Light-induced superconductivity using a photoactive electric double layer

See allHide authors and affiliations

Science  13 Feb 2015:
Vol. 347, Issue 6223, pp. 743-746
DOI: 10.1126/science.1256783

Light switch for superconductivity

The conducting properties of materials sensitively depend on the available carrier concentration. Physicists can vary this concentration by inducing carriers in the material; for example, by placing it next to an ionic liquid in an electric field. Suda et al. instead used a layer of the molecule spiropyran, which changes from a non-ionic to an ionic form when it is irradiated by ultraviolet (UV) light. The authors placed the layer on top of a thin crystal of an organic material. When they shone UV light on the spiropyran, the adjacent material became superconducting, thanks to the carriers induced at the interface.

Science, this issue p. 743


Electric double layers (EDLs) of ionic liquids have been used in superconducting field-effect transistors as nanogap capacitors. Because of the freezing of the ionic motion below ~200 kelvin, modulations of the carrier density have been limited to the high-temperature regime. Here we observe carrier-doping–induced superconductivity in an organic Mott insulator with a photoinduced EDL based on a photochromic spiropyran monolayer. Because the spiropyran can isomerize reversibly between nonionic and zwitterionic isomers through photochemical processes, two distinct built-in electric fields can modulate the carrier density even at cryogenic conditions.

The electric potential difference that usually exists across a phase boundary creates two layers of space charges of different signs, the so-called electric double layer (EDL) (1, 2). In recent years, EDL transistors with ionic liquids or electrolytes as gate dielectrics have been used increasingly (3, 4). Under an applied gate voltage, ions are driven to the surface of the channel materials, forming an EDL that acts as a nanogap capacitor. Using this method to accumulate a large numbers of carriers (~1014 cm−2) (59), field-induced superconductivity has been realized in various materials. However, modulations of the carrier density in EDL transistors have been limited to the high-temperature regime because the ionic motions of ionic liquids or electrolytes are frozen below ~200 K.

Here we observe field-induced superconductivity even at low temperatures by using a photoactive EDL system, a spiropyran monolayer. Spiropyran is a photochromic molecule that can photoisomerize reversibly between the nonionic spiropyran (SP) form and the zwitterionic merocyanine (MC) form upon ultraviolet (UV) (forward reaction) or visible light irradiation (reverse reaction) (fig. S1) (10, 11). Therefore, an EDL-like large electric field across the film can be induced or be eliminated by UV light or visible light irradiation, respectively. The advantage of the photochromic EDL system over the conventional one is that it retains the modulation capability even at low temperatures, because these photochromic reactions proceed with photon-energy dissipation under nonequilibrium conditions. Hence, a highly aligned SP monolayer is a promising candidate for a photoinduced EDL system that can directly switch superconductivity only by photoirradiation. Although there have been attempts to tune the conductivity and even superconductivity using photochromic layers (1217), photoinduced phase transitions (including superconducting transitions) have not been realized yet.

We fabricated a photoactive superconducting device by laminating a thin single crystal of κ-(BEDT-TTF)2Cu[N(CN)2]Br (κ-Br) [BEDT-TTF: bis(ethylenedithio)tetrathiafulvalene] on top of an Al2O3(or HfO2)/Nb:SrTiO3 substrate covered with a self-assembled monolayer of spyropiran derivatives (SP-SAM) (Fig. 1). The device also has a bottom-gate structure for electrostatic carrier doping to finely modulate the carrier density accumulated as a result of photoirradiation. We chose κ-Br as a target material because κ-type BEDT-TTF is a strongly correlated molecular conductor (18) that has a relatively low half-filled carrier density (~2 × 1014/cm2) (19). We recently demonstrated a field-induced superconductivity in a thin single crystal of κ-Br (20).

Fig. 1 Configuration of the κ-Br device.

(A) Schematic cross section of the device; the spiropyran (SP) monolayer can isomerize reversibly by UV and visible light irradiation at the interface (fig. S1). (B) Optical microscope image of the κ-Br single crystal laminated on the substrate. Gold wires were attached to the crystal with carbon paste for standard four-probe measurement. Scale bar, 200 μm.

The surface density of SP-SAM estimated by the comparison of absorption spectra between a monolayer film and a solution (fig. S2) was ~1.8 × 1014 molecules/cm2. The surface density for the close-packed model (2.0 × 1014 molecules/cm2) was already reported (13), indicating that we have obtained a high-coverage SP-SAM. The photoisomerization of SP-SAM at room temperature was monitored by UV–visible absorption spectroscopy (fig. S3). The spectrum of the initial state has a weak absorption peak at 580 nm that is ascribed to the π-π* transition in the MC form. This means that the initial state of SP-SAM is in a SP state-dominant mixed state. After UV irradiation for 5 min, the absorption peak was enhanced, indicating SP-to-MC photoisomerization. After a subsequent irradiation with visible light, the peak almost vanished, indicating MC-to-SP back photoisomerization. SP-to-MC-to-SP photoisomerization was then repeated without any attenuation of the area between the absorption curves. These reversible spectral changes clearly demonstrated reversible photoisomerization of spiropyrans, even after the SAM formation.

Subsequently, we investigated the effects of photoirradiation of UV and visible light on the four-probe resistance of κ-Br devices. Figure 2A shows a time trace of the resistance for the device at 5 K. After brief resistance measurements in the dark, the device was irradiated with UV light. During the UV irradiation, the resistance value drastically decreased and the final state persisted even after the irradiation was stopped. Furthermore, this reduced resistance value recovered to nearly the initial value by visible light irradiation, which also lasted even after the irradiation was stopped (Fig. 2B), showing reversible switching capability. The resistance switching was repeated several times, with negligible attenuation (Fig. 2C). To rule out that the response was caused by photothermal effects, we also conducted control experiments on the κ-Br device without the photochromic layer. We found almost no resistance change during UV and visible light irradiation. Therefore, the reversible resistance switching must be caused by the reversible photoisomerization of SP-SAM.

Fig. 2 Time trace of the resistance for the κ-Br devices with and without SP-SAM under photoirradiation at 5 K.

(A) After the resistance measurements in the dark for 180 s, each device was irradiated with UV light for 180 s. The measurements were continued for 180 s after the irradiation was stopped. (B) After the resistance measurements in the dark for 180 s, each device was irradiated with visible light for 180 s. The measurements were continued for 180 s after the irradiation was stopped. (C) Time trace of the resistance for the κ-Br device with alternate photoirradiation of UV and visible light.

Figure 3A shows the temperature dependence of the resistance after various UV irradiation times at 5 K. In the initial state, the κ-Br device exhibited Mott insulating behavior at low temperature; this behavior is contrary to the superconducting character of the bulk crystal and caused by the tensile strain from the substrate (2022) (details are described in fig. S4). The resistance value decreased with increasing the irradiation time, and when the irradiation time reached 180 s, the device showed an abrupt resistance drop around 7.3 K; the drop was suppressed by a magnetic field of 7 T (fig. S5). This behavior indicates that the system showed a photoinduced phase transition from a Mott insulating phase to a partially superconducting phase.

Fig. 3 Temperature dependence of the resistance and the magnetization for the κ-Br device.

(A) Resistance versus temperature for the κ-Br device after various times of irradiation with UV light. (B) Magnetic susceptibility [zero–field-cooled (ZFC) and field-cooled (FC) measurements] for device 4 in the initial state and UV-irradiated state under a field of 100 Oe.

The photoinduced superconducting transition was also confirmed by magnetic susceptibility measurements (Fig. 3B). In the initial state, small shielding and Meissner effects were observed below ~10 K in zero–field-cooled and field-cooled measurements. After UV light irradiation, the zero–field-cooled curve was strongly enhanced, implying an increase in the superconducting fraction. The superconducting fraction created by UV light irradiation can be estimated by comparing the diamagnetic value of the κ-Br device with that of bulk κ-Br. From this estimation, ~5% of the volume fraction seems to be caused by UV light irradiation, suggesting that the superconducting transition occurs not only at the interface layer but also in the bulk of the device, because the volume fraction of the interface layer (monolayer) is only ~0.3% (the thickness of the κ-Br crystal used for the magnetic measurements was 600 nm, which corresponds to ~300 layers). Because the carriers injected at the interface are unlikely to diffuse into the layers far from the interface, such a bulk phase transition should originate from an interlayer dielectric screening of the Coulomb interaction, where effective repulsion among Mott insulating carriers in the bulk of the device is screened by the mobile superconducting carriers at the interface (23). A similar bulk phase transition has been also observed for the field-induced superconductivity in a single-crystal field-effect transistor (FET) of κ-Br (20).

We now turn to the study of the photoswitching mechanism. κ-Br has been shown to turn superconducting as a result of electrostatic carrier doping (20). It is therefore reasonable to assume that the photoinduced phenomena presented here are also caused by carrier doping by the photoisomerization of SP-SAM. It has been found that, when organic molecules with sizable dipole moments are organized as a SAM, a cooperative effect of charge transfer occurs between the monolayer and the substrate (24, 25). The electric field produced by such a monolayer dipole can exceed the field that can be applied across the gate dielectric in a standard FET configuration. We can calculate the expected electric field inside a SAM through the following relationshipEmbedded Image(1)where N, dmol and ε are, respectively, the surface density, the height of the molecules, and the effective dielectric constant inside the monolayer (26, 27). For our SP-SAM, we assumed that N is 1.8 × 1014 molecules/cm2, ε is between 2 and 3, μmol is 6.4 D for the SP state and 13.9 D for the MC state, and dmol is 2.2 nm (13); conversion efficiency from SP-to-MC state is ~66%. Using these values, we can estimate the difference in the electric fields (Ein) between the SP and MC states as 5.0 ~ 7.5 MV/cm, which is much larger than the electric fields that can be produced by applying a gate voltage using the bottom-gate structure.

To confirm this hypothesis, we performed gate voltage sweeps using the bottom-gate structure with various irradiation times. Figure 4 shows the contour plots of the resistance as a function of temperature and gate voltage for increasing times of UV light irradiation, which was performed at 5 K. The irradiation was stopped during the gate-sweep mode electrical measurement. The κ-Br device without SP-SAM was not affected by UV or visible light irradiation (fig. S9). For κ-Br devices with SP-SAM, the contour plot shifted in the positive voltage direction with elapsed UV-light irradiation time, reflecting a gradual progress of the hole carrier injection, as expected from Eq. 1. After the irradiation time of 60 s, a sudden resistance drop appeared at negative gate voltage (VG ≈ −4 V), showing the (field-effect assisted) photoinduced superconducting transition. When the irradiation time reached 180 s, a superconducting phase could be observed even without gate voltage. The observed voltage shift after 180 s is ~9 V, which leads to the difference in the built-in potential of 4.3 MV/cm. This value is somewhat smaller than the results obtained theoretically by Eq. 1 (5.4 ~ 8.1 MV/cm). One reason for this discrepancy is a decline of the conversion efficiency of SP-to-MC photoisomerization because of the lamination of the crystal; it is well known that the free volume around spiropyrans has an effect on the photochromic response (28).

Fig. 4 Evolution of the contour plot of resistance as a function of temperature and gate voltage during the photoirradiation and extended phase diagram of the κ-Br device.

The contour plots were recorded in initial state (A); UV-irradiated states with irradiation time of 30 s (B), 60 s (C), 90 s (D), 120 s (E), 150 s (F), and 180 s (G); and visible-irradiated state with the irradiation time of 300 s (H). (I) Phase diagram of the κ-Br as a function of temperature and carrier concentration was obtained by overwriting the contour maps (B) to (G) on (A) with appropriate shifts in the gate voltage (irradiation time), because each contour map could be connected smoothly at nearly identical resistance point. SC and MI denote the superconducting and the Mott insulating phases, respectively.

Finally, a phase diagram of the κ-Br as a function of temperature and irradiation time (with applied gate voltage) was obtained by superimposing Fig. 4, A to G, as shown in Fig. 4I. We found continuous shift of the contour plots in Fig. 4, A to G, without noticeable color gradation difference in the overlapping areas. In addition, we observed a portion of a superconducting dome resembling that of the superconducting cuprates (29) and the κ-Br with conventional FET configuration (20). These results indicate that the photoinduced effect was caused not by thermal heating or interface chemical reactions but rather by the (hole) carrier doping. Furthermore, the field-induced carriers and the photoinduced carriers were working as the same type of carriers that are indistinguishable from one another in the κ-Br crystals. This means that SP-SAM EDL can expand the limit of electrostatic carrier doping beyond the density that can be accumulated only by means of a normal FET configuration.

Supplementary Materials

Materials and Methods

Figs. S1 to S9

References and Notes

  1. Acknowledgments: Financial support for this work was provided by Grants-in-Aid for Scientific Research (S) (no. 22224006) from the Japan Society for the Promotion of Science (JSPS).
View Abstract

Navigate This Article