A Solid‐State Intramolecular Wittig Reaction Enables Efficient Synthesis of Endofullerenes Including Ne@C60, 3He@C60, and HD@C60

Abstract An open‐cage fullerene incorporating phosphorous ylid and carbonyl group moieties on the rim of the orifice can be filled with gases (H2, He, Ne) in the solid state, and the cage opening then contracted in situ by raising the temperature to complete an intramolecular Wittig reaction, trapping the atom or molecule inside. Known transformations complete conversion of the product fullerene to C60 containing the endohedral species. As well as providing an improved synthesis of large quantities of 4He@C60, H2@C60, and D2@C60, the method allows the efficient incorporation of expensive gases such as HD and 3He, to prepare HD@C60 and 3He@C60. The method also enables the first synthesis of Ne@C60 by molecular surgery, and its characterization by crystallography and 13C NMR spectroscopy.

S3 reduced using a 1.5 mm stainless steel rod inserted into the 1/16" (1.59 mm) i.d. steel tube. A custom 30 mL manual pump from Sitec (model 759.1100 -as 750.1100 but with internal dead volume reduced to ~1.7 mL, 1000 atm rating) was used to increase the pressure of the source gas (figure S1.2(a)).
The high-pressure reactor was constructed from a 100 mm length of 316L steel coned and threaded tubing with 3/8" external diameter (9.52 mm) and 5.2 mm internal diameter (pressure rating 2400 atm) with a blank (720.2114) at one end and adapter (720.1620) to 1/4" steel tube at the other (figure S1.2(b)). A cut down 5 mm o.d. pyrex NMR tube was usually used to contain the sample of solid open-fullerene 4, although direct packing of the reactor could be used to increase the mass of 4 which could be accommodated.
Pressure was constantly monitored using a Wika model HP-2 pressure sensor (hydrogen compatible, when appropriate) connected to a Hengstler 0735A20000 LED display. An oven modified from a Kugelrohr distillation apparatus and fitted with a Sestos model DIS temperature controller and PT100 thermocouple, was used to heat the reactor. Thermocouples were also attached to the reactor, at each end and in the centre, and the temperature at each position was constantly recorded using a Pico technology USB TC-08 data logger attached to a PC. The reaction temperature stated for each procedure in section S1.4 is from the sensor attached to the centre of the reactor tube.
To achieve higher final pressures the bomb was sometimes cooled in liquid nitrogen while filling (figure S1.2(c)).
The system allowed evacuation to <0.2 mbar before use. When using 3 He gas, following completion of the reaction, as much gas as possible was pumped from the reactor back into the source cylinders, and any residual gas recovered into previously evacuated 1L cylinders. Also with 3 He filling additional components allowed the loaded bomb to be pressure tested with 4 He before evacuation and charging with 3 He.

S1.3 Synthesis of phosphorus ylid 4 from C60
A solution of C60 (6.821 g, 9.473 mmol) and 3,6-bis (6-(tert-butyl)pyridin-2-yl)pyridazine (1.621 g, 4.678 mmol) in 1chloronaphthalene (200 mL) was degassed under dynamic vacuum (approx. 0.2 mm Hg) and sonication. The stirred solution was then warmed to reflux for 42 h under N2, before cooling to room temperature and dilution with toluene (150 mL). This mixture was transferred to a jacketed photo reactor and the water-cooled solution was irradiated with a high-pressure Na lamp (400 W), placed in the middle of the reactor at an approximate distance of 2 cm from the solution, bubbling oxygen at a continuous flow rate of approx. 20 mL min -1 for 1 h. The reaction mixture was then diluted with hexane (200 mL) and poured directly over a silica column (7 cm diameter column, 600 mL of SiO2) packed with hexane:toluene (1:1).
Using an eluent gradient of hexane:toluene (1:1) → 100% toluene, a purple band containing unreacted C60 and 1chloronaphthalene was first collected. From quantitative HPLC assay, the amount of recovered C60 was calculated to be 4.06 g (60%), and the quantity of C60 not recovered was 2.76 g (3.83 mmol). Solvents were removed by vacuum distillation and residual C60 was washed with Et2O (3 × 30 mL) before drying in vacuo. Recovered C60 was of sufficient purity to be used in a repeat reaction.
A second band containing material with Rf = 0.5 in toluene was collected, and removal of solvents in vacuo gave open-fullerene 5 as a brown powder (2.848 g, 70% based on C60 consumption). Spectroscopic data were consistent with the published data. [2] Conversion of 5 to bis(hemiketal) 3 was carried out according to the published method. [2] Conversion of bis(hemiketal) 3 (5.300 g, 4.727 mmol) to phosphorus ylid 4 (5.403 g, 85%) was carried out by scaleup of our previously reported procedure. [3] The structure of 4 was confirmed by X-ray crystallography as described in section S3.1, and all spectroscopic data were consistent with those previously reported. S5 S1.4 Solid-state filling of 4 with in situ trapping of the endohedral species S1. 4

.1 Synthesis of HD@5
A thin-walled pyrex tube was charged with solid open-fullerene 4 (0.726 g, 0.538 mmol) and inserted into a steel high-pressure reactor equipped with pressure intensifier (section S1.2). The reactor was degassed under dynamic vacuum (approx. 0.2 mm Hg) before cooling to −196 °C using liquid nitrogen and charging with HD gas to 26 atm. The gas was then compressed in the reactor to 119 atm, before allowing to warm to room temperature at which point the pressure reached 420 atm.
The reactor was then heated to 140 °C and maintained at this temperature for 14 h, with a stable internal pressure of 520 atm during this time. After cooling to room temperature and slow release of the pressure, the solid residue was purified by flash column chromatography (SiO2 eluted with toluene). Material with Rf = 0.5 in toluene was collected, and removal of solvents in vacuo gave the title compound HD@5 as a brown powder (0.432 g, 75 %) with 83% HD filling.

S1.4.3 Synthesis of H2@5
A thin-walled pyrex tube was charged with solid open-fullerene 4 (0.556 g, 0.412 mmol) and inserted into a steel high-pressure reactor equipped with pressure intensifier (section S1.2). The reactor was degassed under dynamic vacuum (approx. 0.2 mm Hg) and then charged with H2 gas to 121 atm. The reactor was then cooled to −196 °C using liquid nitrogen and the gas compressed to 700 atm. The reactor was warmed to room temperature, at which point the pressure reached 1384 atm, then heated to 186 °C and maintained at this temperature for 2 hours, with a stable internal pressure of 1806 atm during this time. After cooling to room temperature and slow release of the pressure, the solid residue was purified by flash column chromatography (SiO2 eluted with toluene). Material with Rf = 0.5 in toluene was collected, and removal of solvents in vacuo gave H2@5 as a brown powder (0.351 g, 79%). >95% H2 filling was calculated from the 1 H NMR spectrum. 1  All spectroscopic data were consistent with those we have previously published for a 60% filled sample of H2@5. [4]

S1.4.4 Synthesis of 4 He@5
A thin-walled pyrex tube was charged with solid open-fullerene 4 (0.494 g, 0.366 mmol) and inserted into a steel high-pressure reactor equipped with pressure intensifier (section S1.2). The reactor was degassed under dynamic vacuum (approx. 0.2 mm Hg) and then charged with He gas to 72 atm. The gas was compressed to 439 atm, and the reactor was then cooled to −196 °C using liquid nitrogen.

S1.4.6 Synthesis of Ne@5
A thin-walled pyrex tube was charged with solid open-fullerene 4 (0.588 g, 0.436 mmol) and inserted into a steel high-pressure reactor equipped with pressure intensifier (section S1.2). The reactor was degassed under dynamic vacuum (approx. 0.2 mm Hg) and then charged with Ne gas to 37 atm. The gas was compressed in the reactor to 346 atm, and the reactor was then cooled to −196 °C using liquid nitrogen.

S1.5.4 Synthesis of H2@C60
Using the general procedure described in section S1. All spectroscopic data were consistent with those previously reported for H2@C60. [6] 1 H 1 H S20 S1.5.5 Synthesis of 4 He@C60 Using the general procedure described in section S1. 5  Two-step cage closure of open fullerene 3 He@5 was carried out according to our previously reported procedure. [4] The optimized general procedure given in section S1.5.1 was developed later, but we did not apply this method to a repeated synthesis of 3 He@C60 due to the expense and limited availability of 3 He gas.
To a stirred solution of 52% filled 3 He@5 (

S2. Density Functional Theory Calculations
Model structures 2a and 4a were used to represent compounds 2 and 4, in which the 6-tert-butylpyridyl groups were replaced by methyl substituents. Calculations were carried out using Gaussian 09, [8] using the M06-2X functional [9] with cc-pVDZ [10] basis set to locate minimum energy and transition state structures and to characterise them through frequency calculations. The cc-pVTZ [10] basis set with an ultrafine integration grid was used to calculate electronic energies and to correct for Basis Set Superposition Error using the counterpoise method. [11] Thermal corrections to the electronic energy to give the enthalpy at 298 K and 1 atm were derived from frequency calculations at the M06-2X/cc-pVDZ level using the Gaussian freqchk utility. [12] The frequencies were not scaled and low frequency modes were not removed.

S3.1.2 Experimental method
Preparation of a single crystal was made by slow evaporation, after layering a few drops of hexane over a solution of open-fullerene 4 (1 mg) in CHCl3 (1 mL), in a 1.5 mL glass vial. The crystal 0.43×0.19×0.06 mm 3 was mounted on a MITIGEN holder with silicon oil on a Rigaku AFC12 FRE-VHF diffractometer. The crystal was kept at a steady T = 100(2) K during data collection. The structure was solved with the ShelXT [13a] structure solution program using the Intrinsic Phasing solution method and by using Olex2 [14] as the graphical interface. The model was refined with version 2016/6 of ShelXL [13b] using Least Squares minimisation.
A multi-scan absorption correction was performed using CrysAlisPro 1.171.40.37a (Rigaku Oxford Diffraction, 2019) using spherical harmonics as implemented in SCALE3 ABSPACK. The absorption coefficient of this material is 0.237 mm -1 at this wavelength (l = 0.711Å) and the minimum and maximum transmissions are 0.856 and 1.000 The structure was solved in the space group P21/c (# 14) by Intrinsic Phasing using the ShelXT [13a] structure solution S27 program and refined by Least Squares using version 2016/6 of ShelXL. [13b] All non-hydrogen atoms were refined anisotropically. Hydrogen atom positions were calculated geometrically and refined using the riding model. There is a single molecule in the asymmetric unit, which is represented by the reported sum formula. In other words: Z is 4 and Z' is 1.

S3.2.1 Experimental method
Preparation of single dark orange crystals was made by slow evaporation of the nickel(II) octaethylporphyrin/benzene solvate of a sample of Ne@C60 with >99.5% neon filling. [15] A suitable crystal 0.44×0.27×0.10 mm 3 was selected and mounted on a MITIGEN holder with silicon oil on an Rigaku AFC12 FRE-VHF diffractometer. The crystal was kept at a steady T = 100(2) K during data collection. The structure was solved with the ShelXT [13a] structure solution program using the Intrinsic Phasing solution method and by using Olex2 [14] as the graphical interface. The model was refined with version 2016/6 of ShelXL [13b] using Least Squares minimisation.

Refinement:
A dark orange plate-shaped crystal with dimensions 0.44×0.27×0.10 mm 3 was mounted on a MITIGEN holder with silicon oil. X-ray diffraction data were collected using a Rigaku AFC12 FRE-VHF diffractometer equipped with an Oxford Cryosystems low-temperature device, operating at T = 100 (2)  The final completeness is 99.90 % out to 28.499° in Q.