1. Introduction

There has been a rapidly increasing demand in recent years for optoelectronic. devices capable of operating in the THz frequency range for use in a diverse mix of applications such as biomedical imaging, material characterization, high-speed communication and homeland security [1–3]. This has been seen in the development of optical components, like phase shifters [4–7], gratings [8], modulators [9], and filters [10–13]. However, the lack of suitable functional materials has so far restricted the practical development of THz devices. Conventional transparent conducting electrodes and electro-optic (EO) materials, which are needed in voltage-controlled components to provide a longitudinal electric field parallel to the propagating light, only offer a low transmittance at THz frequencies. Moreover, although the current industry standard for transparent inorganic anodes in rigid devices [14], indium tin oxide (ITO) does exhibit superb transparency in the visible range, it is highly absorbing in the THz region (less than 10% transmission rates at 0.2-1.2 THz) [15]. Reducing the thickness of ITO films can enhance their transmittance in the THz region, but this also has the effect of significantly increasing their resistance [16]. Adding to the problem is the fact that the brittle nature of ITO makes it unsuitable for use in flexible applications [17], and that it is usually formed by a sputtering process that is too slow for industrial-scale production [18]. As a result, more novel materials, such as metal films [19], graphene [20], and metamaterials [21] have been explored as potential THz transparent electrodes, but these create problems of their own in terms of overcoming the complexity of fabrication and relatively high production cost. In the case of EO materials, conventional crystals such as KDP and LiNbO₃ do not provide a smooth frequency response across the THz range, and are not suitable for use as THz modulators or other functional devices due to phonon absorption. It is therefore clear that the practical realization of THz modulation requires the development of an easy-to-fabricate, low-cost, high-transparency electrode material and high-quality EO material.

As an organic electrode material, poly (3,4-ethylenedioxythiophene):poly (4-styrenesulfonate) (PEDOT:PSS) shows great promise over other conducting polymers thanks to its high conductivity, good stability, good thin-film transparency, high proper work function as an anode, and an ability to be processed in aqueous solution [22–24]. This has already seen it used as an antistatic coating and hole injection layer in organic light-emitting diodes (OLEDs) and solar cells [25,26], but the conductivity of pristine PEDOT:PSS film (usually no more than 10 S/cm) is well below that of typical ITO (typically around 4000 S/cm). However, by mixing PEDOT:PSS with various solvents like dimethylsulfoxide (DMSO) or ethylene glycol (EG), the conductivity can be improved by up to 2 or 3 orders of magnitude (ca. 1000 S/cm) [27–29], thereby providing a device performance much closer to that of conventional ITO-based devices.

In this report, highly conductive DMSO-doped-PEDOT:PSS (Clevios PH1000) thin films produced by spin coating onto quartz and silicon substrates are evaluated for their potential to be used as low-cost transparent conducting electrodes for a blue-phase liquid crystal (BPLC) modulator operating at THz frequencies. As there is generally a trade-off between the transparency and conductivity of a transparent electrode, in that a greater thickness brings both an increase in charge carriers and the amount of light absorbed inside the film, DMSO doping was employed to improve the conductivity without sacrificing transmittance.

2. Experimental

High resistance silicon and quartz substrates were ultrasonically cleaned in deionized (DI) water with 3% Mucasol for 10 min at 60 °C, and then rinsed thoroughly with DI water. This was followed by ultrasonic cleaning in ethanol, isopropyl alcohol, and DI water for 10 min each step. These cleaning steps were repeated until a haze-free surface was obtained and solvents used were of pro analysis quality. The clean substrates were then treated by UV/ozone for 20 min to improve their hydrophilicity.

A 1.3 wt% PEDOT:PSS solution (PH 1000, Clevios, Leverkusen, Germany) with a 2.5:1 weight ratio of PSS to PEDOT was mixed with 5, 10 or 15 vol% DMSO and stirred for at least 3 h at 90 °C, then stirred for a further 12 h at room temperature to improve the uniformity of the solution for film formation. These mixtures were filtered through a 0.45 μm syringe filter and used to spin-coat a thin layer onto the cleaned substrates. The resulting PEDOT:PSS films were then annealed in air at 130 °C for 30 min, with their final thickness being measured using a step-profiler (Alpha-Step 500, Tencor). The sheet resistance of the films was measured using an in-line four probe method and their surface morphology was analyzed by AFM imaging (TITANIUM Atomic Force Microscope, NT-MDT). The transmittance of the films was measured by a THz time-domain spectrometer system (TDS 1008, BATOP GmbH).

The BPLC materials used in this experiment consisted of 88.78 wt% of positive nematic LC (BPH006, HCCH), 3.14 wt% of chiral dopant (R5011, HCCH), 3.88 wt% of ultraviolet (UV)-curable monomer (TMPTA, HCCH), and 3.93 wt% of cross-linker (RM257, HCCH). The mixture was placed in an isotropic state into an empty LC cell consisting of two quartz substrates with a 15 vol% DMSO-doped PEDOT:PSS film on one side and a cell gap of 104 μm. The samples were then cooled at a rate of 0.2 °C/min to observe the phase transition process under a polarized optical microscope (POM, XPL-30TF). The cells were then irradiated for 5 min during their blue phase (>100 °C) by an ultraviolet light with an intensity of 3 mW/cm².

The THz source of the experimental system used was an optically-pumped THz laser system (SIFIR-50 FSW, Coherent). Here, a 20 Hz sinusoidal-wave AC signal was applied to the modulator between two polarizers, with the output signal being measured by a Golay detector (Golay Cell GC-1P, Tydex).

3. Results and discussion

Details of the film thickness and conductivity achieved with different spin coating speeds and layers are summarized in Table 1. The sheet resistance of ITO samples was also measured using the same method for comparison. Note that as the presence of DMSO affects the solution concentration and viscosity, as well as solvent evaporation during the spin coating process, the thickness of the films decreases from 62 to 52 nm with an increase in DMSO concentration from 5 to 15 vol% despite the same spin speed being used (4000rpm). More importantly though, the films with added solvent show a significant increase in conductivity of up to10670 S/cm with a bi-layer film containing 15 vol% DMSO. There is, however, a decrease in conductivity when the number of spin coated layers is increased to three. With those films containing a DMSO concentration of more than 15 vol%, the conductivities of the single layer films remain nearly constant at about 5000 S/cm. This suggests that the enhanced carrier transport observed in DMSO-doped PEDOT:PSS is due to either an increased carrier delocalization or the screening effects of the dopant [30,31]. Carrier delocalization is known to increase with dopant concentration, but only up to the point at which the dipolar molecules have fully screened the charge impurities. Beyond this, the dipolar molecules act as scattering centers themselves, which would explain the decrease in conductivity at higher DMSO concentrations. Given that the conductivity of DMSO-doped films has been proven to be almost frequency independent [32], this makes it a potentially very useful electrode material.

Table 1. PEDOT:PSS and ITO sample thickness and conductivity

View Table

The surface morphology of PEDOT:PSS electrodes doped with different concentrations of DMSO was investigated using atomic force microscopy (AFM). with Fig. 1 showing AFM height images of single layer films. It is evident from this that the addition of DMSO leads to areas with a greater height than the average film thickness, with this resulting in a gradual increase in root mean square (r.m.s.) roughness of films increased gradually from 1.200 nm to 1.578 nm(1 μm × 1 μm) with an increase in DMSO concentration from 5% to 15 vol%.

 

Fig. 1 2D and 3D AFM height images of PEDOT:PSS doped with (a,b) 5, (c,d) 10, or (e,f) 15 vol% DMSO.

Download Full Size | PDF

As shown in Fig. 2, the transmittance of the DMSO-doped PEDOT:PSS films decreased with an increasing number of PEDOT:PSS layers due to the simple fact that thinner films are more transparent; i.e, the transmittance of conducting polymer films follows Beer’s law. A comparison of the transmittance of single-layer DMSO-doped PEDOT:PSS films in Fig. 2(d), however, reveals a peak transmittance of 54.9, 59.3 and 83.5% with 5, 10 and 15 vol% DMSO, respectively, with corresponding sheet resistances of 11.35, 10.53 and 8.36 Ω. Thus, the 15 vol% DMSO-doped PEDOT:PSS not only has the highest conductivity (5078 S/cm), it also achieves the highest transmittance. What is more, this conductivity exceeds literature values for solvent-doped PEDOT:PSS films (~1418 S/cm) [26] and VPP PEDOT (≤1180 S/cm) [14], which achieved a maximum transmittance of 83.5% at 1.22 THz. The fact that DMSO-doped PEDOT:PSS electrodes can provide a balance between high transparency and high conductivity obviously makes them far more useful than conventional transparent conducting electrodes that must make a tradeoff between optical transmittance and electrical conductivity [33]. All the transmittance we mentioned in this Letter was power transmittance.

 

Fig. 2 Transmittance of single layer, bi-layer and tri-layer PEDOT:PSS films doped with (a) 5, (b) 10 or (c) 15 vol% DMSO. (d) Transmission of single-layer PEDOT:PSS films doped with among 5, 10 or 15 vol% DMSO (σ = 3137, 3678,5078 S/cm, respectively).

Download Full Size | PDF

The high-performance of these DMSO-doped PEDOT:PSS transparent electrodes has the potential to make possible a range of optical devices, with a THz modulator being of particular importance in many applications. Electrically controlled modulators have already been achieved by controlling the alignment of liquid crystals at room temperature, but it is blue-phase liquid crystals (BPLCs) that are of much greater interest due to their unique physics properties such as sub-millisecond response time, lack of reliance on a surface alignment layer and large optical Kerr effect [34,35]. A THz modulator based on DMSO-doped PEDOT:PSS film was therefore constructed by sandwiching a 104 μm BPLC layer between two 500 μm quartz substrates that were spin-coated with 15 vol% DMSO-doped PEDOT:PSS film electrode. We tested the performance of the BPLC modulator at 2.52 THz. Figure 3(a) shows the electrically induced light scattering system used for this BPLC modulator, in which electrical pulses were applied from the voltage amplifier and the response from the Golay detector was fed into an oscilloscope. The waveform of the input and output signals with different polarized directions of the analyzer are shown in Fig. 3(b). Note that the frequency of the output signal is doubled, and when the polarizer and analyzer directions are parallel, the optical transmittance reaches a maximum. Conversely, when the polarizer and analyzer directions are crossed, the optical transmittance is at a minimum, meaning that light scattering does not change the polarized direction. The output signals under different applied voltages with the polarizer and analyzer parallel are shown in Fig. 3(c), from which we see that the output signal is enhanced when the applied voltage is increased. BPLC molecules are anisotropic and dynamically switchable by electric field (E). There are different electric field effects on blue phase, such as lattice orientation effect, Kerr effect, and so on [36]. But the behavior we observed is not the same as the reported effects above. Based on this, by referring to the electrically induced light scattering performance of lanthanum-modified lead zirconate titanate transparent ceramics [38], it is proposed that the scattering centers in the BPLC are refractive index discontinuities at the domain walls, and so when an electric field is applied the inner domains may switch, thus causing a change in transmittance.

 

Fig. 3 (a) Electrically induced light scattering performances measurement system for the BPLC modulator. (b) Waveform of the input signal (black curve) and output signal with different polarized directions of the analyzer (colored curve). (c) Output signal produced under different applied voltages when the polarizer and analyzer directions are parallel.

Download Full Size | PDF

The advantages of using DMSO-doped PEDOT:PSS films as transparent electrodes for a THz modulator can be seen by comparing them with conventional materials such as ITO and novel materials such as graphene. For instance, past research has shown that even single-layer graphene only produces a transmittance of 88% and sheet resistance of 230 Ω at THz frequencies, with additional layers reducing both the sheet resistance and transmittance [20]. In contrast, DMSO-doped PEDOT:PSS electrodes are more likely to provide a balance of high transparency and high conductivity, making them more useful than other transparent conducting electrodes (e.g. ITO, carbon nanotube networks and graphene) that suffer from the classic tradeoff between optical transmittance and electrical conductivity [33]. Current experiments have used ITO nanowhiskers as transparent conductors for liquid-crystal THz phase shifters, where their conductivity is about 254 S/cm; however, fabricating nanostructures over several material templates risks increasing scattering and surface recombination losses. Moreover, the technologies required to take advantage of combining both micro- and nano-scale surface textures are yet to be fully developed. Several studies have also examined the uniformity issue of using nanometer-scaled pillars, and although they significantly enhance device performance, their pattern shapes are still random around the surface [37]. In contrast, DMSO-doped PEDOT:PSS films can be fabricated by a simple spin coating process that is much easier than that of ITO nanowhisks. Furthermore, the transmittance of DMSO-doped PEDOT:PSS electrodes at THz frequencies, as well as their surface morphology and electrical conductivity, are closely related to the film thickness. That is, we can obtain an ideal transmittance and conductivity by simply changing the thickness of PEDOT:PSS films. Another advantage is that DMSO-doped PEDOT:PSS films have a high transmittance over a wide range of frequencies, thereby providing DMSO-doped PEDOT:PSS-based devices with broadband modulation without the need to further tailor the material. To confirm the films could be used in broadband frequency, we tested the BPLC modulator not only at THz frequencies (1.63, 2.52 THz) but also in the visible range (632.8nm). This DMSO-doped PEDOT:PSS films provided a controllable, electrically induced light scattering that presented the feasibility of using these films as a transparent electrode in optoelectronic devices that require transparency in both the visible and THz spectrum.

4. Conclusion

In summary, we have successfully produced a novel DMSO-doped PEDOT:PSS films as transparent electrodes by spin-coating of DMSO/PEDOT:PSS hybrid solution. This study has demonstrated that the transmittance of electrodes based on DMSO-doped PEDOT:PSS films in the THz frequency range is dependent on their thickness. However, a maximum transmittance of 83.5% at 1.22 THz with a single 52 nm-thick film coincides with a high electrical conductivity of 5078 S/cm, meaning that a high transparency and electrical conductivity need not be mutually exclusive. When applied to a THz modulator, the electrically induced light scattering behavior allows for a high level of control, thus making DMSO-doped PEDOT:PSS films an excellent choice as a transparent conductor for EO devices in the visible and THz frequency range.