Carbon-doped freestanding TiO₂ nanotube arrays in dye-sensitized solar cells

Abstract

Dye-sensitized solar cells (DSSCs) were fabricated with both closed- and open-ended freestanding TiO₂ nanotube arrays that were doped with carbon via chemical vapor deposition to improve their electron transport properties. The energy conversion efficiencies of DSSCs increased from 5.07% to 6.21% after doping with a small amount of carbon, i.e., an enhancement of 22.4%. This suggests that the π–π conjugation introduced by carbon doping improved the efficiency of electron transport. However, energy conversion efficiencies of DSSCs with a large amount of carbon doping on the TiO₂ nanotube arrays decreased from 5.07% to 2.87%, with a lower short-circuit current, open circuit voltage, and fill factor because of the reduced level of dye adsorption on the TiO₂ nanotube arrays.

---

Introduction

Dye-sensitized solar cells (DSSCs) have been extensively researched because of their high energy conversion efficiencies and low cost.^1–3 Recently, mesoporous TiO₂ nanoparticle (NP) films have been widely used for DSSCs. However, their efficiencies have been limited because of grain boundaries, defects, and innumerous trapping sites that can cause charge recombination and low electron mobility.

TiO₂ nanotubes have great potential to overcome the problems of TiO₂ NP films because their unique structure enhances electron transport and charge separation by forging direct pathways and accelerating charge transfer between interfaces.^4–15 TiO₂ nanotube arrays obtained from electrochemical anodization are good candidates for rapid electron-transport of materials with reduced charge recombination because of their highly ordered vertically oriented tubular structures, which gives them a number of advantages over TiO₂ NP films.^16–19 Although DSSCs based on TiO₂ nanotube arrays have high potential for improved energy conversion efficiencies, to date DSSCs based on closed-ended TiO₂ nanotube arrays have shown lower energy conversion efficiencies than DSSCs based on TiO₂ NP films. Recently, we demonstrated that open-ended TiO₂ nanotube arrays where the barrier layers have been removed provide higher energy conversion efficiencies for DSSCs.¹¹

Carbon doping has been considered to be a promising means of improving charge separation and electron transport in solar cells owing to the good electrical properties imparted by π–π conjugation. The Han and Li groups have both reported that carbon nanotube composite films with TiO₂ NPs showed enhanced electron transport,^20,21 and the Gao and Lee groups reported that graphene composite films on photoelectrodes showed improved energy conversion efficiency.^22,23 However, it is not easy to incorporate carbon materials into well-ordered vertically aligned TiO₂ nanotube arrays except via the direct growth of carbon materials.

Here, we report that carbon-doped TiO₂ nanotube arrays were used as photoanodes for DSSCs to increase energy conversion efficiency by improving charge separation and electron transport. The carbon was doped on either closed- or open-ended TiO₂ nanotube arrays by chemical vapor deposition (CVD). The performances of the DSSCs fabricated with the closed- and open-ended TiO₂ nanotube arrays with and without carbon doping were compared to study the influence of carbon doping on the energy conversion efficiency of DSSCs.

Experimental

Preparation of closed- or open-ended TiO₂ nanotube arrays

TiO₂ nanotube arrays were fabricated by anodizing titanium foils (Alfa Aesar, 99.7% purity, 2.5 cm × 4.0 cm × 320 μm). Carbon rods were used as the cathode. The electrolyte was composed of 0.8 wt% NH₄F and 2 vol% H₂O in ethylene glycol at 25 °C. A 60 V DC potential was applied to the anodization system for 2 h. The TiO₂ nanotube arrays formed on the Ti foil were annealed at 450 °C for 1 h. After the sample was cooled to room temperature, a second anodization was performed on the annealed samples at 30 V for 10 min. The sample was then placed in 10% H₂O₂ for 24 h to detach the nanotubes from the Ti foil. The process of opening the bottom layer (the closed-ended tip) of the TiO₂ nanotube arrays was performed by ion milling with Ar⁺ bombardment.

Synthesis of carbon-doped closed- or open-ended TiO₂ nanotube arrays on FTO glass

Fluorine doped tin oxide (FTO) glass was used as the electrode and the supporting substrate for TiO₂ nanotube arrays. Titanium diisopropoxide bis(acetylacetonate) (5 wt% in n-butanol) was injected and spin-coated onto the washed FTO glass. The samples were annealed at 450 °C for 1 h to form a compact TiO₂ blocking layer. To attach TiO₂ nanotube arrays, TiO₂ paste (Solaronix, T/SP) was applied to the FTO glass using the doctor-blade method. The closed- or open-ended TiO₂ nanotube arrays were then introduced into the TiO₂ paste. The samples were annealed at 450 °C for 30 min, then placed in a quartz furnace tube that was filled with nitrogen (200 sccm). At the required times, hydrogen and ethylene were injected into the tube at 450 °C and flow rates of 30 and 40 sccm, respectively.

Fabrication of dye-sensitized solar cells

Dye molecules were adsorbed onto photoanodes consisting of carbon-doped or non-doped TiO₂ nanotube arrays (open-ended or closed-ended) using 0.5 mM solutions of N719 ((Bu₄N)₂Ru(dobpyH)₂(NCS)₂, Solaronix) in absolute ethanol at 50 °C for 8 h. The samples were washed several times with absolute ethanol by immersing for 10 min to remove extra physisorbed dye molecules from the TiO₂ nanotube arrays. Counter electrodes were fabricated by drop-casting chloroplatinic acid (H₂PtCl₆) in ethanol onto the FTO glass. A 60 μm-thick hot-melt Surlyn spacer (Solaronix) was placed between the photoanode and the counter electrode, and the electrolyte was composed of 0.7 M 1-butyl-3-methyl-imidazolium iodide (BMII), 0.03 M I₂, 0.1 M guanidium thiocyanate (GSCN), and 0.5 M 4-tert-butyl pyridine (TBP) in a mixture of acetonitrile and valeronitrile (85 : 15 v/v). The electrolyte was injected into the gap between the electrodes, and the cells were sealed to prevent the electrolyte from leaking.

Characterization of dye-sensitized solar cells

The structures of the closed- and open-ended TiO₂ nanotube arrays with and without carbon doping were analyzed using a field emission scanning electron microscope (FE-SEM, JSM-6700F, JEOL Ltd). Current density–voltage measurements were performed using an electrometer (Keithley 2400) under AM 1.5 illumination (100 mW cm⁻²) provided by a solar simulator (1 kW xenon with AM 1.5 filter, PEC-L01, Peccel Technologies). Electrochemical impedance spectroscopy (EIS) was performed using a potentiostat (Solartron 1287) equipped with a frequency response analyzer (Solartron 1260) between 10⁻² and 10⁶ Hz, and were analyzed using Z-View software (Solartron Analytical) with the aid of an appropriate equivalent circuit. The applied bias voltage and the AC amplitude were set at the open circuit voltage (V_oc) of the DSSCs and 10 mV, respectively. The impedance measurements were carried out at open-circuit potential under one sun of AM 1.5 light illumination.

Results and discussion

Fig. 1 shows the fabrication of DSSCs containing carbon-doped TiO₂ nanotube arrays. First, TiO₂ nanotube arrays were prepared by anodization, as shown in Fig. 1A(a). To yield crystalline TiO₂ nanotube arrays, the substrate was annealed and then a second anodization step was performed to prepare freestanding TiO₂ nanotube arrays, as shown in Fig. 1A(b). The resulting freestanding TiO₂ nanotube arrays were transferred onto FTO glass with TiO₂ paste that was coated onto the FTO glass using a doctor blade and then sintered under ambient conditions, as shown in Fig. 1A(c). Carbon was then synthesized by CVD using ethylene gas, with the coating time being used to control the level of doping. The dye (N719) was then adsorbed onto the TiO₂ nanotube arrays. DSSCs were fabricated by assembly of the working electrode (TiO₂ nanotube arrays) and the counter electrode (Pt), as shown in Fig. 1B.

Fig. 1 Overall scheme showing fabrication of DSSCs based on TiO₂ nanotube arrays. (A) (a) First anodization and annealing step, (b) second anodization and detachment from Ti foil to yield freestanding TiO₂ nanotube arrays, (c) transfer of freestanding TiO₂ nanotube arrays to FTO glass, and (d) carbon doping on TiO₂ nanotube arrays by CVD. (B) Structure of a DSSC.

Fig. 2 shows FE-SEM images of the top, side, and bottom views of freestanding TiO₂ nanotube arrays. The upper pore size was approximately 100 nm, as can be seen in Fig. 2a, and the length of the freestanding TiO₂ nanotube arrays was around 20 μm, as shown in Fig. 2b. Fig. 2c shows the bottom layer of the TiO₂ nanotube arrays which was detached from the Ti foil after anodization, which resulted in closed-ended TiO₂ nanotube arrays. The bottom layer of the TiO₂ nanotube arrays could be removed by using a physical etching method such as ion milling,¹¹ with the bottom pore size of the resulting open-ended TiO₂ nanotube arrays being approximately 50 nm, as shown in Fig. 2d. The DSSCs were prepared with closed- or open-ended TiO₂ nanotube arrays that had been doped with carbon materials to improve electron transport.

Fig. 2 FE-SEM images of TiO₂ nanotube arrays from (a) top, (b) side, and (c) bottom views. The bottoms of the tubes after ion milling are shown in (d).

Raman spectra of TiO₂ nanotube arrays and carbon-doped TiO₂ nanotube arrays between 380 and 1800 cm⁻¹, which can be used to characterize the carbon and TiO₂ structures present, are shown in Fig. 3. The B1g (395 cm⁻¹), A1g (515 cm⁻¹) and Eg (635 cm⁻¹) peaks in the Raman spectrum shown in Fig. 3a indicate that the TiO₂ was in the form of an anatase structure.²⁴ After carbon materials were synthesized by CVD, the anatase TiO₂ peaks were maintained and some narrow bands were formed. The G band at 1586 cm⁻¹ represents graphite, while the D band at 1370 cm⁻¹ was caused by the disorderly network of sp² and sp³ carbon clusters. The intensity ratio of D and G bands was about 0.75, suggesting the presence of π–π conjugation from the sp² sites on the carbon materials. This would lead to more efficient electron transport in DSSCs.

Fig. 3 Raman spectra of (a) TiO₂ nanotube arrays and (b) carbon-doped TiO₂ nanotube arrays.

Fig. 4 presents the current density–voltage curves of three different DSSCs based on closed-ended TiO₂ nanotube arrays that were attached to FTO glass using TiO₂ paste, as measured under AM 1.5 sunlight. The open circuit voltages (V_oc), short-circuit currents (J_sc), fill factors (ff), and energy conversion efficiencies (η) of the DSSCs based on closed-ended TiO₂ nanotube arrays are summarized in Table 1. For the DSSC using undoped TiO₂, the energy conversion efficiency was 4.19%, while when carbon doping was performed on the TiO₂via CVD for 30 s, the energy conversion efficiency improved significantly from 4.19% to 5.16%, corresponding to a 23.15% enhancement. However, when carbon doping was performed for 60 s, the energy conversion efficiency decreased from 4.19% to 2.27%, corresponding to 45.82% suppression in the energy conversion efficiency. The dye loading of the closed-ended TiO₂ nanotube array was 145 nmol cm⁻², which was higher than that of the cells with TiO₂ that had been carbon doped for 30 s and 60 s. Compared with the DSSC using undoped TiO₂ nanotube arrays, the J_sc of the cell using TiO₂ that had been carbon doped for 30 s increased from 7.46 mA cm⁻² to 9.69 mA cm⁻², corresponding to a 28.89% enhancement in spite of the slightly reduced dye loading of 140 nmol cm⁻², indicating that electron transport was improved by carbon doping. However, the ff decreased slightly due to the increase in recombination caused by π–π conjugation in the carbon materials. When TiO₂ was doped with a large amount of carbon, the J_sc decreased remarkably from 7.46 mA cm⁻² to 5.08 mA cm⁻² due to the reduced dye absorption. When larger amounts of carbon materials are doped on closed-ended TiO₂ nanotube arrays, the dye loading decreased to 89 nmol cm⁻², indicating that the binding between the dye and the surface of the TiO₂ nanotube arrays was disturbed. Moreover, the ff also decreased, indicating that there was a large degree of π–π conjugation in the carbon materials.

Fig. 4 Current density–voltage curves of DSSCs with closed-ended TiO₂ nanotube arrays (a) without carbon doping, and with carbon doping via CVD for (b) 30 s and (c) 60 s.

Table 1 Photovoltaic properties of DSSCs with and without carbon doping on closed-ended TiO₂ nanotube arrays

DSSCs based on closed-ended TiO2 nanotube arrays	J sc (mA cm^−2)	V oc (V)	ff	η (%)	Dye loading (nmol cm^−2)
No carbon doping	7.46	0.77	0.73	4.19	145
Carbon doping for 30 s	9.69	0.74	0.72	5.16	140
Carbon doping for 60 s	5.08	0.66	0.67	2.27	89

---

Fig. 5 presents the current density–voltage curves of three different DSSCs based on open-ended TiO₂ nanotube arrays measured under AM 1.5 sunlight. The open circuit voltages (V_oc), short-circuit currents (J_sc), fill factors (ff), and energy conversion efficiencies (η) of the DSSCs based on open-ended TiO₂ nanotube arrays are summarized in Table 2. The energy conversion efficiency of DSSCs based on open-ended TiO₂ nanotube arrays improved from 4.19% to 5.07% compared with DSSCs based on closed-ended TiO₂ nanotube arrays, due to the removal of the barrier layer of TiO₂ nanotube arrays as previously reported by our group.²⁵ When carbon doping was performed by CVD for 30 s, the energy conversion efficiency improved from 5.07% to 6.21%, corresponding to a 22.48% enhancement, whereas when CVD was performed for 60 s, the energy conversion efficiency dropped from 5.07% to 2.87%, corresponding to a 43.39% suppression. The trend in the energy conversion efficiencies of the DSSCs based on open-ended TiO₂ nanotube arrays is similar to that seen with the closed-ended TiO₂ nanotube arrays. However, the energy conversion efficiencies of the open-ended TiO₂ nanotube arrays were typically higher than those of the closed-ended TiO₂ nanotube arrays due to the barrier layer effect. The J_sc of the cell using TiO₂ that was carbon doped for 30 s was 11.49 mA cm⁻², a 29.39% enhancement compared to the undoped material (8.88 mA cm⁻²). As this occurred in spite of the lower dye loading of the doped cell, it supports the idea that electron transport was improved by π–π conjugation across a small amount of carbon materials. On the other hand, a large amount of carbon materials not only diminishes the dye adsorption on the TiO₂ nanotube arrays but also decreases the ff of the DSSCs.

Fig. 5 Current density–voltage curves of DSSCs with open-ended TiO₂ nanotube arrays (a) without carbon doping and with carbon doping via CVD for (b) 30 s and (c) 60 s.

Table 2 Photovoltaic properties of DSSCs with and without carbon doping on open-ended TiO₂ nanotube arrays

DSSCs based on open-ended TiO2 nanotube arrays	J sc (mA cm^−2)	V oc (V)	ff	η (%)	Dye loading (nmol cm^−2)
No carbon doping	8.88	0.77	0.74	5.07	151
Carbon doping for 30 s	11.49	0.75	0.72	6.21	143
Carbon doping for 60 s	6.43	0.66	0.67	2.87	93

---

Fig. 6 shows the EIS spectra of DSSCs with and without carbon materials in open-ended TiO₂ nanotube arrays across a frequency range from 10⁻² to 10⁶ Hz. The applied bias voltage and AC amplitude were set at the open circuit voltage (V_oc) of the DSSCs and 10 mV, respectively. Each spectrum contains two semicircles: a small one at high frequency, and a large one at low frequency. The smaller semicircle is assigned to the parallel combination of the resistance and capacitance at the Pt–FTO/electrolyte and the FTO/TiO₂ interfaces, and the larger semicircle is associated with the resistance and capacitance at the dye-absorbed TiO₂/electrolyte interface and the transport resistance. The data were analyzed using an equivalent circuit, and the fit parameters are listed in Table 3. The Ohmic series resistance (R_s) is caused by the sheet resistance, and corresponds to the value of the x-axis where the first semicircle begins on the left-hand side. The values of R_s were similar with and without carbon doping on the TiO₂ nanotube arrays, meaning that carbon doping did not affect the sheet resistance of FTO or the current collector. R₁ is given by the sum of the interfacial resistances of FTO/TiO₂ and Pt/electrolyte, while R₂ is given by the sum of the charge transfer resistance at the dye-adsorbed TiO₂/electrolyte interface and the transport resistance in the TiO₂ films. DSSCs without carbon doping on the TiO₂ nanotube arrays showed higher values of R₁ (4.013 Ω) and R₂ (34.05 Ω) than were found for DSSCs that had been carbon doped for 30 s, where R₁ and R₂ were 2.866 Ω and 27.73 Ω, respectively. This shows that a small amount of carbon decreases the resistance of TiO₂ nanotube arrays due to π–π conjugation. Moreover, the π–π conjugation affects the conduction band of the TiO₂ nanotube arrays, decreasing the V_oc from 0.77 V to 0.75 V. As shown in Table 2, when carbon doping was performed for 60 s, the values of R₁ (4.023 Ω) and R₂ (47.55 Ω) were higher than in undoped DSSCs due to the lower dye loading level. The large amount of carbon materials on the TiO₂ nanotube arrays disrupts the dye adsorption on the TiO₂ nanotube arrays, meaning that charge cannot be generated. This increases the resistance at the dye-adsorbed TiO₂/electrolyte interface and the transport resistance in the TiO₂ films. As for the other doped material, the π–π conjugation affects the conduction band of the TiO₂ nanotube arrays, greatly decreasing the V_oc from 0.77 V to 0.66 V. The V_oc is defined as the difference between the energy levels of the semiconductor and the electrolyte in a solar cell. When TiO₂ nanotube arrays are doped with carbon, the conduction band edge of the TiO₂ nanotube arrays shifts downward, which is referred to as “band bending”. For this reason, too much carbon doping can decrease the V_oc.

Fig. 6 Impedance spectra of DSSCs with open-ended TiO₂ nanotube arrays (a) without carbon doping and with carbon doping via CVD for (b) 30 s and (c) 60 s.

Table 3 Fit results of impedance spectra for DSSCs with and without carbon doping on open-ended TiO₂ nanotube arrays

DSSCs based on open-ended TiO2 nanotube arrays	R s (Ω)	R 1 (Ω)	CPE1 (F)	R 2 (Ω)	CPE2 (F)
No carbon doping	13.46	4.013	7.878 × 10^−6	34.05	2.325 × 10^−3
Carbon doping for 30 s	13.70	2.866	1.268 × 10^−6	27.73	2.872 × 10^−3
Carbon doping for 60 s	13.42	4.023	5.977 × 10^−6	47.55	2.568 × 10^−3

---

Conclusion

We synthesized carbon materials on closed- and open-ended TiO₂ nanotube arrays for use in DSSCs in order to improve electron transport. The energy conversion efficiencies of DSSCs using TiO₂ nanotube arrays that had been carbon doped for 30 s were better than those of cells using TiO₂ that had not been carbon doped. However, the energy conversion efficiencies of DSSCs based on TiO₂ nanotube arrays that had been carbon doped for 60 s were lower than the control cell. Based on the impedance data and the amount of dye adsorption, it can be seen that a smaller amount of carbon materials was more efficient in improving electron transport due to π–π conjugation of the carbon sp² sites, while a large amount of carbon materials not only decreases dye adsorption on the TiO₂ nanotube arrays but also decreases the ff of the DSSCs. We expect that our research can also be applied to photovoltaics such as perovskite solar cells, solid-state dye-sensitized solar cells, or hybrid solar cells.

Acknowledgements

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2016-A002-0048) and by the Bio & Medical Technology Development Program of the National Research Foundation (NRF) & funded by the Korean government (MSIP&MOHW) (2016-A423-0045).

References

1. B. Oregan and M. Gratzel, Nature, 1991, 353, 737–740.
2. S. Ito, T. N. Murakami, P. Comte, P. Liska, C. Gratzel, M. K. Nazeeruddin and M. Gratzel, Thin Solid Films, 2008, 516, 4613–4619.
3. M. Gratzel, Nature, 2001, 414, 338–344.
4. A. Ghicov, S. P. Albu, R. Hahn, D. Kim, T. Stergiopoulos, J. Kunze, C. A. Schiller, P. Falaras and P. Schmuki, Chem. – Asian J., 2009, 4, 520–525.
5. P. Roy, D. Kim, I. Paramasivam and P. Schmuki, Electrochem. Commun., 2009, 11, 1001–1004.
6. J. H. Park, T. W. Lee and M. G. Kang, Chem. Commun., 2008, 2867–2869.
7. Q. W. Chen and D. S. Xu, J. Phys. Chem. C, 2009, 113, 6310–6314.
8. C. J. Lin, W. Y. Yu and S. H. Chien, J. Mater. Chem., 2010, 20, 1073–1077.
9. L. L. Li, Y. J. Chen, H. P. Wu, N. S. Wang and E. W. G. Diau, Energy Environ. Sci., 2011, 4, 3420–3425.
10. C. Rho and J. S. Suh, Chem. Phys. Lett., 2011, 513, 108–111.
11. C. Rho, J. H. Min and J. S. Suh, J. Phys. Chem. C, 2012, 116, 7213–7218.
12. C. M. Ruan, M. Paulose, O. K. Varghese, G. K. Mor and C. A. Grimes, J. Phys. Chem. B, 2005, 109, 15754–15759.
13. S. So, K. Lee and P. Schmuki, Chem. – Eur. J., 2013, 19, 2966–2970.
14. W.-Y. Rho, H. Jeon, H.-S. Kim, W.-J. Chung, J. S. Suh and B.-H. Jun, J. Nanomater., 2015, 247689.
15. J. M. Macak, H. Tsuchiya, A. Ghicov, K. Yasuda, R. Hahn, S. Bauer and P. Schmuki, Curr. Opin. Solid State Mater. Sci., 2007, 11, 3–18.
16. K. Shankar, G. K. Mor, H. E. Prakasam, S. Yoriya, M. Paulose, O. K. Varghese and C. A. Grimes, Nanotechnology, 2007, 18, 065707.
17. G. K. Mor, O. K. Varghese, M. Paulose, K. Shankar and C. A. Grimes, Sol. Energy Mater. Sol. Cells, 2006, 90, 2011–2075.
18. J. Yan and F. Zhou, J. Mater. Chem., 2011, 21, 9406–9418.
19. J. R. Jennings, A. Ghicov, L. M. Peter, P. Schmuki and A. B. Walker, J. Am. Chem. Soc., 2008, 130, 13364–13372.
20. S. Muduli, W. Lee, V. Dhas, S. Mujawar, M. Dubey, K. Vijayamohanan, S.-H. Han and S. Ogale, ACS Appl. Mater. Interfaces, 2009, 1, 2030–2035.
21. J. Liu, Y.-T. Kuo, K. J. Klabunde, C. Rochford, J. Wu and J. Li, ACS Appl. Mater. Interfaces, 2009, 1, 1645–1649.
22. S. R. Sun, L. Gao and Y. Q. Liu, Appl. Phys. Lett., 2010, 96, 3113.
23. Y. B. Tang, C. S. Lee, J. Xu, Z. T. Liu, Z. H. Chen, Z. B. He, Y. L. Cao, G. D. Yuan, H. S. Song, L. M. Chen, L. B. Luo, H. M. Cheng, W. J. Zhang, I. Bello and S. T. Lee, ACS Nano, 2010, 4, 3482–3488.
24. J. Yan, G. Wu, N. Guan, L. Li, Z. Li and X. Cao, Phys. Chem. Chem. Phys., 2013, 15, 10978–10988.
25. W.-Y. Rho, M.-H. Chun, H.-S. Kim, Y.-B. Hahn, J. S. Suh and B.-H. Jun, Bull. Korean Chem. Soc., 2014, 35, 1165–1168.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6nj02615j

‡ These authors contributed equally to the work.

---