Fullerene derivatives with oligoethylene–glycol side chains: an investigation on the origin of their outstanding transport properties

For many years, fullerene derivatives have been the main n-type material of organic electronics and optoelectronics. Recently, fullerene derivatives functionalized with ethylene glycol (EG) side chains have been showing important properties such as enhanced dielectric constants, facile doping and enhanced self-assembly capabilities. Here, we have prepared field-effect transistors using a series of these fullerene derivatives equipped with EG side chains of different lengths. Transport data show the beneficial effect of increasing the EG side chain. In order to understand the material properties, full structural determination of these fullerene derivatives has been achieved by coupling the X-ray data with molecular dynamics (MD) simulations. The increase in transport properties is paired with the formation of extended layered structures, efficient molecular packing and an increase in the crystallite alignment. The layer-like structure is composed of conducting layers, containing of closely packed C60 balls approaching the inter-distance of 1 nm, that are separated by well-defined EG layers, where the EG chains are rather splayed with the chain direction almost perpendicular to the layer normal. Such a layered structure appears highly ordered and highly aligned with the C60 planes oriented parallel to the substrate in the thin film configuration. The order inside the thin film increases with the EG chain length, allowing the systems to achieve mobilities as high as 0.053 cm2 V−1 s−1. Our work elucidates the structure of these interesting semiconducting organic molecules and shows that the synergistic use of X-ray structural analysis and MD simulations is a powerful tool to identify the structure of thin organic films for optoelectronic applications.

(a) Synthesis of tetraethylene glycol monoethyl ether 1: Triethylene glycol (2.25 g, 15 mmol) in 30 mL of anhydrous THF was added dropwise to a suspension of NaH (60% dispersion in mineral oil) (0.78 g, 19.5 mmol) in 15 mL of anhydrous THF at 0 °C. This mixture was stirred for a further 1 h at 0 °C, and then, a solution of 2-ethoxyethyl 4-methylbenzenesulfonate (3.6 g, 15 mmol) in 15 mL of THF was added dropwise. This mixture was allowed to warm to rt for 1 h and then heated to reflux for 12 h. The reaction mixture was cooled to rt and filtered, and all volatile materials were removed by rotary evaporation. The yellow oil was dissolved in toluene (25 mL), and the organic layer was extracted with water. The aqueous layer was extracted with dichloromethane, and the combined organic layer was dried with Na 2 SO 4 ; the solvent was then removed by rotary evaporation. The yellow oil obtained was purified by column chromatography (silica gel, dichloromethane/ethyl acetate 3:1 to 1:1) to give the desired compound 1 as a light yellow oil (2.4 g, 72%).
(b) Synthesis of the tosylate of tetraethylene glycol monoethyl ether 2: Sodium hydroxide (0.16 g, 4 mmol) dissolved in water (1 mL) and compound 1 (0.6 g, 2.7 mmol) in THF (2 mL) were placed in a three-necked, 25 mL round-bottom flask. The mixture was cooled on an ice bath. p-Toluenesulfonyl chloride (0.48 g, 2.6 mmol) in THF (3 mL) was added dropwise to the mixture. The solution was stirred at 0 °C for an additional 3 h and then poured into ice-water (20 mL) and extracted with ethyl acetate. The organic layer was washed with water and dried over Na 2 SO 4 . The solvent was evaporated in vacuo. The crude product obtained was used directly in the next step (0.8 g, 82%).
(c) Synthesis of 'tegylated' benzaldehyde 3: A three-necked, 250 mL round-bottom flask was charged with p-hydroxybenzaldehyde (0.24 g, 1.93 mmol), compound 2 (0.8 g, 2.12 mmol), K 2 CO 3 (0.8 g, 5.8 mmol) and DMF (10 mL). The reaction mixture was stirred overnight at 90 °C. After cooling, the crude reaction mixture was poured into water (100 mL, pH = 2) and extracted with ethyl acetate. The organic layer was washed subsequently with water (3 x 25 mL) and brine (1 x 25 mL) and dried over Na 2 SO 4 . The solvent was evaporated in vacuo. The crude light yellow oil 3 was pure enough to be used directly in the next step (0.49 g, 78%). 1

Fig. S1
GIWAXS patters for a PPEG-1 film (n = 5) freshly prepared and for the same film aged for 1 week. These patterns were acquired using a lab X-ray instrument.

Table S2
Relative Crystallinity (RC) for the C 60 -EG thin films as determined by the analysis of the 001 GIWAXS reflection. RC was deduced from I(χ) polar plots at 001 peak, χ being the polar angle, and the area below the I(χ) × sin(χ) versus χ plots (A c ) were obtained by fullintegration between 0 and 90°. 1 The one of PDEG-1 is normalized to 1.

Computational Methods
The following 4 MD simulation steps were carried out in series on starting configurations with unit cell dimensions of 2 x 2 x 4.5 nm 3 : 1) 2000 K, flat-bottom restraint, 2 ns; 2) 2000 K, 2 ns; 3) gradual cooling from 2000 to 298 K over 2 ns; 4) 298 K, 2 ns. This leads to a total of 8 ns of MD simulations (2 ns per step). In the case of the body centered unit cell: 1) A snapshot taken from previous step 3 at 1000K and the unit cell is doubled in c direction; 2) annealed at 1000K for 2 ns; 3) gradually cooled to 298K over 2 ns; 4) relaxed at 298K for 2 ns. Due to the small size of the unit cell, shorter cutoffs were used for LJ and Coulomb interactions (0.45 nm). However, particle mesh Ewald (PME) method was used for both LJ and Coulomb interactions to account for interactions beyond this cutoff. 2 Weak coupling schemes were used for both temperature and pressure. 3 The pressure was maintained at 1 bar anisotropically for the three lattice cell parameters, a, b, and c with a compressibility of 5x10 -6 bar -1 . Coupling parameters were 1.0 and 0.5 ps for temperature and pressure, respectively. The flat-bottom potential during the first step kept the two C 60 moieties at the top and bottom of the unit cell with respect to the c axis based in order to keep the c-axis spacing the longer one, in accordance with experimental X-ray measurements. The MD simulations were performed using the GROMACS 2018.5 software package. 4 Following the protocol above, 720, 360, 360, and 720 independent MD simulations were run on PDEG-1, PTEG-1, PTeEG-1, and PPEG-1, respectively. Results presented are either distribution or the mean of these independent runs, for each molecule. Interlayer spacing distributions for the C 60 -EG series are given in Figure S3. For the simulated scattering, Gaussian broadening with standard deviations of 0.01 Ang -1 and 0.03 Ang -1 were used in figures corresponding to XRD and GIWAXS, respectively. Files to reproduce the computational work are provided as part of the SI files.