Photoresponse properties based on CdS nanoparticles deposited on multi-walled carbon nanotubes

Abstract

The influence of multi-walled carbon nanotube (i.e. MWCNT) architectures for facilitating charge transport in CdS semiconductor films was investigated. The power conversion efficiency (PCE) of the CdS nanoparticle-deposited multi-walled carbon nanotube nanohybrids (i.e. CdS@MWCNT) could be enhanced when the MWCNT was employed as the conducting scaffold in the CdS nanoparticle based photoelectrochemical cell. The CdS nanoparticles were deposited onto the MWCNT surface to improve photoinduced charge separation and transport of carriers (electrons and holes). The PCE of the CdS@MWCNT nanohybrids solar cell was found to be as high as 4.54%, contrasting sharply with the value (1.67%) for pure CdS nanoparticles. The outcome of the present work might provide a new avenue for designing optoelectronic and light-energy conversion devices.

---

Introduction

Inorganic semiconductor quantum dots (QDs) have been introduced as light harvesters for photovoltaic devices and have stimulated extensive research interest.^1–6 CdS, CdSe, and PbS quantum dots offer an alterable band gap that can easily tune the optical absorption by controlling the particle size.^7–9 Moreover, their high extinction coefficients and multiple exciton generation are in favor of enhanced photoresponse of QDs based optoelectronic devices.^10–12 For example, the CdS nanoparticles can combine with inorganic materials, conjugated polymer, carbon nanotubes (CNTs), or g-C₃N₄ to form photovoltaic or photocatalytic devices that exhibit excellent photoelectric properties. (i) The CdS quantum dots can be used to sensitize TiO₂ nanotube-array films that can facilitate the propagation and kinetic separation of photogenerated charges, and enhanced photocurrent and efficiency have also been found.¹³ The photocurrent density of CdSe-sensitized SnO₂ solar cells can reach up to 17.40 mA cm⁻² when the CdS quantum dots are introduced into the solar cells.¹⁴ (ii) The power conversion efficiency (PCE) of a CdS/CuInSe₂ nanostructured heterojunction thin film can be drastically enhanced when the conjugated polyaniline is injected into the solar cells.¹⁵ Our previous studies also shown that the photoelectric response of the CdS/polyaniline (inorganic/conjugated polymers) nanocomposite were improved when iodine (I₂) and ferroferric oxide (Fe₃O₄) were added into the active layer.^16,17 (iii) Kamat et al. showed that a fast electron transfer from excited CdS into the single-walled carbon nanotube (SWCNT) was successfully achieved after the CdS quantum dots were deposited onto the SWCNT surface. The unprecedented high quantum yields for the generation of photocurrents were obtained using CdS nanoparticles/carbon nanotubes as a hybrid system.¹⁸ (iv) J. G. Yu et al. reported that the photocatalytic rates of the CdS nanowires were significantly enhanced in the presence of the g-C₃N₄ shell, and exhibited excellent photostability even though they were irradiated in a nonsacrifical system.¹⁹ However, the charge transfer and redistribution under light excitation is still a major problem for the above mentioned photovoltaic or photocatalytic devices.²⁰ For example, the unfavorable interface between inorganic nanoparticles and conjugated polymer will limit the solar cell efficiency by exciton dissociation. Therefore, for the inorganic semiconductor, the advent of methods for controlling its morphology and size or combining with other conducting scaffold materials can tune the interface states of inorganic semiconductors, and it may well improve the charge transfer and reduce the charge recombination.^21,22 Based on the above discussion, recently, attaching quantum dots to the surface of carbon nanotubes has been considered as a general strategy to construct innovative architectures via structural modification and morphology control.^23–25 A wide range of carbon nanotubes based nanohybrids have shown promising applications in catalysis,²⁶ sensors,²⁷ and optoelectronic devices²⁸ when the inorganic semiconductor quantum dots are directly grown on the surface of carbon nanotubes or are connected to carbon nanotubes.

Previous studies showed that carbon nanotubes could served as electron acceptors from the excited CdS quantum dots, resulting in a detectable photocurrent through the nanotube device.²⁸ Therefore, in this article, we present a simple attaching non-covalent chemical approach for assembling the multi-walled carbon nanotube (MWCNT) with CdS nanoparticles as an inorganic nanocomposite/carbon material (i.e. CdS@MWCNT) to elucidate its photoresponse properties. In addition, the prepared CdS@MWCNT nanohybrids were characterized by scanning electron microscopy (SEM), X-ray diffraction, Raman spectroscopy, and Ultraviolet-Visible spectroscopy (UV-Vis). Moreover, the photocurrent of the CdS@MWCNT nanohybrids solar cell is expected to be enhanced about three times compared with single CdS nanoparticles. The outcome of the present work may provide further information on enhancing the photoresponse of optoelectronic devices.

Experimental

Preparation of the CdS@MWCNT nanohybrids

All reagents were supplied by Aladdin Chemistry Co., Ltd. and used as received without further purification. A certain quantity of the MWCNT was sonicated in 50 mL methanol solution for 0.5 h under a N₂ atmosphere. Then, the cadmium acetate (0.067 g mL⁻¹) was dissolved in the above methanol solution, and thioacetamide methanol solution (0.018 g mL⁻¹, 50 mL) was subsequently added under a N₂ atmosphere, and the reaction was continued for 12 h at room temperature to produce CdS@MWCNT methanol colloidal suspension. The CdS@MWCNT nanohybrids with various mass percentages of CdS were synthesized by adjusting the molar masses of cadmium acetate and thioacetamide methanol solution under the same conditions (molar ratio of cadmium acetate and thioacetamide methanol solution was constant, as well as the MWCNT mass). The methanol colloidal suspension was obtained by vacuum filtration and was washed by ethanol and distilled water several times to remove the unreacted substance, and then transferred to the desirable substrate. Here, the thickness and size of the film could be controlled by the solution volume and the membrane diameter; the Si/SiO₂ wafer (5 × 5 mm²) with n-type silicon served as the desirable substrate to accept CdS@MWCNT nanohybrid films.

Characterization

SEM (Hitachi S-4800, Japan) and TEM (JEM-2010HR, Japan) were used to observe the prepared products morphology. The XRD spectra were collected by an X-ray diffractometer (D-MAX 2200 VPC, Japan Rigaku Company). The current density versus voltage (J–V) characteristics were recorded using a Keithley 2400 source meter under simulated AM 1.5 G one-sun (100 mW cm⁻²) illumination provided by a solar simulator (69920, 1000 W Xe lamp with optical filter, Oriel). Photoresponse properties were measured in 1 M Na₂SO₃ aqueous solution (pH = 12) under light illumination (λ = 450 nm) using an electrochemical station (CH Instruments 660C, Shanghai Chenhua Inc., China).

Results and discussion

The structural analysis of CdS@MWCNT nanohybrids

The SEM and TEM images of the pure MWCNT bundles and CdS@MWCNT nanohybrids are shown in Fig. 1. It can be seen that a close look at the electrode morphology is provided by these images. Fig. 1a and c show that the MWCNT bundles could serve as a support structure to combine with other semiconductor nanoparticles. The CdS nanoparticles were then deposited onto the MWCNT surface and the diameters were between 5 and 10 nm (Fig. 1b, d and e), roughly the same size of the MWCNT bundles. Separated CdS nanoparticles were grafted onto a single or small bundle of MWCNTs without forming larger size agglomerates (Fig. 1b and d). Good MWCNT nanoparticle adhesion was obtained since the CdS nanoparticles remained on the MWCNT even after washing. Moreover, incomplete coverage was observed after the CdS nanoparticles were attached to the surface of MWCNT bundles. Meanwhile, the MWCNT scaffold buried within the particulate film is thus expected to communicate with the CdS nanoparticles. The density and surface coverage of the hybrids could be controlled by the CdS weight percentage (from 10 to 90 wt%) loaded during the reaction. To make the nanohybrid films, a modest loading (50 wt%) was used because the presence of too many CdS nanoparticles could have prevented MWCNT interconnection and reduced the film conductivity. Also, the photocurrent response as a function of the weight percentage of CdS nanoparticles deposited on MWCNT will be discussed in the following paragraph. In addition, the pure CdS nanoparticles were also prepared by the same synthetic route, which is shown in Fig. 2. The SEM and TEM images show that the diameters of CdS nanoparticles were controlled in 5–10 nm range in accordance with the CdS@MWCNT nanohybrids.

Fig. 1 (a) The SEM images of pristine MWCNT bundles. (b) The SEM images of the CdS@MWCNT bundles with 50 wt% CdS nanoparticles deposited onto the Si/SiO₂ wafer. (c) The TEM image of pristine MWCNT bundles. (d) The TEM image of CdS@MWCNT bundles with 50 wt% CdS nanoparticles. (e) The high resolution TEM (HRTEM) image of CdS nanoparticles coated onto the MWCNT.

Fig. 2 (a) The SEM images of pristine CdS nanoparticles. (b) The TEM images of pristine CdS nanoparticles.

Fig. 3 shows a typical XRD pattern of the MWCNT bundles, CdS nanoparticles, and CdS@MWCNT nanohybrids prepared by our synthetic route. The XRD pattern of the CdS@MWCNT nanohybrids exhibited relatively broad peaks that can be assigned to (111), (220), and (311) planes of cubic CdS with a zinc blende crystal structure.²⁹ In addition, despite the (002) MWCNT reflection being somewhat overlapped by the CdS peaks, it is evidently presented in the XRD pattern of the CdS@MWCNT nanohybrids. To further clarify the structure of the CdS nanoparticles, the high resolution TEM (HRTEM) image is shown in Fig. 1e, the displayed 0.32 nm lattice spacing in the HRTEM image is consistent with the (111) plane of the CdS crystal structure according to the Bragg equation (2d sin θ = nλ, λ = 1.518 Å, n = 2, θ = 26.6°).

Fig. 3 XRD patterns of the MWCNT bundles, CdS nanoparticles and CdS@MWCNT bundles with 50 wt% CdS nanoparticles.

The Raman spectra for the MWCNT bundles, CdS nanoparticles and CdS@MWCNT nanohybrids are exhibited in Fig. 4a. It can be found that two discernible peaks marked with ☆ are observed at bands at 300 and 600 cm⁻¹ for the CdS and CdS@MWCNT, which are assigned to the longitudinal optical phonon mode (1-LO) and its overtone (2-LO) for the CdS nanoparticles, respectively.^30,31 A lower intensity ratio of 2-LO to 1-LO is a result of weak electron phonon interaction at this particle size and further verified that assembled CdS nanoparticles belonged to a cubic structure, corresponding with the XRD results. Furthermore, for the Raman spectrum of CdS@MWCNT nanohybrids, the bands at 1588 cm⁻¹ and 1337 cm⁻¹ could be assigned to an intense peak of G mode and a small disordered-induced peak of D mode (◇ marked), respectively. Interestingly, the relative intensity of the D peak to G peak for the CdS@MWCNT nanohybrids had no noticeable change compared with that of pure MWCNT. This suggests that a desirable nondestructive approach for hybridizing the MWCNT was achieved by our synthetic route described in this article, and the electronic structure of the MWCNT and CdS can be well preserved. Otherwise, a covalent attachment would lead to a drastic increase of D peak intensities.²⁸

Fig. 4 (a) Raman spectra of the MWCNT bundles, CdS nanoparticles and CdS@MWCNT bundles with 50 wt% CdS nanoparticles, which were drop-coated onto a glass slide. (b) UV-Vis absorption spectra of the MWCNT bundles, CdS nanoparticles and CdS@MWCNT bundles with 50 wt% CdS nanoparticles in aqueous suspension.

The UV-Vis absorption of the MWCNT bundles, CdS nanoparticles and CdS@MWCNT nanohybrids in aqueous suspension are shown in Fig. 4b, while the aqueous suspension of MWCNT gives a typical featureless MWCNT absorption. The pure CdS nanoparticles show an additional shoulder peak around 440 nm which is the characteristic of CdS nanoparticles with a 5–10 nm average size,^24,32 which is consistent with the SEM and TEM data. Interestingly, the UV-Vis absorption peak had no change when CdS nanoparticles were coated onto the MWCNT surface. This means that the synthetic route described in this article is a desirable nondestructive approach, and it can preserve the electronic structure of the CdS nanoparticles and MWCNTs.

The photoresponse properties of the CdS@MWCNT nanohybrids

It is well known that the (h)–electron (e) pairs can be formed when CdS nanoparticles are stimulated under light excitation, which will then undergo separation. Finally, the transfer of a conduction-band electron to the electrode leads to the occurrence of photocurrents.¹⁷ These experiments were carried out in a photoelectrochemical cell, where the Si/SiO₂ wafer with a deposit of the target material behaved as the working electrode, a platinum sheet as the paired electrode, and a saturated calomel electrode (SCE) as the reference electrode. The photocurrent generation of the CdS electrode under light illumination was recorded at different excitation wavelengths. The dependence of the photocurrent on the excitation wavelength is shown in Fig. 5a. It can be seen that the onset of photocurrent generation was observed at λ = 450 nm (Fig. 5b) and it closely obeys the UV-Vis absorbance characteristics of the CdS nanoparticles (Fig. 4b). Previous studies showed the CNT films alone exhibit small cathodic photocurrents.³³ However, the photocurrent action spectrum confirms that the light excitation was centered on the CdS nanoparticles, suggesting the observed photocurrent was dominated by the initial excitation of the CdS nanoparticles. This means that the photocurrent of the CdS@MWCNT nanohybrids will also be stimulated by the light illumination at λ = 450 nm.

Fig. 5 (a) The origin of photocurrent generation of the CdS electrode probed at different excitation wavelengths. (b) The photocurrent of the CdS electrode depends on the excitation wavelength. The inset image reveals the test procedure.

Fig. 6a shows the photocurrents of the as-prepared electrodes under light illumination at λ = 450 nm. It was seen that the photocurrents of the MWCNT bundles, CdS nanoparticles and CdS@MWCNT nanohybrids were 0, 2.20, and 6.50 mA cm⁻², respectively, under the illuminated condition (ON). However, the photocurrent values of the as-prepared devices were 0, 0.028, and 0.09 mA cm⁻² under the dark condition (OFF), meaning that the ON/OFF values were 0, 78.57, and 72.22, respectively. The high ON/OFF ratio of the device confirmed that the as-prepared active composite acts as a photodiode.³⁴ It is interesting to note that the CdS films containing a MWCNT network exhibit a roughly 300% higher photocurrent. This means that the MWCNT plays an important role in improving the photoresponse properties within the composite film. Moreover, the ON/OFF cycles of illumination further demonstrated the reproducibility and stability of the photocurrent response of these CdS@MWCNT nanohybrids. In addition, the photoresponse performance of CdS@MWCNT nanohybrids' dependence on CdS loading onto the MWCNT was also proved. The content of the MWCNT was kept constant while the CdS loading was varied. Fig. 6b is the photoresponse of the CdS@MWCNT nanohybrids electrode at different CdS nanoparticle loadings. We found an increase in photocurrent of the CdS@MWCNT nanohybrids with increased CdS loading. The CdS nanoparticles can be effectively dispersed onto the surface of the MWCNT in favor of facilitating charge collection and transportation toward the collecting electrode surface.³⁵ However, the photocurrent was saturated at 50% CdS loadings due to the limitations of light absorption within the CdS film (Fig. 6b and c). Furthermore, a decrease in the photocurrent was observed at higher loadings (70% and 90%). This is caused by the CdS nanoparticles aggregating to clusters, which are not able to make a direct contact with the MWCNT bundles. As a result, the charge carrier cannot effectively transport to the surface of the electrode.

Fig. 6 (a) Photocurrent response to light illumination (λ = 450 nm) of different materials: MWCNT = 0 mA cm⁻², CdS = 2.20 mA cm⁻², and CdS@MWCNT = 6.5 mA cm⁻² (the CdS@MWCNT bundles with 50 wt% of CdS nanoparticles). (b and c) Photocurrent response as a function of the weight percentage of CdS nanoparticles deposited on MWCNTs. The content of the MWCNT was kept constant while the CdS loading was varied.

In addition, the current density versus voltage (J–V) characteristics of as-prepared CdS and CdS@MWCNT electrode are displayed in Fig. 7a. The pure CdS shows an open-circuit voltage (V_oc) of 0.76 V, a short-circuit current density (J_sc) of 2.2 mA cm⁻², and a PCE of 1.67%. The CdS@MWCNT bundles with 50 wt% CdS nanoparticles show a V_oc of 0.71 V, a J_sc of 6.4 mA cm⁻², and a PCE of 4.54%. The results demonstrate that the MWCNT can improve the photocurrent generation in CdS-based photoelectrochemical solar cells.

Fig. 7 (a) J–V characteristics of the MWCNT bundles, CdS nanoparticles and CdS@MWCNT bundles with a 50 wt% of CdS nanoparticles under illumination (100 mW cm⁻²). (b) Fluorescence spectra of the MWCNT bundles, CdS nanoparticles and CdS@MWCNT bundles with a 50 wt% of CdS nanoparticles at light illumination (λ = 450 nm).

The possible mechanism of the photocurrent enhancement in CdS@MWCNT nanohybrids

The results reveal that the PCE of CdS@MWCNT nanohybrids is found to be as high as 4.54% after subsequent incorporation of the MWCNT (electron acceptors), contrasting sharply with the value (1.67%) of the pure CdS. To clarify the enhanced mechanism of photocurrent in the CdS@MWCNT nanohybrids, the fluorescence emission spectra of as-prepared samples are shown in Fig. 7b. Apparently, the fluorescence intensity of the CdS nanoparticles is dramatically quenched when they are deposited onto the surface of the MWCNT (the emission wavelength was 550 nm), confirming the existence of the strong CdS–MWCNT interaction and electron flow in the CdS@MWCNT nanohybrids under light excitation.³⁶ This indicates that the charge transfer between the CdS and MWCNT is responsible for the reversible change of J–V characteristics of the nanohybrids solar cell.³⁷ In other words, the charge transfer between the CdS and MWCNT indeed exists in the nanohybrids solar cell.

It is well known that the band transition of CdS can easily induce the separation of holes and electrons by excitation light used for stimulation (Fig. 8a). It is worth noting that once the charge carriers (hole and electron) are successfully separated, they need to be transported to the corresponding electrodes to provide an external direct current. Otherwise, the recombination of photogenerated hole–electron pairs will happen at particle grain boundaries during charge transport.³⁸ A previous study pointed out that photoinduced electron transfer to CNTs was proposed,³⁹ which means the MWCNT may serve as electron acceptors from the excited CdS (Fig. 8a). The fluorescence spectra show that the emission of CdS is dramatically quenched when it is deposited onto the surface of MWCNT, suggesting a strong excited-state interaction is present in the CdS@MWCNT nanohybrids. A possible pathway for the deactivation of excited CdS is electron transfer to the MWCNT (Fig. 8b). The J–V characteristics of as-prepared CdS and CdS@MWCNT electrode also showed that the flat band potentials, as recorded from the zero current potential (Fig. 7a), were 0.76 and 0.71 V for CdS and CdS@MWCNT films, respectively. Such a shift in the flat band potential is an indication of the electron transfer from CdS to MWCNT as the two systems undergo charge equilibration.³⁵ Moreover, as illustrated in Fig. 8c, the band difference between the lowest unoccupied molecular orbital (LUMO) of CdS and the Fermi level (E_F) of the MWCNT facilitates exciton dissociation and results in a photoinduced electron transfer from the CdS to the MWCNT.^40,41 These results are ascribed to the photoinduced charge separation and electron transfer at the interfaces of the MWCNT and the CdS, leading to an enhanced density of free electrons in the MWCNT. Therefore, combining with the CdS semiconductor and MWCNT bundles has a beneficial role of improving charge separation. In other words, the probability of forming electrons and holes can be greatly enhanced to improve the photocurrent generation.

Fig. 8 The possible mechanism of the photocurrent enhancement in CdS@MWCNT nanohybrids.

Conclusions

To summarize, the deposition of CdS onto MWCNTs was successfully achieved by a simple attaching noncovalent chemical approach. The collaboration of the CdS nanoparticles (electron donor) and MWCNT bundles (electron acceptors) significantly improved the photocurrent response. The PCE of CdS@MWCNT nanohybrids solar cell was found to be as high as 4.54%, contrasting sharply with the 1.67% value of the pure CdS nanoparticles. The outcomes of the present work might provide further information on enhancing the photoresponse of optoelectronic devices. The CdS/MWCNT nanohybrids as potential materials may be applied in innovative functional devices such as optical detectors, sensors, and photovoltaics.

Acknowledgements

Dr C. L. Hu gratefully acknowledges the support of the National Natural Science Foundation of China (Grant No. 51303066), the Foundation of Science and Technology Bureau of Wuhan (Grant No. 2015071704011600).

Notes and references

1. K. S. Leschkies, R. Divakar, J. Basu, E. Enache-Pommer, J. E. Boercker, C. B. Carter and E. S. Aydil, Nano Lett., 2007, 7, 1793.
2. H. A. Atwater and A. Polman, Nat. Mater., 2010, 9, 205.
3. X. Yan, X. Cui, B. S. Li and L. S. Li, Nano Lett., 2010, 10, 1869.
4. J. Tang, K. W. Kemp, S. Hoogland, K. S. Jeong, H. Liu, L. Levina and K. W. Chou, Nat. Mater., 2011, 10, 765.
5. G. Konstantatos, I. Howard, A. Fischer, S. Hoogland, J. Clifford, E. Klem and E. H. Sargent, Nature, 2006, 442, 180.
6. J. Tang and E. H. Sargent, Adv. Mater., 2011, 23, 12.
7. D. R. Baker and P. V. Kamat, Adv. Funct. Mater., 2009, 19, 805.
8. K. S. Leschkies, R. Divakar, J. Basu, E. Enache-Pommer, J. E. Boercker, C. B. Carter and E. S. Aydil, Nano Lett., 2007, 7, 1793.
9. D. D. Wanger, R. E. Correa, E. A. Dauler and M. G. Bawendi, Nano Lett., 2013, 13, 5907.
10. S. A. Mcdonald, G. Konstantatos, S. G. Zhang, P. W. Cyr, E. J. D. Klem, L. Levina and E. H. Sargent, Nat. Mater., 2005, 4, 138.
11. J. H. Bang and P. V. Kamat, ACS Nano, 2009, 3, 1467.
12. G. M. Wang, X. Y. Yang, F. Qian, J. Z. Zhang and Y. Li, Nano Lett., 2010, 10, 1088.
13. W. T. Sun, Y. Yu, H. Y. Pan, X. F. Gao, Q. Chen and L. M. Peng, J. Am. Chem. Soc., 2008, 130, 1124.
14. M. A. Hossain, J. R. Jennings, Z. Y. Koh and Q. Wang, ACS Nano, 2011, 5, 3172.
15. R. A. Joshi, V. S. Taur and R. Sharma, J. Non-Cryst. Solids, 2012, 358, 188.
16. C. L. Hu, S. Y. Chen, S. Peng, X. Q. Liu and J. Y. Liu, Appl. Surf. Sci., 2015, 358, 443.
17. C. L. Hu, Y. J. Chen, X. D. Chen, B. Zhang, J. Yang, J. Y. Zhou and M. Q. Zhang, Chem.–Eur. J., 2012, 18, 1467.
18. I. Robel, B. A. Bunker and P. V. Kamat, Adv. Mater., 2005, 17, 2458.
19. J. Y. Zhang, Y. H. Wang, J. Jin, J. Zhang, Z. Lin, F. Huang and J. G. Yu, ACS Appl. Mater. Interfaces, 2016, 5, 10317.
20. B. Q. Sun, E. Marx and N. C. Greenham, Nano Lett., 2003, 3, 961.
21. F. Vietmeyer, B. Seger and P. V. Kamat, Adv. Mater., 2007, 19, 2935.
22. V. Georgakilas, D. Gournis, V. Tzitzios, L. Pasquato, D. M. Guldi and M. Prato, J. Mater. Chem., 2007, 17, 2679.
23. D. Takagi, H. Hibino, S. Suzuki, Y. Kobayashi and Y. Homma, Nano Lett., 2007, 7, 2272.
24. J. Shi, Y. Qin, W. Wu, X. Li, Z. X. Guo and D. Zhu, Carbon, 2004, 42, 455.
25. Y. Zhu, H. I. Elim, Y. L. Foo, T. Yu, Y. Liu, W. Ji and C. H. Sow, Adv. Mater., 2006, 18, 587.
26. Y. Y. Mu, H. P. Liang, J. S. Hu, L. Jiang and L. J. Wan, J. Phys. Chem. B, 2005, 109, 22212.
27. S. Mubeen, T. Zhang, B. Y. Yoo, M. A. Deshusses and N. V. Myung, J. Phys. Chem. C, 2007, 111, 6321.
28. X. L. Li, Y. Jia and A. Y. Cao, ACS Nano, 2010, 4, 506.
29. X. L. Li, Y. Q. Liu, L. Fu, L. C. Cao, D. C. Wei and Y. Wang, Adv. Funct. Mater., 2006, 16, 2431.
30. V. Sivasubramanian, A. K. Arora, M. Premila, C. S. Sundar and V. S. Sastry, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2006, 31, 93.
31. A. Balandin, K. L. Wang, N. Kouklin and S. Bandyopadhyay, Appl. Phys. Lett., 2000, 76, 137.
32. H. Weller, Angew. Chem., Int. Ed., 1993, 32, 41.
33. S. Barazzouk, S. Hotchandani, K. Vinodgopal and P. V. Kamat, J. Phys. Chem. B, 2004, 108, 17015.
34. S. Yasutomi, T. Morita, Y. Imanishi and S. Kimura, Science, 2004, 304, 1944.
35. A. Kongkanand, R. M. Domínguez and P. V. Kamat, Nano Lett., 2007, 7, 676–680.
36. X. L. Li, Y. Jia, J. Q. Wei, H. W. Zhu, K. L. Wang, D. H. Wu and A. Y. Ao, ACS Nano, 2010, 4, 2142.
37. P. Brown, K. Takechi and P. V. Kamat, J. Phys. Chem. C, 2008, 112, 4776.
38. M. Grätzel, Photoelectrochemical cells, Nature, 2001, 414, 338.
39. D. M. Guldi, G. M. A. Rahman, N. Jux, N. Tagmatarchis and M. Prato, Angew. Chem., 2004, 116, 5642.
40. A. Hagfeldtt and M. Gratzel, Chem. Rev., 1995, 95, 49.
41. W. P. Hou, N. J. Zhao, D. L. Meng, J. Tang, Y. Zeng, Y. Wu, Y. Z. W. Weng, C. G. Cheng, X. L. Xu, Y. Li, J. P. Zhang, Y. Huang, C. W. Bielawski and J. X. Geng, ACS Nano, 2016, 10, 5189.

---