Carbon‐Doped Porous Polymeric Carbon Nitride with Enhanced Visible Light Photocatalytic and Photoelectrochemical Performance

The photocatalytic activity of polymeric carbon nitride (CN) is strongly tied to its surface area, morphology, optical, and electronic properties. Herein, a straightforward approach to adjusting the electronic properties and morphology of CN materials by fine tuning their carbon content while preserving their high surface area is proposed. To do so, supramolecular assemblies based on CN monomers together with an additional carbon‐rich monomer that does not participate in the preorganization are calcinated. The use of a supramolecular assembly as the precursor endows the CN material with a high specific surface area and an ordered morphology, whereas the addition of a carbon‐rich monomer provides light carbon doping. As a result, the new CN exhibits excellent activity as a photoanode material in photoelectrochemical cells and as a photocatalyst for the hydrogen evolution reaction. Detailed studies reveal that the modified CN samples show enhanced charge carrier transfer and separation efficiency, improved light absorption response, a tunable energy band structure, a higher electrochemical surface area, and better electronic conductivity compared with a reference CN.


Introduction
Polymeric carbon nitride (CN) semiconductor materials have attracted great interest as metal-free organic photocatalyst owing to their low cost, high chemical and thermal stability, ease of synthesis, and good catalytic activity. [1] As a photocatalyst, CN has been used for the production of hydrogen (H 2 ) through water splitting, [2] CO 2 reduction, [3] various organic transformations, [4] and the degradation of pollutants. [5] However, the intrinsic drawbacks of CN, especially large grain boundaries, small specific surface area, low density of active sites, high recombination rate of electron-hole pairs, and moderate light absorption, hinder its activity as a photocatalyst. [6] To overcome these challenges, many different strategies have been proposed, such as morphology control, [7] copolymerization, [8] heteroatom doping, [9] and heterojunction formation. [10] Among these, the copolymerization of standard CN precursors with other organic compounds is an effective route to fabricate modified CN polymers with improved photocatalytic activity. [11] Hence, finding suitable comonomers with a diverse chemical composition and structure to alter the band structure and optoelectronic properties of CN is an important task. [12] Carbon doping and nitrogen defects in CN have previously been shown to improve electrical conductivity, charge transfer rate, and photocatalytic activity. [13] Most importantly, carbon doping can increase the size of delocalized π-bond networks, which favors electron transfer, modulates the band structure, and extends the optical absorption range to lower energies. [14] In addition, an asymmetric distribution of C and N atoms on the surface causes charge nonuniformity, which may improve photocatalytic properties. [15] However, the amount of carbon doping must be carefully adjusted to augment the activity: even a small deviation may lead to a nonactive photocatalyst. CM supramolecular complexes-composed of cyanuric acid (C) and melamine (M) connected through hydrogen bonds-are frequently used as a precursor to CN-based materials. [16] The incorporation of doping molecules into the CM selfsupramolecular complex was found to affect the photoelectronic properties of CN. Caffeine, [17] phosphoric acid, [18] barbituric acid, [19] thymine, [20] urea, [21] 2,4-diamino-6-phenyl-1,3,5-triazine (Mp), [7] 2,4,6-triaminopyrimidine (TAP), [22] and chitosan [23] are some of the precursors that were used to tune the properties of CN-based materials. Nevertheless, the incorporation of additives within the hydrogen framework of CM often leads to excessively high carbon levels in the final CN. [24] In this work, we propose a simple method for the synthesis of tetracyanoethylene (TCNE)-modified CN materials by thermally induced copolymerization of a CM supramolecular assembly and TCNE. To do so, we calcinate the CM supramolecular assembly with various amounts of added TCNE under an inert atmosphere. Unlike the incorporation of additional monomer before the supramolecular assembly in solution-typically water, in DOI: 10.1002/aesr.202200035 The photocatalytic activity of polymeric carbon nitride (CN) is strongly tied to its surface area, morphology, optical, and electronic properties. Herein, a straightforward approach to adjusting the electronic properties and morphology of CN materials by fine tuning their carbon content while preserving their high surface area is proposed. To do so, supramolecular assemblies based on CN monomers together with an additional carbon-rich monomer that does not participate in the preorganization are calcinated. The use of a supramolecular assembly as the precursor endows the CN material with a high specific surface area and an ordered morphology, whereas the addition of a carbon-rich monomer provides light carbon doping. As a result, the new CN exhibits excellent activity as a photoanode material in photoelectrochemical cells and as a photocatalyst for the hydrogen evolution reaction. Detailed studies reveal that the modified CN samples show enhanced charge carrier transfer and separation efficiency, improved light absorption response, a tunable energy band structure, a higher electrochemical surface area, and better electronic conductivity compared with a reference CN.
which the solubility of each monomer may differ significantly, the solid-state synthesis enables fine adjusments of the elemental ratio and overall properties to be made easily . The delicate carbon  modification leads to better light absorption response, a tunable  energy band structure, a large specific surface area, and a more   effective charge carrier separation compared with a reference CN  (CN-CM). As a result, TCNE-modified CN demonstrates good activity as a photocatalyst for H 2 evolution and rhodamine B degradation. Furthermore, the new CN shows high performance as a photoanode material in photoelectrochemical cells. Various Scheme 1. CN (CM and CN-CMT x ) synthesis at 550°C from hydrogen-bonded supramolecular complexes and TCNE as the starting reactants.  photo-and electrochemical studies reveal that the modified CN samples exhibit improved charge carrier transfer and separation efficiency, higher electrochemical surface area, and better electronic conductivity.

Synthesis and Characterization of Precursors
Supramolecular assemblies of hydrogen-bonded CM were formed by mixing equal amounts of cyanuric acid and melamine in water. [7,25] We then calcinated mixtures of CM supramolecular preassembly as the CN precursor and TCNE as an additive (denoted as T), achieving the preparation of TCNE-modified polymeric CNs (CN-CMT x ), as shown in Scheme 1. CMT x powders were prepared by mixing CM (1.0 g) with different wt% of TCNE (x ¼ 0, 0.50, 1.0, 1.5, 2.0, 2.5, and 10); the combination of powders was ground using a mortar and pestle to ensure thorough mixing. Scanning electron microscopy (SEM) images of the CM and CMT x precursors ( Figure S1, Supporting Information) demonstrate that the morphologies of the supramolecular assemblies were not altered by the addition of the TCNE precursor. The molecular structure of the mixtures was further characterized via Fourier-transform infrared (FTIR) spectroscopy and X-Ray diffraction (XRD). The FTIR spectra of CM and CMT x disclose similar vibration modes, indicating that the small amount of added TCNE is insufficient to affect the structure of the CM complex ( Figure S2, Supporting Information  Figure S3, Supporting Information). [7] Hence, XRD analysis confirms that the presence of TCNE does not alter the supramolecular structure of CM. TCNE is potentially only bound to the surface.

CN Characterization
The CM and CMT x mixtures were each pyrolyzed at 550°C in nitrogen atmosphere for 4 h to prepare photoactive CNs (CN-CM and CN-CMT x ). TCNE is a carbon-rich material containing only C and N (as an ideal graphitic CN) and four cyanogroups. [26] During calcination, the cyanogroups of TCNE could integrate directly into the CN during the copolymerization of CM with TCNE, resulting in the formation of a C─C bond attached to the heptazine ring. [27] The C─C bond linked to the heptazine ring may break during copolymerization at high temperatures because it is less stable (longer bond length) than a C─N bond, resulting in amino group vacancies, as shown in Scheme 1. [27] The amino group's absence reduces the heptazine ring's  connectivity, resulting in the formation of nitrogen-poor and carbon-rich CN. This caused the CN network structure to fracture, resulting in some defects in CN-CMT x materials. The morphology and microstructure of CN-CM and CN-CMT x were investigated using SEM and transmission electron microscopy (TEM), as shown in Figure 1a-d. When the TCNE content increased, the size of the CN nanosheets decreased and higher porosity (relative to reference CN) developed, indicating that the TCNE copolymerization reaction caused nanosheets' fragmentation and the formation of pores on the CN's surface. The TEM images show that CN-CMT 2 exhibits a much thinner and more porous layer structure than CN-CM. The increased porosity of the CN-CMT x materials could lead to an increase in the amount of catalytically active sites while shortening the diffusion path of photogenerated charge carriers. [28] The formation of CN materials after heating was confirmed by FTIR spectroscopy and XRD analysis ( Figure 2). The FTIR spectra show the typical breathing mode of the triazine units in CN at 809 cm À1 and the characteristic stretching modes for the CN heterocycles from 1200 to 1600 cm À1 (Figure 2a). We assigned the broad band between 3000 and 3600 cm À1 to either -NH 2 groups or surface-adsorbed H 2 O molecules, or both. The XRD patterns ( Figure 2b) of the CNs exhibit two distinct diffractions at 13.0°a nd 26.7°, which we assigned to the (100) and (002) planes, indicative of a layered structure formation. For the CN-CMT x sample, the peak corresponding to (002) is broadened, which is due to the lower interaction between the layers.
The molecular structure of the different CNs was further determined by solid-state NMR (Figure 2c and S4, Supporting Information). [29] The 1 H magic-angle spinning (MAS) NMR spectra reveal two signals with peaks maxima at %9.4 and %4.3 ppm, attributable to the presence of -NH x groups and residual water in the heptazine lattices, respectively ( Figure S4, Supporting Information) [30] ; this indicates the presence of melon-like NH-bridged heptazine-based structures with some -NH and -NH 2 terminations. The slight shift in CN-CMT 2 signals implies that the heptazine units have different chemical environments depending on the orientation of the adjacent layers. As a result, heptazine-based structures with some -NH and -NH 2 groups are involved in intermolecular hydrogen bonding, causing CN-CMT 2 to shift downfield. [31] The 13 C crosspolarization CP MAS NMR spectrum exhibits two clear signals at 164.5 and 157.3 ppm, which we ascribed to carbon bonded to NH 2 groups (N 2 C-NH 2 ) and sp 2 carbon of the C-N bonds (CN 3 ) in the heptazine units, respectively. [32] CN-CMT x and CM-CN show similar peak patterns, indicating the preservation of the basic heptazine structure in CN-CMT x .
The surface chemical environments of C and N were investigated by X-Ray photoelectron spectroscopy (XPS) (Figure S5-S6  and Table S1, Supporting Information). [33] In the CN-CM  samples, the C 1s binding energy signal at 285.1 eV is ascribed to graphite carbon atoms (C═C), whereas the peak at 288.1 eV is attributed to sp 2 carbon (N─C═N) present in the CN backbone. [13a,34] The peak at 289.4 eV belongs to C─O (surface-adsorbed atmospheric oxygen) or C─N bonds. The N 1s peak at 398.6 eV corresponds to sp 2 -hybridized aromatic N in tri-s-triazine rings (C─N═C). We assigned the peak with a high binding energy at 399.5 eV to nitrogen (N─(C)3) at the structural edges of CN. The peak located at 400.9 eV indicates the presence of amino functions (C─NH) originating from the terminal amino groups on the surface. [35] The O 1s XPS spectra of CN can be resolved into two peaks at 531.9 and 533.5 eV, attributed to C─O and adsorbed water, respectively. In the C 1s spectrum of CN-CMT 2 , the three peaks located at the binding energies of 284.9, 288.5, and 290.2 eV are similar to those in the corresponding spectrum of CN-CM ( Figure S6, Supporting Information).
The new peaks at 286.5 and 291.5 eV are assigned to sp 2hybridized C─NH and C═O, respectively. The presence of C═O might stem from an incomplete reaction of oxygencontaining intermediates formed during the pyrolysis of the CM supramolecular assembly, or it could be doped into the bulk structure. [36] The peak area of sp 2 -hybridized C═C for CN-CMT 2 is higher than that of CN-CM, because of the additional carbon atoms introduced into the structure. The N 1s XPS spectrum shows three peaks at 398.7, 399.6, and 400.8 eV, which we ascribed to C─N═C, N─(C) 3 , and C─NH, respectively. The new peak at 401.8 eV is characteristic of amino functional groups (-NH 2 ). Elemental analysis data (Table S2, Supporting Information) unveils only slight changes in the C/N ratio owing to the low wt%. of the dopant. The light absorption properties of CNs were investigated by UV-vis diffuse reflectance spectroscopy (DRS) (Figure 3a). The absorption edge of CN-CMT x is slightly redshifted and the visible light absorption is enhanced compared with the reference CN-CM. This implies that the intrinsic electronic properties of the CN materials were changed by the alteration of their elemental composition: it leads to a decrease in the bandgap of the CN-CMT x samples (Figure 3b), improving light harvesting. The photoluminescence (PL) intensity of all CN-CMT x samples is dramatically reduced compared with the reference CN-CM, indicating an alternative nonradiative path for the photogenerated electron-hole pairs (Figure 3c). We attributed this to the creation of defect levels and to the nanosheet structure of doped CN-CMT x , which lead to a shorter charge carrier migration path and improved electronic conductivity, thus suppressing the photoinduced recombination of electron-hole pairs. [37] To further study the lifetime of photogenerated charge carriers in the samples, we also performed time-resolved fluorescence measurements ( Figure S7, Supporting Information). CN-CMT 2 presents a faster exponential decay than CN-CM, with an average   Figure S8, Supporting Information). The CB position of CN-CMT x is downshifted and remains more above the reduction potential of H þ /H 2 . N 2 adsorption-desorption isotherms ( Figure S9, Supporting Information) indicate that all the CN-CMT x samples have a higher specific surface area than the reference CN-CM (S A , Figure 4a). The measured S A of CN-CMT 2 is 86.2 m 2 g À1 , almost twice that of the reference CN (46.6 m 2 g À1 ). The total pore volume of CN-CMT 2 (2.4 cm 3 g À1 ) is five times larger than that of CN-CM (0.52 cm 3 g À1 ).

Photocatalytic Activity
The photocatalytic hydrogen evolution experiment was performed under a white light-emitting diode (LED) illumination in a water-triethanolamine (TEOA) mixture with Pt as a cocatalyst ( Figure 4b). Among all the modified CN-CMT x samples, CN-CMT 2 exhibited the highest photocatalytic H 2 evolution activity (1757 μmol h À1 g À1 )-an almost fourfold increase over that of CN-CM (422 μmol h À1 g À1 )-as shown in Figure 4b. We envision that the enhancement in activity is due to the elemental modification, which improves the electronic properties of the material, and to the increased surface area of the porous CN structure, which provides more sites for cocatalyst deposition and H 2 production. [38] CN-CMT 2 demonstrates excellent stability during H 2 evolution (Figure 4c and S10a, Supporting Information), still evolving a steady amount of H 2 after four cycles under the same reaction conditions. A wavelength-dependent H 2 generation experiment of CN shows that the variation in apparent quantum efficiency (AQE) coincides with the variation in absorption in the UV-vis spectrum ( Figure S10b, Supporting Information and Figure 3a). For CN-CMT 2 , the AQE at 405 nm is estimated to be about 6%. The activity of the CN samples as photocatalysts for pollutant degradation was measured using the RhB dye. [7] A proposed photocatalytic RhB dye degradation scheme over CN is shown in Figure S11, Supporting Information. As shown in Figure 4d, after 15 min of the photocatalytic reaction, CN-CM showed relatively poor RhB degradation rates. In contrast, the photocatalytic activity of the CN-CMT x samples was significantly enhanced, particularly CN-CMT 2 , which achieved full RhB degradation within 15 min. Finally, we prepared CN-CM and CN-CMT 2 electrodes for photoelectrochemical cells by the doctor-blade method ( Figure S12, Supporting Information). [33] Photocurrent density measurements ( Figure 5) at 1.23 V versus RHE in 0.1 M KOH reveal a higher and more stable photocurrent response for the CN-CMT 2 electrode than for reference electrode CN-CM,  indicating an enhanced charge separation and improved electronic structure. The photocurrent responses were obtained under ten cycles of illuminated/dark conditions. Figure 5a shows that CN-CMT 2 exhibits an improvement in photocurrent density, reaching 60 μA cm À2 , significantly higher than that of CN-CM (21 μA cm À2 ). In the presence of a hole scavenger (TEOA), CN-CMT 2 exhibits a high current density of 160 μA cm À2 compared with CN-CM (68 μA cm À2 ) (Figure 5b). The enhancement in photogenerated charge separation and electron conductivity of the TCNE-modified CN-CMT 2 electrode is further demonstrated by the improved stability of the modified CN films ( Figure S13, Supporting Information). The electrochemical impedance spectroscopy (EIS) results imply an increase in hole transfer to the solution for the CN-CMT 2 samples compared with CN-CM, which in turn lead to longer electron lifetime (Figure 5c). [36a] The improvement in conductivity was supported by electrochemically active surface area (ECSA) measurements of the CN films (Figure 5d), which also showed higher values for the CN-CMT 2 films. The higher ECSA reflects the participation of a larger proportion of the CN surface in the photoelectrochemical process, which increases electron collection.

Conclusion
We introduced a facile method to synthesize a porous, highly photoactive carbon-doped CN by the surface modification of a supramolecular assembly as the reactant. To do so, we used TCNE as an additive that, upon thermal condensation, modifies the structural, electronic, and catalytic properties of CN. The new TCNE-modified CN photocatalyst exhibited an improved absorption in the visible range, a modulated energy band structure, an efficient separation of photogenerated charge carriers, and a high surface area leading to more active sites. The new CN-CMT 2 presents a significant enhancement in photocatalytic activity with a hydrogen evolution rate of 1757 μmol h À1 g À1 and a high QE of 6% at 405 nm. Furthermore, CN-CMT 2 exhibits high and stable photocurrent densities when used as a photoanode material in photoelectrochemical cells. (Photo)electrochemical studies reveal a better charge separation under illumination, owing to the increased electrochemical surface area, good electronic conductivity, and improved hole extraction to the solution of CN-MT 2 .
Synthesis of Polymeric CN Materials: The CM complex was prepared by mixing cyanuric acid (C, 1 g, 7.75 mmol) and melamine (M, 0.98 g, 7.75 mmol) in a 1:1 molar ratio in 40 mL DI water in an automatic shaker. The resulting powder was filtered and washed several times with water and then dried at 60°C in a vacuum oven. CMT x powders were then prepared by mixing CM (1.0 g) with x wt% of TCNE (x ¼ 0.50, 1.0, 1.5, 2.0, 2.5, and 10). Each mixture was ground until well mixed using an agate mortar and pestle and then calcined at 550°C for 4 h under N 2 atmosphere (heating rate of 2°C min À1 from room temperature to target temperature). Afterward, the CN materials were collected and labelled as CN-CMT x . A CN-CM reference was prepared using the same procedure from the CM complex (free of TNCE).
Photocatalytic H 2 Evolution Reaction (HER): The production of hydrogen by CN materials was measured in a Schlenk flask thermally regulated with a cooling system and using a white LED array as the irradiation source, 3 wt% Pt as a cocatalyst, and triethanolamine (TEOA) as a hole scavenger. Photocatalytic hydrogen evolution experiments were conducted under an inert argon atmosphere at a constant temperature of 25°C. In a 50 mL Schlenk flask, 15 mg of CN was suspended in a solvent mixture (19 mL) composed of water and TEOA in a 9:1 v/v ratio and 19.6 μL of a H 2 PtCl 6 solution (8 wt% in water). The mixture was then kept for 30 min in the dark under a constant argon flow to reach thermal equilibrium; the reaction was subsequently started by switching on a 100 W white LED array (Bridgelux BXRA-50C5300; λ > 410 nm). The amount of evolved gas in the headspace of the reactor was evaluated by gas chromatography (Agilent 7820 GC System). In this system, upon visible light irradiation, CN materials generated electron-hole pairs that migrated to the catalyst's surface. The valence band (VB) holes in the VB reacted with TEOA when the electron-hole pairs separated spatially, while photoexcited electrons reduced the molecular Pt salt to Pt (0), allowing subsequent electrons to reduce water molecules into H 2 .
The apparent quantum yield (AQY) for H 2 evolution was determined using mounted LEDs (Thor Labs, Model no. M405L4, M430L4, and M455L4 equipped with a 405, 430, and 450 nm bandpass filter, respectively). Quantum yield measurements were performed in a sealed reactor connected to an argon line and an Agilent 7820 GC system. The reactor was continuously purged with argon (in the dark) to remove the existing hydrogen, nitrogen, oxygen, and other gases, and the purging process was monitored by automatic sampling every 11 min. After purging, the LED was switched on, and the amount of produced hydrogen gas was recorded; the integrated area was used further for the calculation of the average quantum yield (AQY%), which was calculated as follows: AQY ¼ N e /N p Â 100% ¼ 2 M/N p Â 100%, where N e is the amount of reaction electrons, N p is the amount of incident photons, and M is the amount of H 2 molecules.
Photocatalytic Tests: The photocatalytic activity of the CN materials was determined by performing rhodamine B (RhB) dye degradation under a white 100 W LED (Bridgelux BXRA-50C5300; λ > 410 nm) illumination. RhB (in 20 mL water, 20 mg L À1 ) and the photocatalyst (10 mg) were mixed in a glass vial and placed in the dark under continuous stirring until an adsorption-desorption equilibrium was reached. After a specific time of illumination duration, the samples were collected. RhB concentration was monitored by measuring the optical absorbance (utilizing the absorption maximum of RhB, λ max ¼ 554 nm), which was plotted as a normalized concentration C/C 0 .
Characterization: SEM images of the CN electrodes were obtained using a JEOL JSM-7400 F high-resolution SEM, equipped with an field emission gun source, operated at an accelerating voltage, U 0 ¼ 3.5 kV (after sputtering with %10 nm Pt using a Quorum Q150T ES system). TEM images were obtained using a Tecnai (FEI) T12 G 2 TWIN microscope at U 0 ¼ 120 kV. UV-vis absorption and DRS were measured on a Cary 100 spectrophotometer in transmittance mode (10 mm quartz cuvettes) or equipped with a diffuse reflectance accessory (DRA). Fluorescence measurements were carried out on a FluoroMax 4 spectrofluorometer from Horiba Scientific. FTIR was performed on a Thermo Scientific Nicolet iS5 FTIR spectrometer (equipped with a Si ATR). XRD (XRD) were obtained using a PANalytical's Empyrean diffractometer equipped with a position-sensitive detector X'Celerator. Data was collected with a scanning time of %7 min for 2θ ranging from 5°to 60°using Cu Kα radiation (λ ¼ 1.54178 Å, 40 kV, 30 mA). XPS data were obtained from an X-Ray photoelectron spectrometer ESCALAB 250 ultrahigh vacuum (1 Â 10 À9 bar) device with an Al Kα X-Ray source and a monochromator. The X-Ray beam size was 500 μm, survey spectra were recorded with a pass energy (PE) of 150 eV, and high-energy resolution spectra were recorded with a PE of 20 eV. To correct for charging effects, all spectra were calibrated relative www.advancedsciencenews.com www.advenergysustres.com to a carbon C 1 s peak positioned at 284.8 eV. The XPS results were processed using the Thermo Scientific Avantage software. Elemental analysis data for carbon, nitrogen, and hydrogen (CNH) was collected using a Thermo Scientific Flash Smart elemental analyzer OEA 2000. Nitrogen sorption measurements and Brunauer-Emmet-Teller (BET) specific surface area calculations were performed on a Quantachrome NOVAtouch NT LX 3 system. Solid-state NMR experiments were carried out in a Bruker Avance III 500 MHz narrow-bore spectrometer, using a 4 mm double-resonance MAS probe at a spinning rate of 8 KHz. 13 C CP MAS experiments were carried out using a 2.5 μs 1 H 90°pulse, a 2 ms mixing time, and a 3 s recycle delay between acquisitions. 1 H direct excitation experiments were carried out using a 2.5 μs 90°pulse and a 2 s recycle delay. Time-resolved fluorescence measurements were performed using a time-correlated single-photon counting setup with Fluorolog Modular Spectrofluorometer. The PL decay curves were monitored at their maximum emission wavelength under the excitation of light with a wavelength of 370 nm. The triple-exponential function was used to fit the decay curves. Intensity-averaged fluorescence lifetime [τ] was calculated using the equation: τ ¼ P i¼n i¼1 a i τ 2 i = P i¼n i¼1 a i τ i where τ i is the lifetime and a i is the amplitude of the i th component.
CM and CMT x film preparation: For the preparation of the CM and CMT 2 supramolecular pastes, CM powder (1.0 g) without and with TCNE (2 wt%) were blended thoroughly in ethylene glycol (2.0 mL). The pastes were then doctor -bladed onto FTO-coated glass (using 2 scotch tape layers (L) to control the film thickness and achieve a homogeneous coating). Then, the films were subsequently dried at 90°C on a hotplate and finally transferred into a closed (but not sealed) 16 mm-diameter glass test tube. All the FTO substrates were cleaned by successive sonication in an aqueous detergent solution (1% m/v Alconox), acetone, and ethanol, and were finally dried before use. For the calcination process, melamine powder (0.2 g) was placed at the bottom of a test tube, in the middle of which two electrodes (FTO substrates) were then placed. The test tube was purged with N 2 for 5 min and covered with Al foil. Finally, the samples were heated up to 550°C at a ramp rate of 5°C min À1 and kept at 550°C for 4 h under a flow of N 2 to obtain the CN-CM and CN-CMT 2 electrodes.
All electrochemical measurements were performed using a threeelectrode system on an Autolab potentiostat (Metrohm, PGSTAT302N). Pt foil (1 cm 2 ) and Ag/AgCl (saturated KCl) were used as the counter and reference electrodes, respectively. The electrolyte was either a 0.1 M KOH aqueous solution (pH ¼ 13.1) or a 0.1 M KOH solution containing 10% v/v TEOA. All the potentials were converted to RHE values using the equation V RHE ¼ V Ag/AgCl þ 0.059 Â pH þ 0.197 (V). Photocurrents were measured at 1.23 V versus (vs) RHE under the illumination of a solar simulator (Newport 300 W Xe arc lamp, equipped with an AM 1.5G and water filters) at a power density of 100 mW cm À2 , which was calibrated using a thermopile power meter (Model 919-P, Newport), that is, 1 sun conditions. Nyquist plots of the samples were measured in the frequency range from 100 kHz to 10 mHz at an applied voltage of 1.23 V versus RHE. Mott-Schottky plots of CN were measured in 1 M Na 2 SO 4 at a 2.0 kHz frequency.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.