Preparation of graphite-like carbon nitride nanoflake film with strong fluorescent and electrochemiluminescent activity

Abstract

The preparation, characterization, fluorescence (FL) and electrochemiluminescence (ECL) of graphite-like carbon nitride nanoflake particles (g-C₃N₄ NFPs) and nanoflake films (g-C₃N₄ NFFs) have been reported. Highly water-dispersible g-C₃N₄ NFPs with a height of ∼5 to 35 nm and a lateral dimension of ∼40 to 220 nm have been extracted from bulk g-C₃N₄ materials by chemical oxidation. New, stable and defined g-C₃N₄ NFFs can be easily obtained by drying NFPs on certain hydrophilic substrates such as glass or electrode surfaces. Both g-C₃N₄ NFPs and g-C₃N₄ NFFs have good FL activities, i.e. they can give strong blue light (435 nm) emission under UV light (365 nm) excitation. The as-prepared g-C₃N₄ NFFs on a glassy carbon electrode exhibit strong non-surface state ECL activity in the presence of reductive–oxidative coreactants, including dissolved oxygen (O₂), hydrogen peroxide (H₂O₂) and peroxydisulfate (S₂O₈²⁻) and give rise to blue light emission (435 nm), which is the same as the wavelength of FL. The non-surface state ECL mechanisms of the g-C₃N₄ NFF–coreactant systems have been studied and discussed in detail.

---

1. Introduction

Carbon nitride materials have experienced a renaissance of activity in recent years due to their unique properties such as high hardness, wear resistance, chemical inertness, wide-band optical transparency, stable electron field emission, and especially biocompatibility.^1–3 They can exist in several allotropes that have diverse properties, however, the graphite phase carbon nitride (g-C₃N₄) with the tris-s-triazine tectonic unit is regarded as the most stable allotrope at ambient conditions, and has the smallest direct band gap (2.7 eV) due to the π-conjugated graphitic planes formed by the sp² hybridization of carbon and nitrogen.⁴ g-C₃N₄ has shown several potential applications as a metal-free catalyst for Friedel–Craft reactions,⁵ as a reactant for metal nitride nanoparticle synthesis,⁶ and as a photocatalyst for hydrogen production.⁷

It has been found that the nano-sized g-C₃N₄ not only provides a high surface area but also improves the performance of the material.⁸ Accordingly, growing attention has been paid on the synthesis of the nano-sized g-C₃N₄. To date, the nano-sized g-C₃N₄ was usually obtained by the so-called hard-template technique using expensive mesoporous silica,⁴ porous anodic alumina oxide (AAO) membranes,⁹ and layered clays.¹⁰ The preparation of the nano-sized g-C₃N₄ without using a template has not been reported yet. In this work, we have prepared highly water-dispersible g-C₃N₄ nanoflake particles (NFPs) by an up-down method, i.e. by chemically oxidizing the bulk g-C₃N₄ with nitric acid. This method shows many advantages, including low cost, low-temperature operation, and simple experimental performance, and thus would be important in the large-scale production and extensive application of nano-sized g-C₃N₄. Additionally, it has been found that g-C₃N₄ NFPs can be easily made into g-C₃N₄ nanoflake films (NFFs) by drying the g-C₃N₄ NFP suspension on certain substrates. The as-prepared g-C₃N₄ exhibits good fluorescent and electrochemiluminescent properties, suggesting their promising applications in display devices and chemical sensors.

2. Experimental

2.1. Chemicals

Dicyanamide was purchased from Sigma. Potassium persulfate (K₂S₂O₈) and hydrogen peroxide (H₂O₂) were obtained from Fuchen Chemical Reagent Co. (Tianjin, China). All other reagents were of analytical reagent grade and used without further purification. The 0.1 M pH 7.0 phosphate buffer solution (PBS) was used as the electrolyte. Doubly distilled water was used throughout the work.

2.2. Instrumentation

The preparation of bulk g-C₃N₄ was carried out in a tube furnace (GSL 1400X, Kejing Materials Technology Lt. Co., Hefei, China). Atomic force microscopy (AFM) images were obtained by tapping-mode on a Nanoscope IIIa Digital Instruments with NSC15 tips (silicon cantilever, MikroMasch). X-Ray power diffraction (XRD) patterns were measured with a Japan Rigaku D/max-3C using Cu Kα radiation. To investigate the optical properties of the g-C₃N₄ NFP aqueous solution, the photoluminescent spectrum was recorded on a fluorescence spectrophotometer (Cary Eclipse, Varian). Both the FL spectrum of the g-C₃N₄ bulk material and the ECL spectra of g-C₃N₄ NFFs were recorded with a steady/transient state fluorescence spectrophotometer (Edinburgh Instruments, FLS920). UV-vis absorption was characterized by a UV/vis/NIR spectrophotometer (Lambda 750). Elemental analysis for g-C₃N₄ bulk material and nanoflakes were performed on an organic elemental analyzer (Vario MICRO). The field emission scanning electron microscope (FE SEM) image of the g-C₃N₄ NFF coated on a glassy carbon (GC) electrode was captured by using a Nova NanoSEM 230 field-emission microscope (FEI, USA). Electrochemical (EC) and ECL measurements were carried out on an EC and ECL detection system (MPI-E, Remex Electronic Instrument Lt. Co., Xi'an, China) equipped with a three-electrode system, i.e. a GC disc working electrode (0.0707 cm²), a Pt wire counter electrode, and a Ag/AgCl (3 M KCl) reference electrode.

2.3. Synthesis of g-C₃N₄ NFPs

First, the g-C₃N₄ bulk material was prepared following the previously reported literature.⁷ Briefly, the g-C₃N₄ was prepared by heating dicyanamide for 4 h to 550 °C and kept at this temperature for another 4 h in air. Then the g-C₃N₄ NFPs were synthesized as follows: (1) 1 g of bulk g-C₃N₄ powder was ground well with a mortar and a pestle and was then put into 100 mL 5 M HNO₃ and refluxed for 24 h. After naturally cooling to room temperature, no fluorescence was found in the transparent supernatant due to the poor dispersability of g-C₃N₄ NFPs in highly acidic solution. (2) The refluxed product was in turn centrifuged (2770g) and washed with doubly distilled water four times. The supernatant was collected until its pH reached 7 and concentrated on a rotary evaporator at 60 °C under reduced pressure, resulting in a milk-like suspension. The mass concentration of the g-C₃N₄ NFP suspension was calculated by weighing the powder dried from a certain volume of the suspension.

2.4. Preparation of g-C₃N₄ NFFs

Before preparing the g-C₃N₄ NFFs, substrates including glass slides and GC electrodes were thoroughly cleaned. The glass substrates were cleaned in turn with HCl, NaOH, H₂O, and ethanol, while the GC electrodes (3.0 mm in diameter) were polished in turn with 0.3 and 0.05 μm aluminum slurry, rinsed thoroughly with redistilled water, and then sonicated in doubly distilled water for 2 min. A certain volume of g-C₃N₄ NFP suspension (1 mg mL⁻¹) was dropped on the cleaned substrates and resulted in the formation of stable g-C₃N₄ NFFs on the glass slides and GC electrodes by drying the g-C₃N₄ NFP suspension at room temperature in the air.

3. Results and discussion

3.1. Morphology and FL properties of g-C₃N₄ NFPs and g-C₃N₄ NFFs

The concentrated g-C₃N₄ NFP suspension is a milk-like suspension (see photo (b) in Fig. 1A) and it is very stable (i.e. it can keep unchanged for more than 6 months under room temperature). In contrast, the untreated g-C₃N₄ bulk material cannot be dispersed in aqueous solution to form a homogenous suspension (see photo (a) in Fig. 1A). That the g-C₃N₄ NFPs exhibit highly water-dispersible might result from the formation of –COO⁻ and –NH₃⁺ on the surface of g-C₃N₄ NFPs by acid treatment. This viewpoint is supported by the elemental analysis results which show that oxygen and hydrogen contents are increased after chemical oxidation (data not shown). The diluted g-C₃N₄ NFPs suspension (photo (c) in Fig. 1A) shows bright blue luminescence under excitation with a 365 nm UV light (photo (d) in Fig. 1A), with a fluorescence quantum yield (FLQY) of 3.0% (see the ESI†). The FLQY of the as-prepared NFPs is similar to those of carbon quantum dots without good surface passivation.^11–13 The XRD patterns of the g-C₃N₄ bulk material (curve (e) in Fig. 1B) and g-C₃N₄ NFPs (curve (f) in Fig. 1B) are compared in Fig. 1B. It is apparent, that there is no evident different in XRD patterns between the bulk and NFP materials, indicating that the chemical oxidation treatment with concentrated HNO₃ does not destroy the structure of carbon nitrides, i.e. the graphite-like structure of g-C₃N₄ is retained after chemical oxidation. Although there is no evident change in the structure of g-C₃N₄, the FL peak of g-C₃N₄ is blue-shifted from 465 nm associated with bulk material (see curve (g) in Fig. 1C) to 435 nm associated with g-C₃N₄ NFPs (see curve (h) in Fig. 1C) after chemical treatment, suggesting a significant change in particle size.¹⁴ Additionally, g-C₃N₄ NFPs in aqueous solution has a large UV absorption peak at 296 nm and a shoulder absorption band around 365 nm (see curve (i) in Fig. 1C). Further particle size characterization for the obtained g-C₃N₄ NFPs were carried out by SEM and AFM (see Fig. 2). The SEM images obtained for g-C₃N₄ NFPs deposited on the surface of GC electrode (sections A and B in Fig. 2) indicate that g-C₃N₄ NFPs are nanoflakes with an average size less than 100 nm. The AFM data clearly show that g-C₃N₄ NFPs are nanoflakes with a height of ∼5 to 35 nm and a lateral dimension of ∼40 to 220 nm (see sections C and D in Fig. 2). Furthermore, the SEM images present a closely packed assembly of g-C₃N₄ NFPs, suggesting that thin g-C₃N₄ nanoflake films (i.e. g-C₃N₄ NFFs) can be formed easily on substrates such as the surface of the GC electrode.

Fig. 1 Comparison of g-C₃N₄ nanoflake particles (NFPs) with g-C₃N₄ bulk material. (A) The dispersion of g-C₃N₄ bulk material (a) and concentrated g-C₃N₄ NFPs (b) in water, and the photos of diluted g-C₃N₄ NFP solution under white light (c) and 365 nm UV light (d); (B) XRD patterns of g-C₃N₄ bulk material (e) and g-C₃N₄ NFPs (f); (C) FL emission spectra of g-C₃N₄ bulk material (g) and g-C₃N₄ NFPs (h), and UV-vis spectrum of g-C₃N₄ NFPs in aqueous solution (i).

Fig. 2 SEM images of g-C₃N₄ NFPs film on a glassy carbon substrate with scale bar of 5 μm (A) and 1 μm (B), and AFM image of g-C₃N₄ NFPs (C) and the statistical height distribution of g-C₃N₄ NFPs based on the measurement of 226 particles (D).

The excellent film-forming ability of the g-C₃N₄ NFPs was verified by the following experiments: the g-C₃N₄ NFP solution was dropped on the cleaned glass slides or glass disk templates and kept dry at room temperature in the air. It was found that a white, defined (without any cracks), and semitransparent film (i.e. g-C₃N₄ NFF) was coated on the glass surface (see Fig. 3A), exhibiting an excellent film-forming ability of the g-C₃N₄ NFPs. The as-prepared g-C₃N₄ NFFs can give bright blue FL emission when exposed to a UV light of 365 nm (see Fig. 3B). Interestingly, the g-C₃N₄ NFFs are very stable in aqueous solution as soon as the dry films are prepared from the g-C₃N₄ NFPs. The latter are highly water-dispersible (see Fig. 1). The shape and FL intensity of g-C₃N₄ NFFs coated on a glass template, and immersed in doubly distilled water can remain unchanged for more than 30 days (see sections C and D in Fig. 3). It is still not clear why the g-C₃N₄ NFFs do not dissolve back into g-C₃N₄ NFPs. Probably, the excellent stability of the g-C₃N₄ NFFs results from some interactions between –COOH and NH₂ group at the surfaces of the g-C₃N₄ NFPs, for example, hydrogen bond interaction, or formation of –CONH– linkage by dehydration during drying of the g-C₃N₄ NFPs. The effects of environmental conditions such as metal ions, ionic strength and pH on the stability of g-C₃N₄ NFFs were investigated (Fig. S1–S3†). The g-C₃N₄ NFFs coated on glass substrates can remain stable under the above different environmental conditions, although its FL activity may be quenched by some metal ions such as copper(II) ion (Fig. S1†), high ionic strength (Fig. S2†) and high pH (Fig. S3†). Moreover, the quenched FL can be reversibly restored by rinsing the g-C₃N₄ NFFs with pure water. It is apparent that the high stability of the as-prepared g-C₃N₄ nanoflake films provide an easy and effective way for the development of g-C₃N₄ NFF-based sensors, display devices and photocatalysts.

Fig. 3 The photos of the g-C₃N₄ NFFs on the glass substrate exposed to air under a white light (A), and under a 365 nm UV light (B); the fluorescence photos (under 365 nm UV light excitation) obtained for the g-C₃N₄ NFF on a glass disk template as soon as it was immersed in the aqueous solution (C), and after it was immersed in the aqueous solution for 30 days (D).

3.2. Electrochemiluminescent (ECL) properties of g-C₃N₄ NFFs

3.2.1. ECL of g-C₃N₄ NFFs in the presence of dissolved O₂. ECL,^15,16 and electrochemical responses of g-C₃N₄ NFFs at GC electrode in air-saturated aqueous solution (pH 7.0 PBS) were investigated simultaneously, and the experimental results were shown in Fig. 4 (see curves (c) and (c′)). In the cathodic polarization process, two irreversible cathodic waves with peaks at −0.65 V and −1.33 V were observed for the g-C₃N₄ NFF-modified GC electrode (Fig. 4c). The cathodic wave with a peak at −0.65 V can be attributed to the electrochemical reduction of dissolved O₂ for it disappears after the removal of oxygen from the solution (see Fig. 4b). The peak at −1.33 V can be assigned to the electrochemical reduction of g-C₃N₄ NFFs, as this cathodic wave cannot be observed in the absence of g-C₃N₄ NFFs (Fig. 4a). The simultaneously recorded ECL response shows that a strong ECL signal appears when the potential is more negative than the onset potential (−0.83 V) of the reduction potential of g-C₃N₄ NFFs, and reaches a maximum ECL intensity at −1.07 V (Fig. 4c′), suggesting that electron injection into the g-C₃N₄ NFF semiconductor is essential for the ECL of g-C₃N₄ NFFs. It should be noted here, neither the redox peak nor the ECL response can be observed for the g-C₃N₄ NFFs in the positive potential range from 0 to 1.2 V (before the electrolysis of water), indicating that the g-C₃N₄ NFFs are stable in the positive potential range, and suggesting that the initial potential of cathodic polarization can be set at a random potential between 0 and 1.2 V (e.g. 0 V or +1.0 V).

Fig. 4 Simultaneous ECL and chemical responses of g-C₃N₄ NFFs on GC electrode in various aqueous solutions: (a) bare GC electrode in the air-saturated solution; (b) g-C₃N₄ NFF-modified GCE in the N₂-saturated solution; (c) g-C₃N₄ NFF-modified GCE in the air-saturated solution; (d) g-C₃N₄ NFF-modified GCE in O₂-saturated solution. All aqueous solutions contained 0.1 M and pH 7.0 PBS. The potential scan rates were all 100 mV s⁻¹. Inset: comparison of ECL intensities at different electrode materials (GC, Au, ITO and Pt) with the same area (0.0707 cm²).

It has been revealed that dissolved O₂ is necessary for the ECL, since no ECL signal can be observed after the dissolved O₂ is removed from the air-saturated solution by N₂ (see Fig. 4b′), and a much higher ECL intensity can be produced if the aqueous solution is saturated by pure O₂ (see Fig. 4d′). In other words, dissolved O₂ serves as a coreactant in the ECL of g-C₃N₄ NFFs. Additionally, besides on the GC electrode, the g-C₃N₄ NFFs also show good ECL activities on Au and indium tin oxide (ITO) electrodes (inset in Fig. 4). However, the g-C₃N₄ NFFs gave very weak ECL emission on the Pt electrode, due to an unknown ECL mechanism.

3.2.2. ECL of g-C₃N₄ NFFs in the presence H₂O₂ and S₂O₈²⁻. Besides the above-mentioned dissolved O₂, other reductive–oxidative reagents, such as H₂O₂ and S₂O₈²⁻ can also act as coreactants of the g-C₃N₄ NFF ECL (see Fig. 5b and c). For comparing the ECL activities of dissolved O₂, H₂O₂ and S₂O₈²⁻, the concentrations of H₂O₂ and S₂O₈²⁻ were controlled to be 0.258 mM, which equals the concentration of dissolved O₂ in an air-saturated solution. From Fig. 5, it is known that the coreactant ECL activity of H₂O₂ is slightly higher than that of dissolved O₂, while the coreactant ECL activity of S₂O₈²⁻ is much (20–27 times) higher than those of dissolved O₂ and H₂O₂. Although the ECL signals of the g-C₃N₄ NFF–O₂ and the g-C₃N₄ NFF–H₂O₂ systems are obviously weaker than that of the g-C₃N₄ NFF–S₂O₈²⁻ system, they are sensitive enough and may have some promising applications in novel biosensors since O₂ and H₂O₂ are well known reactants or products in various enzymatic reactions. Additionally, the as-prepared g-C₃N₄ NFFs can give much (ca. 65 times) higher ECL intensity than previously reported carbon quantum dots (CQDs),¹⁷ in the presence of the same concentration of S₂O₈²⁻ (see Fig. S4 in the ESI†), indicating that g-C₃N₄ NFFs are new nano-sized luminophores with very strong ECL activity.

Fig. 5 ECL responses obtained for g-C₃N₄ NFF-modified GC electrodes in various aqueous solutions: (a) air-saturated PBS (with concentration of ca. dissolved 0.258 mM O₂); (b) 0.258 mM H₂O₂ (free of dissolved O₂); and (c) 0.258 mM K₂S₂O₈ (free of dissolved O₂). All the aqueous solutions contained 0.1 M PBS (pH 7.0), and the potential scan rates were all 100 mV s⁻¹.

3.2.3. Non-surface state ECL mechanism of g-C₃N₄ NFF–coreactant systems. It has been reported that ECL of many semiconductor nanoparticles (NPs), such as Si NPs,¹⁸ CdSe/ZnSe NPs,¹⁹ Ge NPs,²⁰ and CQDs,^17,21,22 result from electron transfer in surface states. In the surface state ECL, the maximum ECL wavelengths are usually red-shifted compared with the maximum FL wavelengths, and the ECL intensities are very sensitive to surface passivation.²² It is of considerable research interest to know whether the ECL of g-C₃N₄ NFF consisting of g-C₃N₄ nanoflakes is associated with the surface states or not. In other words, is there any difference in wavelength between the FL and ECL of g-C₃N₄ nanomaterials?

The ECL spectrum of g-C₃N₄ NFFs was obtained by using S₂O₈²⁻ as the coreactant, since the g-C₃N₄ NFF–S₂O₈²⁻ ECL system could give stable and sufficiently strong ECL emission upon applying continuous pulse potentials. The pulse potentials were switched between −1.0 V (where the maximum ECL signal can be reached) and +1.0 V (where g-C₃N₄ NFFs is stable). A pulse period of 2 s was chosen, i.e. 1 s for application of +1.0 V and 1 s for application of −1.0 V, since a very stable ECL pulse signal could be obtained under the above pulse potential conditions. The obtained ECL spectrum is shown in Fig. 6B, which indicates that the maximum emission wavelength of the g-C₃N₄ NFF ECL system is ca. 435 nm. The ECL intensity was strong enough to allow direct observation of blue light emission on the surface of the GCE electrode surface (see inset of Fig. 6B). For comparison, the FL spectrum and FL photo of g-C₃N₄ NFPs is also shown in Fig. 6 (see section A). Apparently, the comparison of Fig. 6A and B suggests that the shape and the maximum wavelength of the ECL spectrum of the g-C₃N₄ NFF–coreactant system are almost the same as those of the FL spectrum of g-C₃N₄ nanoflakes. It is apparent, on the basis of above-mentioned luminescent spectra, that we can draw following two conclusions: (1) g-C₃N₄ NFFs serve as the luminophores in ECL reactions; (2) the ECL emission is associated with electron transfer in the core of the nanomaterials (i.e. g-C₃N₄ NFFs) rather than their surface states. In other word, the ECL of g-C₃N₄ the NFF–coreactant system is a non-surface state ECL.

Fig. 6 Comparison of FL (A) and ECL (B) spectra of g-C₃N₄ nanostructures. The experimental conditions for FL spectrum: the concentration of g-C₃N₄ NFP aqueous solution was 15 μg mL⁻¹; excitation wavelength was 360 nm. The experimental conditions for ECL spectrum: the continuous pulse potential was switched between +1 V (for 1 s) and −1 V (for 1 s); the testing solution contained 0.15 M K₂S₂O₈ and 0.6 M PBS (pH 7.0). Inset in part A: the FL photo of the g-C₃N₄ NFP aqueous solution illuminated by a UV beam of 365 nm. Inset in part B: the ECL image of the g-C₃N₄ NFF modified GC electrode.

According to the above investigation, the ECL mechanisms of the g-C₃N₄ NFF–coreactant systems can be proposed in Fig. 7. The ECL reactions may involve four processes, i.e. electron injection, hole donor production, hole injection, and electron–hole annihilation producing ECL emission. The energy of the electron on GCE is raised with the cathodic polarization of potential, and is high enough to be injected into the conducting bands (CB) of g-C₃N₄ NFF:

g-C₃N₄ + e⁻ → g-C₃N₄˙⁻ (R1)

Fig. 7 Proposed ECL reaction mechanism for g-C₃N₄ NFF–coreactant systems.

Meanwhile, hole donors (i.e. free radicals) are produced from the reductive–oxidative coreactants, such as dissolved O₂, H₂O₂ and S₂O₈²⁻:

O₂ + 2H₂O + 2e⁻ → 2H₂O₂ (R2)

H₂O₂ + e⁻ → OH⁻ + ˙OH (R3)

S₂O₈²⁻ + e⁻ → SO₄²⁻ + SO₄˙⁻ (R4)

The holes are subsequently injected into the valence bands (VB) of g-C₃N₄ NFF by the interaction of the free radicals (˙OH or SO₄˙⁻) with g-C₃N₄ NFF:

˙OH + g-C₃N₄˙⁻ → g-C₃N₄* + OH⁻ (R5)

SO₄˙⁻ + g-C₃N₄˙⁻ → g-C₃N₄* + SO₄²⁻ (R6)

Finally, the electrons injected in the conducting band annihilates with the holes in the valence band to form the excited state of the g-C₃N₄ nanoflakes that emits blue light when they go down to their ground states:

g-C₃N₄* → g-C₃N₄ + hν (R7)

4. Conclusions

Highly water-dispersible g-C₃N₄ nanoflake particles (i.e. g-C₃N₄ NFPs) with a height of ∼5 to 35 nm and a lateral dimension of ∼40 to 220 nm has been extracted from the g-C₃N₄ bulk material by chemical oxidation. The g-C₃N₄ NFPs exhibit excellent film-forming ability, on the basis of which new, defined and stable g-C₃N₄ nanoflake films (i.e. g-C₃N₄ NFPs) can be easily prepared on hydrophilic substrates such as glass slides and GCE surfaces. Both g-C₃N₄ NFPs and g-C₃N₄ NFFs have good FL activities, giving strong blue light (435 nm) emission under UV light (365 nm) excitation. The g-C₃N₄ NFFs on a glassy carbon electrode exhibit strong ECL activity in the presence of reductive–oxidative coreactants including dissolved O₂, H₂O₂ and S₂O₈²⁻ and give rise to blue light emission (435 nm). The ECL can be attributed to a non-surface sate ECL since the shape of the ECL spectrum and the maximum ECL emission wavelength are almost the same as those of FL spectrum. The ECL mechanisms of the g-C₃N₄ NFF–coreactant systems are proposed to involve the following processes: electron injection into the conducting bands of g-C₃N₄ NFFs; electrogeneration of hole donors (˙OH or SO₄˙⁻) on the electrode; hole injection into the valence bands of g-C₃N₄ NFFs; and the annihilation of electron–hole pairs to produce ECL emission. The excellent properties of g-C₃N₄ NFFs, such as easy preparation, excellent stability in solution, good FL activities, and high ECL sensitivity towards dissolved O₂, H₂O₂, and S₂O₈²⁻ enable the new g-C₃N₄ NFFs nanomaterials to have promising applications in biosensors, the study of free radicals, and new display devices.

Acknowledgements

This study was financially supported by National Natural Science Foundation of China (21075018), the Program for New Century Excellent Talents in Chinese University (NCET-10-0019), the National Basic Research Program of China (2010CB732400), and the Program for Changjiang Scholars and Innovative Research Team in University (no. IRT1116).

Notes and references

1. E. Kroke and M. Schwarz, Coord. Chem. Rev., 2004, 248, 493.
2. D. C. Cameron, Surf. Coat. Technol., 2003, 169–170, 245.
3. F. Z. Cui and D. J. Li, Surf. Coat. Technol., 2000, 131, 481.
4. A. Thomas, A. Fischer, F. Goettmann, M. Antonietti, J. O. Müller, R. Schlögl and J. M. Carlsson, J. Mater. Chem., 2008, 18, 4893.
5. F. Goettmann, A. Fischer, M. Antonietti and A. Thomas, Angew. Chem., Int. Ed., 2006, 45, 4467.
6. A. Fischer, M. Antonietti and A. Thomas, Adv. Mater., 2007, 19, 264.
7. X. C. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domenand and M. Antonietti, Nat. Mater., 2008, 8, 76.
8. Y. J. Zhang, A. Thomas, M. Antonietti and X. C. Wang, J. Am. Chem. Soc., 2009, 131, 50.
9. X. H. Li, J. S. Zhang, X. F. Chen, A. Fischer, A. Thomas, M. Antonietti and X. C. Wang, Chem. Mater., 2011, 23, 4344.
10. G. F. Jiang, C. H. Zhou, X. Xia, F. Q. Yang, D. S. Tong, W. H. Yu and S. M. Liu, Mater. Lett., 2010, 64, 2718.
11. X. Xu, R. Ray, Y. Gu, H. J. Ploehn, L. Gearheart, K. Raker and W. Scrivens, J. Am. Chem. Soc., 2004, 126, 12736.
12. J. Zhou, C. Booker, R. Li, X. Zhou, T. K. Sham, X. Sun and Z. Ding, J. Am. Chem. Soc., 2007, 129, 744.
13. H. Liu, T. Ye and C. Mao, Angew. Chem., Int. Ed., 2007, 46, 6473.
14. M. Groenewolt and M. Antonietti, Adv. Mater., 2005, 17, 1789.
15. M. M. Richter, Chem. Rev., 2004, 104, 3003.
16. W. Miao, Chem. Rev., 2008, 108, 2506.
17. Y. Dong, N. Zhou, X. Lin, J. Lin, Y. Chi and G. Chen, Chem. Mater., 2010, 22, 5895.
18. Z. Ding, B. M. Quinn, S. K. Haram, L. E. Pell, B. A. Korgel and A. J. Bard, Science, 2002, 296, 1293.
19. N. Myung, Y. Bae and A. J. Bard, Nano Lett., 2003, 3, 1053.
20. N. Myung, X. Lu, K. P. Johnston and A. J. Bard, Nano Lett., 2004, 4, 183.
21. L. Y. Zheng, Y. W. Chi, Y. Q. Dong, J. P. Lin and B. B. Wang, J. Am. Chem. Soc., 2009, 131, 4564.
22. Y. Dong, R. Wang, H. Li, J. Shao, Y. Chi, X. Lin and G. Chen, Carbon, 2012, 50, 2810.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c2nr32248j

---