Successively Regioselective Electrosynthesis and Electron Transport Property of Stable Multiply Functionalized [60]Fullerene Derivatives.

With the recent advance in chemical modification of fullerenes, electrosynthesis has demonstrated increasing importance in regioselective synthesis of novel fullerene derivatives. Herein, we report successively regioselective synthesis of stable tetra- and hexafunctionalized [60]fullerene derivatives. The cycloaddition reaction of the electrochemically generated dianions from [60]fulleroindolines with phthaloyl chloride regioselectively affords 1,2,4,17-functionalized [60]fullerene derivatives with two attached ketone groups and a unique addition pattern, where the heterocycle is rearranged to a [5,6]-junction and the carbocycle is fused to an adjacent [6,6]-junction. This addition pattern is in sharp contrast with that of the previously reported biscycloadducts, where both cycles are appended to [6,6]-junctions. The obtained tetrafunctionalized compounds can be successively manipulated to 1,2,3,4,9,10-functionalized [60]fullerene derivatives with an intriguing "S"-shaped configuration via a novel electrochemical protonation. Importantly, the stability of tetrafunctionalized [60]fullerene products allows them to be applied in planar perovskite solar cells as efficient electron transport layers.

1. Introduction

Over the past two decades, functionalized [60]fullerene (C60) derivatives have attracted wide attention because of their promising applications in materials, nanotechnology, and biological sciences [1-4]. Particularly, some C60 derivatives, represented by [6,6]-phenyl-C61-butyric acid methyl ester (PCBM), have exhibited excellent superiors in perovskite solar cells (PSCs) as electron transport layers (ETLs) [5-7]. However, the number of regioisomers increases dramatically with the number of addends and thus causes the problem of tedious chromatographic separation for individual regioisomer. Only one isomeric monocycloadduct is usually formed, up to 8 regioisomeric biscycloadducts have been isolated from cycloaddition reactions of C60 [8]. On the other hand, the regioisomers of tetrafunctionalized C60 derivatives reported most commonly are 1,2,3,4-isomers (A) [9-14], 1,4,11,15-isomers (B) [15-18], 1,2,4,15-isomers (C) [19-23], and 1,2,3,16-isomers (D) [24-29], while for the hexafunctionalized C60 derivatives, the most frequently encountered regioisomers are 1,2,3,4,5,6-isomers (E) [30-32] and 1,2,4,11,15,30-isomers (F) [15–18, 33–35] (Figure 1). Although the elegant templated multifunctionalizations of fullerenes have been devised to realize high regioselectivity [8, 36], the regiocontrol on the formation of a specific isomer of biscycloadducts and multicycloadducts is still a daunting task.

Figure 1

Addition patterns of tetra- and hexafunctionalized [60]fullerene derivatives.

The electrophilic-to-nucleophilic reactivity reversal of fullerenes and their derivatives caused by electrochemical reduction opens a new territory in fullerene chemistry and has demonstrated increasing importance in the efficient synthesis of novel fullerene derivatives [37], including the abovementioned types A, C, and D [11, 19, 25–29]. It was interestingly found that the same dianionic [60]fulleroindoline (1a2–, vide infra) behaved differently toward alkylating and acylating reagents. For example, the reaction of 1a2– with benzyl bromide gave both mono- and dibenzylated 1,2,3,16-adducts [25], while its reaction with benzoyl chloride could afford only monoacylated 1,2,3,16-adducts [28]. The attempted synthesis of diacylated 1,2,3,16-adducts failed and remains challenging. In our recent work [38], we disclosed the synthesis of tetra- and hexafunctionalized [60]fullerene derivatives with unprecedented addition patterns from the reaction of 1a2– with 1,2-bis(bromomethyl)benzene. Even though we could capture the intermediate leading to the cyclized 1,2,4,17-adduct, the instability of this unique cyclized product precluded its characterization by 13C NMR and single-crystal X-ray analysis. Bearing the aforementioned different reactivity of the same dianionic species toward alkylating and acylating reagents and instability of the previously obtained tetrafunctionalized product in mind, our continuous efforts in electrochemical functionalization of fullerene derivatives [25, 28, 29, 38–41] stimulated us to investigate the reaction of the dianionic [60]fulleroindolines 1a–c2– with phthaloyl chloride in order to contrast their reactivity behavior and to see if the unprecedented tetrafunctionalized cis-3′ isomers (type G) and “S”-shaped hexafunctionalized products (type H) bearing two acyl groups can be generated in the present case [42]. It turns out that these two unique types of tetra- and hexafunctionalized products can be successfully synthesized and are stable up to 107−275°C. Intriguingly, the reactivity behaviors of 1a2– toward 1,2-bis(bromomethyl)benzene and phthaloyl chloride are quite different, and chemical properties of their anionic tetrafunctionalized products also behave divergently. Importantly, the stability of the current tetrafunctionalized products allows them to be utilized in planar perovskite solar cells so as to investigate their electron transport properties.

2. Results and Discussion

2.1. Electrosynthesis of Tetra- and Hexafunctionalzied Fullerene Derivatives

The dianionic species of [60]fulleroindolines 1 can be obtained by controlled potential electrolysis (CPE) and have ring-opened structures after acceptance of two electrons [25, 28]. For these ring-opened structures, the most negatively charged carbon atom among the fullerene skeleton is located at the para position of the aryl substituents. The reaction of 12– with acyl chlorides proceeded via a SN2 rather than a SET process [28]. Therefore, we surmise that if phthaloyl chloride is chosen to react with 12–, a similar SN2 pathway would afford the anionic intermediate I, followed by the intramolecular SN2 ring-closure process via C–N bond formation [25, 28] to generate 2 with the heterocycle rearranged to a [5,6]-junction and the carbocycle anchored to a [6,6]-junction (Figure 2). It is noteworthy that our thus designed 1,2,4,17-functionalized C60 derivatives 2 have a unique cis-3′ addition pattern (Figure 1), which is in sharp contrast with that for the reported typical cis-3 adducts where both cycloadditions occur at [6,6]-junctions [8].

Figure 2

Synthetic design for the unprecedented cis-3′ regioisomer.

[60]Fulleroindolines 1a–c were synthesized according to our reported procedure [43]. The cyclic voltammograms (CVs) of 1a–c were very similar and showed an irreversible second redox process (Figures ), hinting that the C–N bond cleavage occurred after receiving two electrons [25–29, 38]. It turned out that the reaction of the dianionic species of 1a–c with phthaloyl chloride indeed afforded the desired cis-3′ regioisomers 2a–c. The cyclization of 1a2–, which was obtained by CPE at –1.24 V vs. SCE, with phthaloyl chloride was chosen to screen the optimal reaction conditions (for details, see the text and in the Supplementary Materials). It was found that the reaction of 1a2– with 20 equiv. of phthaloyl chloride in ortho-dichlorobenzene (ODCB) at 0°C for 2 h generated 2a in 40% yield. Similarly, the employment of substrate 1b bearing one methoxy group on the phenyl ring and substrate 1c containing two methoxy groups on the phenyl ring afforded 2b and 2c in 41% and 48% yields, respectively (Figure 3).

Figure 3

Reaction of the electrochemically generated 12– with phthaloyl chloride. Unless otherwise specified, all the reactions were performed with 0.02 mmol of 12– and 0.4 mmol of phthaloyl chloride in 25 mL of ODCB at 0°C for 2 h under an argon atmosphere. aIsolated yield. b10 equiv. of phthaloyl chloride.

Both the first and second redox processes in the CVs of 2a–c were irreversible (Figures ), suggesting that they could be further electrochemically derivatized. In an attempt to obtain the hexafunctionalized fullerene derivative by protonation of 2a2– generated from 2a by CPE at –1.20 V with trifluoroacetic acid (TFA), only the protonated 1,2,3,4-adduct IIa [44] of 1a2– was isolated due to the facile deacylation of the dianionic species and fast concomitant protonation under our conditions. Fortunately, we discovered that with the addition of 1 equiv. of TFA before electroreduction of 2a, the hexafunctionalized fullerene derivative 3a was isolated in 40% yield along with a trace amount of IIa. Further increasing the amount of TFA was detrimental to the product yield. Similarly, products 3b and 3c could be obtained by the electrochemical protonation of 2b and 2c in 33% and 32% yields, respectively (Figure 4). It is believed that product 3 is generated by a highly efficient process of stepwise one-electron reduction and protonation of 2 to give intermediates III and IV, followed by another one-electron reduction to afford V and final protonation (Figure 4). Alternatively, 3 might be generated by a reversal of the sequence as shown in Figure 4 with the first protonation at the carbon atom next to the ketone group. The success for the formation of 3 is probably ascribed to that the presence of TFA facilitates the sequential one-electron reduction and concomitant protonation and thus prohibits the deacylation.

Figure 4

Syntheses of 3a, 3b, and 3c. Unless otherwise specified, all the reactions were performed with 0.01 mmol of 2, 0.01 mmol of TFA, and CPE at –1.20 V, at 25°C under an argon atmosphere. aIsolated yield. bCPE at –1.13 V.

It is worthwhile and illustrative to compare the different reactivity behaviors of the dianionic [60]fulleroindoline 1a2– toward the present phthaloyl chloride and the previously investigated 1,2-bis(bromomethyl)benzene and to contrast the physical and chemical properties of the formed multiply functionalized fullerene derivatives [38]. A monoalkylated 1,2,3,16-adduct, which verified the assumed addition preference at the para position of the aryl substituent, could be isolated if the reaction of 1a2– with 1,2-bis(bromomethyl)benzene proceeded for a short time (10 min, 0°C) and then quenched with TFA. In contrast, the attempts to intercept the anionic intermediate Ia with TFA failed, reflecting that the ring-closure process of Ia was highly rapid to generate 1,2,4,17-adduct 2a. Unlike 1,2-bis(bromomethyl)benzene, phthaloyl chloride reacted with 1a2– at higher temperature afforded only stable 2a, while the isomeric 1,2,3,4-adduct could not be identified. Product 2a was thermally stable up to 168°C (), yet the tetrafunctionalized product from 1,2-bis(bromomethyl)benzene was unstable, tended to decompose, and was partially rearranged to the more stable isomeric 1,2,3,4-adduct. Another difference between these two counterparts was that the acid TFA must be added before the electroreduction of the tetrafunctionalized 2a to successfully form the hexafunctionalized 3a due to the fast deacylation of 2a2–, while TFA could be added as the proton source after the generation of the dianonic tetrafunctionalized product from 1,2-bis(bromomethyl)benzene.

2.2. Characterizations

All new products 1b, 1c, 2a–c, and 3a–c were fully characterized by MALDI-TOF HRMS, 1H NMR, 13C NMR, FT-IR, and UV-Vis spectroscopies. Particularly, the two doublets around 6 ppm with a coupling constant of 2.3 Hz in the 1H NMR spectra of products 3a–c indicated that they contained two fullerenyl protons in 1,4-arrangement [44]. The HMBC spectrum of 3a clearly showed that the proton (6.09 ppm) at C2 (57.04 ppm) correlated with C1 (61.02 ppm) and C3 (68.32 ppm) and that the proton (6.02 ppm) at C10 (56.03 ppm) correlated with C9 (80.56 ppm) (Figures ), indicating that these protons and carbons were adjacent. Furthermore, the assigned structures of 2b and 3b were established by the single-crystal X-ray diffraction analyses (Figure 5). This is the first time that the assignment of 1,2,4,17-adducts was confirmed by single-crystal structure.

Figure 5

(a) ORTEP diagram of one enantiomer of 2b with thermal ellipsoids shown at 50% probability. (b) ORTEP diagram of one enantiomer of 3b with thermal ellipsoids shown at 50% probability. The chloroform molecules are omitted for clarity.

The single crystal of 2b was obtained through slow diffusion of methanol into a chloroform solution of 2b at 4°C. Figure 5(a) displays the X-ray single-crystal diagram for one of the two enantiomers (0.5 : 0.5) of 2b, where a heterocycle is bonded to C60 through a Caryl atom and a N atom at C4 and C17 sites, respectively, and two ketone groups are attached to C1 and C2 sites, respectively. The four functionalized fullerene carbon atoms are uplifted from the spherical surface notably because of their sp3 characters with the bond lengths of 1.560(16) Å and 1.594(10) Å for the C1–C2 and C4–C17 bonds, respectively. The bond lengths for C2–C3, C3–C4, and C1–C6 are 1.559(14) Å, 1.444(12) Å, and 1.509(15) Å, respectively, which are within the range of typical C–C single bond lengths, whereas the C5–C6 bond has a bond length of 1.380(15) Å, thus possessing double bond character. The resolved single-crystal structure unambiguously demonstrates that the molecular structure of the obtained 1,2,4,17-adduct has the cis-3′ addition pattern. The single crystal of 3b was obtained through slow evaporation of a chloroform solution of 3b at 4°C. The X-ray single-crystal diagram for one of the two enantiomers (0.5 : 0.5) of 3b are illustrated in Figure 5(b) and resembles that of 2b except that two additional hydrogen atoms are attached to C2 and C10 atoms. These two carbon atoms bearing hydrogen atoms are also uplifted from the spherical surface notably because of their sp3 characters with the bond lengths of 1.613(7) Å and 1.556(10) Å for the C1–C2 and C9–C10 bonds, respectively. The bond lengths of 1.646(7) Å, 1.601(13) Å, and 1.579(10) Å for C1–C9, C2–C3, and C3–C4, respectively, indicate that they are C–C single bonds; meanwhile, bond lengths for C5–C6 and C11–C12 are 1.358(10) Å and 1.363(11) Å, thus showing double bond character. Intriguingly, the resolved single-crystal structure unequivocally reveals that the molecular structure of the obtained 1,2,3,4,9,10-adduct [45] has a unique “S”-shaped configuration.

2.3. Applications in Perovskite Solar Cells

Although products 2a–c and 3a–c bear a heterocycle fused to a [5,6]-junction of C60, they are thermally stable up to 107−275°C, as determined by thermogravimetric analyses (TGA) (Figures ). Given that fullerene derivatives such as PCBM have strong electron-accepting ability and thus have been popularly applied as ETLs of planar PSCs [46, 47], we next applied two representative highly soluble fullerene products 2a and 2b as novel ETLs of regular-structure (n-i-p) PSC devices with configurations of ITO/ETL/Cs0.05FA0.83MA0.12PbI2.55Br0.45 perovskite/Spiro-OMeTAD/Au, in which 2,2′,7,7′-tetrakis(N,N-di-p-methoxyphenylamine)-9,9′-spirobifluorene (Spiro-OMeTAD) was used as the hole transport material (Figure 6(a)). For comparison, devices without ETL and with commonly used PCBM ETL were also fabricated [46, 47]. The current density-voltage (J−V) curves of the PSC devices based on different ETLs with optimized thicknesses measured under one sun illumination are shown in Figure 6(b), and the measured photovoltaic parameters, including open-circuit voltage (Voc), short-circuit current (Jsc), fill factor (FF), power conversion efficiency (PCE), series resistance (Rs), and shunt resistance (Rsh) of the best performance devices, are summarized in Table 1. The control device without ETL showed a Voc of 1.10 V, a Jsc of 17.92 mA cm−2, an FF of 54.71%, and a PCE of 10.77%. When 2a and 2b were incorporated as ETLs, the device performance enhanced obviously. The 2a-based device exhibited an increased PCE of 13.81%, calculated from a Voc of 1.09 V, a Jsc of 20.82 mA cm−2, and an FF of 60.65%. Upon using 2b as ETL, PCE of the device increased further to 14.04%, which approached that of the PCBM-based device (14.49%). These results showed the considerably good electron transport properties of 2a and 2b.

Figure 6

(a) Schematic structure of the n-i-p PSC device with or without (w/o) fullerene derivative as the ETL. (b) J−V curves of the n-i-p PSC devices without ETL and with PCBM, 2a, and 2b as the ETL measured under illumination of an AM 1.5 solar simulator (100 mW cm−2) in air. The scanning direction was from open-circuit voltage to short-circuit (reverse), and the scan speed was 100 mV s−1.

Table 1

Device parameters of PSCs without ETL and with 2a, 2b, and PCBM as ETLs.

ETL	V oc (V)	J sc (mA cm−2)	FF	PCE (%)	R s (Ω cm2)	R sh (Ω cm2)
—	1.10	17.92	54.71	10.77	7.7	297.6
2a	1.09	20.82	60.65	13.81	6.4	515.4
2b	1.08	20.56	63.12	14.04	5.0	487.1
PCBM	1.11	22.45	58.37	14.49	6.0	304.2

It is known that n-i-p PSC devices based on the conventional TiO2 ETL usually suffer from severe current-voltage hysteresis [7]. To examine the hysteresis of J−V curves of our devices based on 2a or 2b ETL, we measured the J−V curves in different scan directions (Figure 7), and the corresponding device parameters are given in Table 2. The control device without ETL showed severe hysteresis with a hysteresis index, defined as [PCE(reverse) − PCE(forward)]/PCE(reverse) [48], of 20.6%. Upon incorporating 2a or 2b ETL, the device exhibited negligible hysteresis with a small hysteresis index of 1.80% or 1.51%, respectively. This is similar to the case of PCBM ETL. Such a dramatic suppression of the hysteresis may come from the improved electron transport due to the strong electron-accepting ability of 2a and 2b, resulting in suppressed charge accumulation at the perovskite/ITO interface [49, 50]. Therefore, these results along with the comparable PCE to PCBM reveal the promising applications of 2a and 2b in PSCs.

Figure 7

The J-V curves of the PSCs using different fullerene derivatives, (a) without ETL, (b) PCBM, (c) 2a, and (d) 2b as the ETL in different scan directions.

Table 2

Photovoltaic parameters of the best performance devices without ETL and with 2a, 2b, and PCBM as ETL in different scan directions.

ETL	Direction	V oc (V)	J sc (mA cm−2)	FF (%)	PCE (%)	Hysteresis index (%)
—	Reverse	1.11	16.58	56.49	10.36	20.60%
	Forward	1.07	15.81	48.53	8.23
2a	Reverse	1.09	20.20	57.90	12.78	1.80%
	Forward	1.09	20.22	56.89	12.55
2b	Reverse	1.12	19.05	61.94	13.26	1.51%
	Forward	1.12	19.30	60.17	13.06
PCBM	Reverse	1.11	22.45	58.37	14.49	2.40%
	Forward	1.10	21.86	58.90	14.14

3. Conclusion

In summary, we have achieved an efficient and regioselective synthesis of the diacylated products of fulleroindolines 1a–c by the reaction of the electrochemically generated dianionic 1a–c2– with phthaloyl chloride. The obtained tetrafunctionalized fullerene products are 1,2,4,17-adducts 2a–c and have a unique cis-3′ addition pattern where the heterocycle is rearranged to a [5,6]-junction and the carbocycle is appended to a [6,6]-junction. Intriguingly, 1,2,4,17-adducts can be successively protonated to provide hexafunctionalized fullerene products 3a–c by a stepwise one-electron reduction and protonation of 2a–c, which are 1,2,3,4,9,10-adducts and possess an intriguing “S”-shaped addition pattern. The tetra- and hexafunctionalized fullerene derivatives have been fully characterized by spectroscopic data and single-crystal X-ray diffraction analyses. Both 2a–c and 3a–c are stable up to 107−275°C, and representative fullerene products show considerably good electron transport performance in planar perovskite solar cells. This study paves the way to regiocontrolled synthesis of novel multifunctionalized fullerene derivatives toward applications in energy conversion.

4. Materials and Methods

4.1. General Procedure for Synthesis of –

The dianionic 12– was obtained by electroreduction from [60]fulleroindoline 1 (0.02 mmol) at –1.24 V by CPE and then reacted with phthaloyl chloride (0.40 mmol or 0.20 mmol). After being stirred at 0°C for 2 h, the resulting mixture was directly filtered through a silica gel (200–300 mesh) plug with CS2/CH2Cl2 (1 : 1 v/v) to remove the supporting electrolyte and insoluble materials and then evaporated in vacuo to remove the solvent. Next, the residue was further separated on a silica gel column (300–400 mesh) with CS2/CH2Cl2 as the eluent to afford 2 as an amorphous brown solid along with unreacted 1.

4.2. General Procedure for Synthesis of –

The mixture of 2 (0.01 mmol) and TFA (0.74 μL, 0.01 mmol) was dissolved in ODCB containing 0.1 M TBAP and then electroreduced by CPE at –1.20 V or –1.13 V. The potentiostat was turned off after the theoretical coulomb was reached. The resulting mixture was directly filtered through a silica gel (200–300 mesh) plug with CS2/CH2Cl2 (1 : 1 v/v) to remove the supporting electrolyte and insoluble materials and then evaporated in vacuo to remove the solvent. Next, the residue was further separated on a silica gel column (300–400 mesh) with CS2/CH2Cl2 as the eluent to afford 3 as an amorphous red-brown solid along with a minor byproduct II.

4.3. Device Fabrication of Perovskite Solar Cells

The patterned ITO-coated glass was cleaned by sequential ultrasonic treatment in detergent, deionized water, acetone, and isopropanol for 15 min and then treated with ultraviolet-ozone for 20 min. The PCBM (20 mg mL−1 in ODCB) or representative fullerene derivative (2a and 2b, saturated solution in ODCB) was deposited on the ITO substrates by spin coating at 2000 rpm for 60 s. The as-spun films were annealed at 100°C for 10 min. Next, Cs0.05FA0.83MA0.12PbI2.55Br0.45 perovskite precursor solution (1.3 M dissolved in dimethyl sulfoxide and N,N-dimethylformamide with a volume ratio of 2 : 8, with molar ratios of PbI2/PbBr2, 1.1 : 0.2; FAI : MABr, 1 : 0.2; CsI/(FAI+MABr), 0.05 : 0.95; PbI2/FAI, 1.1 : 1; and PbBr2 : MABr, 1 : 0.2) was spin-coated onto the substrates with a two-step procedure. The first step was 2000 rpm for 10 s with an acceleration of 200 rpm. The second step was 6000 rpm for 30 s with an acceleration of 2000 rpm. At 15 s before the end of the second procedure, 100 μL chlorobenzene (CB) was dropped on the spinning substrate. The substrate was then immediately transferred on a hotplate and heated at 100°C for 60 min. After the perovskite films were cooled down to room temperature, the hole transport layer was deposited on top of the perovskite film by spin coating at 3000 rpm for 30 s using a CB solution which contained 73.2 mg mL−1 of Spiro-OMeTAD and 28.8 μL mL−1 of tert-butylpyridine, as well as 18.8 μL mL−1 of bis(trifluoromethane)sulfonimide lithium salt (Li-TFSI, 520 mg mL−1 in acetonitrile). Finally, the device was transferred into a vacuum chamber (10−6 torr), and an Au electrode (ca. 55 nm thick) was thermally deposited through a shadow mask to define the effective active area of the device (0.10 cm2).