Does a nitrogen matter? Synthesis and photoinduced electron transfer of perylenediimide donors covalently linked to C 59 N and C 60 acceptors †

The ﬁ rst perylenediimide (PDI) covalently linked to an azafullerene (C 59 N) is described. PDI-C 59 N and PDI-C 60 dyads where PDI acts as an electron-donor moiety have been synthesized by connection of the balls to the PDI 1-bay position. Photoexcitation of the PDI unit in both systems results in formation of the charge-separated state by photoinduced electron transfer from the singlet excited state of the PDI moiety to the C 59 N or to the C 60 moiety. The charge-separated state has a lifetime of 400 ps in the case of PDI-C 59 N and 120 ps for the PDI-C 60 dyad in benzonitrile at 298 K. This result has signi ﬁ cant implications for the design of organic solar cells based on covalent donor – acceptor systems using C 59 N as an electron acceptor, indicating that longer-lived charge-separated states can be attained using C 59 N systems instead of C 60 systems.


Introduction
The seeking for efficient molecular systems, based on the covalent combination of suitable and versatile building blocks, capable to mimic the photoinduced electron transfer that naturally occurs in photosynthesis, has been a constant in material science. 1 In a simplistic definition, a molecular system which combines an electron-acceptor moiety with an electron-donor counterpart, able to be excited by the action of light, can be considered as a potential photosynthetic system, and therefore useful for photovoltaic applications.[60]Fullerene, due to its extraordinary electron-acceptor character and its fascinating chemical versatility, has been one of the most utilized electron-acceptor building blocks 2 for that purpose and, thus, several efficient artificial photosynthetic [60]fullerenebased systems have been described, using different chromo-phores as photoexcitable electron-donor counterparts. 3erformances of these donor-acceptor materials can be improved by selecting an appropriate electron-donor counterpart, but also by replacing one or more carbon atoms of the fullerene cage by heteroatoms, affording the so-called heterofullerenes. 4The substitution of a tetravalent carbon atom of the three dimensional network by a trivalent nitrogen atom is one of the most interesting alterations.This simple modification leads to azafullerene C 59 N, a fullerene-based building block with an improved electron-acceptor character, when it is compared to that of pristine C 60 fullerene. 5A number of different azafullerene-based donor-acceptor covalent compounds have been previously described, showing interesting photophysical properties that make them ideal candidates to be incorporated in organic photovoltaic devices.Hirsch et al. described the synthesis of an azafullerene-ferrocene dyad that shows strong electronic coupling between the ferrocene and azafullerene moieties in the ground state. 6Additional C 59 Nbased dyads carrying photoactive units such as porphyrin 7 and pyrene 8 were also prepared.More recently, azafullerene-electron-donor dyads, using phthalocyanine 9 and corrole 10 units, have been reported to undergo photoinduced electron transfer upon selective irradiation of the electron-donor moiety.
On the other hand, perylenediimides (PDI) comprise an outstanding family of perylene derivatives that, due to an extreme synthetic versatility and an easily tunable electronic character, have become one of the most promising classes of molecular materials in organic photovoltaic devices. 11In this context, we can find only a few examples where PDI derivatives † Electronic supplementary information (ESI) available: Characterization spectra of the new compounds.See DOI: 10.1039/c5nr00308c a Área de Química Orgánica Instituto de Bioingeniería, Universidad Miguel Hernández, Elche, Spain.E-mail: asastre@umh.es;Fax: +34 966658408; Tel: +34 966658351 have been successfully employed as an electron-donor moiety in fullerene-based systems, eventually undergoing photoinduced electron and/or energy transfer. 12ere we present the synthesis, characterization and photophysical properties of a new photoexcitable molecular system, consisting of an electron-donor PDI moiety, carrying a pyrrolidinyl substituent in the 1-bay position, which also acts as a linker to the C 59 N counterpart (Scheme 1).The obtained donor-acceptor PDI-C 59 N dyad 1, upon selective excitation of the PDI subunit in benzonitrile, shows efficient intramolecular photoinduced electron transfer, generating a 400 ps-lived charge-separated state (CSS).Together with two covalently linked porphyrin-C 59 N dyads, 7 this is the longest CSS lifetime value obtained so far for a C 59 N-based donor-acceptor dyad in a polar solvent.Moreover, in order to study the effect of the nitrogen atom on the photophysics, the analogous PDI-C 60 dyad 2 (Scheme 1) was synthesized, as reference material, obtaining a 3-times shorter CSS lifetime than for the PDI-C 59 N dyad 1, thus, highlighting the role of nitrogen in the photoinduced electron-transfer processes.
All new compounds were fully characterised using standard analytical techniques such as 1 H and 13 C NMR, HR-MS, FT-IR, electronic absorption and photoluminescence.Fig. 1 compares Scheme 1 Synthetic routes for preparing PDI-C 59 N 1 and PDI-C 60 2 dyads as well as PDI-based reference material 6. the aromatic region of the 1 H NMR spectra for 1, 2 and 6 in CDCl 3 as solvent.Taking as reference the C1 of the PDI core (the substituted one), and comparing the signals of analogous hydrogens in compounds 1, 2 and 6, it is worth noting that the chemical displacement of the nearby hydrogens (H 2 , H 11 and H 12 ) is significantly shifted in dyad's spectra, while the contrary trend is observed in the displacement of H 5 , H 6 , H 7 and H 8 .For example, the H 2 signal goes from 8.69 to 9.19 and 9.18 ppm (in 2 and 1 respectively), while H 6 goes from 8.45 to 8.40 ppm.Furthermore, the fullerene influence is slightly larger when C 59 N is present instead of C 60 , and it can be recognised in the H 12 displacement, which changes from 8.11 ppm in 6, to 8.29 and 8.40 ppm in dyads 2 and 1, respectively.Fig. 2 shows the aliphatic regions of the spectra, where some differences can also be found due to the presence of the covalently attached fullerene.For example, the signals corresponding to the hydrogens labeled as H B , H C and H E , which appear as broad signals in the spectrum of PDI-based 6, split into two in the corresponding spectra of PDI-fullerene dyads 1 and 2. All these differences in the 1 H NMR spectra of dyads and reference material can be explained by considering the electronwithdrawing character of the fullerene sphere, which is more intense in the case of C 59 N as compared with C 60 , and also assuming a pseudo-fixed molecular conformation in PDI-fullerene dyads caused by an electrostatic interaction between the PDI and the fullerene moiety including a concave-convex interaction 3l,15 that causes a different influence of the sphere on the chemical shift of the pyrrolidine-linker geminal protons.

Photophysical and electrochemical properties
Fig. 3 shows the overlapped UV-vis spectra of dyads 1 (black line) and 2 (blue line) and C 60 5a, C 59 N 4b and PDI 6 as reference compounds using benzonitrile as solvent.The absorption spectra of PDI-C 59 N 1 and PDI-C 60 2 dyads are characterised by two sets of absorption maxima around 430 and 638 nm.The band centered at 430 nm can be attributed to transitions 1-0 and 0-0 in the PDI species, 16 while the one centred at 638 nm consists of a broad charge transfer (CT) band that usually appears in bay amino-substituted PDI compounds.The latter band appears as a consequence of an electron transfer between the electron-donor pyrrolidine and the electron-acceptor PDI units, 17 and therefore is very sensitive to changes in the electronic character of the entire molecule.When those spectra are compared to the absorption profile of reference compound PDI 6, some differences can be found.One of these differences, directly related to the electronic interaction between the electroactive moieties, is a bathochromic displacement of the band centered at 632 nm, which now appears at 638 nm in both PDI-C 60 and PDI-C 59 N.This fact reveals the electronic influence of the fullerene sphere on the perylene subunit in the ground state, as it has been already pointed out in the comparison of 1 H NMR spectra (cf.Fig. 1 and 2).
On the other hand, Fig. 4 shows the fluorescence emission spectra of PDI-C 59 N 1 and PDI-C 60 2 dyads, when selectively excited at 429 nm.As can be seen, the fluorescence spectrum of 6 clearly shows three different emission bands, centered at 480, 512 and 691 nm.When the emission spectra of PDI-C 59 N and PDI-C 60 dyads 1 and 2 are compared to that of the reference compound, the first and second bands are partially  quenched while the third one is extinguished.The latter is indicative of intra-dyad electronic interactions between the two components of the dyad (i.e., PDI and fullerene) at the excited state.The strong fluorescence quenching of the 691 nm band in both 1 and 2 as compared to 6, supports electron and/or energy transfer as the decay mechanism of the PDI-centred singlet excited state.This effect is slightly more pronounced in the case of PDI-C 59 N dyad 1 (see inset Fig. 4).
The electrochemical data of PDI-C 59 N 1 and PDI-C 60 2 dyads, as well as 6, 4b and 5b as reference materials for PDI-C 59 N, and C 60 , respectively, in benzonitrile using Fc/Fc + as an internal standard are listed in Table 1 (see differential pulse voltammograms in ESI, Fig. S13 †).The measurements for PDI 6 reveal a one-electron oxidation and two one-electron reduction processes centred at 0.55, −1.15 and −1.35 V respectively, while the measurements for fullerene reference compounds show two one-electron reversible reduction processes at −1.01 and −1.42 V.The voltammograms of dyads 1 and 2 are very similar, in spite of the nitrogen atom in dyad 1, and can be considered as a combination of the already mentioned electrochemical processes assigned to PDI and fullerene electroactive moieties.A slight influence of the fullerene sphere (either C 59 N or C 60 ) over the PDI moiety can be recognised, and it is highlighted by a 20-30 mV anodic displacement of the oxidation wave and a 40 mV cathodic displacement of the second reduction process, which is attributed to the PDI counterpart.As contrary to previous results reported by us, 9,10 the similar first reduction values obtained for PDI-C 60 and PDI-C 59 N might come from the use of different solvents.Based on the redox data shown in Table 1 and neglecting the Coulombic interactions, the electrochemical band gap for 1 is calculated as 1.59 eV (1.61 eV for 2).
Femtosecond laser flash photolysis on a deaerated benzonitrile solution of 1 and 2 indicates the formation of a chargeseparated state, providing solid evidence for the electron-transfer deactivation mechanism in 1 and 2, after selective excitation of the perylene moiety.The transient absorption spectrum taken at 1 ps after laser pulse irradiation of the absorption band of the PDI moiety at 490 nm exhibits an absorption maximum at 900 nm, which is assigned to the singlet excited state of C 59 N of PDI-C 59 N dyad 1 (Fig. 5a).No indication of the PDI singlet excited state was observed in spite of the selective excitation of PDI, indicating that the rate of energy transfer from the singlet excited state of the PDI moiety to the C 59 N moiety is too fast to be followed even with the use of femtosecond laser flash photolysis (fwhm = 130 fs).Then, electron transfer from PDI to the singlet-excited state of C 59 N takes place to form PDI •+ -C 59 N •− as the charge-separated state.The transient absorption band decayed with a rate constant of 5.5 × 10 10 s −1 giving rise to an absorption band at λ max = 1030 nm due to the C 59 N •− radical anion, 9,10 which indicates the formation of the charge-separated state.From the decay of the absorption band at 1030 nm, a 400 ps lifetime (k = 2.5 × 10 9 s −1 ) was obtained for the charged-separated state (Fig. 5b), being the first example of a charge-separated state in a donoracceptor dyad based on PDI and azafullerene.
When benzonitrile was replaced by toluene, the rate constant of charge separation became much larger (3.9 × 10 11 s −1 ), whereas the rate constant of charge recombination became smaller (1.3 × 10 9 s −1 ) as shown in Fig. S14 (ESI †).The faster charge separation in toluene than in benzonitrile may result  from the smaller solvent reorganization energy of toluene than that of the more polar benzonitrile in the Marcus normal region, where the charge separation becomes faster with decreasing the solvent reorganisation energy of electron transfer.1d,18 By the same token, the slower charge recombination in toluene than in benzonitrile results from the smaller solvent reorganisation energy of toluene than that of benzonitrile in the Marcus inverted region, where the charge recombination becomes slower with decreasing the solvent reorganisation energy of electron transfer.1d, 18 On the other hand, a similar transient absorption spectral change was observed for PDI-C 60 dyad 2 as shown in Fig. 6a.The transient absorption recorded at 1 ps after femtosecond laser pulse irradiation is assigned to the singlet excited state of C 60 , indicating that ultrafast energy transfer from the singlet excited state of PDI to the C 60 moiety also occurs in PDI-C 60 .Then, electron transfer from PDI to the singlet excited state of C 60 takes place to form the charge-separated state.The rate constant for electron transfer is determined to be 8.6 × 10 10 s −1 from the rise of the absorbance at 1000 nm (inset of Fig. 6b).This rate is slightly faster than that of PDI-C 59 N (vide supra).The lifetime of the charge-separated state was determined to be 120 ps from the decay of absorbance at 1000 nm as shown in Fig. 6b.It is worth to mention the much shorter lifetime of the CSS in PDI-C 60 , less than one third of that in PDI-C 59 N, pointing to the remarkable influence of the nitrogen atom in C 59 N, which is not explained only by an increased electron-accepting character as the reduction potentials of 1 and 2 indicate.In both cases, the decay of the charge-separated state by back electron transfer yields the ground state rather than the triplet excited state of PDI as indicated by no triplet transient absorption of 1 and 2 in nanosecond laser flash photolysis measurements (Fig. S14 in ESI †).

General information
All chemicals were reagent grade, purchased from commercial sources, and used as received, unless otherwise specified.
Column chromatography: SiO 2 (40-63 mm) TLC plates coated with SiO 2 60F254 were visualized by UV light. 2 and 6 crude products were finally purified by a Combiflash Rf chromatography system (Teledyne Technologies, Inc., Thousand Oaks, CA), while 1 crude product was finally purified by preparative HPLC (Japan Analytical Industry Co Ltd, utilising a Buckyprep 20 × 250 column with toluene as eluent at 10 mL min −1 flow rate).NMR spectra were measured with a Bruker AC 300.UV/vis spectra were recorded with a Helios Gamma spectrophotometer.Fluorescence spectra were recorded with a PerkinElmer LS 55 Luminescence Spectrometer and IR spectra with a Nicolet Impact 400D spectrophotometer.Mass spectra were obtained from a Bruker Microflex matrix-assisted laser desorption/ionization time of flight (MALDI-TOF).

Electrochemistry
Differential pulse voltammetry measurements were performed in a conventional three-electrode cell using a µ-AUTOLAB type III potentiostat/galvanostat at 298 K, over PhCN deaerated sample solutions (∼0.5 mM), containing 0.10 M tetrabutylammonium hexafluorophosphate (TBAPF 6 ) as supporting electrolyte.A glassy carbon (GC) working electrode, Ag/AgNO 3 reference electrode and a platinum wire counter electrode were employed.Ferrocene/ferrocenium was used as an internal standard for all measurements.

Conclusions
We have synthesized for the first time a dyad where a perylenediimide unit is connected to C 59 N. Taking advantage of the electron donor character of the mono pyrrolidinyl PDI derivative synthesised, the PDI-C 59 N dyad undergoes efficient intramolecular photoinduced electron transfer to afford a 400 pslived CS state in PhCN.PDI-C 60 has also been synthesized, as a reference compound, affording a 3 times shorter CSS lifetime than the PDI-C 59 N dyad to address the influence of the nitrogen in the photoinduced electron-transfer process.The present results have significant implications for the design of organic solar cells based on covalent donor-acceptor systems using C 59 N as the acceptor and shows that the longer-lived charge-separated states can be attained using C 59 N systems instead of C 60 systems.
HR MALDI-TOF (negative mode) experiments clearly confirm the structure of PDI-C 59 N and PDI-C 60 dyads 1 and 2, respectively, showing base peaks due to the molecular ions at 1476.463 amu and 1474.302amu, with isotopical distributions that exactly match the simulated isotope patterns for C 106 H 52 N 4 O 6 and C 107 H 51 N 3 O 6 , respectively (see ESI, Fig. S1 †).

Table 1
Fig. 5 (a) Transient absorption spectra of C 59 N-PDI (0.5 mM) in deaerated PhCN after femtosecond laser excitation at 510 nm; (b) time profile of the absorbance at 1030 nm.Inset: decay time profile for the short time range.Gray line is drawn by the double exponential curve fitting.

Fig. 6
Fig. 6 (a) Transient absorption spectra of C 60 -PDI (0.5 mM) in deaerated PhCN after femtosecond laser excitation at 510 nm; (b) Time profile of the absorbance at 1000 nm.Inset: rise and decay time profile for the short time range.Gray line is drawn by the double exponential curve fitting.