Design of molecules for specific functions and their structural control in a nanoscale have attracted much interest for device downsizing as well as development of complex and multi-functional devices. In order to obtain nanostructures in various dimensions, supramolecular approaches are quite useful because of the structural flexibility. Furthermore, appropriate design of the constituent molecules may amplify and/or combine their specific functions in the supramolecular systems by proper molecular orientation and intermolecular interaction. Tetrathiafulvalene (TTF) is a strong electron donor, and its complexes with various kinds of electron acceptors are known to become semiconductors, metals and superconductors. In our previous studies, we found that introduction of appropriate substituents with a hydrogen-bonding group to the TTF moiety resulted in formation of its one-dimensional (1D) stacks to form nanofibers. Charge-transfer (CT) complexes of the TTF derivatives were also found to form 1D structures, which was considered as conducting nanowires useful for molecular electronics. In addition, if we can control reversible conformational changes of the molecules in nanowires by external stimuli such as light, electric or magnetic field and heat, the induced motions can be applied for molecular machines as well as switching devices. In this study, we synthesized the TTF derivatives having various side chains as shown in Figure 1A and their self-organization was investigated. Also, the CT complexes were prepared by mixing the TTF derivatives with various acceptor molecules and their electrical properties were evaluated. Characteristics of the mono-TTF compounds 1a–1d are as follows. For all compounds, π-π interaction between TTF moieties is expected. In addition, 1b–1d have phenyl rings which also induce π-π interactions between them. For 1c, cinnamoyl groups in adjacent molecules can react to form a cyclobutane ring by UV irradiation, and this reaction may be useful to fix the supramolecular structures. Amide groups for intermolecular hydrogen-bonding in 1b and 1d are also important to form molecular stacking arrays. Chiral centers in these compounds were introduced for the helical structure formation. The terminal ferrocene moiety in 1d is electrochemically active species, which can be utilized as a stimulus-responsive part. The CT complexes were prepared by mixing the TTF derivatives and F2TCNQ or F4TCNQ. In any combinations, new absorption bands were observed between 600 and 900 nm, suggesting formation of the CT complexes. The nanostructures of the TTF derivatives and their CT complexes were prepared by casting the corresponding toluene solution on HOPG substrates. Although 1b formed nanofibers, 1a, 1c and 1d became amorphous films or nanodots. Among these three compounds, 1a and 1c do not have relatively strong interactions like hydrogen bonding in their substituents. This indicates that the π-π interaction between TTF moieties is not enough to form the 1D nanostructure in general. In the case of 1d with amide groups, the bulky ferrocene groups at the end of the substituents seems to weaken the intermolecular hydrogen bonding, and the 1D nanostructure was not fabricated. Therefore, in order to intensify the π-π interaction of the core part, we synthesized compounds 2a–2d in which TTF units are dimerized with a disulfide bond. The crystallinity of 2a–2d was improved in comparison with the corresponding mono-TTF compounds. By SEM observation, it was confirmed that 2a and 2d formed nanoribbons having the width of about 800 nm and nanofibers having the width of about 200 nm, respectively (Figure 1B). This result suggested that intermolecular interaction was enhanced by the dimerized TTF structure with stronger π-π stacking. Thus, nanofiber formation became possible even in the compound with bulky substituents. In addition, the CT complexes of 2a–2d with F2TCNQ or F4TCNQ were prepared and their nanostructures were observed. Except for the CT complex using 2c, we confirmed the nanofiber formation. All of the CT complexes were found to show conductivity at the semiconductor level of 10-2–10-3 S m-1. Figure 1
Read full abstract
Dec 16, 2019
Young‐Jae Jin + 2
Open Access