ELECTROPOLYMERIZED THIOPHENE-BASED DEVICES FOR PHOTOCATALYTIC HYDROGEN GAS HARVESTING

Emmeline Kao1, Qiaohao Liang1, Goran Rez-Kallah Bertholet1, Jianan Lu1, and Liwei Lin1*

1Berkeley Sensor and Actuator Center, University of California, Berkeley, USA ABSTRACT

We present thiopene-based devices made by spin-coating and electropolymerization for usage in solar-powered, photocatalytic hydrogen gas (H2) harvesting. Two innovative claims achieved in this work include: (1) first demonstration of electropolymerized photoelectrochemical devices for water splitting, and (2) drastically improved performance of EP-PEC devices over spin-coated PEC H2 harvesters, achieving >0.5V improvement in onset voltage (Von, bias voltage needed to produce photocurrent), with Von of 0V vs. Ag/AgCl. As such, this work opens up a new class of material and device fabrication for cheaper and efficient PEC H2 harvesting systems. KEYWORDS Photoelectrochemistry; water splitting; organic semiconductor; electropolymerization; hydrogen gas harvesting; INTRODUCTION

PEC water splitting devices, working ideally, will passively (without bias voltage) convert sunlight into electron/hole pairs that will oxidize water (2H2O + 4h+ → 4H+ + O2) and reduce resulting H+ ions (4e- +4H+ → 2H2 ), as shown in Figure 1. As opposed to that of a photovoltaic (PV) cell, a PEC water splitting semiconducting electrode uses generated photocurrent to

catalyze red/ox reactions in aqueous medium. Since water splitting is easily scalable with the area of semiconducting material used, it is ideal for large scale hydrogen harvesting. In contrast, achieving large scale hydrolysis using PVs requires multiple cells to be stacked in series, increasing resistivity across the H2 production system, and decreasing conversion efficiency [1]. However, efficient conversion of solar energy into fuel has been difficult to accomplish. Ideal water splitting materials have been difficult to fabricate because they require a specific band structure and excellent photocatalytic properties: (1) band gap (Eg) must be small to absorb the solar spectrum effectively; (2) Eg must be positioned correctly to allow for the red/ox reactions necessary to split water (conduction band above reduction potential, and valence band below oxidation potential); (3) materials must be able to withstand a highly oxidizing (aqueous) environment; and (4) materials must exhibit high charge separation efficiency in contact with an aqueous electrolyte. Moreover, materials often exhibit poor Von (require large bias voltages to observe photocurrent) due to poor conductivity, recombination, and undesirable band structure [2].

Recent approaches to increasing photocurrent and decreasing onset voltage attempt to increase conversion efficiency, absorbing more of the solar spectrum by using organic dye-sensitizers or catalysts. Dye-sensitizers, small-Eg organic semiconductor dopants, enhance

Figure 1: Conceptual illustration of a PEC water splitting system. Light induces electron/hole pair generation, oxidizing water into hydrogen ions and oxygen gas. The generated electrons reduce the hydrogen ions at the photocathode, producing hydrogen gas. We increase PEC efficiency through (1) material choice (optimal band structure for PEC water splitting) and (2) increased charge separation (through electropolymerization of n-type thiophene, limiting minority carrier distances by polymer reorientation).

performance by increasing light absorption; the bulk material provides a stable base for the moisture-sensitive organic material [3]. While effective, the hybrid material is still unable to absorb as much of the solar spectrum as a purely organic, low-Eg material. The addition of hydrogen evolution or oxygen evolution catalysts to water splitting materials is also crucial to increasing photocurrent and decreasing onset voltage. Indeed, some semiconducting materials show little observable photocurrent without catalyst. Only recently have devices with organic material as bulk been realized as effective water splitting photoelectrodes [4].

Though cheaper, thiophene-based materials, such as poly-3hexylthiophene (P3HT), have been used widely in photovoltaic applications, P3HT proves difficult to use as a water splitting material because of its susceptibility to moisture. Previous work using P3HT (Eg=~2.2 eV) attempt to stabilize the material with oxide coatings such as titanium dioxide (TiO2; Eg=3.05 eV). However, this presents a trade-off between stability and performance. Thicker TiO2 passivation layers allow for more stable hydrogen generation (> 3h), but also limit solar absorption and decrease conductivity, lowering photocurrent.

By electroplating polythiophene (PT), we achieve three distinct advantages over P3HT: (1) PT, a more chemically stable derivative of P3HT, is able generate photocurrent with excellent Von without the use of a passivation layer (Figure 1b.1); (2) by electroplating, hydrophilic aromatic groups align to the substrate surface, decreasing minority charge carrier distance and subsequently increasing conductivity (Figure 1b.2); and (3) by electroplating, we achieve a texturized surface, allowing for increased hydrogen production at low

Figure 2: Fabrication of an electropolymerized thiophene-based electrode. First, a seed layer is deposited at 2V and allowed to polymerize at 1.95V for 60 s on a conductive substrate. After the material is electropolymerized, the substrate is glued and electrically insulated on a glass substrate. Finally, electrical contact is established using conductive tape and insulated wiring. The sample is passivated using epoxy to isolate the active material and to ensure that only the active material is in contact with the aqueous solution.

Figure 3: (Left) Photo of fully passivated device using conductive copper tape, insulated wire, and epoxy on a glass substrate. (Right) Photo of 3-electrode electrochemical set-up. Electrochemical measurements (Linear sweep voltammetry) are conducted in 0.5M H2SO4 vs. Ag/AgCl with a platinum wire counter electrode.

Table 1: Parametric variation of spin speed, concentration, solvent, and substrate, with resulting effects on photocurrent. Due to poor charge separation of spin-coated P3HT, most devices generate no measurable photocurrent. Devices with detectable photocurrent show very large onset voltages.

voltages. Thus, electroplated PT, with favorable band structure for both solar absorption and photocatalysis and increased charge separation efficiency for vertical transfer of electrons, is able to achieve Von=0 V vs. Ag/AgCl, even without use of a surface oxygen or hydrogen evolving catalyst (Figure 1b).

METHODS Device Fabrication

The proposed film fabrication maximizes charge separation by reorienting the polymer relative to the substrate (Figure 1b). By electroplating to avoid the alignment of hydrophobic hexyl groups, we limit travel distance of minority carriers. The device fabrication process is shown in Figure 2.

A reference P3HT electrode is fabricated by spin coating P3HT dissolved in 1,2 dichlorobenzene or chloroform at varying concentrations from 20–30 mg/mL overnight while heated to 70oC. Spin rate was also varied from 400–800 RPM. The substrates were varied from p-type silicon wafer to quartz to isolate the photocatalytic properties. The electrode is then thermally annealed at 200oC for 30 min. to allow for polymer reorientation. The different spin-coating parameters are shown in Table 1, with resulting Von for different fabrication recipes.

Stainless steel is used as the substrate for electropolymerized PT. The conductive substrate is placed in a plating solution composed of 0.1 M tetrabutylammonium hexafluorophosphate and 0.1 M thiophene monomer solution in acetonitrile with a platinum wire as counter electrode. A two-step electropolymerization process was developed: (1) 10-second seed layer deposition with Vbias = 2V vs. Ag/Ag+; and (2) a 60-second film deposition with Vbias = 1.95V vs. Ag/Ag+. Electrical contact is established using copper tape and insulated wire, the substrate is passivated onto a glass substrate, and is immersed in aqueous electrolyte for photoelectrochemical testing (Figure 3).

Electrode Testing

The electrode is tested using a three-electrode setup. Electrodes were tested electrochemically using a Ref 600

Gamry Potentiostat. Linear Sweep Voltammetry was conducted in dark and light conditions in an isolated chamber lined with Thorlabs Blackout Hardboard. The samples were tested vs. Ag/AgCl from -1 V to 0V or 0V to 1V (depending on majority carrier) with a platinum counter electrode in 0.5 M H2SO4 to promote hydrogen evolution (Figure 3). RESULTS AND DISCUSSION

Spin-coated P3HT shows uniform and smooth coverage, with different coloring depending on concentration of P3HT in 1,2 dichlorobenzene/chloroform (Figure 4). Scanning electron microscopy (SEM) images of spin-coated P3HT show uniform thickness and conformal coating of the substrate up to 20µm (Figure 5a). In contrast, EP-PEC devices show relatively uneven coverage due to non-conformal seed layer deposition. SEM images of electropolymerized PT show significant surface texturization (Figure 5b).

Figure 5: Fabrication results: SEM images of (a) spin-coated film, with relatively conformal and uniform coverage; (b) electropolymerized polythiophene film, showing increased texturization of the film, most likely due to uneven polymerization during seed layer formation. Uneven polymerization is hypothesized to be a result of point defects on the surface of the conductive substrate (stainless steel shim).

b) a)

Figure 4: Spin-coated P3HT on quartz and p-type silicon wafers (left) and electropolymerized PT (right). Spin-coated P3HT shows relatively uniform coverage, with varying colors depending on concentration (darker corresponds to higher concentrations). Electropolymerized PT shows relatively uneven coverage due to uneven seed layering.

The devices are tested using a Xenon lamp as the light source with a scan rate of 50 mV/s. Figure 6 shows photocurrent measurements of the spin-coated P3HT device and the EP-PEC devices. While both devices exhibit photocurrent, spin-coated devices showed poor onset voltage (~-0.5 vs. Ag/AgCl), even under parametric variation (see Table 1). The EP-PEC device shows Von < 0 V vs. Ag/AgCl. We hypothesize that superior onset voltage is a result of increased charge separation efficiency. Since orientation of the polymer relative to the substrate dictates the direction of electron flow, as shown in Figure 2, by electropolymerizing thiophene monomer, which precludes alignment of hydrophobic hexyl groups and therefore allows electron flow perpendicular to the substrate surface, we limit travel distance of minority carriers relative to the substrate. By increasing conductivity and thus charge separation, solar energy is more efficiently converted into H2, as shown by the excellent onset voltage (0V vs. Ag/AgCl) for the EP-PEC device as compared to the poor onset voltage of the spin-coated device.

Additionally, the use of electropolymerization allows for increased texturitzation of the electrode device, by taking advantage of nucleation during the seed layer step. Under SEM, we observe a highly texturized surface on top of the flat conductive substrate. As nucleation tends to occur at point defects on the substrate, the surface of electropolymerized PT exhibits increased surface area. Moreover, electroplating has the potential to open pathways toward PT deposition onto porous conductive structures such as activated carbon structures, carbon nanotube forests, and nanowire arrays. By using a highly porous material, one can increase light absorption by using the geometry of the material to harvest reflected light. Porous materials are also able to increase active material while maintaining small minority carrier distances, preserving increased charge separation efficiency.

CONCLUSIONS

This work presents first demonstrations of electropolymerized organic semiconducting materials (PT) for PEC water splitting. As such, these devices show

drastic improvement in Von over conventional spin-coated P3HT devices due to superior charge separation and texturization using the electropolymerization method. As compared to spin-coated P3HT, with Von = ~-5.5 V vs. Ag/AgCl, EP-PEC devices achieve Von = 0 V vs. Ag/AgCl. As such, electroplating has the potential to open pathways toward more robust deposition of organic semiconductor PEC materials onto porous conductive substrates. ACKNOWLEDGEMENTS

This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. 1106400. The authors would like to thank Dr. Jinwoo Bae for invaluable conversation and resources. The authors would also like to thank Winston Ou. REFERENCES [1] Zou, Zhigang, et al. "Direct splitting of water under

visible light irradiation with an oxide semiconductor photocatalyst." Nature, vol. 414, pp. 625-627, 2001.

[2] Zhao, Y., et al. "Oxidatively stable nanoporous silicon photocathodes with enhanced onset voltage for PEC proton reduction." Nano letters, vol. 15, pp. 2517-2525, 2015.

[3] Li, Fusheng, et al. "Organic dye-sensitized tandem photoelectrochemical cell for light driven total water splitting." J. Amer Chem Society, vol. 137, pp 9153-9159, 2015.

[4] Haro, Marta, et al. "Toward Stable Solar Hydrogen Generation Using Organic Photoelectrochemical Cells." J. Phys. Chem. C vol. 119, pp 6488-6494, 2015.

CONTACT

* L.L. tel: +1-510-6435495; lwlin@berkeley.edu

Figure 6: Photocurrent results for P3HT. Poor conductivity results in late onset voltage in spin-coated devices. However, by electropolymerizing onto a conductive substrate, and isolating hydrophilic thiophene groups, we increase conductivity and achieve good onset voltage. Additionally, texturization by electropolymerization improves performance.