Kinetically stabilized C60–toluene solvate nanostructures with a discrete absorption edge enabling supramolecular topotactic molecular exchange

Kinetically stabilized C60–toluene solvate nanostructures with a discrete absorption edge enabling supramolecular topotactic molecular exchange

CARBON 7 4 ( 2 0 1 4 ) 4 4 –5 3 Available at www.sciencedirect.com ScienceDirect journal homepage: www.elsevier.com/locate/carbon Kinetically stab...

2MB Sizes 0 Downloads 8 Views

CARBON

7 4 ( 2 0 1 4 ) 4 4 –5 3

Available at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/carbon

Kinetically stabilized C60–toluene solvate nanostructures with a discrete absorption edge enabling supramolecular topotactic molecular exchange Moumita Rana, R.R. Bharathanatha, Ujjal K. Gautam

*

Nanomaterials and Energy Laboratory, New Chemistry Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur P.O., Bangalore 560064, India

A R T I C L E I N F O

A B S T R A C T

Article history:

Nanosized fullerene solvates have attracted widespread research attention due to recent

Received 9 December 2013

interesting discoveries. A particular type of solvate is limited to a fixed number of solvents

Accepted 1 March 2014

and designing new solvates within the same family is a fundamental challenge. Here we

Available online 12 March 2014

demonstrate that the hexagonal closed packed (HCP) phase of C60 solvates, formed with m-xylene, can also be stabilized using toluene. Contrary to the notion on their instability, these can be stabilized from minutes up to months by tuning the occupancy of solvent molecules. Due to high stability, we could record their absorption edge, and measure excitonic life-time, which has not been reported for any C60 solvate. Despite being solid, absorbance spectrum of the solvates is similar in appearance to that of C60 in solution. A new absorption band appears at 673 nm. The fluorescence lifetime at 760 nm is 1.2 ns, suggesting an excited state unaffected by solvent–C60 interaction. Finally, we utilized the unstable set of HCP solvates to exchange with a second solvent by a topotactic exchange mechanism, which rendered near permanent stability to the otherwise few minutes stable solvates. This is also the first example of topotactic exchange in supramolecular crystal, which is widely known in ionic solids. Ó 2014 Elsevier Ltd. All rights reserved.

1.

Introduction

The fullerene (C60) solvates and their nanostructures are fast emerging as a new class of functional materials due to their interesting mechanical and transport properties [1–5]. The ease of solution processibility has lead to diverse forms of C60-solvate nanostructures [6–11], which can easily release the solvent molecules to yield highly crystalline C60 nanostructures [12]. The possibility of spin-coherent transport has been recently demonstrated [13]. Among the different

* Corresponding author. E-mail address: [email protected] (U.K. Gautam). http://dx.doi.org/10.1016/j.carbon.2014.03.001 0008-6223/Ó 2014 Elsevier Ltd. All rights reserved.

solvates, hexagonal close packed (HCP) phase is the most interesting one because of its ability to form nanowires effortlessly and exhibit high conductivity as well as stability under pressure [14]. Wang et al. have discovered that some HCP solvates become ultra-incompressible under high pressure with an ability to indent diamond [4]. The nanowires of HCP C60. m-xylene exhibit an electron mobility of 11 cm2 V1 s1 [15], which is much higher than that of 0.1 cm2 V1 s1 in pristine C60 and comparable to common high-mobility organic semiconductors [15,16]. It has also been reported that solvent

CARBON

7 4 (2 0 14 ) 4 4–53

incorporation enhances the excitonic recombination rates and luminescence in HCP solvates manifold, although investigation on their optical properties is quite limited [14]. Thus it is not known how the nature and the origin of absorption processes as well as excitonic carrier lifetime change when the inter C60 interactions are modified by the incorporated solvent molecules. The exploration of these interesting properties is particularly hindered due to the lack of a variety of solvates within the HCP family. Only few mono or disubstituted benzene solvents such as m-xylene, m-dichlorobenzene, etc. form stable HCP solvates [17]. The C–H  p interactions between the two –CH3 groups of xylene and C60 probably stabilize the molecular arrangement in the crystal [18]. m-dichlorobenzene also forms HCP solvates due to the chloro-methyl exchange rule [19]. However, designing a new solvate using a different solvent system is quite challenging due to the complicated nature of their formation, which depends on a number of weak interactions involving several molecules in solution and also due to lack of understanding on conformity of the interacting groups inside a crystal. Examination of the crystal structure of m-xylene-C60 HCP solvate reveals that the solvent molecules occupy the cavities having a threefold symmetry [20]. The methyl groups of m-xylene are entirely disordered in three allowed directions about the symmetry axis, which leads us to hypothesize that toluene as well as mesitylene too should form these important HCP solvates. Toluene, interestingly is known to form monoclinic solvates while mesitylene tends to forms orthorhombic solvates [6,21]. Their HCP solvates have not been found, except for one instance, wherein a transiently stable toluene compound was observed [22]. Taking all these into account, there is not only a possibility to obtain more HCP solvates, but also an opportunity to investigate their stability, which is crucial for their existence. While the exploration of new HCP solvates and their properties is necessary, an interesting strategy to obtain C60 cosolvate can be envisioned while pondering upon the instability of the toluene solvates. Unstable solvates are usually stored in the mother liquor, since it suppresses their decomposition due to a dynamic equilibrium between the solvate and the mother liquid [23]. Such equilibrium leads to anomalous solubility of C60 in toluene [24,25]. The equilibrium constant will be larger for less stable solvates. Therefore, if a batch of unstable toluene solvates is placed in surrogate mother liquor which contains structurally similar second solvent, there exists a distinct possibility of replacement of some solvent molecules from the initial solvate crystal with the second solvent. The resulting cosolvate would therefore have two different solvent molecules occupying an identical crystallographic position within the crystal. To the best of our knowledge, such cosolvates are not known in the case of C60 and due to phase separation cannot be obtained by conventional precipitation methods due to phase separation. Being motivated by the recent interesting findings on the HCP C60 solvates, and the rare possibility to look into their stability, we have systematically investigated the growth and properties of C60 toluene solvates. Further, we have explored their optoelectronic properties, such as absorbance, photoluminescence (PL) and excitonic life-time. Thus, from carefully controlled reactions, we show that their stability may be

45

improved from a few minutes up to a period of over months, which is as good as the other robust solvates. The occupancy of the solvent molecules in the solvate crystal and stoichiometry were found to be the key to their stability. The absorbance spectrum is similar in appearance to that of C60 in solution, while the PL spectrum closely matches that of pristine C60 powder. The absorption bands appear at energies intermediate to that of C60 in solution and in solid. A new absorption band appeared at 1.68 eV. The fluorescence lifetime measurements reveal that lifetime of the exciton is 1.2 ns, which is similar to pure C60 [26]. The observations suggest that the solvent–C60 interactions can influence the electronic transitions related to the ground state only, and not of the excited state. To the best of our knowledge, this is the first instance where a fluorescence lifetime and absorbance of any C60 solvate-crystal is investigated. Finally, we took advantage of the labile nature of the toluene molecules in the unstable set of solvate-nanostructures and carried out topotactic exchange reactions with another suitable solvent. This, for the first time, has led to the C60-cosolvate nanostructures wherein two types of solvent molecules occupy crystallographically identical positions. The formation of cosolvate has rendered near permanent stability to the otherwise few minute stable toluene-solvates. Notably, this would also be the first example to exhibit molecular topotactic exchange, i.e. exchange of a moiety by another in a crystalline system maintaining the surrounding exactly the same, just like the well-known ionic compounds [27] exhibiting the same, which is unlikely in supramolecular systems [28,29].

2.

Experimental section

2.1.

Preparation of the toluene solvates

Fullerene (C60, 99.5%), toluene (99.8%), isopropanol (IPA, P99.5%) were purchased from Sigma–Aldrich and were used without further purification. C60 was dissolved in toluene by maintaining concentrations of 0.45, 0.9, 1.8 mg/ml for different sets of reactions. The 1D solvate-nanocrystals were obtained at 5 °C by two methods (i) liquid–liquid interfacial precipitation (LLIP) [30] and (ii) ultrasonic liquid–liquid interfacial precipitation (ULLIP). Prior to the mixing, C60 solution and IPA were kept inside an incubator set to a desired temperature for 1 hour for thermal equilibration. In case of LLIP method, 8 ml IPA was added to 2 ml C60 solution with a rate of 2.5 ml/minute and kept at the predetermined temperature until complete precipitation. In the ULLIP method, The C60 solution in toluene and the IPA were mixed using a bath sonicator which was filled with water pre-set to a desired temperature. The reaction schemes are shown in Fig. S1.

2.2.

Characterization

The samples were characterized by field emission scanning electron microscopy (FESEM), transmission electron microscopy (TEM), powder X-ray diffraction (PXRD), fourier transform infrared spectroscopy (FT-IR), raman spectroscopy, thermogravimetric analysis (TGA), ultraviolet-visible (UV-Vis) spectroscopy, photoluminescence spectroscopy (PL).

46

CARBON

2.3.

7 4 ( 2 0 1 4 ) 4 4 –5 3

Topotactic solvent exchange

For topotactic exchange, the solvate structures were first separated from the mother liquor and exposed to a mesitylene and IPA mixture with volume ratio (1:4), which is already saturated with dissolved C60 (so that possibility of dissolution of C60 from the solvate can be excluded). This was kept inside the incubator for 1 h at 25 °C. The products were taken out and used for further characterizations. See Supporting information for experimental details.

3.

Results and discussion

3.1.

C60–toluene solvates and nanostructures

The precipitation of C60–toluene solvate nanowires is accompanied by a complete decoloration of the solvent mixture. We observed that 2–6 days are required to form the nanowires, where higher concentration leads to faster rates. Under all conditions, the solvates are generated in a one dimensional (1D) wire-like form. In Fig. 1a, we show an FESEM image of the C60 solvates obtained after 3 days by using 1.8 mg/ml of C60 at 5 °C. As seen, the elongated and straight solvate wires have blunt tips with a diameter of 0.9–1.5 lm and high aspect ratios of 450. The morphology of the solvates was greatly influenced by the concentration of the precursor C60 and hence the kinetics. Their length was marginally reduced when 0.9 mg/ml of C60 was used (Fig. 1b). When we further reduced the concentration to 0.45 mg/ml, the reaction was

much slower and the precipitation completed only after 6 days, yielding two distinct types of microrods. About one third of these rods, which were found near to the interface, have much larger diameter and uniquely appeared tubular on one side and a sharp edge on the other side (Fig. 1c). The others looked simple rod-like, with well formed hexagonal edges (Fig. S2). In Fig. 1d, we show the average diameters and aspect ratios for these products, establishing that the effect of concentration on diameter is more pronounced than on length. Diameter concurrently reduces with the increase in C60 concentrations, reaching 600 nm at the highest employed C60 concentrations. Precipitation from toluene is generally employed to obtain pure C60 structures in face centered cubic (FCC) phase. On the other hand, toluene forms monoclinic C60 solvates when prepared by evaporation techniques. We employed powder XRD technique to examine the purity and crystal phase of the different C60-assemblies. All 1D structures obtained by us yielded the same diffraction pattern, as shown in Fig. 2a, which is different from the usual FCC phase of the pristine C60 and the monoclinic toluene solvates [6,31]. Instead, the diffraction patterns of our solvates match the HCP phase of the stable m-xylene-C60 solvates with a space group of P63 [14,20,22] (Fig. S3). Fig. 2b and c displays the crystal structure of the C60-m-xylene solvate. Fullerenes with mirror symmetry are arranged in hexagonal layers along the c axis, creating two distinct cites to be occupied by the solvent molecules. A thermally disordered solvent molecule occupies trigonal

(a)

(a)

(b)

50 µm

200 µm

(d)

200 µm

20

0

15

150

10

300

5

450

0

0.45

0.9

1.8

(c)

Aspect Ratio

Diameter (µm)

(c)

(b)

600

C60 conc. in Toluene (mg/ml) Fig. 1 – FESEM images of 1D C60 solvates synthesized from LLIP method using toluene/IPA system with C60 concentrations in toluene (a) 1.8 mg/ml, (b) 0.9 mg/ml, (c) 0.45 mg/ml. Inset images show the characteristic features of the tip of the corresponding 1D structures. (d) a plot showing distribution of diameter and aspect ratio of 1D structures obtained with different C60 concentration in toluene. (A colour version of this figure can be viewed online.)

Fig. 2 – (a) PXRD pattern of the 1D solvates synthesized by LLIP method. The vertical gray lines at the bottom correspond to the diffraction peak positions simulated from C60-m-xylene solvate single crystal data (b) the crystal structure of C60-m-xylene solvate along C axis: the solvent molecules with static disorder [20] are represented by blue and red colours posses site symmetry of C3 and 63 respectively. (c) A 3-dimensional view of the 63 solvent site and orientation of C60 molecules around it. Here, red and green colors correspond to the position of molecules at Z = and respectively. (A colour version of this figure can be viewed online.)

CARBON

prismatic cavities (site 1, Fig. 2b: blue colored molecules) between layers with a 3-fold symmetry while heavily disordered solvent molecules reside along a 63 screw axis (site 2, Fig. 2b: red1 coloured molecules) [20]. The orientation of C60 and solvent molecules around 63 axis is shown in Fig. 2c. Comparison of the XRD pattern of as synthesized solvate and simulated pattern based on the m-xylene solvate structure is shown in Fig. 2a. Thus all the diffraction peaks in our PXRD pattern can be indexed on a space group P63 with lattice parameters of a = 2.34 nm and c = 1.00 nm. The presence of toluene in the nanowires was confirmed by using IR spectroscopy and mass spectroscopy analysis (Fig. S4). In order to investigate their growth direction, detailed transmission electron microscopy analysis of the solvate nanowires was performed. Fig. 3a shows a typical TEM image of a nanowire. All nanowires have uniform diameter and density throughout, as inferred from uniform color contrasts. The SAED pattern (Fig. 3c) recorded at the middle of the structure reveals the single-crystalline nature of the wire. All diffraction spots can be indexed on a HCP unit cell. A high-resolution energy filtered TEM image of the structure is shown in Fig. 3b, revealing the (1 0 0), (0 0 1), (1 0 1) planes with lattice ˚ respectively. Based on SAED spacing of 20.26, 10.28, and 9.2 A and HRTEM analysis of various nanowires, we determined their growth direction to be along the (0 0 1). A model of a nanowire is shown in Fig. 3d. The toluene molecules in site 1 as well as in site 2 stand perpendicular to the growth direction forming 1D channels along the nanowires. The success in the synthesis of HCP C60 solvate-crystal is dependent on the diffusion dynamics and the solvation properties of C60 in solution, which is influenced by the reaction temperature. When we carried out the reaction at 25 °C, C60 precipitated in the pristine FCC crystal-structure with octahedron morphology (Fig. S5a and Sb). At 15 °C, the precipitated product contained a mixture of both FCC and HCP phases (Fig. S6). Pure solvate with an HCP phase was obtained only at 5 °C. The formation of such supramolecular assemblies is a complex process with delicate balance of diverse weak intermolecular forces. Each C60 is solvated by the toluene molecules in solution due to the week polar interactions [32]. However, when IPA is added to C60 solution, this molecular network changes due to homogenous mixing and induces precipitation of C60. Each C60 molecule must associate itself with the solvent molecules prior to precipitation, the nature of which determines the precipitated product. We believe that an adduct of C60–toluene pair is the possible precursor building block for the HCP solvate phase, while individual C60 molecules nucleate to the FCC phase. Since these weak interactions are temperature sensitive and stable only at lower temperatures, the solvates too form at lower temperatures. The binding of the toluene molecules to C60 in the solvate nanowires is comparable to the isostructural m-xylene solvates [14]. From TG analysis (Fig. 4a), we found the solvates are stable up to 58 °C and decompose thereafter with the release of toluene. The C60 to toluene ratio is estimated from a total weight loss of 9.3% at complete decomposition to be 1:0.83, which is marginally substoichoimetric from an optimum value

1

47

7 4 (2 0 14 ) 4 4–53

(a)

0.5 µm

(c)

(b) 10.28 Å (001)

9.2Å (101)

200 301

100

101

5nm

20.26 Å (100)

001

10-1

-101

-10-1

-200

00-1

a

(d)

o c b Fig. 3 – (a) Low resolution TEM image of C60 nanorod synthesized by LLIP method. (b) Energy filtered HRTEM image of the selected region of the nanorod as shown in 4a. The growth direction was determined to be (001) (c) SAED pattern of the nanorod. (d) A model of the nanowire showing the molecular orientation of C60 (blue) and solvent molecules (red and green) along the growth direction (001). The red and green colours correspond to the 1D solvent channels with 63 and C3 site symmetry respectively. (A colour version of this figure can be viewed online.)

of 1. The differential Friedman isoconversional analysis (Eq. (1)) was employed to evaluate the activation energies associated with the decomposition of the solvates [33].   da Ea ln ð1Þ ¼C dt RTa where a is the extent of decomposition, (da/dt) stands for rate of decomposition, constant C is related to the frequency factor or the stability of the nascent nuclei and Ea represents activation energy of decomposition. Since decomposition is an exothermic process and mechanisms involved at different stages of decomposition may be different [34], we used a temperature region around the decomposition onset (initial 1.5% decomposition) to evaluate the corresponding activation energy. Fig. 4b is a plot of ln(da/dt) vs. reciprocal T and from linear fit to these plots, Ea for solvate decomposition onset have been derived as 188.2 kJ/mole.

3.2.

Formation of HCP solvate nanorods

We could drastically reduce the size of the solvate nanowires and yet maintain the 1D morphology by inducing ultrasonic

For interpretation of color in Fig. 2, the reader is referred to the web version of this article.

48

CARBON

7 4 ( 2 0 1 4 ) 4 4 –5 3

(a)

(b)

Fig. 4 – (a) A plot consisting weight loss and extent of decomposition vs. temperature for the sample synthesized by LLIP method using C60 concentration of 0.45 mg/ml. (b) A Friedmann plot of corresponding TGA data considering the initial weight loss (highlighted area in 6a) that was used to estimate activation energy for decomposition. (A colour version of this figure can be viewed online.)

(a)

Fig. 5 – FESEM image of the HCP solvate nanorods obtained by ULLIP method using C60 concentration of 0.9 mg/ml. The inset images show well defined hexagonal tips of the nanorods. (A colour version of this figure can be viewed online.)

assisted nucleation as well as fast diffusion of solvent and antisolvent. Fig. 5 shows an FESEM image of the solvate nanorods obtained by ultrasonication of 2 ml 0.9 mg/ml of C60 in toluene with 8 ml of IPA at 5 °C for 5 min. The rods are extremely uniform in size, measuring about 600 nm in diameter, 10 lm in length, and hexagonal cross-sections. XRD and TEM images reveal that the nanorods are single-crystalline in nature (see Supporting information Fig. S7). We have investigated the effect of reactant concentrations on this synthesis and found that the changes in the shape and size of the nanorods are minimal. Moreover, as will be discussed later in Section 3.5, these structures are relatively unstable as compared to the ones obtained by LLIP method, which is important for the observation of the topotactic exchange phenomena.

3.3.

Stability of the solvates-nanostructures

These C60 solvate structures offer a suitable system to investigate the nature of solvate stability, since the nanowires are rather unstable and convert to pristine FCC phase in a time bound manner. This has been used to advantage to prepare

(b)

(c)

Fig. 6 – Time dependent PXRD patterns of the samples synthesized by LLIP method using C60 concentration in toluene (a) 1.8 mg/ml, (b) 0.45 mg/ml. The peak at 10.8° and 11.3° define the presence of FCC and HCP phases respectively. The decomposition of HCP phase to FCC can be clearly observed by the evolution of 10.8° peak. (c) TGA of the samples synthesized by LLIP method using C60 concentration 1.8 mg/ml (green), 0.45 mg/ml (red). (A colour version of this figure can be viewed online.)

various pure 1D C60 nanostructures [35]. Lesser vapor pressure of toluene was reported as the reason behind the

CARBON

instability of the solvates [22]. However this is not convincing and we now found that even in the case of toluene alone, stability may be tailored from a few minutes up to months. Fig. 6a shows the PXRD pattern of the C60–toluene solvate nanowires obtained by using a high concentration (1.8 mg/ ml) of toluene solutions. A small FCC peak at 10.8° appears ever after 30 minutes of the separation of the nanowires from the mother liquor. The FCC peak slowly grows and after 24 h, a considerable amount of the solvates converts to the pristine structure. When the same sample was checked after 10 days, the sample mostly consists of pristine C60. When the solvate structures were obtained by using a C60 solution of smaller concentration (0.45 mg/ml), the structures were extremely stable. Fig. 6b shows the time dependent evolution of the PXRD patterns and as can be seen, the solvates-wires were stable up to 1 month without the appearance of an FCC peak. The higher stability of the solvates obtained using low concentration of precursor is concurrent with the slow kinetics of their crystallization (2 days at 1.8 mg/ml and 6 days at 0.45 mg/ml). The incorporation of solvent molecules in the solvate-crystals is probably partial at a higher rate of crystallization. To compare the amounts of toluene present in both these samples, we have carried out TGA analysis just after their separation from mother liquid (Fig. 6c). It shows that even though both samples decompose around the same temperature, the sample obtained using 1.8 mg/ml toluene contains considerably less toluene molecules (8.4% weight loss, toluene: C60 = 0.74) than the sample obtained from 0.45 mg/ ml toluene (9.3% weight loss, toluene: C60 = 0.83). As seen in the image, the solvent molecules at site 1 are relatively static and disorder is expected only in the positioning of methyl substituents. The same molecules at site 2 are more labile. The decomposition of the solvate crystal is associated with the formation of nuclei of the pristine C60 FCC phase, which must commence with collapse of local crystal structure around a vacant toluene position. Therefore the occupancy of toluene molecules and their distribution in the two solvent sites in the crystal must be two crucial parameters for stability of the solvate nanostructures.

3.4.

49

7 4 (2 0 14 ) 4 4–53

edges in particular, which lie in the visible spectrum of light and potentially useful for chemical interactions has not been investigated so far [38–40]. Herein, we first consider these weak transitions which are forbidden due to the icosahedral symmetry of C60 and commence only due to the Herzberg– Teller (HT) vibronic coupling between the ground and excited electronic states [26]. These transitions are prominent only when C60 is in a solution, but feeble and structureless in a solid C60 crystal [36]. Remarkably in the case of our solvates, we found that while the OA spectra retain the features of a C60 solution, their PL properties are similar to that of solid C60. The OA spectrum of the as synthesized solvate nanorods in the range of 500–800 nm is seen in Fig. 7a. The spectrum exhibits a structured absorbance with an onset of absorption at 730 nm, followed by six well-resolved optical transitions. The C60 solution in toluene exhibits a similar OA spectrum, but is blue shifted by about 17 nm (Fig. 7a). In contrast, the OA spectrum of pristine C60 is relatively featureless and red-shifted to 751 nm and linearly increases until 600 nm. A comparison of these three OA spectra suggests that the absorbance of the solvate nanostructures is (i) somewhat equally but oppositely shifted from that of the pristine and C60 solutions and (ii) the vibronic coupling behavior of C60 molecules is similar in the solvates and in solution. We further recorded the Raman spectra of the solvates after the absorbance and

(a)

Optical properties

The C60 solvate crystals provide an interesting system to investigate optoelectronic properties because the C60-solvent interactions are substantially modified due to the loss of their translational motion, unlike in a solution. The optical properties of solid C60 are closely related though to its solutions due to the molecular nature of solids [36]. C60 molecules form a face centered cubic close packing system in pristine C60, interacting through van der Waals forces; whereas in solution, each C60 molecule is surrounded by solvent molecules alone. In this solvate structure, the distance between adjacent C60 molecules increases due to the presence of solvent molecules. Here the interaction between solvent and C60 molecule is weak C–H  p in addition to the van der Waals interaction between adjacent C60 layers. Hence, these changes in symmetry as well as interactions make the optical properties of the solvates system more interesting. Few studies were done to explore the optical absorption (OA) and photoluminescence (PL) properties of the solid solvates [14,37]. Their absorption

(b)

(c)

Fig. 7 – (a) Absorption spectra of C60 dissolved in toluene, C60–toluene solvate nanorods and pristine C60 in the wave length range 500 nm to 800 nm. (b) Raman spectra of pristine C60 (black) and C60 toluene solvate (red) after optical measurements, showing no change in the position of pentagonal pinch mode. (c) A schematic energy level diagram depicting the HT transitions in these samples. (A colour version of this figure can be viewed online.)

50

CARBON

7 4 ( 2 0 1 4 ) 4 4 –5 3

luminescence studies to ascertain that the observed shifts are intrinsic to the solvate and not due to photopolymerization of C60 (Fig. 7b) [41–43]. No shift of the Ag(2) pentagonal pinch mode at 1467 cm1 was observed, and the novel optical properties of the C60 solvates can therefore be ascribed to the C60-solvent interactions within the crystal [44]. Due to the similarity of the electronic structure of the solvate C60 and isolated C60 molecules in solution, we classify the energy levels of the solvate C60 by the same irreducible representations of the group Ih. The lowest electronic excited state of molecular C60 has an electronic con-figuration of h9uf11u and can be singlet or triplet in nature [45]. The triplet state has a lower energy and much lower transition probability and therefore transitions observed in the absorption edge involve S ! S transitions. The OA spectrum of the C60 solvate crystals exhibits six distinct transitions at 673, 642, 611, 588, 564 and 516 nm (star marked in Fig. 7a). In spite of the obvious shifts, this is quite similar to that of the OA spectra of C60 solution in toluene and solid C60 at low temperature [36]. We therefore index the peaks at 642 and 611 nm to symmetry equivalent transitions involving singlet 1Ag ground state and 1 F1g excited state (Fig. 7a). The peaks observed at 588 and 564 nm arise out of overtone and combination bands [46,47]. Interestingly, we also observe a prominent low energy electronic transition at 673 nm which is absent in the spectra of both in solid C60 as well as in solution phase. Such a peak has not been observed before for C60 related materials. Singlet–triplet transitions are expected to appear at lower energies as compared to singlet–singlet transitions. However they should be further red-shifted up to 681 nm at least and much lower in intensity unlike in the present case [48]. Since assignment of this peak is more complicated due to the degeneracy associated with a number of HT transitions and additional vibronic features may arise due to the restricted translational motion of the molecules in the crystal-solvate, further investigations are required to ascertain the origin of this new peak. We estimated the optical band-gap of the C60–toluene solvates from the absorbance spectrum using the Tauc equation [44]. pffiffiffiffiffiffiffiffi ahm ¼ Cðhv  Egap Þ ð2Þ where Egap corresponds to the Tauc gap of the material, hm is the photon energy, m is absorption coefficient and C is a constant [49]. As seen in Fig. S8, the toluene solvates exhibit a band-gap of 1.68 eV, which is 0.07 eV blue-shifted with respect to the precursor pristine C60 band-gap of 1.61 eV. The red-shift observed in both the OA spectrum of the solvates and pristine C60 in comparison to C60 solution (Fig. 7a) could be caused by the same factors. The shifts to lower energy in C60 crystal originate from the C60–C60 interactions due to close proximity, which result in broader electronic band structures and decrease in the band-gap [36]. A similar change in band-gap was recently shown from first principle investigation electronic structure calculations in the case of C60-cubane cocrystals also [50]. The nearest distances between two C60 centers ˚ in a FCC crystal and in our solvates are about 9.9 and 10.0 A respectively. The other inter C60 distances are comparatively larger in the solvate. Thus the lowest unoccupied molecular orbital energy level of C60–toluene solvate should lie within

that of solid C60 and isolated molecules in solution, as described schematically in Fig. 7c. Alternatively the solvates behave as an extended C60 crystal, moderately expanded by an interpenetrating matrix of solvent molecules. We further estimated the Urbach energy (U) for the solvates be 61.39 meV (see Supporting information). This is markedly larger than 37 meV at T = 293 K and comparable to 59 meV at 470 K observed for pristine C60 films [51]. This Urbach energies are also comparable to the Si:H phases [52]. Larger U indicates more disorder in the system which the extra solvent molecules may account for. A PL spectrum of the C60–toluene solvate acquired using an excitation wavelength of 600 nm is shown in Fig. 8a. We found that the spectra were indifferent to the excitation wavelength, and contain two major emission peaks centered at 707 and 763 nm. This is slightly blue shifted when compared with the PL spectra of pristine C60 crystals [53]. We first exclude the possibility of photopolymerization due to irradiation from Raman measurements. The emission band observed at 763 nm is attributed to zero-phonon line (ZPL), originated from the direct exciton polaron recombination from S1 (h9uf1u) to S0 (h10 u ) [26]. The new emission band centered at 707 nm is specific to the toluene solvate crystals and not reported for C60 crystals or C60-solutions. This emission possibly originates from a state unique to the toluene solvates because PL is associated with electrons in the LUMO level where Jahn– Teller distortions are non-negligible. The electronic transitions bands in UV and PL spectra of C60 are related by mirror symmetry [36]. Even though it is not clear at present, it is probable that the new absorbance band observed at 670 nm (Fig. 7a) and this PL band are related. In C60 based systems, such as supramolecular assemblies with Zn-porphyrin [54], the flourescence lifetime (s) are significantly altered due to the stronger interactions. Since the interaction between toluene and C60 is responsible for the existence of the HCP solvate phase, it may also influence excitonic recombination properties. In the absence of such report on any C60 solvate, we carried out s measurements using time resolved PL measurements near 760 nm at room temperature (26 °C) and compared with the conventional C60 based chromophores. Fig. 8b shows the decay profiles for the solvate material which was fitted by a single

(a)

(b)

Fig. 8 – (a) Photoluminescence spectrum and (b) lifetime decay plot for C60–toluene solvate nanorods. The experimental lifetime data can be fitted with a single exponential decay with time constant of 1.2 ns. (A colour version of this figure can be viewed online.)

CARBON

7 4 (2 0 14 ) 4 4–53

exponential decay with a constant of 1.2 ns. This is in good agreement with the same for recombination of self trapped excitons and shows that exciton life time remain unaffected due to interaction with solvent molecules [26,53]. In comparison, of C60 Zn-porphyrin monomers exhibit slower s values of 2.2 ns, which as converted to nanorods and nanotubes becomes much faster of 0.18 ns (85% decay) and 0.39 ns (65% decay) respectively.

3.5. Towards controlled co-solvate nanostructures by topotactic exchange of the solvent molecules It would be interesting to utilize the weak stability of the toluene solvates as an advantage. Even though the solvate nanocrystals are unstable at room temperature and decompose easily when left dry by the continuous release of the solvent molecules, these remain extremely stable even for months when kept in the mother liquor. This happens because the solvent molecules are constantly being released and reabsorbed by the solvate crystals in the mother liquid. As a result, there is a dynamic equilibrium between the solvent molecules in the mother liquid and within the solvate-crystals. We have taken advantage of the fact that some solvates are less stable than the others, therefore must have a higher equilibrium constant and faster rate of solvent absorption and desorption in mother liquor. Therefore if the solvent molecules in the mother liquor are replaced by a suitable molecule, this would lead to substitution of the solvent molecules in the solvate crystal. The solvent occupation sites in the C60–toluene solvates are associated with C3 symmetry

51

wherein methyl substituent of toluene is distributed over the three symmetry equivalent positions (Fig. 2b). We have therefore chosen mesitylene to exchange toluene since the threefold symmetry should allow it to effortlessly fit in these C60 intermolecular sites. Besides we have chosen the less stable nanorods (obtained by the ULLIP method) since these may have higher exchange rates. Indeed, we could demonstrate that such topotactic exchange leads to core-shell solvate structures (Fig. 9a), wherein shell is rich in the second solvent. This success is remarkable because synthesis of C60-solvates with more than one solvent of similar nature has remained impossible. Based on a number of controlled experiments we realized that it is so because the precipitation rate and associated nucleation phenomena in two solvents are distinct. More interestingly, the solvate crystals obtained after topotactic exchange shows an exceptional enhancement of stability from a few minutes to over a month. The topotactic exchange was carried out at 25 °C for 1 hour by dispersing the solvate nanorods in mother liquor consisting of isopropanol mixture of mesitylene instead of toluene. To ascertain the presence of mesitylene in the nanorods, we recorded infrared spectra (IR) before and after exchange (Fig. 9b and c). The peak observed at 729 cm1 in the pure unsubstituted nanorods is ascribed to the C–H out of plane bending mode of toluene [55]. The peak is marginally shifted by 2 cm1 from that of pure solvent (see Supporting information). The C–H out of plan banding vibration in pure mesitylene appears at 841 and 680 cm1 (Fig. S10a). After topotactic exchange, these peaks are shifted to 830 and 682 cm1 respectively. We also observe an additional

(a)

(b)

(c)

Fig. 9 – (a) A schematic depicting cross section of a nanorod undergoing topotactic exchange of solvent molecules on its surface leading to a core shell structure. FTIR spectra of the nanorods (b) before and (c) after topotactic exchange. The peaks corresponding to toluene and mesitylene are indicated by open and filled circles respectively. (A colour version of this figure can be viewed online.)

52

CARBON

7 4 ( 2 0 1 4 ) 4 4 –5 3

vibrational band at 697 cm1. Even though an in-depth study of the shifts and the new peaks are under progress, these unambiguously establish the presence of both the toluene and mesitylene in the crystal nanorods. The formation of the mesitylene shell enhances the stability of the solvate nanorods over 1000 fold. Fig. S11 shows the contrasting stability of the solvate and the co-solvate nanorods. The solvate nanorods begin to decompose soon after their removal from the mother liquor and a noticeable FCC diffraction peak at 10.8° appears after 1 h in the PXRD pattern. On the other hand, the nanorods obtained post-topotactic-exchange reactions are extremely stable. We have examined the PXRD pattern of the nanorods kept dry in air after one month of their preparation which revealed that they are still resistant to decomposition. The enhanced stability should result from the stronger interactions in co-solvates. Since mesitylene has three –CH3 groups, all close to the neighboring C60 molecules in the solvate crystal, the associated C–H  p interactions are stronger in comparison to that in the toluene case with only one –CH3 group.

4.

Conclusion

We have demonstrated that the extremely labile solvate of C60–toluene, which belongs to the highly conducting HCP C60 solvate family, can be stabilized by at least two different approaches. Slower kinetics of synthesis leads to the higher occupancy of the solvent sites in the solvate-crystal and increases their stability. The stable solvates begin to decompose at 58, 20 °C below than the stable m-xylene solvates and have decomposition activation energy of 188 kJ/mole. We found that these solvates exhibit interesting optoelectronic due to the incorporation of solvent molecules, in spite a pronounced red shift, absorbance spectra of the solvatecrystals are quite similar in appearance to that of C60 in solution. Unlike the featureless spectrum of pristine C60, these exhibit a stepwise absorption in the visible spectrum of light arising out of the forbidden singlet to singlet transitions. We also observed a novel absorption band centered at 673 nm. The excitonic lifetime of a C60 solvate is measured for the first time, which for the emission band at 760 nm is comparable to pure C60 and measures 1.2 ns. Finally we have designed a scheme to utilize the labile nature of the toluene molecules in the unstable HCP solvates. Therein, we have exchanged and incorporated a second type of solvent molecule by a topotactic exchange mechanism. This is the first example where a C60-cosolvate has been obtained as designed. The exchange of toluene with mesitylene is expected to take place near the surface of the solvate nanowires forming a core-shell structure. This led to almost complete suppression of their decomposition at ambient conditions without changing molecular arrangements. Moreover, this phenomenon demonstrates a unique possibility of supramolecular topotactic solvent exchange, very common in ionic solids.

5.

Notes

The authors declare no competing financial interest.

Acknowledgments Dr. Ujjal K. Gautam gratefully thanks the Department of Science and Technology, India for Ramanujan Fellowship. Moumita Rana is thankful to CSIR, India for SRF.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.carbon. 2014.03.001. R E F E R E N C E S

[1] Skabara PJ, Arlin J-B, Geerts YH. Close encounters of the 3D kind – exploiting high dimensionality in molecular semiconductors. Adv Mater 2013;25(13):1948–54. [2] Corley DA, He T, Tour JM. Two-terminal molecular memories from solution-deposited C60 films in vertical silicon nanogaps. ACS Nano 2010;4(4):1879–88. [3] Briseno AL, Mannsfeld SCB, Ling MM, Liu S, Tseng RJ, Reese C, et al. Patterning organic single-crystal transistor arrays. Nature 2006;444(7121):913–7. [4] Wang L, Liu B, Li H, Yang W, Ding Y, Sinogeikin SV, et al. Longrange ordered carbon clusters: a crystalline material with amorphous building blocks. Science 2012;337(6096):825–8. [5] Kaur N, Gupta S, Jindal VK, Dharamvir K. Pressure induced transformations in condensed and molecular phases of C60. Carbon 2010;48(3):744–55. [6] Korobov MV, Stukalin EB, Mirakyan AL, Neretin IS, Slovokhotov YL, Dzyabchenko AV, et al. New solid solvates of C60 and C70 fullerenes: the relationship between structures and lattice energies. Carbon 2003;41(14):2743–55. [7] Masuhara A, Tan Z, Ikeshima M, Sato T, Kasai H, Oikawa H, et al. Cyclic transformation in shape and crystal structure of C60 microcrystals. CrystEngComm 2012;14(22):7787–91. [8] Sathish M, Miyazawa K, Hill JP, Ariga K. Solvent engineering for shape-shifter pure fullerene (C60). J Am Chem Soc 2009;131(18):6372–3. [9] Geng J, Zhou W, Skelton P, Yue W, Kinloch IA, Windle AH, et al. Crystal structure and growth mechanism of unusually long fullerene (C60) nanowires. J Am Chem Soc 2008;130(8):2527–34. [10] Shrestha LK, Yamauchi Y, Hill JP, Miyazawa K, Ariga K. Fullerene crystals with bimodal pore architectures consisting of macropores and mesopores. J Am Chem Soc 2013;135(2):586–9. [11] Bae E, Kim ND, Kwak BK, Park J, Lee J, Kim Y, et al. The effects of fullerene (C60) crystal structure on its electrochemical capacitance. Carbon 2010;48(13):3676–81. [12] Tan Z, Masuhara A, Kasai H, Nakanishi H, Oikawa H. Thermal-induced shape transformation of solvated C60 microcrystals. Carbon 2013;64:370–6. [13] Zhang X, Mizukami S, Kubota T, Ma Q, Oogane M, Naganuma H, et al. Observation of a large spin-dependent transport length in organic spin valves at room temperature. Nat Commun 2013;4:1392. [14] Wang L, Liu B, Yu S, Yao M, Liu D, Hou Y, et al. Highly enhanced luminescence from single-crystalline C60Æ1mxylene nanorods. Chem Mater 2006;18(17):4190–4. [15] Li H, Tee BCK, Cha JJ, Cui Y, Chung JW, Lee SY, et al. Highmobility field-effect transistors from large-area solutiongrown aligned C60 single crystals. J Am Chem Soc 2012;134(5):2760–5.

CARBON

7 4 (2 0 14 ) 4 4–53

[16] Polander LE, Tiwari SP, Pandey L, Seifried BM, Zhang Q, Barlow S, et al. Solution-processed molecular bis(naphthalene diimide) derivatives with high electron mobility. Chem Mater 2011;23(15):3408–10. [17] Yao M, Andersson BM, Stenmark P, Sundqvist B, Liu B, Wagberg T. Synthesis and growth mechanism of differently shaped C60 nano/microcrystals produced by evaporation of various aromatic C60 solutions. Carbon 2009;47(4):1181–8. [18] Suezawa H, Yoshida T, Ishihara S, Umezawa Y, Nishio M. CH/ pi interactions as disclosed on the fullerene convex surface. A database study. CrystEngComm 2003;5(93):514–8. [19] Desiraju G, Sarma J. The chloro-methyl exchange rule and its violations in the packing of organic molecular solids. J Chem Sci 1986;96(6):599–605. [20] Ramm M, Luger P, Zobel D, Duczek W, Boeyens JCA. Static disorder in hexagonal crystal structures of C60 at 100 K and 20 K. Cryst Res Technol 1996;31(1):43–53. [21] Zhou Y, Zhou W. Growth mechanism of C60/mesitylene nanowires. CrystEngComm 2012;14(4):1449–54. [22] Minato J, Miyazawa K. Solvated structure of C60 nanowhiskers. Carbon 2005;43(14):2837–41. [23] Mullin JW, editor. Crystallization. Butterworth-Heinemann; 2001. [24] Ruoff RS, Malhotra R, Huestis DL, Tse DS, Lorents DC. Anomalous solubility behaviour of C60. Nature 1993;362(6416):140–1. [25] Kadish KM, Ruoff RS. Fullerenes: recent advances in the chemistry and physics of fullerenes and related materials. 1997: The Electrochemical Society; 1997. [26] Dresselhaus MS, Dresselhaus G, Eklund PC. Fullerenes. J Mater Res 1993;8(08):2054–97. [27] Jain PK, Amirav L, Aloni S, Alivisatos AP. Nanoheterostructure cation exchange: anionic framework conservation. J Am Chem Soc 2010;132(29):9997–9. [28] Zhao Z, Zhang W, Ren P, Han X, Mu¨ller U, Yilmaz B, et al. Insights into the topotactic conversion process from layered silicate RUB-36 to FER-type zeolite by layer reassembly. Chem Mater 2013;25(6):840–7. [29] Kyprianidou EJ, Papaefstathiou GS, Manos MJ, Tasiopoulos AJ. A flexible Cd2+ metal organic framework with a unique (3,3,6)-connected topology, unprecedented secondary building units and single crystal to single crystal solvent exchange properties. CrystEngComm 2012;14(24):8368–73. [30] Miyazawa K, Kuwasaki Y, Obayashi A, Kuwabara M. C60 nanowhiskers formed by the liquid–liquid interfacial precipitation method. J Mater Res 2002;17(01):83–8. [31] Jin Y, Curry RJ, Sloan J, Hatton RA, Chong LC, Blanchard N, et al. Structural and optoelectronic properties of C60 rods obtained via a rapid synthesis route. J Mater Chem 2006;16(37):3715–20. [32] Gallagher SH, Armstrong RS, Lay PA, Reed CA. Solvent effects on the electronic spectrum of C60. J Phys Chem 1995;99(16):5817–25. [33] Friedman HL. Kinetics of thermal degradation of charforming plastics from thermogravimetry. Application to a phenolic plastic. J Polym Sci Part C: Polym Symp 1964;6(1):183–95. [34] Jorda JL, Jondo TK. Barium oxides: equilibrium and decomposition of BaO2. J Alloys Compd 2001;327(1–2):167–77. [35] Tan Z, Masuhara A, Kasai H, Nakanishi H, Oikawa H. Multibranched C60 micro/nanocrystals fabricated by reprecipitation method. Jpn J Appl Phys 2008;47:1426. [36] Wang Y, Holden JM, Rao AM, Eklund PC, Venkateswaran UD, Eastwood D, et al. Optical absorption and

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47] [48]

[49]

[50]

[51]

[52] [53]

[54]

[55]

53

photoluminescence in pristine and photopolymerized C60 solid films. Phys Rev B 1995;51(7):4547. Wakahara T, Miyazawa K, Nemoto Y, Ito O. Diameter controlled growth of fullerene nanowhiskers and their optical properties. Carbon 2011;49(14):4644–9. Creegan KM, Robbins JL, Robbins WK, Millar JM, Sherwood RD, Tindall PJ, et al. Synthesis and characterization of C60O, the first fullerene epoxide. J Am Chem Soc 1992;114(3):1103–5. Schuster DI, Baran PS, Hatch RK, Khan UA, Wilson RS. The role of singlet oxygen in the photochemical formation of C60O. Chem Commun 1998(22):2493–4. Lee J, Hong S, Mackeyev Y, Lee C, Chung E, Wilson LJ, et al. Photosensitized oxidation of emerging organic pollutants by Tetrakis C60 aminofullerene-derivatized silica under visible light irradiation. Environ Sci Technol 2011;45(24):10598–604. Zhou P, Rao AM, Wang K-A, Robertson JD, Eloi C, Meier MS, et al. Photo-assisted structural transition and oxygen diffusion in solid C60 films. Appl Phys Lett 1992;60(23):2871–3. Rao AM, Zhou P, Wang K-A, Hager GT, Holden JM, Wang Y, et al. Photoinduced polymerization of solid C60 films. Science 1993;259:955–7. Cornett DS, Amster IJ, Duncan MA, Rao AM, Eklund PC. Laser desorption mass spectrometry of photopolymerized fullerene (C60) films. J Phys Chem 1993;97(19):5036–9. Eklund PC, Rao AM, Zhou P, Wang Y, Holden JM. Photochemical transformation of C60 and C70 films. Thin Solid Films 1995;257(2):185–203. Negri F, Orlandi G, Zerbetto F. Quantum-chemical investigation of Franck–Condon and Jahn–Teller activity in the electronic spectra of Buckminsterfullerene. Chem Phys Lett 1988;144(1):31–7. Dong Z-H, Zhou P, Holden JM, Eklund PC, Dresselhaus MS, Dresselhaus G. Observation of higher-order Raman modes in C60 films. Phys Rev B 1993;48(4):2862–5. Saito S, Oshiyama A. Cohesive mechanism and energy bands of solid C60. Phys Rev Lett 1991;66(20):2637–40. Leach S, Vervloet M, DesprksDespre`s, Bre´heret E, Hare JP, John Dennis T, et al. Electronic spectra and transitions of the fullerene C60. Chem Phys 1992;160(3):451–66. Hasegawa S, Nishiwaki T, Habuchi H, Nitta S, Nonomura S. Optical energy gap and below gap optical absorption of fullerene films measured by constant photocurrent method and photothermal deflection spectroscopy. Fullerene Sci Technol 1995;3(2):163–78. Zo´lyomi V, Koltai J, Ku¨rti J, Pekker S. The electronic band structure of fullerene-cubane cocrystals. Phys Status Solidi B 2008;245(10):2018–21. Gotoh T, Nonomura S, Watanabe H, Nitta S, Han D. Temperature dependence of the optical-absorption edge in C60 thin films. Phys Rev B 1998;58(15):10060–3. Skumanich A. Optical absorption spectra of carbon 60 thin films from 0.4 to 6.2 eV. Chem Phys Lett 1991;182(5):486–90. Shin HS, Yoon SM, Tang Q, Chon B, Joo T, Choi HC. Highly selective synthesis of C60 disks on graphite substrate by a vapor-solid process. Angew Chem Int Ed 2008;120(4):705–8. Hasobe T, Sandanayaka ASD, Wada T, Araki Y. Fullereneencapsulated porphyrin hexagonal nanorods. An anisotropic donor-acceptor composite for efficient photoinduced electron transfer and light energy conversion. Chem Commun 2008;29:3372–4. Stuart BH. Infrared spectroscopy: fundamentals and applications. Wiley; 2004.