1
Control of Electronic Property of C60 Fullerene via Polymerization

Nobuyuki Aoki

Chiba University, Graduate School of Science and Engineering, Department of Materials Science, 1-33 Yayoi, Inage, Chiba, 263-8522, Japan

1.1 Introduction

1.1.1 History of Polymerization of C60 Fullerene

Fullerenes, the spherical cage molecules composed of carbon atoms, were discovered by Kuroto and Smalley coworkers in 1985 [1]. After the development of a large-scale synthesis method by Kraetschmer et al. in 1990 [2], they became usable not only in vacuum but also in atmospheric conditions. C60 is the most popular molecule in the fullerenes composed of 60 carbon atoms as shown in Figure 1.1a. It has semiconductor characteristics having a bandgap. The experimental values have been shown as typically 1.5-1.8?eV for the highest occupied molecular orbital-lowest unoccupied molecular orbital () gap [3, 4]; however, the value varies within the range of 1.43-2.35. The electrical properties of fullerene are based on the bandgap. It shows n-type semiconductor characteristics; the activation energy is less than half of the bandgap and is close to the half value only at high temperatures. At around room temperature, the activation energy is in the range of 100-200?meV, which relates to activation from the donner-like state within the optical gap [5]. The transport properties are strongly affected also by the orientation state of the crystalline fullerene. On decreasing the temperature from room temperature, the first kink of the conductivity is observed at around 260?K related to the restriction of the orientation angle of rotation of C60 molecules [6]. The second kink at 90?K relates to the glass transition of the orientation angle.

Figure 1.1 Structural model of a C60 fullerene molecule (a) and a fullerene dimer having a dumbbell structure (b).

The conductivity drastically falls by several orders of magnitude, mostly becoming insulating due to absorption of oxygen molecules [7]. This is due to the formation of deep level trap sites lying 0.7?eV below the bottom of the conduction band. Therefore, most of the transport measurements of the semiconducting properties are done in vacuum conditions. An inert atmosphere such as argon, nitrogen, or helium also helps maintain the conductivity. If the sample was exposed in air once, heating at 160-180?°C in vacuum is necessary to recover the conductivity by desorbing oxygen from the thin film of fullerene [8].

Crystalline C60 fullerene exhibits face centred cubic () structure due to van der Waals interaction. Intermolecular interactions change the electrical and optical properties. A possibility of intermolecular coupling was observed first by photo-irradiation using laser light [9]. UV-visible light illumination read the photopolymerization of the C60 molecules in oxygen-free condition since oxygen hinders the reaction by forming photoexcited triplets [10]. Such a process occurs only above 260?K since a random orientation of the rotation is essential for the polymerization process.

Photo-transformation takes place by [2+2] cycloaddition reaction mechanism [11], where faced double molecular bonds are broken and a four-member ring is formed as shown in Figure 1.1a.

Such a dimer structure is called "dumbbell type" polymerization. For the occurrence of the polymerization reaction, the following requirements can be summarized:

1. The C60 molecules should be situated close enough to each other (an application of pressure assists this situation).
2. They must be rotating freely.
3. Their double bonds must be faced in parallel.
4. Certain external energy that opens the double bond must be applied (photo-excitation, thermal agitation, plasma, electron beam [] absorption, pressure application, etc.).
5. A four-atom carbon ring is formed.

After forming the intermolecular bonding, the mean intermolecular length, typically 1.0?nm in fcc structure, shortens to 0.01-0.03?nm. And then, the free rotation of the molecule stops, and solubility in polar organic solvents such as toluene, xylene, hexane, and so on is lost. For the photochemical reaction between C60 molecules, the following reaction scheme is proposed [11]. Although the [2+2] cycloaddition of neutral C60 molecules is thermally forbidden due to the Woodward-Hoffmann rules, this type of reaction is photochemically allowed between an excited and a ground state molecule. The interaction of the singly occupied p*-orbital of the photoexcited molecule with the unoccupied p*-orbital of the ground state molecule, as well as the interaction of the singly and doubly occupied p-orbitals gives rise to a symmetrically allowed and energetically favorable transition state. The simplified orbital interactions of two C?C double bonds are illustrated in Figure 1.2. Such a polymerization process in a solid or a thin film of C60 can also take place by other means: application of high pressure at high temperature, intercalation of alkali metals, plasma treatment, EB irradiation, and so on. If a negative ion reacts with a neutral ground state molecule, the interaction of the singly occupied and the vacant p*-orbitals results in a lower energy transition state, similarly to the photochemical mechanism.

Figure 1.2 Schematic of the photochemical reaction (a) and energy diagrams of frontier orbital interactions of reactants in [2+2] cycloaddition reactions (b).

For the intermolecular bond formation, two types of possibilities are proposed. In one configuration, the two C atoms shared by the two adjacent hexagons in a C60 are covalently bonded to the C atoms that are shared by the two hexagons in the adjacent C60 (66/66 bond); in another, the two C atoms shared by a hexagon and a pentagon in a C60 are bonded to the adjacent C60 (66/65 bond) shown in Figure 1.3b. It is known that the 66/66 bond is more stable thermodynamically than the 66/65 one; therefore, the 66/65 one must be a very rare case [12, 13]. However, existence of a C60 polymer composed of 66/65 bonds (Figure 1.3b) is the only model to explain the metallic property of the polymer having a two-dimensional rhombohedral () structure realized by high temperature and high pressure application [14].

Figure 1.3 (a) 66/66 and (b) 66/65 bonding structures.

Hence, we can control the electrical properties of C60 fullerene depending on the polymerized structure. In this chapter, after showing the basic electrical properties of a pristine C60 fullerene, various kinds of polymerization regimes and the electrical properties will be introduced.

1.1.2 Electronic Property of Pristine C60 and n-Type FET Action

The first C60 field effect transistor () was reported by Paloheimo et al. [15] in 1993 and developed by Haddon et al. [16] in 1995. They clarified that C60 FET works as an n-type transistor and exhibits fairly good carrier mobility around 0.1?cm2 V-1 s-1. The highest mobility is more than 10?cm2 V-1 s-1 as reported by Li et al. in 2012 [17]. The properties are severely dependent on the environment. After being exposed in air or oxygen, the conductance suddenly drops and it becomes almost insulating [7]. Therefore, most of the transport experiments are performed in vacuum conditions. The conductance also depends on temperature, so that there might be a large amount of charge trapping states, in other words donor-like states, in the pseudo gap of C60 FET [18]. In general, these states usually come from structural disorders at crystalline defect or grain boundary, as well as from polaronic disorders introduced by guest impurities. The latter would be an effect of oxygen adsorption, by which the conductance decreases drastically. Therefore, transport measurements of a C60 FET require an oxygen-free environment such as vacuum or inert gas, or covering by a passivation layer [19]. Figure 1.4 shows typical current-voltage (I-V) curve and transfer curves of a C60 thin film FET. The FET structure is fabricated on a SiO2 layer of heavily doped Si wafer. The electrical contacts are performed by Au. The channel length and width are 5 and 100?µm, respectively. The field effect mobility, µ, in the low field region can be estimated from the following equation [20]:

where ID is the drain current, L is the channel length, W is the channel width, COX is the gate capacitance of oxide layer, VDS is the drain voltage, VGS is the gate voltage, and VT is the threshold voltage. Considering the threshold voltage to be 15?V, the mobility is estimated as 0.1?cm2 V-1 s-1.

Figure 1.4 Typical transistor curves of a C60 thin film FET (a). Transfer curve and estimated mobility (b).

The temperature dependence of a C60 FET shows nearest neighbor hopping, which depends on T-1 at temperatures lower than room...