvette of 10 Â 10 mm was used. All spectra were corrected for the instrument response. The C 70 -R samples were measured at 375 nm, while all other samples were measured upon photoexcitation at 430 nm. The experiments were performed at RT. Fluorescence quantum yields were determined by the comparative method using meso-tetraphenyl porphyrin (H 2 P) and C 60 in toluene as reference, respectively.
Singlet Oxygen Calculation: Triplet quantum yields of the donoracceptor systems DL1 and DL2 were performed in toluene, anisole, and THF and calculated by the comparative method using pure C 60 in toluene as the reference system. The emission spectra of the obtained singlet oxygen at around 1275 nm were recorded using a Fluorolog-3 (Horiba Jobin Yvon) spectrometer equipped with a 450 W xenon lamp, NIR accessories, a high-pass filter (λ ¼ 780 nm), an integration time of 20 s, single-grating monochromators and using a slit width of 3 nm and a quartz glass cuvette of 10 Â 10 mm. As photoexcitation wavelength 415 nm was chosen.
Femtosecond Transient Absorption Spectroscopy: Femtosecond transient absorption studies were performed with laser pulses (1 kHz, 150 fs pulse width) from an amplified Ti/sapphire laser system (Model CPA 2110, Clark-MXR Inc.; output 775 nm). For excitation wavelengths of 387 and 430 nm, a nonlinear optical parametric amplifier (NOPA) was used to generate ultrashort tunable visible light pulses out of the pump pulses. The transient absorption pump probe spectrometer (TAPPS) can be described as a two-beam setup in which the pump pulse is used as an excitation source for transient species and the delay of the probe pulse is exactly controlled by an optical delay rail. As the probe (white-light continuum), a small fraction of pulses stemming from the CPA laser system was focused by a 50 mm lens into a 2 mm thick sapphire disc. The transient spectra were recorded using fresh argon-saturated solutions in each laser excitation. All experiments were performed at 298 K in a 2 mm quartz cuvette.
Nanosecond Transient Absorption Spectroscopy: Nanosecond transient absorption studies were performed with laser pulses (1 kHz, 150 fs pulse width) from an amplified Ti/sapphire laser system (Model CPA 2110, Clark-MXR Inc.; output 775 nm). For excitation wavelengths of 387 and 430 nm, a NOPA was used. The TAPPS uses an excitation source for transient species and a separate white-light laser. The width of the white probe pulse is 1 ns, offering the possibility to investigate excited state kinetics up to several hundred microseconds. The transient spectra were recorded using fresh argon-saturated solutions in each laser excitation. All experiments were performed at 298 K in a 2 mm quartz cuvette.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.