Novel carbon nitride composites with improved visible light absorption synthesized in ZnCl₂-based salt melts

Abstract

Poly(triazine imide)-based carbon nitride materials with BET surface areas up to 200 m² g⁻¹ were synthesized in ZnCl₂ containing salt melts without the use of hard templates. We found that the composition, structural order, optical properties and morphology of the products can be adjusted by careful selection of synthesis parameters. The nature of the salt eutectic and precursor concentration in the melt have an especially large influence, with ZnCl₂ being a reactive solvent. This novel synthesis route provides access to easily processable materials with improved optical absorption in the visible range that can be used as composite photocatalysts, CO₂ adsorbents or nanocomposite fillers.

---

Introduction

Graphitic carbon nitride materials are light element composed polymeric semiconductors which have recently found numerous applications as photo-¹ and electrocatalysts,² oxidation catalysts,³ catalyst supports and nanocomposite fillers,⁴ and are reported to be promising even for photovoltaic applications.⁵ For most applications, carbon nitrides with a special morphology and relatively high surface areas are needed. Typically, this is achieved by using various silica templates, which need to be removed after the synthesis, mostly by reaction with an HF source.⁶ The latter step however requires additional safety precautions and put severe restrictions on scalability of carbon nitrides production. Alternative methods to adjust morphology and surface areas of final products include, for example, supramolecular preorganization of monomers,⁷ “soft-templating”⁸ and solvothermal synthesis.⁹ Besides, C₃N₄-related materials such as crystalline poly(triazine imides) (PTI) that have well-defined morphologies and possess increased surface areas can be prepared using LiX/KX salt melts (X = Cl, Br) as reaction media.¹⁰

In general, molten salt(s) can serve as a solvent for high-temperature materials synthesis, as catalyst (for example, ZnCl₂ in trimerization reactions) but also as “soft template” for tailoring micro- and mesoporosity of the resulting products. Previously, ZnCl₂ and ZnCl₂-containing melts were successfully used for the synthesis of porous covalent triazine-based frameworks¹¹ and porous functional carbons.¹²

In this contribution, we further explore the utility of salt melt synthesis for the preparation of carbon nitrides, with the main focus on increasing the surface areas of products, as well as on controlling their structure and morphology by varying the synthesis parameters. We report on the preparation of new carbon nitride hybrid materials using zinc chloride containing eutectic mixtures, their characterization by means of X-ray diffraction (XRD), elemental analysis (EA), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), FTIR-spectroscopy and gas adsorption, and discuss potential applications.

Experimental section

Chemicals and materials

Lithium chloride (99%), sodium chloride (99.5%), potassium chloride (99%), cesium chloride (99%) and Rhodamine B (95%) were purchased from Sigma Aldrich. Zinc chloride (98%) was purchased from Alfa Aesar, melamine (99%) from Acros Organics and dicyandiamide (98%) from Merck. All the chemicals were used without further purification.

Synthesis procedure

Salts and melamine were ground together in a glovebox (mBraun Unilab, O₂ < 0.1 ppm, H₂O < 0.1 ppm) under argon atmosphere according to the eutectic compositions (Table 1). Reaction mixtures (∼5–10 g) were transferred into porcelain crucibles and covered with lids. Crucibles were placed in oven and heated under constant nitrogen flow (15 mL min⁻¹, Scheme S1a†). The crude products were removed from the crucibles and washed first with ethanol (50–100 mL), then with deionized water (50–100 mL) and finally with 1 M HCl solution (50–100 mL). Each step was carried out for 24 hours. Final products were isolated by filtration, thoroughly washed with deionized water (500 mL) and dried in a vacuum oven at 50 °C for 15 h.

Table 1 Composition and melting points of eutectic salt mixtures used in this paper

Salt mixture A/B(/C)	A content, wt%	B content, wt%	C content, wt%	Molar content of ZnCl2	Melting temperature, °C (ref. 17)
ZnCl2	100	—	—	1.00	318
LiCl/ZnCl2	8.5	91.5	—	0.770	294
NaCl/ZnCl2	23.7	76.3	—	0.580	270
KCl/ZnCl2	36.3	63.7	—	0.490	230
CsCl/ZnCl2	48.3	51.7	—	0.575	263
NaCl/KCl/ZnCl2	10.7	13.8	75.5	0.600	203

---

Reference g-C₃N₄ was prepared by heating DCDA with the ramp of 2.3 °C min⁻¹ up to 550 °C and subsequent holding at this temperature for 4 h under constant nitrogen flow (15 mL min⁻¹) in a covered crucible. The final product was thoroughly ground.

Dye photodegradation experiments

For photodegradation experiments Rhodamine B (RhB) was used as a model dye. 5 mg of a catalyst were dispersed in 5 mL of RhB solution in deionized water (10 mg L⁻¹) in a cuvette. The suspension was agitated in dark for 30 minutes in order to reach dark-adsorption equilibrium, and then illuminated with a LED module emitting at 420 nm (12 W, OSA Opto Lights) from the fixed distance of 5 cm. The RhB degradation was monitored as the decrease of its concentration over irradiation time using UV/Vis spectrophotometer. For this purpose, after a certain irradiation time t, an aliquot of 300 μL of reaction mixture was taken, diluted with 1.7 mL of deionized water, kept for 30 minutes in dark to allow for catalyst precipitation, then UV-Vis absorption spectrum was measured. For calculation of the rate constants, dye absorbance (at 550 nm) after dark equilibration (t = 0 min), A₀, and at t = 60 min, A, were used. Rate constants k were calculated as follows:

k = ln (A/A₀) × (60 min)⁻¹

Characterization

Powder X-ray diffraction patterns were measured on a Bruker D8 Advance diffractometer with Cu_Kα radiation (λ = 0.15148 nm) equipped with a scintillation counter detector applying a step size 0.05° 2θ and counting time of 3 s per step. FT-IR spectra were recorded on a Varian1000 FT-IR spectrometer equipped with an attenuated total reflection unit with diamond applying a resolution of 4 cm⁻¹. Nitrogen adsorption–desorption measurements were performed after degassing the samples at 150 °C for 20 hours using a Quantachrome Quadrasorb SI-MP porosimeter at 77.4 K. CO₂ and N₂ adsorption–desorption isotherms at 273, 283 and 303 K were conducted on a Quantachrome Autosorb-1MP instrument after prior degassing. The specific surface areas were calculated by applying the Brunauer–Emmett–Teller (BET) model to adsorption isotherms (N₂ at 77.4 K) for 0.05 < p/p₀ < 0.3 using the QuadraWin 5.05 software package. Elemental analysis was accomplished as combustion analysis using a Vario Micro device. SEM images were obtained on a LEO 1550-Gemini microscope. Optical absorbance spectra of powders were measured on a Shimadzu UV 2600 equipped with an integrating sphere. The absorption spectra of RhB solutions were recorded on a T70 UV/VIS spectrophotometer (PG instruments Ltd.). The emission spectra were recorded on LS-50B, Perkin Elmer instrument. The excitation wavelength was 300 nm. EDX investigations were conducted on a Link ISIS-300 system (Oxford Microanalysis Group) equipped with a Si(Li) detector and an energy resolution of 133 eV. X-ray photoelectron spectroscopy (XPS) was performed on a Multilab 2000 (Thermo) spectrometer equipped with Al Kα anode (hν = 1486.6 eV). All spectra were referenced to the C 1s peak of adventitious carbon at 285.0 eV. For quantification purposes, survey at a pass energy of 50 eV and high-resolution spectra at pass energy of 20 eV were recorded and analyzed by XPS Peak 4.1 software (written by Raymund Kwok). The spectra were decomposed assuming line shapes as sum functions of Gaussian (80%) and Lorentzian (20%) functions. Raw areas determined after subtraction of a Shirley background¹³ were corrected according to following sensitivity factors¹⁴ (C 1s: 0.25; N 1s: 0.42; O 1s: 0.66; Cl 2p: 0.73; Zn 3p: 0.75). Etching Ar⁺ bombardment was performed at 2 kV and 18 mA of ionic current.

Results and discussion

General considerations

The synthesis of graphitic carbon nitride (g-C₃N₄) usually starts from dicyandiamide (DCDA) and proceeds by a number of sequential condensation processes that occur between ∼230 and 550 °C (Scheme 1a). We selected zinc chloride as a main component of the salt melt due to the two following reasons. First of all, ZnCl₂ is a moderate-strength Lewis acid and was expected to solubilize the precursor and at least some of the condensation intermediates owing to strong Lewis acid–base interactions. Besides, it has a low melting temperature that can even be lowered to T_m < 300 °C in eutectic mixtures with alkali metal chlorides (see Table 1). On the other hand, melamine was chosen as a carbon nitride precursor due to the fact that its condensation temperature towards melam is estimated to be roughly 335 °C (ref. 15) (Scheme 1a). That way, we hope to ensure melting of the solvent before the onset of the melamine condensation reaction.

Scheme 1 (a) Idealized condensation scheme of dicyandiamide (DCDA) to graphitic carbon nitride; (b) idealized structure of poly(triazine imide)/Li⁺Cl⁻.¹⁶

Initial investigations

Initial investigations were conducted at a fixed precursor to salt mixture ratio (1 : 5 by weight), while the composition of the salt mixture was varied by choosing different alkali metal chlorides.

The synthesis and work-up procedure yielded colored products, which were analyzed by EA, EDX and XPS with regard to their composition (Table 2). The C/N weight ratios in the products range from 0.59 to 0.61 and are slightly higher than the value for reference g-C₃N₄ (0.58) but slightly lower than the stoichiometric value for C₃N₄ (0.64). This might be explained by incomplete amide condensation and the corresponding presence of terminal NH/NH₂-groups. The sum of C, N, and H as obtained from EA is typically lower than 100%, usually around 82–87%. The remaining 13–18 wt% are Zn (3–5%), O (10–15%) and Cl (1–3%) ions as analyzed by EDX spectroscopy and XPS for selected samples. Zinc could be removed completely by washing the composites with 10 M HCl. This step is however accompanied by additional protonation of the product surface¹⁸ and was hence not employed in the default recipe. The presence of oxygen in products is partially the result of the aqueous work-up of the reaction mixtures but also due to the strong adsorption of water and CO₂ under ambient conditions.

Table 2 Composition of materials synthesized in different ZnCl₂ containing salt melts (EA data), and calculated BET surface areas of products

Melt	C, wt%	N, wt%	H, wt%	C/N, wt%	100 − (C + N + H)	Color	SBET, m^2 g^−1
Ref. bulk C3N4	35.0	60.1	2.06	0.58	2.8	Yellow	10
LiCl/ZnCl2	30.6	50.8	2.70	0.60	15.9	Pale yellow	20
NaCl/ZnCl2	29.6	49.1	3.04	0.60	18.3	Brown	193
KCl/ZnCl2	30.1	49.2	3.06	0.61	16.0	Yellow	42
CsCl/ZnCl2	30.2	49.8	3.13	0.59	16.5	Yellow	68
NaCl/KCl/ZnCl2	29.7	50.3	3.13	0.59	16.9	Brown	71
Pure ZnCl2	31.5	52.5	2.86	0.60	13.1	Pale yellow	58

---

XRD and FTIR spectroscopy investigations showed that the synthesized materials have features typically observed for previously described carbon nitride polymers, but differ from each other with respect to the degree of structural order. Fig. 1a illustrates that products prepared in NaCl/ZnCl₂, CsCl/ZnCl₂ and NaCl/KCl/ZnCl₂ eutectics are mostly X-ray amorphous as they show only broad halos at 2θ ∼ 15° and 27°, LiCl/ZnCl₂-derived material shows quite defined reflections corresponding to higher crystallinity, while the product obtained from pure ZnCl₂ is characterized by outstanding structural order. The main 2θ reflection around 27° is present in the diffractograms of all materials and corresponds to interplanar stacking of carbon nitride layers. Another reflection observed for most of the materials (except for NaCl/KCl/ZnCl₂ and NaCl/ZnCl₂ eutectics) at 2θ ∼ 12–15° is related to in-plane structural packing motifs. The appearance of XRD patterns of products prepared in ZnCl₂ and LiCl/ZnCl₂ suggests that these materials are poly(triazine imide)-like as corroborated by recent investigations of acid washed poly(triazine imides) obtained from LiCl/KCl.¹⁹ KCl/ZnCl₂-derived material seems to be composed of two phases characterized by different crystallinity.

Fig. 1 (a) WAXS patterns and (b) FTIR spectra of bulk-C₃N₄ and products synthesized in salt melts containing ZnCl₂. Vertical line indicates 1350 cm⁻¹ position.

In principle, crystallinity of carbon nitrides prepared in different ZnCl₂-containing melts increases in the row NaCl/KCl/ZnCl₂ ∼ NaCl/ZnCl₂ < CsCl/ZnCl₂ < KCl/ZnCl₂ < LiCl/ZnCl₂ < bulk g-C₃N₄ < ZnCl₂, and therefore can be tuned by varying the nature of the salt eutectic used for the synthesis.

Fig. 1b shows that the main IR absorption bands displayed by the synthesized materials are similar to those of reference g-C₃N₄. Namely, the large absorption band at 1200–1650 cm⁻¹ corresponds to stretching vibrations of CN heterocycles, a strong vibration at ∼800 cm⁻¹ is attributed to deformation vibrations of triazine or tri-s-triazine ring, and the broad band between 2400 and 3400 cm⁻¹ indicates stretching vibrations of surface hydroxyl groups (OH–) and terminal and residual amino-groups (NH₂–, NH–). The presence of the distinct absorption band at ∼1350 cm⁻¹ in ZnCl₂-, LiCl/ZnCl₂- and KCl/ZnCl₂-derived materials suggests that they are built of triazine rather than of tri-s-triazine rings²⁰ that is in good agreement with the XRD data. Besides, the amorphous nature of NaCl/ZnCl₂, CsCl/ZnCl₂ and NaCl/KCl/ZnCl₂-derived products causes broadening and subsequent overlap of the absorption bands, while improved structural order (products from pure ZnCl₂, LiCl/ZnCl₂ and KCl/ZnCl₂) results in more defined vibration peaks.

Amorphous CsCl/ZnCl₂–C₃N₄ sample was investigated by XPS in order to unravel the structure of the carbon nitride polymer and intercalated Zn²⁺ species. The C 1s spectrum (Fig. 2a) shows a main carbon species with a binding energy of 288.2 eV corresponding to CN₃ bonds.²¹ In addition to the C 1s peak of adventitious carbon at 285.0 eV, there is a weak peak at 283.3 eV which results from carbide impurities in the sample. The deconvolution of the N 1s signal reveals three contributions (Fig. 2b): at 401.2 eV (NH_x groups), 399.7 eV (C–N=C) and 397.6 eV (deprotonated N-sites). The number of contributions as well as their chemical shifts matches the calculated values for N 1s electrons in PTI prepared in LiCl/KCl.²² The C/N ratio is then calculated as a ratio of the area of the main carbon peak to the total area of nitrogen peaks. The value of 0.59 fits the one for poly(triazine-imide) reported by Wirnhier et al.^10a and differs from fully condensed heptazine-based C₃N₄ (0.64). Furthermore, the ratio of amino-nitrogen atoms (NH_x groups) to ring nitrogen atoms (C–N=C) calculated as a ratio of the corresponding peak areas, is equal to 0.49 and is close to the theoretical value of 0.5 for the idealized structure of PTI ((C₃N₃)₂(NH)₃). For heptazine-based structure this value would be much lower because only uncondensed terminal amino-groups would contribute to the XPS spectrum and only weak peaks of –NH_x are reported for this case.²¹ All these findings prove that the carbon nitride component of the composite is based on poly(triazine imide).

Fig. 2 (a) C 1s and (b) N 1s XPS spectra of carbon nitride synthesized in CsCl/ZnCl₂.

XPS investigations further confirmed the presence of O, Zn and Cl in the sample. O 1s signal (Fig. S1a†) consists of three contributions: at 529.3 eV (assigned to O–Zn bonds²³), 531.5 eV (–OH groups)²⁴ and 533.0 eV (adsorbed water).²⁵ The ratio between these three contributions is 1 : 16 : 8. The presence of Zn–O bonds is further confirmed by Zn 3p signal (Fig. S1b†) which shows two main and almost equal contributions: at 88.4 eV for Zn–O bonds²⁶ and at 89.6 eV for Zn–Cl and Zn–OH bonds. The peak of Cl 2p (Fig. S1c†) at 198.1 eV, by-turn, indicates the presence of Zn–Cl bonds. Basing on these findings, we conclude that the intercalated Zn species are mainly zinc oxide and zinc oxychloride. The latter one is the product of hydrolysis of ZnCl₂ during the work-up procedure. The qualitative differences between surface and bulk of the product were analyzed by comparing the XPS spectra before and after Ar-ion bombardment of the sample which allows removing the surface layer of the material. The Cl 2p signal after ion bombardment reveals two peaks that correspond to different Zn–Cl bonds (Fig. S2a†). The contribution at 198.2 eV is due to zinc oxychloride while the one at 199.5 eV indicates the presence of zinc chloride²⁷ which didn't hydrolyze being entrapped in the structure. The presence of ZnCl₂ in the bulk of the product is also supported by Zn 3p signal (Fig. S2b†) which shows now an additional contribution at 91.4 eV due to Cl–Zn–Cl bonds. The amount of zinc oxychloride in the bulk of the product is much higher than the amount of zinc oxide, though at the surface their quantities were almost even. This finding together with the fact that the amount of Zn in the bulk of the material is at least 3 times higher than at the surface illustrates the difficulty to remove the intercalated Zn species ascribed to the diffusion limitations. In contrast, the oxygen content in the bulk of the product is at least two times lower than at the surface that is explained by the removal of surface hydroxyl groups and surface-adsorbed water. The weight content of the elements calculated from XPS data is in good agreement with the values obtained from EA, ICP and EDX measurements (Table S1†).

The morphology of the prepared materials was investigated by SEM. It was found to be fairly homogeneous within each of the samples and variant depending on the nature of alkali metal chloride constituting the salt melt (Fig. 3). Carbon nitrides are obtained as crystalline densely-packed nanosheets of 100–200 nm in diameter from pure ZnCl₂ and LiCl/ZnCl₂ eutectic melts. NaCl/ZnCl₂, CsCl/ZnCl₂ and NaCl/KCl/ZnCl₂ melts give amorphous nanoparticles with the diameters of 50 ± 10 nm, 60 ± 10 nm and 40 ± 10 nm, respectively. KCl/ZnCl₂ eutectic gives rise to the mixed product morphology containing larger sheets (1–2 μm in diameter), and spherical nanoparticles aggregates (∼300 nm in diameter).

Fig. 3 Representative SEM images of carbon nitrides synthesized in (a) ZnCl₂, (b) LiCl/ZnCl₂, (c) KCl/ZnCl₂, (d) NaCl/ZnCl₂, (e) CsCl/ZnCl₂ and (f) NaCl/KCl/ZnCl₂.

ZnCl₂ salt melt derived materials have higher specific surface areas as determined by Brunauer–Emmett–Teller (BET) analysis of the gas adsorption data compared to the reference g-C₃N₄ (see Table 1). However, the values are much lower than those reported for other types of materials prepared in ZnCl₂ melts (CTFs¹¹ and carbons¹²). The appearance of nitrogen adsorption–desorption isotherms suggests the absence of any kind of inner porosity in final MCl/ZnCl₂–C₃N₄ products, which is typical for carbon nitrides prepared in salt melts.^10a Hence the estimated surface area values are reflecting only the external surface area of material particles, in good agreement with the results of SEM investigations.

The structural order and morphology of the products seem to correlate with the amount (χ) of ZnCl₂ in the salt melt at a fixed melamine to salt weight ratio (1 : 5) (see Table 1). Thus, at χ(ZnCl₂)∼0.6, products are obtained as amorphous spherical nanoparticles. Being a strong Lewis acid, ZnCl₂ interacts with cyano-groups of precursors and intermediates. This interaction facilitates their solubilization in the salt melt, but is also the source of the discussed structural complexity. The second component of the salt mixture (MCl, M = Li⁺, Na⁺, K⁺, Cs⁺) allows tuning the solubility of the reaction intermediates in the salt melt and thus influences the onset of “polymer–salt melt” phase separation. Li⁺ and K⁺ cations favor the solubilization of the reaction intermediates that results in slower phase separation and allow the growth of crystalline poly(triazine imides). Na⁺ and Cs⁺ seem to cause rather quick supersaturation of the melt with oligomeric intermediates that leads to the formation of numerous amorphous nanoparticles, which cannot crystallize into more extended networks.

Influence of synthesis parameters

In order to control morphology and crystallinity of salt melt derived carbon nitrides, the influence of various reaction parameters on product properties was studied using NaCl/ZnCl₂ as an example eutectic. Among these are synthesis temperatures (400, 450, 500, 550, 600 °C) (Fig. S3 and Tables S2, S3†), heating rates (2.5, 5, 10, 20, 40 °C min⁻¹) (Fig. S4 and Table S4†), holding times (2, 4, 6, 8, 10 hours) (Fig. S5 and Table S5†), precursor to salt ratios (1 : 1, 1 : 2, 1 : 5, 1 : 10, 1 : 20) (Fig. S6 and Table S6†), synthesis atmospheres (N₂, air, ampoule) and type of C₃N₄ precursors (DCDA, melamine). The results of these studies are briefly discussed below, but for more detailed information the reader is referred to the ESI.†

We found that the microstructure of products is mainly defined already at 400 °C. This means that the final morphology is determined by precipitation and crystallization of intermediary condensates at the point of vanishing solubility in the salt melt. No significant changes in FTIR spectra and WAXS diffractograms can be observed when the synthesis temperature is further increased from 400 to 600 °C (Fig. S3†). However, the degree of condensation of the polymeric networks increases with raising the reaction temperature, resulting in slightly improved C/N ratios (0.59 to 0.61) and narrowing the band gap of the resulting semiconductors, as already visually observed as an intensification of products colors. In the case of applying a one-step heating procedure (Scheme S1b†), materials obtained at 550 °C are characterized by slightly higher values of the BET surface areas if compared to those synthesized at 600 °C (144 and 115 m² g⁻¹, respectively) while providing the same C/N ratio (0.61, Table S3†), thus 550 °C was selected as the temperature of choice for running the condensation. This is in good agreement with previous observations however on bulk condensation of C₃N₄ precursors.

Variations of the heating rate and holding time (Fig. S4 and S5†) have only minor influence on products structures, though the best products compositions characterized by a C/N ratio of 0.61 were obtained at 10° min⁻¹ ramp and 6 hours holding time. Other studied parameters gave lower C/N ratios in the products (Tables S4 and S5†).

As the morphology is given by a product precipitation at rather early stages of the condensation, the concentration of the precursor in the melt must have crucial impact on the morphology and crystallinity of ZnCl₂-derived carbon nitrides. Here, three different regimes were established: both low (1 : 1 precursor to salts weight ratio) and high (1 : 10, 1 : 20 wt ratios) melamine concentrations give highly crystalline products, while concentrations in between (1 : 2, 1 : 5 wt ratios) lead rather to amorphous polymeric structures (Fig. S6†). The crystalline species have sheet-like morphology characterized by different diameters: 0.3–1 μm for 1 : 1 ratio, 1.5–2 μm for 1 : 10, and 50–80 nm for 1 : 20. The intermediary concentrations yield nanoparticles 50–100 nm in size (Fig. 4). The BET surface areas of products after acidic work-up vary between 60 and 190 m² g⁻¹ (Table S6†).

Fig. 4 SEM images of materials synthesized in NaCl/ZnCl₂ salt melts using different precursor to salt ratios: (a) 1 : 1, (b) 1 : 2, (c) 1 : 5, (d) 1 : 10, (e), (f) 1 : 20.

We can only speculate about the origin of this morphology changes. We suggest that at high precursor concentrations, a molecular complex between melamine and ZnCl₂ is initially formed, which consumes all ZnCl₂. Further transformations do not occur in a solution anymore but in solid state, and rather the structure of complexes between ZnCl₂ and intermediates determines the high order in the final products. Upon decreasing melamine concentration, the initially formed molecular complex is redissolved at higher temperatures, but while polymerizing quickly reaches supersaturation and precipitates as amorphous nanoparticles. In some cases (NaCl/ZnCl₂, CsCl/ZnCl₂, NaCl/KCl/ZnCl₂) even spinodal decomposition of the polymer–salt melt phase is assumed to take place, yielding highly microporous materials (see below). At low precursor concentration, demixing of phases occurs slowly, and the polymer (PTI) has enough time to crystallize. It ejects the solvent because the cohesion energy of carbon nitride is higher than the secondary valence to the metal center.

Products prepared in pure ZnCl₂

In order to elaborate on the reactive nature of pure ZnCl₂ and the role of the second component of the eutectic salt mixture to moderate this reactivity, we also investigated the condensation of melamine in pure ZnCl₂ at different precursor concentrations. Materials derived from pure ZnCl₂ melts have typically dense structures with heterogeneous morphologies, which complicate the removal of the intercalated Zn ions, especially at low precursor to salt ratios (1 : 1, 1 : 2). Some illustrative SEM images of products obtained from ZnCl₂ melts at different precursor concentrations are shown in Fig. 5, while Fig. S7a† presents typical WAXS patterns of composites. Similar to NaCl/ZnCl₂ eutectic, three regimes determining the product morphology and crystallinity are observed. At 1 : 1 ratio, both nanoparticle and nanosheet morphologies can be recognized, at 1 : 2 ratio the resulting material is obtained as amorphous spherical nanoparticles, while at 1 : 5 and 1 : 10 ratios the products are mostly crystalline, organized into sheet-like structures covering broad range of shapes and dimensions (from the nano- to micrometer range). As a consequence, the BET surface areas of those products are typically low.

Fig. 5 Selected SEM images of materials synthesized in ZnCl₂ using different precursor to salt ratios: (a) 1 : 1, (b) 1 : 2, (c) 1 : 5 and (d) 1 : 10.

Due to the very strong donor–acceptor interactions of ZnCl₂ with the precursor, zinc cyanamide (also called zinc carbodiimide, Zn(CN₂)) is formed as a by-product when performing the reaction in pure ZnCl₂. At 1 : 10 precursor to salt ratio, the yield of ZnCN₂ is ∼50%. When the ratio is further lowered down to 1 : 20, solely ZnCN₂ is formed (Fig. S7a and b†). These results prove the precoordination between the salt and the monomer via N–Zn donor–acceptor bonds formation and point out the alternative reaction pathway in the case of ZnCl₂ if compared to LiX/KX (X = Cl, Br) melts. LiX/KX melts only weakly solubilize poly(triazine imides), therefore phase demixing occurs early, and big crystalline structures can grow from that melts. On the contrary, strong binding of Zn to condensation intermediates ensures their solubilization, and at later condensation stages the reaction progresses towards the thermodynamically more stable Zn(CN₂) rather than to C₃N₄. Thus, ZnCl₂ should be always considered as a reactive solvent in the case of carbon nitride synthesis.

Zn(CN₂) can be easily removed by washing the product mixture with diluted acid, or converted to zinc oxide by re-heating the composite in air (2 hours at 400 °C; note that contrary to carbon nitride Zn(CN₂) is not oxidation stable). In this way, PTI/ZnO composites can be obtained. Fig. 6 shows a TEM image and SAED pattern of such a composite containing ∼40 wt% C₃N₄ and ∼60 wt% ZnO (crystallite size: 13 ± 2 nm) and having a BET surface area of 270 m² g⁻¹ (Fig. S8†). The photocatalytic activity of this material is discussed below.

Fig. 6 TEM image and SAED pattern of PTI/ZnO composite.

Optical absorption and emission properties of ZnCl₂-derived carbon nitrides are compared with those of g-C₃N₄ in Fig. 7a and b, respectively. As it can be seen, NaCl/ZnCl₂- and CsCl/ZnCl₂-derived materials absorb visible light over a wide wavelength range. If compared to g-C₃N₄, the absorption band edge of these composites is obviously smeared to longer wavelengths. The appearance of the absorption spectra of NaCl/ZnCl₂–C₃N₄, CsCl/ZnCl₂–C₃N₄ and PTI/ZnO composite (Fig. S9†) is typical for dyade materials, such as carbon@TiO₂.²⁸ In a dyade, two components form a joint electronic system, and the altered absorption spectrum likely points out at charge transfer between carbon nitride and Zn containing species (such as ZnO). At the same time, the steady state photoluminescence of the products from NaCl/ZnCl₂ and CsCl/ZnCl₂ excited at 300 nm is negligible, that implies that the relaxation of excitons occurs via non-radiative pathways: charge transfer at heterojunction and/or non-radiative recombination at the impurity states in the crystal lattice.

Fig. 7 (a) UV-visible light diffuse reflectance spectra and (b) steady-state photoluminescence spectra of g-C₃N₄ and MCl/ZnCl₂–C₃N₄ (M = Li, Na, K, Cs) excited at 300 nm.

The product obtained from LiCl/ZnCl₂ melt on the other hand absorbs less visible light and is characterized by slightly stronger emission than g-C₃N₄. Both absorption and emission spectra indicate that this semiconductor has a wider band-gap compared to the reference carbon nitride. This is in good agreement with the assignment that this structure is based on triazine imide moieties, as previously reported salt melt products.^19,22 The optical properties of the material synthesized in KCl/ZnCl₂ lie in between those of LiCl/ZnCl₂ and NaCl/ZnCl₂- or CsCl/ZnCl₂-derived composites, that is in agreement with XRD and SEM data discussed above.

Composites obtained after only aqueous work-up of the reaction mixtures

Materials, which were only washed with water, but not with 1 M HCl showed a C/N weight ratio of 0.70–0.85. This suggests the formation of some carbonaceous species other than C₃N₄ (e.g. ZnCO₃). This goes along with an overall low content of C, N, H (sum: 40–50 wt%, Table S7†) and the presence of other elements. Small amounts of zinc cyanamide, zinc cyanide and zinc oxychloride might also be present, according to XRD studies (Fig. S10†). Acid washing step efficiently removes all these by-products, so that the final C/N ratio is ∼0.6 again, as mentioned above.

The water-washed MCl/ZnCl₂-derived composites typically possess higher specific surface areas than the acid-treated ones. The most striking example is NaCl/ZnCl₂–C₃N₄, whose BET surface area dropped from 700 m² g⁻¹ (value after water washing) to 193 m² g⁻¹ after the acidic treatment (Fig. S11†). Such behavior may be explained by the collapse of the internal structure during acidic work-up or the removal of high surface area impurities, such as nanoparticles onto the hybrid. In the first case, the composite material after water washing may be envisaged as a poor, amorphous analogue of metal organic frameworks, in which triazine imide (oligomeric) moieties play the role of organic ligands, which are interconnected by Zn²⁺ ions, ZnCl₂, ZnO clusters, or other species. Such microporous composite materials show very promising adsorption properties, which will be discussed later.

Applications

Among potential applications for at least some of ZnCl₂-derived carbon nitride composites, we would like to underline the following two: CO₂ adsorbents and photocatalysts.

The highly microporous NaCl/ZnCl₂-derived C₃N₄ composite (water-washed only) was investigated with respect to its CO₂ adsorption properties (Fig. 8 and S12, S13†). It shows quite high CO₂ uptake of 3.6 mmol g⁻¹ at 0 °C and 2.5 mmol g⁻¹ at 30 °C. The steep rise of the amount of adsorbed CO₂ at low pressures indicates the presence of rather small micropores, which could indeed be interesting for separation applications. The N₂ uptake at 0 °C and 30 °C was also studied and initial calculation with respect to the CO₂/N₂ selectivity α were undertaken. Ideal adsorbed solution theory (IAST)²⁹ gives α = 225 (0 °C) and α = 100 (30 °C) at 1 bar and a gas composition of CO₂/N₂ of 0.15/0.85 that reflects an approximate composition of a flue gas. Such selectivities are indeed competitive to some reported zeolites³⁰ and MOFs³¹ given also the simplicity of synthesis and purification procedures as well as the relatively high product yield of 50% (based on total C content in the precursor and the product).

Fig. 8 CO₂ and N₂ adsorption–desorption isotherms of microporous NaCl/ZnCl₂-derived C₃N₄ composite (after water washing step).

The photocatalytic properties of ZnCl₂-derived carbon nitrides were evaluated in a model reaction assay, the photodegradation of Rhodamine B (RhB) under blue light (λ = 420 nm) irradiation. Here, two selected examples of remarkable performance are the microporous NaCl/ZnCl₂-derived C₃N₄ hybrid (water-washed), again, and the C₃N₄/ZnO (40 wt%/60 wt%) nanocomposite. The estimated rate constants of RhB degradation accomplished by these two photocatalysts are k₀ = 21 × 10⁻³ min⁻¹ and k₀ = 12 × 10⁻³ min⁻¹, respectively; these are ∼10 and ∼5 times higher than the one observed for the reference g-C₃N₄ (2.2 × 10⁻³ min⁻¹, Fig. S14†). We mainly attribute these enhanced activities to the increased surface areas of the composites and improved absorption of visible light due to the formation of the dyadic structures. Additionally, the C₃N₄/ZnO hybrid is characterized by an improved crystallinity of ZnO NPs that results in better transport of the photo-generated and transferred electrons to the surface-adsorbed RhB and O₂ molecules.

Conclusions

In summary, condensation of melamine in ZnCl₂-containing eutectic salt melts gives rise to a broad range of novel carbon nitride-based composite materials with the improved absorption in visible light range due to the formation of dyadic system between C₃N₄ and ZnO clusters or other Zn²⁺ containing species. Unlike LiX/KX (X = Cl, Br), ZnCl₂ plays a role of the reactive solvent during the synthesis of carbon nitrides, and binds strongly to the condensation intermediates. Adjustment of precursor concentration and a proper selection of the alkali metal chloride constituent of MCl/ZnCl₂ melt give a possibility to change the on-set of phase demixing, tune the interactions strength between the condensation intermediates and the solvent and vary the solubility of the intermediates in the melt. Overall, one can direct the reaction from zinc cyanamide to both crystalline poly(triazine imides) or MOF-like hybrid materials. The latters have surface areas up to 700 m² g⁻¹ and turned out to be interesting as highly performing CO₂ adsorbents and photocatalysts.

Acknowledgements

The authors want to gratefully acknowledge the Max Planck Society, Dr J. Hartmann for EDX measurements and Dr G. Clavel for TEM measurements.

Notes and references

1. (a) X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2009, 8(1), 76–80; (b) K. Maeda, X. Wang, Y. Nishihara, D. Lu, M. Antonietti and K. Domen, J. Phys. Chem. C, 2009, 113(12), 4940–4947.
2. M. Shalom, S. Gimenez, F. Schipper, I. Herraiz-Cardona, J. Bisquert and M. Antonietti, Angew. Chem., 2014, 126(14), 3728–3732.
3. (a) Y. Wang, J. Zhang, X. Wang, M. Antonietti and H. Li, Angew. Chem., 2010, 122(19), 3428–3431; (b) Y. Wang, Y. Di, M. Antonietti, H. Li, X. Chen and X. Wang, Chem. Mater., 2010, 22(18), 5119–5121; (c) Y. Wang, H. Li, J. Yao, X. Wang and M. Antonietti, Chem. Sci., 2011, 2(3), 446–450; (d) F. Su, S. C. Mathew, G. Lipner, X. Fu, M. Antonietti, S. Blechert and X. Wang, J. Am. Chem. Soc., 2010, 132(46), 16299–16301.
4. M. Naffakh, V. López, F. Zamora and M. A. Gómez, Soft Mater., 2010, 8(4), 407–425.
5. (a) Y. Zhang, T. Mori, L. Niu and J. Ye, Energy Environ. Sci., 2011, 4(11), 4517–4521; (b) Y. Zhang, T. Mori, J. Ye and M. Antonietti, J. Am. Chem. Soc., 2010, 132(18), 6294–6295.
6. (a) M. Groenewolt and M. Antonietti, Adv. Mater., 2005, 17(14), 1789–1792; (b) K. Kailasam, J. D. Epping, A. Thomas, S. Losse and H. Junge, Energy Environ. Sci., 2011, 4(11), 4668–4674; (c) F. Goettmann, A. Fischer, M. Antonietti and A. Thomas, Angew. Chem., Int. Ed., 2006, 45(27), 4467–4471.
7. (a) M. Shalom, S. Inal, C. Fettkenhauer, D. Neher and M. Antonietti, J. Am. Chem. Soc., 2013, 135(19), 7118–7121; (b) Y.-S. Jun, E. Z. Lee, X. Wang, W. H. Hong, G. D. Stucky and A. Thomas, Adv. Funct. Mater., 2013, 23(29), 3661–3667.
8. (a) Y. Wang, X. Wang, M. Antonietti and Y. Zhang, ChemSusChem, 2010, 3(4), 435–439; (b) Y. Zhang, Z. Schnepp, J. Cao, S. Ouyang, Y. Li, J. Ye and S. Liu, Sci. Rep., 2013, 3, 2163.
9. (a) Q. Guo, Y. Xie, X. Wang, S. Lv, T. Hou and X. Liu, Chem. Phys. Lett., 2003, 380(1–2), 84–87; (b) Q. Guo, Y. Xie, X. Wang, S. Zhang, T. Hou and S. Lv, Chem. Commun., 2004, 1, 26–27.
10. (a) E. Wirnhier, M. Döblinger, D. Gunzelmann, J. Senker, B. V. Lotsch and W. Schnick, Chem.–Eur. J., 2011, 17(11), 3213–3221; (b) S. Y. Chong, J. T. A. Jones, Y. Z. Khimyak, A. I. Cooper, A. Thomas, M. Antonietti and M. J. Bojdys, J. Mater. Chem. A, 2013, 1(4), 1102–1107.
11. (a) P. Kuhn, M. Antonietti and A. Thomas, Angew. Chem., Int. Ed., 2008, 47(18), 3450–3453; (b) P. Kuhn, A. Forget, J. Hartmann, A. Thomas and M. Antonietti, Adv. Mater., 2009, 21(8), 897–901; (c) P. Kuhn, A. Thomas and M. Antonietti, Macromolecules, 2008, 42(1), 319–326.
12. N. Fechler, T.-P. Fellinger and M. Antonietti, Adv. Mater, 2013, 25(1), 75–79.
13. D. A. Shirley, Phys. Rev. B: Solid State, 1972, 5(12), 4709–4714.
14. C. D. Wagner, L. E. Davis, M. V. Zeller, J. A. Taylor, R. H. Raymond and L. H. Gale, Surf. Interface Anal., 1981, 3(5), 211–225.
15. (a) M. J. Bojdys, Über neue Allotrope und Nanostrukturen von Karbonitriden, Universität Potsdam, Potsdam, 2009; (b) B. V. Lotsch and W. Schnick, Chem.–Eur. J., 2007, 13(17), 4956–4968.
16. K. Schwinghammer, B. Tuffy, M. B. Mesch, E. Wirnhier, C. Martineau, F. Taulelle, W. Schnick, J. Senker and B. V. Lotsch, Angew. Chem., Int. Ed., 2013, 52(9), 2435–2439.
17. G. J. Janz, C. B. Allen, J. R. Downey Jr and R. P. T. Tomkins, Physical Properties Data Compilations Relevant to Energy Storage, National Standard Reference Data System, New York, 1978.
18. Y. Zhang, A. Thomas, M. Antonietti and X. Wang, J. Am. Chem. Soc., 2008, 131(1), 50–51.
19. Y. Ham, K. Maeda, D. Cha, K. Takanabe and K. Domen, Chem.–Asian J., 2013, 8(1), 218–224.
20. V. G. Manecke and D. Wöhrle, Macromol. Chem. Phys., 1968, 120(1), 176–191.
21. A. Thomas, A. Fischer, F. Goettmann, M. Antonietti, J.-O. Muller, R. Schlogl and J. M. Carlsson, J. Mater. Chem., 2008, 18(41), 4893–4908.
22. E. J. McDermott, E. Wirnhier, W. Schnick, K. S. Virdi, C. Scheu, Y. Kauffmann, W. D. Kaplan, E. Z. Kurmaev and A. Moewes, J. Phys. Chem. C, 2013, 117(17), 8806–8812.
23. M. C. Biesinger, L. W. M. Lau, A. R. Gerson and R. S. C. Smart, Appl. Surf. Sci., 2010, 257(3), 887–898.
24. N. Boshkov, K. Petrov, S. Vitkova, S. Nemska and G. Raichevsky, Surf. Coat. Technol., 2002, 157(2–3), 171–178.
25. B. J. Coppa and R. F. Davis, Appl. Phys. Lett., 2003, 82(3), 400–402.
26. B. R. Strohmeier and D. M. Hercules, J. Catal., 1984, 86(2), 266–279.
27. J. C. Klein and D. M. Hercules, J. Catal., 1983, 82, 424–441.
28. L. Zhao, X. Chen, X. Wang, Y. Zhang, W. Wei, Y. Sun, M. Antonietti and M.-M. Titirici, Adv. Mater., 2010, 22(30), 3317–3321.
29. A. L. Myers and J. M. Prausnitz, AIChE J., 1965, 11(1), 121–127.
30. F. Akhtar, Q. Liu, N. Hedin and L. Bergstrom, Energy Environ. Sci., 2012, 5(6), 7664–7673.
31. K. Sumida, D. L. Rogow, J. A. Mason, T. M. McDonald, E. D. Bloch, Z. R. Herm, T.-H. Bae and J. R. Long, Chem. Rev., 2011, 112(2), 724–781.

---

Footnote

† Electronic supplementary information (ESI) available: O 1s, Zn 3p and Cl 2p XPS spectra of CsCl/ZnCl₂–C₃N₄, schemes of heating procedures; WAXS patterns, FTIR spectra, EA and surface areas of products synthesized at different temperatures, heating rates, holding times and ratios; WAXS patterns and FTIR spectra of materials made in pure ZnCl₂; diffuse reflectance spectra of NaCl/ZnCl₂–C₃N₄ and C₃N₄/ZnO composites, CO₂ and N₂ adsorption–desorption experimental isotherms and predictions of gas uptake of CO₂ and N₂ at 303 K, Rhodamine B degradation data. See DOI: 10.1039/c4ra08236b

---