Highly efficient synthesis of [60]fullerene oxides by plasma jet.

Atmospheric pressure nonequilibrium plasma jet has been applied to the synthesis of [60]fullerene oxides (C60On) for the first time. C60O and C60O2 were produced and isolated in high yields up to 44% and 21%, respectively. The structural assignment of C60O was confirmed by comparison with the reported spectroscopic data. Theoretical calculations of 13C NMR chemical shifts for eight isomers of C60O2 were performed and compared with the experimental data to assign the most possible structure for the obtained C60O2 dominantly as an e isomer.

Introduction

Since the discovery of [60]fullerene (C60), various C60 derivatives have been synthesized due to their broad potential applications in materials science, biology and nanoscience [1-7]. Among these derivatives, the study of fullerene oxides (C60On) has been a popular area of research for decades. Particularly, the monoxide C60O is the best characterized oxidation product of C60, and exhibits unique properties [8-11]. In 1992, Cox, Smith and co-workers reported the photochemical synthesis of C60O in 16% yield by UV irradiation of an oxygenated benzene solution of C60 in the presence of benzyl [12]. Nearly at the same time, Foote, Whetten and co-workers described the formation of C60O in 4% yield from the oxidation reaction of C60 with dimethyldioxirane [13]. In 1995, the Balch group disclosed that the oxidation reaction of C60 with m-chloroperoxybenzoic acid (m-CPBA) afforded C60O in 30% yield and C60O2 as the cis-1 isomer in 8% yield [14]. In addition, the formation of C60O and C60On (n ≥ 2) was identified by mass spectroscopy or high-performance liquid chromatography (HPLC), yet no isolated yields were reported in the oxidation of electrochemically generated anionic C60 [15] and ozonolysis of C60 [16-18]. Interestingly, C60O has been employed as the precursor for further functionalizations, leading to the formation of C120O [19], 1,3-dioxolane derivatives of C60 [20], indoline derivatives of C60 [21], and 1,2-perfluorophenylfullerenol [22]. Similarly, the cis-1 isomer of C60O2 has been used as the precursor for bis-1,3-dioxolane derivatives of C60 [20]. Despite these achievements, new approach to obtain C60O and C60O2 in higher yields is still highly demanding. Furthermore, the efficient synthesis and characterization of other isomeric C60O2 except cis-1 isomer are still unknown until now.

Recently, atmospheric pressure nonequilibrium plasma jet (APNPJ) has attracted extensive attention due to its wide range of applications in plasma biology, healthcare, medicine as well as surface and materials processing [23-28]. Because APNPJ generates plasma in open space rather than in confined gaps, there are no restrictions on the size of the objects to be treated, and short lifetime active species and even charged particles can easily reach the employed objects [29-33]. To the best of our knowledge, the application of APNPJ to fullerene chemistry has not been reported until now. Herein, we report the synthesis of C60O and C60O2 by APNPJ for the first time. This novel method features simple operation and high efficiency.

Experimental

The schematic view of the experiment set-up is shown in figure 1. It consisted of two coaxial glass tubes with a hollow steel needle in the centre of the inner tube. The hollow steel needle with a tip radius of about 100 µm served as a high-voltage (HV) electrode, which was directly connected to a pulsed direct current (DC) power supply, and was also used for carrying and guiding the working gas flow. There was a nozzle outlet at the end of the inner tube with an inner diameter of approximately 2 mm. The distance between the needle tip and the nozzle was about 5 mm. The glass tube was fixed into a round-bottom flask with a distance of about 3 mm between the nozzle and the solution surface. The working gas, argon (99.99%) passed through the hollow steel needle to generate the plasma jet, while O2 (99.9%) was introduced into the round-bottom flask, serving as the surrounding gas and oxygen source. Before turning on the power supply, Ar and O2 were allowed to flow into the flask for 5 min. Then, a high pulsed DC voltage was applied to the HV electrode, and a plasma jet was generated at the end of the needle inside the glass tube. The synthesis process was carried out by applying the plasma jet to the chlorobenzene solution of C60.

Figure 1.

Schematic view of the experimental set-up.

Results and discussion

Synthesis of C60On

Initially, the applied discharge voltage of 3.5 kV was fixed, Ar and O2 flow rates were kept at 0.2 and 0.8 l min−1, respectively. Under the above APNPJ conditions, a solution of C60 (7.2 mg, 0.01 mmol) in 3 ml of chlorobenzene was treated at 10°C for 10 min to give a claret-red solution along with brown precipitates, which did not dissolve in commonly used organic solvents [14,16]. The solution was analysed by HPLC on a Buckyprep column (4.6 × 250 mm) (table 1, entry 1).

Table 1.

Optimization of the plasma jet-promoted reaction of C60 with O2a.

entry	Ar/O2 (l min−1)	V kV−1	reaction temp. (°C)	reaction time (min)	HPLC peak area for C60Ob	HPLC peak area for C60O2b
1	0.2/0.8	3.5	10	10	7552	1528
2	0.2/0.8	3.5	0	10	9017	1953
3	0.2/0.8	3.5	−10	10	8915	1985
4	0.2/0.8	3.5	0	15	8586	2526
5	0.2/0.8	3.5	0	5	5924	658
6	0.2/0.8	4.0	0	10	5937	1408
7	0.2/0.8	3.0	0	10	6346	755
8	0.2/0.4	3.5	0	10	9273	2643
9	0.2/0.2	3.5	0	10	7949	2080
10	0.3/0.4	3.5	0	10	9023	2188
11	0.2/0.4	3.5	0	25	3190	3354
12	0.2/0.8	3.5	0	25	5205	3764
13	0.2/0.8	3.5	−10	25	5944	5362
14	0.2/0.8	3.5	−20	25	6718	6306
15	0.2/0.8	3.5	−30	25	2554	4008
16	0.2/0.8	3.5	−20	30	3104	4376
17	0.2/0.8	3.5	−20	20	6882	4157

aUnless otherwise noted, the reaction was performed using 7.2 mg of C60 in anhydrous chlorobenzene (3 ml) at the set temperature under the plasma conditions for the designated time.

bThe HPLC peak area with an injection of 8 µl of the reaction mixture.

Structural assignments

As shown in figure 2a, three peaks were observed for the reaction mixture. Peaks I and II could be conclusively assigned as unreacted C60 and the known epoxide C60O, respectively, by comparison of their retention times with those of the authentic samples. The structure of C60O was further unambiguously confirmed by comparison of its HR-MS (MALDI-TOF), 13C NMR, UV-vis and FT-IR spectra with those reported previously [12-14]. A series of work has been done to identify the possible structure for peak III. Interestingly, peak III had a different retention time from that of the cis-1 isomer of C60O2, instead it was eluted out at the same retention time as that of the unknown minor isomer of C60O2 produced by the oxidation of C60 with m-CPBA (figure 2b) [14]. The product contained in peak III was separated out by semi-preparative HPLC equipped with a Buckyprep column (10 × 250 mm) to allow structural analysis. The fraction for peak III was evaporated in vacuo and isolated as a brown powder, which had poor solubility in common good solvents including o-dichlorobenzene (ODCB) for fullerenes. The high-resolution mass spectrum HR-MS (MALDI-TOF) of the product showed a strong ion peak at 751.9878, which matched with the calculated molecular ion peak (751.9893, [M]+) of dioxide C60O2. The infrared spectrum for peak III in a KBr pellet revealed no bands above 1550 cm−1 and thus no C–H or C=O moiety was present. The peaks for the fullerene cage were observed at 1427, 1182, 563 and 525 cm−1. There were also other peaks due to the derivatization of the fullerene skeleton. The UV-vis spectrum of a given bisadduct of C60 depends mostly on the addition pattern rather than the nature of the addend. Hence, it is possible to use the UV-vis spectra as a diagnostic tool for the structural assignment of the newly synthesized bisadducts. The UV-vis spectrum for peak III exhibited a broad absorption band near 466 nm, and was almost the same as that of analogous e regioisomers of C62(COOEt)4 and C60(NCOOEt)2 [34-36]. The 13C NMR spectrum for peak III in CS2/CDCl3 with chromium(III) tris(acetylacetonate) as the relaxation reagent was obtained (figure 3, also see electronic supplementary material, figures S8 and S9). There were about 30 dominant peaks for the sp2 carbons of C60 in the 13C NMR spectrum, suggesting that peak III predominantly consisted of one C60O2 isomer. The attempt to separate the possible isomers from each other in peak III by the recycling preparative HPLC on a Buckyprep or Buckyprep-M column failed (see electronic supplementary material, figures S13 and S14).

Figure 2.

(a) HPLC trace for the reaction mixture treated by APNPJ under conditions in entry 1 of table 1. (b) Black line: HPLC trace for the reaction mixture of C60 and m-CPBA; red line: HPLC trace after addition of the isolated C60O2 under the APNPJ conditions to the reaction mixture of C60 and m-CPBA. (c) HPLC trace giving the highest yield of C60O under APNPJ conditions. (d) HPLC trace giving the highest yield of C60O2 under APNPJ conditions. The mobile phase on the Buckyprep (4.6 × 250 mm) column was toluene (1 ml min−1).

Figure 3.

Comparison of the calculated 13C NMR spectrum of the e isomer of C60O2 with the experimental 13C NMR spectrum of peak III.

In order to gain further insight into the molecular structure for peak III, theoretical calculations of the 13C NMR chemical shifts based on gauge-including atomic orbitals (GIAO) [37] for eight isomers of C60O2 were performed at the B3LYP/6-311+G (2df, 2pd) level with the Gaussian 09 program [38] and compared (see figure 3 and electronic supplementary material, figure S15). The linear correlation between the calculated 13C NMR chemical shifts and experimentally obtained chemical shifts of cis-1 isomer of C60O2 was very good (R = 0.999, see electronic supplementary material, figure S16), indicating that the calculated 13C NMR chemical shifts based on GIAO were reliable. The comparison of the number and relative position of signals between computational and experimental 13C NMR spectra suggested that only the calculated 13C NMR spectrum of the e isomer could match with that of peak III well, consistent with the conclusion from the above-mentioned UV-vis spectrum analysis.

Optimization of reaction conditions

With the preliminary result (table 1, entry 1) in hand, the reaction conditions were optimized by changing the reaction time, reaction temperature, discharge voltage and gas flow rate to obtain the highest yield of C60O, and the results are summarized in table 1. The reaction temperatures and reaction times were varied, and it was found that 0°C and 10 min were optimal (table 1, entries 2–5). The discharge voltage was also examined, increasing or decreasing the voltage did not improve the yield (table 1, entries 6 and 7). Next, the gas flow rate was further investigated. A combination of 0.2 and 0.4 l min−1 for Ar and O2 resulted in the highest yield of C60O (table 1, entry 8). However, further decreasing the O2 flow rate was not beneficial to the reaction efficiency (table 1, entry 9). When the flow rate of the working gas was increased to 0.3 l min−1, a slightly decrease in product yield was observed (table 1, entry 10). The plasma plume was too weak to initiate the reaction when the Ar flow rate was 0.1 l min−1. Therefore, the optimal reaction conditions for the synthesis of C60O were determined as follows: the applied discharge voltage of 3.5 kV, the Ar flow rate of 0.2 l min−1, the O2 flow rate of 0.4 l min−1, the reaction temperature of 0°C and the reaction time of 10 min. In order to achieve a higher yield of C60O2, the reaction time was prolonged from 10 to 25 min; a slightly higher yield was obtained (table 1, entry 11). Increasing the O2 flow rate from 0.4 l min−1 to 0.8 l min−1 further improved the yield of C60O2 (table 1, entry 12). It was found that decreasing the reaction temperature to −20°C afforded the highest yield of C60O2 (table 1, entries 13−15). Further increasing or decreasing the reaction time was not beneficial to the formation of C60O2 (table 1, entries 16 and 17). As a result, the optimal reaction conditions for C60O2 were achieved by lowering the temperature to −20°C and prolonging the reaction time to 25 min as well as increasing the O2 flow rate to 0.8 l min−1 (table 1, entry 14). Under the optimum reaction conditions, the isolated yields for C60O and C60O2 were 44% (figure 2c) and 21% (figure 2d), respectively. Notably, this method could be successfully applied to a scale-up reaction. The reaction of C60 (36 mg, 0.05 mmol) under the optimal conditions for the synthesis of C60O could afford 37% yield after prolonging the reaction time to 45 min. Likewise, treatment of C60 (36 mg, 0.05 mmol) with the plasma jet at −20°C for 90 min provided C60O2 in 11% yield. In addition, we had also preliminarily studied the oxidation of C70 under the plasma conditions. Unfortunately, much less of C70 was converted when compared with C60, probably due to its lower reactivity (see electronic supplementary material, figures S17−S19). Therefore, the oxidation of C70 by plasma jet was not investigated in more details.

Optical emission spectra and reaction mechanism

To shed light on the mechanism of the plasma-promoted formation of C60On, optical emission spectroscopy was employed to identify the reactive species generated by the Ar/O2 APNPJ. Figure 4 shows the comparison of the optical emission spectra of both Ar and Ar/O2 plasma jets in the range of 650−860 nm. Two additional emission peaks, which were assigned to the reactive atomic oxygen species O* (777.2 nm, 3p5P–3s5S) and O* (844.6 nm, 3p3P–3s3S), were detected in the Ar/O2 APNPJ [39-42]. The generation of the reactive oxygen species as well as the production of C60On are shown in equations (3.1)−(3.5) [43-46]. Firstly, the plasma jet initiated the reaction and produced active species including activated Ar*, electrons and so on. Then, these active species collided with dioxygen to produce the active atomic oxygen species (equations (3.1)−(3.4)), which reacted with C60 in solution to produce C60O and C60O2.

Figure 4.

Optical emission spectra in the range of 650–860 nm: (a) Ar, (b) the mixture of Ar and O2.

Different from the dominant formation of the cis-1 isomer of C60O2 by the oxidation with m-chloroperoxybenzoic acid, the e isomer of C60O2 was selectively generated under our plasma jet conditions, indicating that their reaction mechanisms should be different. Previous theoretical study showed that the cis-1 isomer was the most stable and the e isomer was the second most stable thermodynamically among the eight isomers of C60O2 [47]. Therefore, the selective formation of the e isomer of C60O2 under our plasma jet conditions should be governed by a kinetic process, most probably due to very reactive oxygen atom species.

Conclusion

In summary, we have successfully developed the plasma jet-promoted synthesis of fullerene oxides. Under the respectively optimized conditions, C60O and C60O2 were isolated in 44% and 21% yields, which are the highest reported so far. More importantly, the generated C60O2 under the plasma jet conditions dominantly consisted of the e isomer, which exhibited very different regioselectivity compared to the cis-1 isomer as the major product of C60O2 formed from the oxidation of C60 with m-chloroperoxybenzoic acid. This is the first time that plasma jet is applied to fullerene chemistry, and this work may be of interest to synthetic chemists for developing the unique plasma technique to promote chemical reactions.

General procedure and characterization data

Synthesis of C60O as the major product: under the applied discharge voltage of 3.5 kV, the Ar flow rate of 0.2 l min−1, the O2 flow rate of 0.4 l min−1, the chlorobenzene solution (3 ml) of C60 (0.01 mmol, 7.2 mg) was treated with the plasma jet at 0°C for 10 min, and colour of the solution turned from purple to claret-red. The resulting solution from five runs was combined and filtrated through a silica gel plug with carbon disulfide as the eluent to remove any insoluble material and then evaporated in vacuo. The residue was dissolved in toluene and separated by recycling preparative HPLC on a Buckyprep column (10 × 250 mm) with a mixture of toluene and n-hexane (1 : 1 v/v) as the eluent, giving unreacted C60 (10.9 mg, 31%), C60O (15.9 mg, 44%) and C60O2 (4.0 mg, 11%).

A scale-up reaction of C60O as the major product: under the applied discharge voltage of 3.5 kV, the Ar flow rate of 0.2 l min−1, the O2 flow rate of 0.4 l min−1, the chlorobenzene solution (15 ml) of C60 (0.05 mmol, 35.8 mg) was treated with the plasma jet at 0°C for 45 min, and colour of the solution turned from purple to claret-red. The resulting solution was filtrated through a silica gel plug with carbon disulfide as the eluent to remove any insoluble material and then evaporated in vacuo. The residue was dissolved in toluene and separated by recycling preparative HPLC on a Buckyprep column (10 × 250 mm) with a mixture of toluene and n-hexane (1 : 1 v/v) as the eluent, giving unreacted C60 (15.4 mg, 43%), C60O (13.5 mg, 37%) and C60O2 (1.8 mg, 5%).

13C NMR (100 MHz, CDCl2CDCl2 with chromium(III) tris(acetylacetonate) as a relaxation reagent all 4C unless indicated) δ 145.17 (8C), 145.06, 144.94 (2C), 144.23, 143.96, 143.85, 143.74, 143.40 (2C), 142.96, 142.95, 142.43 (2C), 142.25, 142.05, 141.83, 140.78, 89.53 (2C, sp3-C of C60); FT-IR ν/cm−1 (KBr) 1537, 1504, 1458, 1427, 1378, 1310, 1243, 1183, 1080, 876, 769, 742, 674, 625, 598, 568, 526, 498, 435; UV-vis (toluene) λmax/nm (log ε) 424 (3.19), 496 (3.07); UV-vis (CHCl3) λmax/nm (log ε) 255 (5.13), 323 (4.57), 412 (3.47), 420 (3.35), 494 (3.26); HR-MS (MALDI-TOF) m/z calcd for C60O [M]+ 735.9944, found 735.9909.

Synthesis of C60O2 as the major product: under the applied discharge voltage of 3.5 kV, the Ar flow rate of 0.2 l min−1, the O2 flow rate of 0.8 l min−1, the chlorobenzene solution (3 ml) of C60 (0.01 mmol, 7.2 mg) was treated with the plasma jet at −20°C for 25 min, and colour of the solution turned from purple to claret-red. The resulting solution from five runs was combined and filtrated through a silica gel plug with carbon disulfide as the eluent to remove any insoluble material and then evaporated in vacuo. The residue was dissolved in toluene and separated by recycling preparative HPLC on a Buckyprep column (10 × 250 mm) with a mixture of toluene and n-hexane (1 : 1 v/v) as the eluent, affording C60 (4.5 mg, 13%), C60O (11.6 mg, 32%) and C60O2 (7.8 mg, 21%).

A scale-up reaction of C60O2 isomers as the major product: under the applied discharge voltage of 3.5 kV, the Ar flow rate of 0.2 l min−1, the O2 flow rate of 0.8 l min−1, the chlorobenzene solution (15 ml) of C60 (0.05 mmol, 36.5 mg) was treated with the plasma jet at −20°C for 90 min, and colour of the solution turned from purple to claret-red. The resulting solution was filtrated through a silica gel plug with carbon disulfide as the eluent to remove any insoluble material and then evaporated in vacuo. The residue was dissolved in toluene and separated by recycling preparative HPLC on a Buckyprep column (10 × 250 mm) with a mixture of toluene and n-hexane (1 : 1 v/v) as the eluent, affording unreacted C60 (7.4 mg, 20%), C60O (10.6 mg, 28%) and C60O2 (4.0 mg, 11%).

13C NMR (100 MHz, CS2/CDCl3 with chromium(III) tris(acetylacetonate) as a relaxation reagent) δ 147.36, 147.16, 147.01, 146.59, 146.33, 146.10, 145.73, 145.36, 145.33, 145.25, 145.16, 144.86, 144.73, 144.45, 144.37, 144.34, 144.14, 144.09, 144.06, 143.93, 143.56, 143.50, 143.31, 143.23, 142.99, 142.80, 142.54, 142.43, 141.76, 141.46, 139.14, 90.96 (sp3-C of C60), 89.84 (sp3-C of C60), 89.26 (sp3-C of C60); FT-IR ν/cm−1 (KBr) 1539, 1507, 1456, 1427, 1376, 1305, 1240, 1182, 1085, 1036, 964, 796, 770, 746, 625, 563, 525, 497, 432; UV-vis (toluene) λmax/nm (log ε) 466 (3.73); UV-vis (CHCl3) λmax/nm (log ε) 250 (5.12), 301 (4.72), 466 (3.63); HR-MS (MALDI-TOF) m/z calcd for C60O2 [M]+ 751.9893, found 751.9878.