Electrical and optical properties of electrospun TiO₂-graphene composite nanofibers and its application as DSSC photo-anodes

Abstract

The present study reports the electrospinning of TiO₂-graphene composite nanofibers to develop conductive nano-fiber mats using polyvinylpyrrolidone as a carrier solution. This carrier solution was sublimated at 450 °C to attain a complete conducting continuous nanofibrous network. It was observed during the annealing that as the graphene content was increased to 1 wt% the continuous fiber morphology was lost. Annealing did not have any impact on the fiber diameter (∼150 nm) or morphology as the graphene content was maintained between 0.0–0.7 wt%. The surface porosity of these samples was found be in the range of 45–48%. The presence of graphene in TiO₂ nanofibers was confirmed using Raman spectroscopy. Photoluminescence spectroscopy showed excitonic intensity to be lower in graphene-TiO₂ samples indicating that the recombination of photo-induced electrons and holes in TiO₂ can be effectively inhibited in the composite nanofibers. Fluorescence spectroscopy was used to confirm this phenomenon where blue and quenched emissions were observed for the electrospun TiO₂ nanofibers and composite fibers, respectively. Conductivity measurements showed the mean specific conductance values obtained for TiO₂-graphene composites to be about two times higher values than that of the electrospun TiO₂ fibers. Assembling these TiO₂-graphene fiber composites as photoanodes in dye sensitized solar cells, an efficiency of 7.6% was attained.

---

Introduction

Carbon materials such as graphite,¹ carbon black,² mesoporous carbon,³ carbon nanotube (CNT),^4,5 and graphene^6–8 have attracted great deal of attention for potential applications in various optoelectronic devices.^9,10 Graphene is a mono-layer structure of two-dimensional graphite. Single layered graphene is highly transparent and absorbs ∼2.3% of incident white light¹¹ making it highly suitable for photo-anode applications in dye sensitized solar cells (DSSCs). Its versatility in terms of high conductivity, mechanical strength and energy gap tunability,¹² compared to different semiconductor oxides like TiO₂, makes it a highly desirable secondary phase in composite form. Inclusion of graphene into semiconductor oxides like TiO₂ has another advantage of high surface area (theoretical value of 2630 m² g⁻¹)¹³ which can be beneficial in improving the interfacial contact in the composite. Previous studies have shown that the direct interaction between TiO₂ and graphene can be used to prevent the re-aggregation of the exfoliated sheets of graphene.¹⁴ This phenomenon has been mainly attributed to the redox ability of the excited TiO₂ phase. Owing to the observation of charge injection from excited TiO₂ to restore the conjugated π-bonded carbon, it can be expected that the presence of graphene in such composite can enhance the electron-accepting ability of grapheme, thereby enhancing the electron transport and thus impede the charge recombination of excited TiO₂.¹⁵ These properties can be beneficial for the applications of TiO₂ in solar cells, especially in dye-sensitized solar cells (DSSCs). There have been several reports of incorporating graphene into TiO₂ using several routes like hydrothermal,^16,17 electrospinning,¹⁸ water-phase synthesis,¹⁹ solvothermal reaction,²⁰etc. In the present study, we report electrospinning of different graphene content TiO₂ as continuous nanofibers. The amount of graphene in the composite is varied from 0.0–1.0 wt% to obtain morphologically stable electrospun nanofibrous mats with increasing conductivity. Furthermore, on the application front, we report the first examples of employing these composites as photo-anodes in DSSC applications. The preliminary results show an efficiency of 7.6% for TiO₂-graphene based photoanodes as compared to 6.3% for bare TiO₂ electrospun fibers.

Experimental

All the chemicals used in the present study were of analytical grade obtained from Sigma Aldrich. 0.5 g of polyvinylpyrollidone (PVP) was completely dissolved in 3 ml of methanol solution. 1 ml of glacial acetic acid and titanium isopropoxide was introduced into the PVP solution and completely dissolved under constant stirring at room temperature. Graphene powder (Quantum Corporation, India) in different weight proportion (0.2, 0.5, 0.7 and 1 wt%) was added separately to this solution mixture and rigorously stirred to obtain uniform blend solutions. 5 ml of the each blend solution was taken in a syringe and fed into an electrospinning set up (Zeonics System, India). Electrospinning was performed for 4 h at 28 °C at an applied potential of 12 kV by keeping a tip to collector distance of 14 cm and a flow rate of 1 ml h⁻¹. The relative humidity of the chamber was kept at ∼55%. The electrospinning was done over Al foils with dimensions 8 cm × 8 cm × 0.2 mm. After electrospinning, these samples were annealed at 450 °C for 1 h to sublimate binder PVP from the electrospun mats. Based on the graphene content in the TiO₂, the electrospun samples were designated as G-T-0.2, G-T-0.5, G-T-0.7 and G-T-1 where G, T and the numbers represent graphene, TiO₂ and the weight % of graphene, respectively.

Morphological and structural analyses for the processed nanofibrous mats were done using field emission scanning electron microscopy-energy dispersive X-ray (FESEM-EDX) (JEOL JSM 6490 LA) and transmission electron microscopy (TEM, JEOL, Japan). Phase analyses were performed using X-ray diffractometry (XRD, X'Pert PRO Analytical), Raman spectroscopy (Witec confocal Raman -300 AR) and X-ray photoelectron spectroscopy (XPS, Kratos Analytical, Ultra axis). Image analysis software (Gmdh, USA) was used to determine the diameters and average surface pore sizes of the electrospun fibers.

Viscometer (Brookfield Corporation DV-II + Pro) was used to estimate the viscosity of the polymeric blend solution. The photo-luminescent (PL) spectra of the samples were obtained using spectrophotometer (Horiba Jobin Yvon Fluoroman-4) at an excitation wavelength of 350 nm. The fluorescent microscope imaging of the samples was performed using Olympus BX51 under an excitation wavelength of ∼350 nm.

Scanning electrochemical microscope ((SECM, Nanotech, Munich), using Pt probe (25 μm²) and lithium perchlorate as a conducting medium) was used for measuring the current voltage (I–V) characteristics and subsequently the specific conductance (G, S cm⁻²).

These experimental data were analyzed using a conventional ranking method^21,22 to yield the Weibull parameters, for the complete characterization of the statistical properties of the measured specific conductance. This was done by arranging the results measured at each point from the lowest to the highest. The f^th result in the set of n = 20 data was assigned a cumulative probability of occurrence, P_f, which was calculated with

The measured specific conductance (S cm⁻², (G), and the cumulative probability, P_f, were then analyzed, by using a simple, least-square regression method, according to the alternative form of the well-known two-parameter Weibull distribution equation

where m and Go are the Weibull modulus and the mean specific conductance, respectively.

As an application, G-T-0.7 electrospun fibers were employed as photo-anodes in the DSSC. Prior to the DSSC assembly, photo-anodes comprising of an electrospun TiO₂-graphene layer on FTO (fluorine doped tin oxide) glasses were soaked in 0.3 mM ruthenium(II) dye (N719, Solaronix) in an acetonitrile and ter-butanol solution for 12 h and coupled with the platinized FTO counter electrode. A paraffin spacer was introduced between the two electrodes to prevent the occurrence of a short-circuit. Physical clamping of these two electrodes was done and an electrolyte comprising of 0.01 M lithium iodide, 0.001 M iodine, 0.6 M butyl methyl imadizolium iodide and 0.5 M tertiary butyl pyridine was locally introduced between the electrodes. Post this procedure, the internal spacing between electrodes was sealed off to prevent any electrolyte leakage. The current–voltage (I–V) characteristics of the assembled DSSCs were measured under an illumination of one sun using a solar simulator (Newport, Oriel class A). Keithley 2400 digital source meter was used to measure the photo-current (Isc) and open-circuit Voltage (Voc) under an applied external potential scan for an exposed area of 0.2 cm².

Results and discussion

In the present study it was observed that the mixing of PVP with different proportions of TiO₂ and graphene resulted in viscous solutions suitable for the electrospinning process. The viscosity of the blend solutions varied from 2.6–3.2 Pa·s as the graphene content increased from 0.0—1.0 wt%.

Fig. 1(a–e) shows the SEM images of the annealed electrospun TiO₂ fibers with different graphene content. No charring of the fibers occurred during the sublimation of PVP (sublimation temperature ∼380 °C^23,24) during annealing at 450 °C. It was also observed that addition of graphene in the given proportions (i.e. 0.0–1.0 wt%) did not have much effect on the TiO₂ fiber morphology or the diameter. However, as the graphene content approached 1 wt% it tends to lose its continuous morphology as seen in Fig. 1(e). This could be mainly attributed to the difference in thermal expansion mismatch between graphene (coefficient of thermal expansion (CTE: −8.0 × 10⁻⁶/C)²⁵ and TiO₂ (CTE: 4.5 × 10⁻⁶/C))^26,27 resulting in the generation of residual stress and breakage. The presence of a graphene phase was not revealed by SEM possibly due to its low phase contrast and concentration in the matrix. Since the maximum content of graphene that could be incorporated in TiO₂ fibers while retaining its continuous fiber morphology during post annealing was for G-T-0.7 samples; these samples were used for further characterization.

Fig. 1 SEM images of annealed electrospun (a) TiO₂ fibers, (b) G-T-0.2, (c) G-T-0.5, (d) G-T-0.7 and (e) G-T-1.0

Fig. 2(a) shows the TEM image of G-T-0.7 samples after annealing. The presence of graphene incorporation into TiO₂ nanofibers was observed (see arrow in Fig. 2(a)). The evidence on crystallinity of the TiO₂ was confirmed from the HR-TEM analysis as shown in Fig. 2(b). HR-TEM analysis showed that the lattice fringes of TiO₂ exhibited an interplanar spacing of ∼0.35 nm and it corresponds to (101) plane of anatase phase. The presence of graphene phase in the G-T-0.7 was confirmed using Raman spectroscopy (Fig. 3). Raman spectrum of G-T-0.7 samples showed the presence of D and G bands of graphene indicating successful incorporation of graphene into the TiO₂ matrix.

Fig. 2 TEM images of annealed electrospun (a) G-T-0.7 samples showing the graphene embedment in TiO₂ fibers (b) HR-TEM analysis of TiO₂ in G-T-0.7 fibers.

Fig. 3 Raman spectrum of G-T-0.7 samples in comparison with electrospun TiO₂ nanofibers.

The XRD analysis confirmed the presence anatase TiO₂ phase (Fig. 4) in the composite. XPS studies were employed to determine the elemental composition of the electrospun samples (Fig. 5). Table 1 shows the elemental atomic concentration of Ti, O and C present in G-T-0.7 and as prepared TiO₂ fiber samples. Interestingly, carbon was detected in both samples indicating that trace amount of carbon was left behind during sublimation of PVP. However, it is anticipated that this carbonization effect would have been similar in both samples as they were subjected to similar annealing conditions. G-T-0.7 samples exhibited carbon content to be ∼15% higher than as compare to as-prepared TiO₂ fiber samples, showing incorporation of grapheme in TiO₂ matrix.

Fig. 4 The XRD analysis of G-T-0.7 samples.

Fig. 5 The XPS analysis of TiO₂ nanofibers and G-T-0.7 samples.

Table 1 XPS analysis showing the elemental atomic concentration of G-T-0.7 and as-prepared TiO₂ fiber samples

Samples	C 1s	Ti 2p	O 1s
TiO2 fiber	13.46	41.05	45.49
G-T-0.7	28.46	31.70	39.81

---

Average diameter of the non annealed G-T-0.7 fibers was measured as ∼220 nm and exhibited a normal distribution curve from 50 to 900 nm and while annealed fibers was measured as ∼150 nm (Fig. 6a & b) It was found that ∼75% of the annealed fibers showed diameter size below 150 nm. The average surface pore size values calculated was found to be higher for the annealed G-T-0.7 samples (9 ± 4.72 μm) when compared to the non annealed G-T-0.7 samples (6 ± 3.37 μm) (Fig. 7a & b). The surface porosity measurements using an image analysis model²⁸ from SEM images revealed values of ∼45 ± 4% and ∼48 ± 5% before and after annealing G-T-0.7 samples respectively.

Fig. 6 Plot of fiber diameter distribution of G-T-0.7 samples (a) before and (b) after annealing.

Fig. 7 Plot of surface pore size distribution of G-T-0.7 samples (a) before and (b) after annealing.

Fig. 8 shows the PL spectra of the as electrospun TiO₂ and the G-T-0.7 excited at 350 nm. Both the samples showed similar PL pattern with strong emission band at 467 nm and 510 nm respectively. However in G-T-0.7 samples, the excitonic PL intensity was found to be lower which indicated that the recombination of photo-induced electrons and holes in TiO₂ can be effectively inhibited in the composite nanofibers. The inhibition effect can be explained from the view of stepwise structure of energy levels constructed in G-T composite, as shown in Fig. 9. The conduction band of TiO₂ is −4.2 eV and valence band is −7.4 eV (vs vacuum)²⁹ while graphene is a zero band gap material.¹⁰ Such energy levels can be beneficial for photo-induced electrons to transfer from the TiO₂ conduction band to the graphene, which could efficiently separate the photo-induced electrons and prevent charge recombination in the electron-transfer process, this phenomenon can be beneficial for photovoltaics. (Fig. 10a & b) shows the fluorescent microscope imaging of TiO₂ and G-T-0.7 samples respectively, for a scan area of 0.5 cm². These images confirm the blue and quenched emissions observed from the PL studies for as electrospun TiO₂ and G-T-0.7 samples respectively.

Fig. 8 PL spectra of the as electrospun TiO₂ and the G-T-0.7 excited at 350 nm.

Fig. 9 Schematic diagram of the energy levels of TiO₂ and graphene.

Fig. 10 Fluorescent microscope imaging of a) TiO₂ and b) G-T-0.7 samples for a scan area of 0.5 cm².

In the present study, SECM was used to investigate the charge transfer mechanisms across the TiO₂-graphene interfaces. (Fig. 11a & b) shows the spatial mapping of the surface current in electrospun TiO₂ and G-T-0.7 samples by keeping the distance between the probe and the conducting surface constant. It was seen that for a scanned area of 100 μm², the topography established by the current peaks in G-T-0.7 samples was more pronounced than as electrospun TiO₂ on a given current scale.

Fig. 11 Spatial mapping of the surface current in annealed a) electrospun TiO₂ and b) G-T-0.7 samples.

Fig. 12a shows the I–V curves characterized in SECM mode. Fig. 12a readily shows quantitatively that the graphene inclusion increased conductivity of TiO₂. However it should be noted that the SECM is a localized representation (scan area: 100 μm²) of the electrical characteristics of the sample surface. Thus I–V values at numerous points were recorded on different samples (4 samples × 5 measurements) of electrospun TiO₂ and G-T-0.7 and plotted as Weibull curves (Fig. 12b) to get a wider perspective of the electrical nature of these fiber mats in both cases. The specific conductance G₀ measured from Weibull plots (Fig. 12b) showed higher values for the G-T-0.7 samples ((0.32 ± 0.01) × 10⁻⁶ S cm⁻²) compare to the electrospun TiO₂ samples ((0.15 ± 0.04) × 10⁻⁶ S cm⁻²). It was interesting to note that the scattering in specific conductance data indicated was similar in both samples, meaning that the dispersion of graphene into the TiO₂ fiber matrix was uniform.

Fig. 12 a) I–V curves characterized in SECM mode for different TiO₂-graphene fiber composites and b) Weibull plot comparison of the specific conductance of annealed electrospun TiO₂ and G-T-0.7 samples.

It is hypothesized that the presence of “metallic” graphene in the TiO₂ matrix can lead to vast changes in the carrier mobility when the charge carriers are in the electronic vicinity of the graphene islands. This is because of the large relativistic mobility of electrons and holes (Dirac fermions) in graphene. Graphene-TiO₂ interface represents an electronic contact between a wide band gap semiconductor and narrow band gap semiconductor. Such interfaces can produce deep depletion of the charge carriers at the interface with substantial band bending. This will effectively produce a low concentration of interface carriers and effectively inhibit the possibility of back injection, thereby making charge flow unidirectional. This phenomenon can be highly desirable for photovoltaic stand point especially when graphene bridges contacts TiO₂ islands.

Fig. 13 compares the photo-current density-voltage curves of the assembled DSSC employing G-T-0.7 and bare TiO₂ fibers as photo-anodes. The open-circuit voltage (Voc), short-circuit photo-current density (Jsc), fill factor (FF) and efficiency (η) for the G-T-0.7 samples were found to be 0.71 V, 16.2 mA cm⁻², 66% and 7.6% as compared to 0.71 V, 13.9 mA cm⁻², 63% and 6.3% for bare TiO₂ fibers, respectively. For both systems, the Voc was found to be similar with a significant improvement in both Jsc values for G-T-0.7 samples resulting in ∼21% increase in η values. The TiO₂-graphene fiber system seems to be promising for DSSC applications, however a detailed study elucidating the role of different graphene content in TiO₂ is required, which is currently being pursued by our group and is beyond the scope of the present study.

Fig. 13 Current–voltage (I–V) characteristics of DSSC using electrospun TiO₂ fibers and G-T-0.7 samples.

Conclusions

The present study reports the electrospinning of conductive TiO₂-graphene composites in combination with PVP as a carrier solution resulting in well-defined and structurally stable fibers. These electrospun fibers were heat treated at 450 °C to obtain a highly porous network. The presence of graphene phase in the TiO₂ nanofibers was confirmed using Raman spectroscopy. The excitonic PL intensity was found to be lower in graphene-TiO₂ samples indicating that the recombination of photo-induced electrons and holes in TiO₂ can be effectively inhibited in the composite nanofibers. The fluorescence spectroscopy images confirmed the blue and quenched emissions observed for the electrospun TiO₂ nanofibers and composite fibers respectively. The Weibull plot showed the mean specific conductance values obtained for TiO₂-graphene composites to be about two times higher values than that of the electrospun TiO₂ fibers. Employing these samples as photoanodes in DSSCs, a significant improvement in Jsc values was observed resulting in η = 7.6%.

Acknowledgements

Department of Science and Technology (DST), Government of India is gratefully acknowledged for their financial support under the NATAG program monitored by Dr G. Sundararajan.

References

1. X. Hu, K. Huang, D. Fang and S. Liu, Mater. Sci. Eng., B, 2011, 176, 431–435.
2. Y. S. Park and H. K. Kim, Curr. Appl. Phys., 2011, 11, 989–994.
3. E. Ramasamy and J. Lee, Carbon, 2010, 48, 3715–3720.
4. J. Zhang, X. Li, W. Guo, T. Hreid, J. Hou, H. Su and Z. Yuan, Electrochim. Acta, 2011, 56, 3147–3152.
5. C. T. Hsieh, B. H. Yang and J. Y. Lin, Carbon, 2011, 49, 3092–3097.
6. M. Y. Yen, M. C. Hsiao, S. H. Liao, P. I. Liu, H. M. Tsai, C. C. M. Ma, N. W. Pu and M. D. Ger, Carbon, 2011, 49, 3597–3606.
7. W. Hong, Y. Xu, G. Lu, C. Li and G. Shi, Electrochem. Commun., 2008, 10, 1555–1558.
8. L. Wan, S. Wang, X. Wang, B. Dong, Z. Xu, X. Zhang, B. Yang, S. Peng, J. Wang and C. Xu, Solid State Sci., 2011, 13, 468–475.
9. H. Zhu, J. Wei, K. Wang and D. Wu, Sol. Energy Mater. Sol. Cells, 2009, 93, 1461–1470.
10. T. H. Tsai, S. C. Chiou and S. M. Chen, Int. J. Electrochem. Sci., 2011, 6, 3333–3343.
11. R. R. Nair, P. Blake, A. N. Grigorenko, K. S. Novoselov, T. J. Booth, T. Stauber, N. M. R. Peres and A. K. Geim, Science, 2008, 320, 1308.
12. M. Y. Han, B. Ozyilmaz, Y. B. Zhang and P. Kim, Phys. Rev. Lett., 2007, 98, 206805–206808.
13. Y. Shao, J. Wang, H. Wu, J. Liu, I. A. Aksay and Y. Lin, Electroanalysis, 2010, 22, 1027–1036.
14. G. Williams, B. Seger and P. V. Kamat, ACS Nano, 2008, 2, 1487–1491.
15. J. Liu, J. Tang and J. J. Gooding, J. Mater. Chem., 2012, 22, 12435–12452.
16. J. Shen, B. Yan, M. Shi, H. Ma, N. Li and M. Ye, J. Mater. Chem., 2011, 21, 3415–3421.
17. J. Shen, M. Shi, B. Yan, H. Ma, N. Li and M. Ye, Nano Res., 2011, 4, 795–806.
18. V. J. Babu, A. S. Nair, P. Zhu and S. Ramakrishna, Mater. Lett., 2011, 65, 3064–3068.
19. C. Zhu, S. Guo, P. Wang, X. Li, Y. Fang, Y. Zhai and S. Dong, Chem. Commun., 2010, 46, 7148–7150.
20. K. Zhou, Y. Zhu, X. Yang, X. Jiang and C. Li, New J. Chem., 2011, 35, 353–359.
21. K. Sujith, A. M. Asha, P. Anjali, N. Sivakumar, K. R.V. Subramanian, S. V. Nair and A. Balakrishnan, Mater. Lett., 2012, 67, 376–378.
22. A. M. Asha, K. Sujith, P. Anjali, N. Sivakumar, K. R.V. Subramanian, S. V. Nair and A. Balakrishnan, J. Nanosci. Nanotechnol., 2012, 12, 1–6.
23. M. Chen, H. Qu, J. Zhu, Z. Luo, A. Khasanov, A.S. Kucknoor, N. Haldolaarachchige, D. P. Young, S. Wei and Z. Guo, Polymer, 2012, 53, 4501–4511.
24. C. Lau and Y. Mi, Polymer, 2002, 43, 823–829.
25. D. Yoon, Y. W. Son and H. Cheong, Nano Lett., 2011, 11, 3227–3231.
26. D. R. Hummer, P. J. Heaney and J. E. Post, Powder Diffr., 2007, 22, 352–357.
27. S. Liu, B. Liu, K. Nakata, T. Ochiai, T. Murakami and A. Fujishima, J. Nanomater., 2012, 1–5.
28. A. Mikrajuddin and R. Khairurrijal, Indonesian J. Phys., 2009, 20, 37–40.
29. G. Zhu, L. K. Pan, T. Xu, Q. F. Zhao and Z. Sun, J. Alloys Compd., 2011, 509, 7814–7818.

---