Peat-derived nitrogen-doped porous carbons as photothermalassisted visible-light photocatalysts for water splitting

BAI Jin-peng, XIAO Nan,, SONG Xue-dan, XIAO Jian, QIU Jie-shan,2,

（1. State Key Lab of Fine Chemicals, Liaoning Key Lab for Energy Materials and Chemical Engineering, PSU-DUT Joint Center for Energy Research, School of Chemical Engineering, Dalian University of Technology, Dalian 116024, China;

2. College of Chemical Engineering, Beijing University of Chemical Technology, Beijing 100029, China）

Abstract： Photocatalytic H2 evolution is considered one of the most important processes for H2 production. Carbon materials are potential candidates for large-scale and cost-effective photocatalytic water splitting, yet their activity needs to be further improved.We report the synthesis of nitrogen-doped porous carbons using peat moss as a precursor and urea as a nitrogen source. The properties of carbons as photothermal-assisted visible-light photocatalysts were investigated. Due to the photothermal effect, the system temperature increased quickly to 55 °C in 15 min under visible light irradiation, which subsequently helps increase the photocatalytic activity by about 25%. It has been found that the crystallinity and nitrogen content of the carbon materials can be changed by changing the carbonization temperature, and these have an impact on their photocatalytic activity. A peat-derived carbon carbonized at 800 °C, with a N content of 4.88 at.% and an appropriate crystallinity has an outstanding photocatalytic activity with a high H2 evolution rate of 75.6 μmol H2 g−1 h−1 under visible-light irradiation.

Key words: Visible-light；Photothermal-assisted；Carbon based photocatalyst

1 Introduction

Environment pollution and energy crisis as a result of excessively using fossil fuels are huge challenges that human being is facing[1,2]. To mitigate these problems, various technologies, including photocatalytic H2evolution reaction (HER), photoelectrochemical HER, and electrocatalytic HER, which can generate clean, safe, and economical energy, have received lots of attention[3-5]. Although their performance cannot meet the practical requirements, photocatalysis is still considered as the most promising candidate for H2production due to its low cost and mere consumption of solar energy[2,6]. Therefore, it remains challenging to improve the performance of photocatalysts.

As a catalytic process, not only their intrinsic properties, but also the reaction conditions can affect the performance of photocatalysts[7-9]. Meng et al.[10]reported that the photocatalytic reaction rates of ZnO,g-C3N4and TiO2can be increased by about 111.2%,36.1% and 45.7% with increasing the temperature from 4 to 45 °C. Jiang et al.[7]reported that the photocatalytic reaction rates of carbon-based photocatalysts can be increased by about 100% with increasing the temperature from 20 to 50 °C. However, in practical applications, photocatalytic reactions may be carried out at quite low ambient temperature, e.g., in the winter of middle and high latitude. In that case, the additional energy is required to offset the influence of low ambient temperature, which will reduce the energy benefits from photocatalysis.

Photothermal effect, which spontaneously converts solar energy into heat, can increase the temperature of photocatalysis system and is regarded as effective and energy saving methods. Compared with TiO2and C3N4, carbon materials exhibit prominent photothermal effect because of their strong light absorption ability, especially in red band[11,12]. Meanwhile, recent researches have indicated the potential of carbon materials in photocatalytic CO2reduction and water splitting[7,13]. However, the research on carbon-based photocatalysts is still at its nascent stage, and how to improve their performance still remains a big challenge[11].

The major steps of the photocatalytic water splitting process include photon absorption, photoexcited charge separation, charge diffusion and transport and electrocatalytic reaction[14-16]. Therefore, the photocatalysis can be considered to be electrocatalysis where electrocatalytic components induce the water splitting driven by the potential shifts caused by the photocatalyst (photon absorber)[17]. Therefore, the improvement of the photocatalytic activity of carbonbased catalysts is supposed to start from enhancing the absorption of light, photoexcited charge separation,shortening the charge diffusion and transport paths, as well as boosting the electrocatalytic activity[18]. From the point of view of the carbons’ microstructure, the high crystallinity of a photocatalyst can improve the efficiency of charge diffusion and transport[11,19], as well as reduce the number of recombination centers,which avail the diffusion of photoexcited electron-hole from bulk to surface to initiate redox reactions[4]. Meanwhile, abundant heteroatoms doped structural defects can enhance the catalytic properties of carbons[11]. However, the construction of a highperformance carbon-based photocatalyst is challenging, because there is a trade-off between the two mentioned factors, i.e., the increase of crystallinity is generally accompanied by the decrease of heteroatom doping. In view of this, a systematic study and deep understanding on the effect of the microstructure on the photocatalytic performance of carbon materials is quite necessary.

Biomass, possessing abundant functional groups and hierarchical morphologies, is a sustainable and renewable precursor for porous and heteroatom-doped carbon materials. Peat moss, the most abundant biomass, has been applied in producing functional carbons[20]. Herein, we use peat moss to synthesize Ndoped carbons, and investigate the influences of the crystallinity and nitrogen content on the carbons’ photocatalytic performance.

2 Experimental

2.1 Materials and synthesis

Prior to the synthesis, the impurities including sticks and stalks in the peat moss were thoroughly picked out. In a typical procedure, 2 g urea was dissolved in 50 mL deionized water. Then, 2 g peat moss was dispersed in the urea solution, stirred for 4 h, and dried at 100 °C in vacuum. After that, the obtained mixture was carbonized at desired temperature (750-950 °C) in nitrogen. After cooled down, the products were washed with NaOH solution (1 mol L−1),hydrochloric acid (2 mol L−1) and NH4HF2solution(4 mol L−1) in sequence to remove the impurities, and followed by drying at 110 °C in vacuum. The obtained samples were named as PMNC-A, where A refers to the carbonization temperature. For comparison, the counterpart was synthesized under the same condition without the addition of urea, and denoted as PMC.

2.2 Material characterization

The scanning electron microscopy (SEM, Hitachi SU8200) and the transmission electron microscopy (TEM, Philips Tecnai G220) were used to characterize the structures and morphologies of the samples. Thermo Scientific Nicolet iN10 was used to measure the Fourier transform infrared (FT-IR) spectra. X-ray diffraction (XRD) patterns were measured by using a X-ray diffractometer (Rigaku D/Max-2400)with CuKα (λ= 0.154 06 nm). X-ray photoelectron spectroscopy (XPS) measurement was performed on a Thermo VG Scientific ESCALAB 250 spectrometer.The diffuse reflectance spectra (DRS) were measured by a Hitachi F-7000 high-performance spectrophotometer, equipped with a diffuse reflectance accessory.Raman spectra were obtained by using a Raman microscope (Thermo Fisher Scientific DXR) with 532 nm laser excitation. The thermogravimetric analyzer (STA449F3 NETZSCH, Germany) was used to conduct thermogravimetry (TG). Nicolet iS50 was used to measure the gas products from raw material.

2.3 Photothermal effect and water splitting test

The photothermal effect and photocatalytic water-splitting were evaluated via a Xenon-lamp (PLSSXE 300) equipped with a UV-cutoff filter (420-780 nm). Typically, 5 mg of the samples were dispersed in 80 mL of triethanolamine aqueous solution(10 vol%). Then the photodeposition of ～2 wt% Pt cocatalyst was executed by directly adding H2PtCl6·6H2O into the reaction solution. An infrared camera (FOTRIC 237) was used to capture the temperature profiles.

Photoelectrochemical measurements were conducted using a three-electrode cell equipped with an electrochemical analyzer (Chenhua CHI760E). The reference electrode was Ag/AgCl (saturating KCl)electrode and the counter electrode was a platinum wire. The Xenon lamp coupled with a cutoff filter (420-780 nm) was used as the visible light source. The electrolyte was 0.1 mol L−1Na2SO4. The H2production was measured by a thermal conductivity detector(Agilent GC6890N). Linear sweep voltammetry(LSV) test was carried out from 0.2 to 1.5 V vs RHE at a 50 mV s−1scan rate. The electrochemical impedance spectroscopy (EIS) was carried out at −0.4 V (vs.Ag/AgCl) over the frequency range of 10−1-106Hz.The conversion between the reference electrode and the reversible hydrogen evolution electrode is shown in Eq (1).

2.4 Density functional theory (DFT) calculation

In this work, all calculations were performed by using the Gaussian 16 package[21]. The geometry optimization and binding energy of the H2O molecule in different catalysts were calculated using DFT with B3LYP hybrid functional[22]and 6-31G(d) basis sets.The dispersion correction was considered by using D3BJ[23,24].

3 Results and discussion

SEM was performed to investigate the morphology and structure of PMNC. The micronsized carbon particles, which inherit the structure of peat moss,present a hierarchical pore structure (Fig. 1a and b).Abundant pore channels are beneficial to light harvesting and water transmission. The TEM images(Fig. S1) demonstrate the presence of interior cavity, which is able to enhance the light harvesting through scattering and reflection of internal multi-light[25-27].Meanwhile, the interior cavity also can promote charge carrier separation by decreasing the diffusion length of the charge carrier[7,28]. As shown in Fig. S2,with the increase of the carbonization temperature, the morphology of PMNC do not change significantly.

FT-IR spectra of the peat moss and PMNC are shown in Fig. 2a. The peak at 670 cm−1belongs to―NH2, 1 630-1 700 cm−1for C＝O stretching, 2 800-3 100 cm−1for C―H vibration and 3 300-3 500 cm−1for the vibration of N―H bond[5,29]. The FT-IR results indicate that the peat moss is rich in amino groups, which is expected to be partially reserved in the derived carbons. The ―NH2characteristic peaks in FT-IR spectra of PMNC are still obvious, although their intensities are weakened with the increase of the carbonization temperature.

The elemental compositions and chemical states of the PMNC were investigated by XPS. The survey XPS spectra as shown in Fig. 2b indicate the presence of C, N and O. As shown in Fig. 2c, the high-resolution C 1s can be fitted into five peaks, which are sp2C(C＝C/C―C), sp3C (C―C), C―OH/C―N, C＝O,and O―C＝O located at 284.6, 285.3, 286.3, 287.6 and 289.0 eV, respectively[30]. Fig. S3 shows the deconvoluted C 1s spectra of the other PMNC samples.The relative atomic percentage of sp2C increases with increasing the carbonization temperature, suggesting the increased crystallinity degree of the carbons(Table S2).

The structures of PMNC were investigated by XRD and Raman. As shown in Fig. 3a, the XRD patterns of PMNC present a broad peak located at ～26°and a weak peak at ～42°, which can be assigned to the (002) and (100) planes of carbon, respectively.With increasing the carbonization temperature, the XRD patterns exhibit gradually narrowed peaks and decreased averaged002values, indicating the increased crystallinity. Raman was carried out to further investigate the effect of the heat treatment temperature on the microstructure of PMNC (Fig. 3b).Two peaks in the Raman spectra areGband(1 580 cm−1) andDband (1 380 cm−1). The peak intensity ratio (ID/IG) progressively decreases from 1.09 to 0.89 with increasing the carbonization temperature from 750 to 950 °C, indicating the increased degree of crystallinity, which consists with the XRD and XPS results[31]. A higher crystallinity will promote more photoexcited electron-hole pairs to diffuse from bulk to surface to initiate redox reactions.

The relative atomic percentage of PMNC was investigated by XPS (Table S1). Notably, the nitrogen content gradually decreases from 5.04% to 1.91%with increasing the carbonization temperature to 950 °C. The deconvoluted N 1s spectrum of PMNC-800 (Fig. 2d) can be fitted into pyridinic N, amino,pyrrolic N and graphitic N, located at 398.4, 398.6,399.6 and 400.7 eV, respectively[32,33]. For the sake of illustrating the evolution of N-containing groups, the deconvoluted N 1s spectra of the other 4 PMNC photocatalysts are shown in Fig. S4 and the contents of various nitrogen species are summarized in Table S4.As shown in Fig. 4, the graphitic N has been well preserved through the temperature range from 750 to 950 °C. In contrast, the contents of pyridinic N,amino, and pyrrolic N, which are relatively stable below 850 °C, decrease significantly at temperature above 900 °C. Besides, as shown in Fig. S5, the contents of amino in PMC are quite low, indicating that the addition of urea is crucial to the formation of amino.

The contact angles (CAs) were measured as the indicator of wettability, which is an important property of the photocatalyst. The CA of PMNC-800 (43°)is much smaller than that of PMC (112°), indicating the improved wettability of the former by introducing urea, which increases the carbons’ polarity through N doping (Fig. S6). Meanwhile, the droplet of water could infiltrate the PMNC within 560 ms. It is reasonable to assume that the unique highly opened porous structure and N doping can improve the wettability of carbon and consequently promote the catalytic performance.

To investigate the formation of N groups at high temperature, TG of peat moss was carried out. As shown in Fig. S7a, the weight loss below 100 °C can be ascribed to the evaporation of water, while the weight loss at 226 °C is due to the pyrolysis of peat moss. The carbon yield of peat moss is about 32 wt.%.The FT-IR spectra of gas products from PMC(Fig. S7b and c) show peaks at 670 and 1 630-1 700 cm−1belonging to ―NH2and C＝O stretching,respectively, indicating that a part of the ―NH2and C＝O are converted into volatile small molecule during pyrolysis. As a result, the N and O contents of PMC gradually decrease with increasing the carbonization temperature. The raw material for PMNC exhibits a quite different TG curve, which start from 150 °C(Fig. 5a). The initial weight loss can be ascribed to the pyrolysis of urea to release NH3,which results in the presence of the absorption peaks centered at 963 and 932 cm−1. It is interesting that, the peak at 1 630-1 700 cm−1of gas products from PMNC is quite weak in contrast to that of PMC, suggesting the reaction between C＝O and NH3, which may be the reason for N doping by adding urea.

Prior to investigating the photocatalytic performance of PMNC, the cocatalyst dosage of the catalytic system is optimized. Fig. S8a shows the effect of the cocatalyst Pt loading percentage on the photocatalytic performance of PMNC-800 under visible light irradiation. With increasing the Pt loading from 0.5 wt.% to 2 wt.%, the H2production increases from 40.8 to 75.6 μmol h−1g−1. Further increasing Pt content to 3 wt.% leads to a slight reduction in the H2evolution rate, which may be due to the changes in the size and density of Pt nanoparticles[34,35]. Therefore, 2 wt.% Pt loading was adopted as a model cocatalyst dosage for research. The catalyst dosage is also optimized. As shown in Fig. S8b, the H2evolution per gram of PMNC first basically remains unchanged and then reduces with increasing the amount of catalyst. This may be that the catalyst dosage has reached a saturation point. Thus, 5 mg of catalyst was used to perform the experiments. The photothermal effect of catalysts was investigated. As shown in Fig. 6a, the temperature of pure water rises to 33.5 °C after 25 min visible light irradiation. When 5 mg g-C3N4was added, the temperature rises from 19 to 37.1 °C(Fig. 6b), indicating the poor photothermal effect of pale yellow g-C3N4. In Fig. 6c, system temperature rises to 54.6 °C within 15 min, indicating the excellent photothermal effect of PMNC which can effectively rise the system temperature within a short time.The effect of temperature on the sample’s photocatalytic performance was studied by carrying out the photocatalytic reaction at room temperature and 55 °C. As shown in Fig. S8c, the photocatalytic reaction rates of PMNC are increased by about 25% with increasing the temperature from room temperature to 55 °C.

Fig. 7a shows the photocatalytic H2evolution rate over PMNC samples under the optimum conditions. Among the five tested photocatalysts, PMNC-800 shows the highest H2evolution rate (75.6 μmol H2g−1h−1), which is 2.25 and 4.03 times of those of PMNC-750 and PMNC-950, respectively. Generally, both the microcrystalline structure and chemical composition can influence the performance of carbon-based photocatalysts[13]. Therefore, the superior photocatalytic performance of PMNC-800 may be attributed to the well balanced crystallinity and N doping. The catalyst shows excellent stability and still retains more than 92% of the initial activity after 100 h (Fig. 7b).After the 100-h stability test, SEM was performed to investigate the morphology and structure of PMNC.As shown in Fig. S9a, obvious changes cannot be observed in the morphology and structure of PMNC.Meanwhile, the N states of PMNC was investigated by XPS (Fig. S9b). As shown in Fig. S9c, after the stability test, the N contents reduce slightly. Therefore, the excellent stability of PMNC can be attributed to the stable morphology and nitrogen atom.

The optical absorption property and band structure of PMNC were also investigated. The UV-vis DRS in Fig. S10a illustrate that the light absorption ability of PMNC is better than those of most of the reported g-C3N4photocatalysts[36,37]. Fig. S10b-f reveal that the bandgaps of PMNC-750, PMNC-800, PMNC-850, PMNC-900, and PMNC-950 are 1.99, 2.08, 2.11,2.04 and 1.85 eV, respectively. The corresponding valence bands (VB) of PMNCs obtained from XPS are 1.41, 1.52, 1.68, 1.58 and 1.19 eV (Fig. S11). According to the formulaECB=EVB−Eg, the conduction band (CB) potential of PMNC are calculated to be−0.58, −0.56, −0.43, −0.46 and −0.66 eV[38], and the band structure diagram is illustrated in Fig. 8.

LSV method was used to testify the effect of doped N on the photocatalytic performance of carbon materials. As shown in Fig. S12a, the current density of PMNC-800 is much higher than that of PMC-800.The EIS was conducted to appraise the carrier mobility of the photocatalysts. It can be observed (Fig.S12b) that PMNC-800 possesses a smaller semicircle than PMC-800. The transient photocurrents of PMNC-800 and PMC-800 in 40-sec light on/off cycles are shown in Fig. S12c. The much larger photocurrent density of the former indicates its superior photoelectrochemical performance due to the higher nitrogen doping, which can improve the separation of excitons and reduce the resistance for interfacial charge carriers’ transfer. Besides, as shown in Fig. S12d, PMNC-800 exhibits an evidently decreased PL emission intensity with respect to PMC-800, indicating the suppressed electron-hole pair recombination rate by N doping. Consequently, the H2evolution rate of peat moss derived carbon has been greatly improved(around 8.4 times) after N doping (Fig. S12e). The results have confirmed that the introduction of N can dramatically enhance the catalytic properties of carbons.

The effect of the carbonization temperature on the photocatalytic performance of PMNC was also investigated. As shown in Fig. 9a, the LSV curve of PMNC-800 exhibits the highest current density among all the samples. The current density of PMNC for water oxidation can be further boosted upon the visible light irradiation (Fig. S13). However, the current density increments of various samples are quite different. Specifically, PMNC-800 shows the biggest current density increment, indicating the most sensitive photoresponse. As shown in Fig. 9b, PMNC-800 possesses a medium-sized semicircle in all the samples.As expected, interfacial charge carriers’ transfer resistance of PMNC-950 is the smallest due to its highest crystallinity, however, its photocatalytic activity is quite poor. Same as the current density for water oxidation, the resistance for interfacial charge carriers’transfer can be further improved upon the visible light irradiation (Fig. S14). These findings illustrate that PMNC has the ability of photo response. Both the increase in the N content and crystallinity can boost the photo response of carbon materials. However, PMNC-800 shows the best photo response, which means that the balance of the crystallinity and defects content is crucial to the carbon-based photocatalysts. The transient photocurrents for five photocatalysts in 40 sec light on/off cycles are shown in Fig. 9c. As expected,the photocurrent intensity increase sequence of these five catalysts is: PMNC-950< PMNC-900< PMNC-750< PMNC-850< PMNC-800, which consists with the photocatalytic H2production rates. PMNC-800 shows the highest activity due to its balanced crystallinity and nitrogen doping[13].

To further interpret the experimental results, the H2O-adsorption abilities at assorted sorts of N-doping sites were simulated, as shown in Fig. 10. A graphene sheet is adopted to model the interaction between PMNC and the H2O molecule due to its similar structure to PMNC. The adsorption energy between ideal graphene layer and H2O molecule was first investigated, which is amounted to be −0.24 eV (Fig. 10a). In the case of N-doping, the N dopant and adjacent carbon atoms become more active so that the binding energy with H2O are immensely increased to −0.55,−0.37, and −0.38 for pyridinic N, amino, and pyrrolic N, respectively (Fig. 10b-e). The DFT calculation confirms that the tuning of the N dopants can significantly boost the catalytic performance, which is consistent with the performance results. In addition,amino N can prolong the lifetime of excited electrons which has been proved previously[29]. Therefore, the high content of amino N and pyridinic N can improve the photocatalytic performance of PMNC.

4 Conclusion

A novel PMNC was prepared with peat moss as the precursor. Thanks to the excellent photothermal effect of PMNC, the system temperature can be elevated to 55 °C within 15 min under visible light irradiation, by which, the photocatalytic activity is increased by about 25%. N doping can not only improve the efficiency of electrocatalytic HER, but also effectively separate the photogenerated carriers and suppress their recombination. The photocatalytic activity of optimized sample is increased by 4.03 times through balancing its crystallinity and nitrogen content. As a result, the PMNC shows performance in visible-light-driven water splitting with a H2evolution rate of 75.6 μmol H2g−1h−1. This work provides a practical and facile method to produce nitrogen-doped carbons, which provides foundation for carbon materials in photochemical applications.

Acknowledgements

National Natural Science Foundation of China(U2003216); Fundamental Research Funds for the Central Universities of China (DUT20LAB131).