Home >  About ISSP >  Publications > Activity Report 2018 > Xxxx Group

Electrostatic Charge Carrier Injection into a Charge Ordered Organic Insulator

Tajima Group

Charge carrier injection into a strongly correlated insulating state often leads to drastic phenomena. Previously, charge carrier injection into such materials was achieved by chemical substitution. However, the structural inhomogeneity, which originates from the random arrangement of chemically substituted atoms, is not avoidable. In contrast to the conventional chemical method, electrostatic (ES) charge carrier injection using a field-effect-transistor (FET) structure has recently been focused. This method does not introduce structural inhomogeneity, and precise and repeatable control of the carrier concentration can be realized by varying the gate voltage.

Fig. 1. (a) Schematics of typical FET device structure measured in this study. (b) Temperature dependence of S-D conductivity with zero gate voltage for four devices, A-D. Thickness (t) of α-ET2I3 crystals are summarized in the table (inset).

Fig. 2. (a)VSG dependence of ISD at various temperatures for device B. Data are normalized by ISD for VSG = 0 at each temperature. (b)Transfer characteristics of the four devices at approximately 58 K.

Here, we report the ES charge carrier injection effect on an organic charge ordered (CO) material α-ET2I3 using a FET structure [1]. ET denotes BEDT-TTF = bis (ethylenedithiolo) tetrathiafulvalene. In the CO state, the strong electronic correlation plays an essential role as well as the Mott insulator. At ambient pressure, α-ET2I3 has a metal-insulator (MI) transition at 135 K, which is associated with charge ordering. Figure 1(a) shows the typical device structure that was measured in this study. The devices are prepared by laminating a thin single-crystal flake [typical dimensions of 100 mm x 50 mm x (thickness: t ≈ 200-600 nm)] on a highly-doped Si substrate with thermally-oxidized SiO2 layer (500 nm). The SiO2 layer and the highly-doped Si act as a gate insulator and a gate (G) electrode, respectively. The S and D electrodes are formed by Au evaporation in vacuum (thickness ~30 nm) through a metal mask. Figure 1(b) shows the temperature dependence of the S-D conductivity for each devices. The steep decease in the conductivity corresponds to the MI transition of α-ET2I3. Below 110 K, the conductivities approximately exhibit thermal activation behavior: α(T) is proportional to exp(-Δ/kBT) (Δ: activation energy, kB: Boltzmann constant). The activation energies are approximately 550-650 K, depending on the devices.

Figure 2(a) shows gate voltage dependence of normalized S-D current [ISD/ISD(VSG = 0)] at various temperatures for device B. Below the MI transition temperature (TMI), ISD increases (decreases) for VSG > 0 (< 0): n-type transfer characteristics are observed. In bulk crystals, the dominant carriers above TMI are holes; however this changes to electrons below TMI, so that n-type behavior would be reasonable. However, in the present study, we have observed bipolar characteristics in certain devices [see Fig. 2(b)], whereas only n-type behavior in the insulating state has previously been reported [2]. As can be seen in Fig. 2(b), the hole transport for VSG strongly depends upon the device under investigation. Moreover, even in the same device, bipolar behavior, in particular that related to hole transport, is suppressed by the repetition of measurements and cooling-heating cycles. This means that the device dependence of hole transport is not sensitive to the device structure, but is, instead, dominated by the conditions of the (α-ET2I3)-Au and/or (α-ET2I3)-SiO2 interfaces. These two interface conditions affect the contact resistance and the channel conductivity, respectively, and they determine the transport properties of the hole channel.

The electric field effect observed in our present study [ISD/ISD(VSG=0) ≈1.7 at the maximum] is rather small compared with the result of recently reported organic FET that uses a κ type ET Mott insulator [3]. The difference between these two devices may be interpreted as the presence or absence of a structural change in the two insulating states. In the case of α-ET2I3, the structural transition associated with charge ordering is observed, and stabilization of the CO state by the low-temperature structure has been suggested theoretically [4]. Therefore, the electronic phase transition is hard to occur, in contrast to in the Mott insulator, where lattice distortion at the MI transition is absent. Because the CO state has a non-uniform charge distribution (depending on each site), change in the crystal structure would be induced by the charge inhomogeneity. On the other hand, one electron is localized per site in the Mott insulating state, so that the charge distribution is homogeneous.


References
  • [1] M. Kimata, T. Ishihara, and H. Tajima, J. Phys. Soc. Jpn to be published.
  • [2] H. M. Yamamoto et al., Physica B: Condensed Matter 404, 413 (2009).
  • [3] Y. Kawasugi et al., Phys. Rev. Lett. 103, 116801 (2009).
  • [4] Y. Tanaka, and K. Yonemitsu: J. Phys. Soc. Jpn. 77, 034708 (2008).
Authors
  • M. Kimata, T. Ishihara, and H. Tajima