Redispersion and Self-Assembly of C60 Fullerene in Water and Toluene.

This work aims at assessing the influence of two different solvents, bidistilled water and toluene, on dispersions of carbon-based engineered nanomaterials, namely, fullerenes, and their self-assembly behavior. The obtained self-assembled carbon-based materials were characterized using UV-vis spectrophotometry and transmission electron microscopy techniques. The results obtained were unexpected when toluene was used for dispersing fullerene C60, with the formation of two different types of self-assembled structures: fullerene C60 nanowhiskers (FNWs) and a type of quasispherical nanostructure. The FNWs ranged between 1 and 6 μm in length, whereas the quasispherical fullerene C60 nanoaggregates ranged between 10 and 50 nm in diameter. Aggregates obtained in toluene showed a well-formed crystal structure. When using water, the obtained aggregates were amorphous and showed a no well-defined shape. Their sizes ranged between 20 and 40 nm for nanosized structures and between 0.4 and 4.8 μm for micron-sized self-aggregates.

Introduction

Engineered carbon-based nanomaterials are being used in a large number of biomedical applications, such as drug-delivery systems,[1] cancer therapy,[2] and theranostics, as well as in electronic devices. It is well known that buckminsterfullerene or fullerene C60 is a truncated icosahedron (I) formed by sp2-hybridized atoms of carbon. Its peculiar external shape is due to a combination of 20 hexagons and 12 pentagons.

Allotropic forms of carbon show significant properties, such as optoelectrical, optoelectronic, and redox properties; effective n-doping of organic semiconductors; and high conductivity,[3] which make of them one of the most promising elements in chemistry. Self-assembled nanostructures show electron-transfer improvements when compared with planar electron acceptors. This property allows the development of organic electron devices (e.g., field-effect transistors and solar cells)[4,5] and optoelectronic devices.

Different syntheses approaches of self-assembled fullerene derivatives have been reported, such as supramolecular approaches,[6,7] controlled precipitation,[8−12] template methods,[13] and photoassisted growth.[11] In this article, we report on a method for obtaining of fullerene C60 self-assembled nanostructures that avoids liquid–liquid interface precipitation (LLIP) and several other issues associated with previously reported preparation routes.[14] Instead of using isopropyl alcohol as a solvent in the LLIP method[9] or modifications for faster preparation, as reported by Jin et al.,[15] we chose a simple method based on self-assembly. This new method yielded a bimodal population of carbon-based materials and finally enabled us to assess the aggregation behavior in two solvents. Our aim in this work was to explore the self-assembly behavior as a function of solvent redispersion nature (solubility) and the structural implications in the final products.

Supramolecular assemblies are defined as chemical structures organized by means of noncovalent bonds, for instance, hydrogen bonds,[16,17] π–π interactions,[18−20] van der Waals interactions,[21,22] and coordination bonds.[23,24] Our supramolecular self-aggregates involve quasispherical aggregates of fullerene C60 nanoparticles (FC60 NPs), with a diameter ranging from 10 to 50 nm; on the other hand, we also found fullerene C60 nanowhiskers (FNWs),a with a length ranging from 1 to 6 μm.

Results and Discussion

Typical transmission electron microscopy (TEM) low-magnification images of the samples obtained by redispersing C60 in bidistilled water are shown in Figure . It appears that in this dispersion medium C60 fullerenes form amorphous aggregates. Additionally, spontaneous production of a slight brownish yellow coloration was observed in the aqueous solution of fullerene C60.

Figure 1

TEM images of fullerene C60 redispersed in bidistilled water. Scale bar: (a) 1 μm, (b) 100 nm, and (c) 50 nm.

A large variety of particles sizes and shapes were identified. The particles show irregular edges and differences in contrast due to different thicknesses. Despite the small size of the fullerene seeds (1–2 nm in diameter), a mixture of large (micron-sized) and small (20–40 nm) structures was obtained. The internal structure of the aggregates was investigated by electron diffraction, and the obtained data showed a diffraction pattern that corresponds to that of an amorphous structure. This result clearly points toward aggregates that have a disordered internal structure.

Quite a different rearrangement was observed when C60 fullerenes were redispersed in toluene. This different behavior was confirmed and characterized using UV–vis spectroscopy, TEM, high-resolution electron microscopy (HRTEM), and electron energy-loss spectroscopy (EELS). In this case, the aggregates of fullerene C60 showed a crystalline face-centered cubic (fcc)[27] structure, forming two different types of aggregates, namely, FC60 NPs and FNWs.

Figure shows TEM images of the self-assembled nano- and micro-structures obtained on redispersing fullerene C60 in toluene. Clearly, the self-assembled nanostructures, including FC60 NPs (20–40 nm) and FNWs (0.4–4.8 μm), have a bimodal shape and size. FNWs have often been obtained using the LLIP method; from now on, we will report on the different behaviors as a function of solvent dispersion nature.[28]

Figure 2

TEM images of fullerene C60 redispersed in toluene forming (a) FC60 NPs and (b) FNWs, and (c) selected area electron diffraction (SAED) pattern obtained from a single nanowhisker.

UV–vis absorbance spectra of FNWs at several concentrations in toluene are shown in Figure . As can be seen, we found a small peak near λAbs = 407 nm, but the maximum absorbance intensity was observed at about 335–336 nm; this is different from previously reported data on pure fullerene C60 redispersed in several solvents, like hexane, wherein a maximum λAbs near 329 nm was reported.[29] The two peaks found in the UV–vis absorption spectra were assigned to (a) the effects exerted by the environment on fullerene C60 molecules (λAbs = 335–336 nm) and (b) the vibronic structure and effects of the surrounding environment on fullerene C60 molecules (λAbs = 407 nm).[30] The UV–vis spectra obtained with water as the solvent showed maximum peaks at 265 and 345 nm, as previously reported by Scharff et al.[31] We associated this small shift in the spectra, when compared to those with toluene, to the hydration of fullerene C60 by water molecules. The heating caused by sonication in small areas of the initial fullerene C60–toluene solution can induce a change in the small amount of fullerene C60 seeds to a point at which the polymerization process starts between neighboring molecules. Following sonication at room temperature, enhanced by sunlight (UV–vis radiation), the formation of crystalline FNWs and FC60 NPs was confirmed by optical and transmission microscopy observations, consistent with the findings of previous works.[32] Aggregation of the fullerene C60 nanoparticles continued, which finally rendered a nanowhisker structure; upon storing the C60 nanoparticles for 2 weeks at 2 °C, we obtained FNWs as large as 60.13 μm, as measured by electron microscopy.

Figure 3

UV–vis normalized absorbance spectra of fullerene C60 in toluene at several concentrations (10–50 μL aliquots in 2 mL of toluene). Inset: Enhanced image of the peak found at 407 nm. Naked-eye observation of C60 redispersed in toluene.

HRTEM images showing a core–shell structure of FNWs are shown in Figure a. In this image, the growth axis of the core (Figure b) was identified to be in the ⟨011⟩ crystalline direction; the corresponding Fourier-transform (FT) displaying the [011] zone axis of the cubic fcc crystalline structure of the FNW core is shown in Figure c.

Figure 4

(a) HRTEM image showing the core–shell structure of the FNWs. (b) HRTEM lattice image showing the structure of the core obtained from the area depicted in (a). (c) FT of (b) with the indexes of the spots calculated on the basis of the fcc crystalline structure. (d, e) Growth axes of the core and shell, respectively. (f) FT obtained from (e), indexed on the basis of the body-centered tetragonal (bct) crystalline structure.

However, analyses of the FT obtained from the shell area confirmed the [010] zone axis, which established the ⟨200⟩ crystalline direction as the growth axis of the shell parallel to the larger axis of the nanowhiskers and confirmed the presence of a bct tetragonal structure on the shell. Coincidentally, as was previously reported by Miyazawa et al.,[33] the same growth axis, ⟨200⟩, was determined previously in the bct tetragonal structure of the FNWs. In the present case, HRTEM images also show clear differences between the inner structure (core cubic crystalline structure) and shell (tetragonal crystalline structure). HRTEM images and FT also confirmed a core@shell structure formed by two different crystalline structures, with planes and growth axes different in both the core and shell. These differences were also previously observed in nanowhiskers by Miyazawa[34] and have been reported in the LLIP method of FNWs synthesis. We also found a difference of 0.5% d in the Miller indices when compared to previously reported data on pristine fullerene C60 crystals.[35]

The influence of solvent[36] on the redispersion and coassembly of fullerene[37] has been previously reported, and its significance has also been underlined,[38] but here, we intend to stress on the influence of electronic interactions on the assembly process. EELS allowed us to determine the nature of C–C bonds present in the internal structure of C60 fullerenes found in both FC60 NPs and FNWs. As this bimodal population was formed as consequence two different self-assembly pathways, it was important to be sure that these two different types of particles had the same electronic configuration, as evidenced by the indexing of SAED. Results are shown in Figure , and they suggest that the bonds established between C60 fullerenes as building units are responsible for the formation of both aggregates.

Figure 5

Electron energy-loss spectra of self-aggregated FC60 NPs (black solid circles) and FNWs (blue solid circles).

Basically, the presence of shoulders with the same shape in EELS confirmed the existence of σ and π bonds in the internal structure of the FC60 NPs and in the FNWs.[39] In view of this, it is also clear that the structure of individual fullerene molecules remained unchanged within a nanowhisker.

Conclusions

In this work, we studied the influence of solvent in a simple one-step method to redisperse the fullerene C60 avoiding the use of a polymeric stabilizer, by means of self-assembly of these carbon-based nanomaterials. In toluene, we obtained two different types of crystalline structures: FNWs, having a length of 1–6 μm, and FC60 NPs, with a diameter of 10–50 nm. In water, we obtained two amorphous populations, one with a size of 20–40 nm and the second with a size of 0.4–4.8 μm, which are larger than the former. Both optical and structural characterizations confirmed the formation of these self-assembled structures. The formation of these shape allotropes of carbon can be due to steric and electrostatic rearrangements of the fullerene C60 involved in the redispersion as a function of solvent nature. Previous work of Hughes et al. underlined that C60 shows very low solubility in water, whereas the solubility of C60 in toluene is 2.8 mg/mL, which is an important factor to take into account in our experiments and in further studies.[40]

Experimental Section

Reagents

C60 fullerene was purchased from Sigma-Aldrich at the maximum commercially available purity of 99.5% (Germany),[41] with no further treatment. Toluene was commercially obtained from Panreac (Spain) (99.9%). Bidistilled water showing 0.10–0.50 μS cm–1 conductance was used for synthesis.

Carbon-Based Nanomaterial Redispersion: Synthesis of FNWs

In a one-step simple ultrasound-assisted synthesis method, we used bidistilled water and toluene to test the influence of the nature of two solvents on the redispersion of carbon-based nanomaterials.[42,43] Briefly, different amounts of C60 fullerene (0.5–40 mg mL–1/6.94 × 10–4–5.55 × 10–2 M) were redispersed as saturated solutions of engineered nanomaterials in bidistilled water and toluene. We did not perform any further treatment; no filtration of the stock solutions was carried out; and we did not use any other polymer or surfactant to stabilize the suspension or obtain pristine carbon nanomaterials. Using a Branson 3510 ultrasonic bath, frequency 40 kHz, redispersion of the C60 fullerene was carried out over 5 min at 298 K, and then, the solutions were stored at room temperature until morphological observations were carried out. First, all fullerene C60 samples in water and toluene were exposed to ultrasonication at room temperature (298 K) under sunlight (it is known that the crystalline structures of fullerene C60 trend to grow under UV–vis radiation). These samples were stored at room temperature until morphological measurements by HRTEM were carried out (ca. 1 h). Then, the samples were stored at 2 °C and light-protected for 3 weeks until morphological characterization was carried out again. Samples for UV–vis absorbance spectroscopy were freshly prepared using the same ultrasound bath. Characterization was carried out in triplicate for every measurement to confer reproducibility of the experiments. TEM analyses were performed using a JEOL JEM1010 TEM operating at an accelerating voltage of 100 kV, whereas HRTEM was performed using a JEOL JEM2010F HRTEM operating at 200 kV. EEL spectra were recorded following standard and well-known methods. In brief, measurements were carried out at an operation voltage of 200 kV in the STEM mode using a Gatan GIF Quantum spectrometer, with an energy resolution of 1.75 eV (FWHM zero loss peak), 0.5 eV/channel energy dispersion, and an EELS collection semiangle of 16 mrad. To avoid the contribution of the grid carbon foil, the EELS spectra were collected from areas of the sample situated over a hole. The background EEL spectra were subtracted using standard routines of Digital Micrograph. Samples for TEM and HRTEM were prepared by dropping and evaporating the obtained colloid on top of a carbon-coated copper grid.

The authors emphasize that the sonication time was selected on the basis of macroscopic observations. The observation of a change in the turbidity and color of the solution after sonication for 5 min suggested a change in the optical properties of the solution; this was the reason behind exploring the optical properties and morphology. Further experiments (either kinetically or thermodynamically controlled) are necessary to clarify and explain the formation process of the above-mentioned structures in detail.