Novel 2D photocatalyst of copper-doped carbon quantum dot CD(Cu) loaded with ultrathin Ni-MOL for degradation of tetracycline Open Access

Recently, chemical pharmaceuticals (antibiotics) received more and more attention due to their negative effects and possible risks to ecosystems (Li et al. 2018). Tetracycline (TC), a typical category of sulfonamide antimicrobial agents, is applied widely in the field of human medicine. The environmental concentration is relatively low (ng/L-mg/L). Some techniques have been used to eliminate TC from water, including adsorption, liquid membrane, photocatalysis and so on (Barhoumi et al. 2016). Photocatalysis is considered to be one of the most progressive methods to remove TC because it is environmentally friendly, and highly stable and cheap. Many photocatalysts have been used to degrade antibiotics from water, including g-C₃N₄, TiO₂ and CdS. However, most of those photocatalysts are facing two challenges in application: they can only absorb ultraviolet region photoenergy and the rate of recombination of photogenerated holes with electrons is high. Thus, designing a photocatalyst with high efficiency of photoelectron-hole separation ability and effectively harvesting visible light has attracted more and more attention in the last few decades (Li et al. 2011).

Owing to low bandgap energy and excellent photo energy harvesting efficiency, organic nanostructured semiconductors have attracted more and more attention. Large numbers of organic-based semiconductor photocatalysis have been constructed with conjugated porphyrins as institutional building blocks to build solid-state soft materials. Metalloporphyrin have garnered special attention due to several important characteristics for photocatalysis, including higher triplet state quantum yields, long-lived excited states, and intense visible absorption (Zhu et al. 2020). The heterogenization of functional groups with long-range-ordered structures in crystalline frameworks will reflect different photocatalytic properties when porphyrins are used as functional molecules in homogeneous systems. Micheroni et al. have constructed porphyrin metal-organic frameworks (MOF) using of 5,15-bis(4-carboxyphenyl) porphyrin (H₂BCPP) porphyrin linkers (Micheroni et al. 2018). Johnson et al. reported that MOFs with octatopic linkers and tetratopic linkers porphyrin possessed a distinct structure (Johnson et al. 2014). Most of those studies focused on the construction of large conjugate system, however, few studies focused on derivatives of porphyrin, such as based-porphyrin carbon quantum dots. Carbon dots (CDs) is an innovative carbon nanomaterial with several advantages, such as good electron conductivity, low toxicity, adjustable fluorescent emission, low cost and facile functionalization. Especially, the up-conversion luminescence characteristic of CDs permits it to directly capture the near-infrared portion of solar energy and convert it to visible light. Compared with traditional carbon source CDs, based-porphyrin CDs possess some intrinsic advantages, such as high N/C atomic ratios, low carbonization temperature and low melting point (Wang et al. 2020a, 2020b). More importantly, it can give rise to metal atomically dispersed in CDs. In recent years, metal-doping have emerged as a new hotspot of activity in photocatalysis with the highest catalytic efficiency. Moreover, as a result of altering electronic and energy level structure, the unique activity of the ligand atoms may be enhanced, resulting in further adjustment in adsorbing visible photoenergy. Compared with other transition metal, copper (Cu) can be easily reduced to zero value and is often used as a substitute for precious metals. In addition, Cu is a variable metal with +1 and +2 valence, which is favorable for the transfer of photocarriers in photocatalysts. However, to our knowledge, the combination of Cu and based-porphyrin CDs has not been investigated yet.

However, it is not convenient to separate CDs from solution due to the characteristics of high dispersion and high stability. Hence, supporting CDs on solid materials is very essential. Recently, the combination of semiconductor materials and MOFs has become a new research trend, which has three significant advantages. Firstly, it avoids the compounding of photo-generated carries by forming Schottky heterojunctions between MOFs semiconductors and precious metal centers. Secondly, it can absorb more visible light by effectively reducing the energy band gap. It is possible to form surface plasmon resonance (SPR) on contact surfaces between the MOF and the semiconductor material. Thirdly, with a high specific surface area and numerous reaction sites, MOFs can hasten the accumulation of pollutants on the composite surface and increase the degradation rate (Pelaeza et al. 2012). Therefore, loading CDs onto MOFs could dramatically enhance the photocatalytic efficiency via two pathways. The first way is that CDs can accelerate the separation of photogenerated electron-holes due to their good electro-conductivity. The second one is that, due to their up-conversion luminescence, CDs can convert the near-infrared light to visible light, thus the light utilization efficiency can be greatly enhanced (Huang et al. 2017). In addition, it is an ideal support for CDs on MOFs with plentiful active spots and extremely large area of specific surface. Especially, two-dimensional (2D) MOFs possess more active sites and loading capacity, in comparison with three-dimensional (3D) MOFs. However, to our knowledge, constructing 2D MOFs supporting CDs to enhance the photocatalytic performance has not been reported.

Herein, in this work, a hybrid photocatalyst combining Ni-MOL (one of MOFs, Ni-metal organic layers) and CDs derived from porphyrin was prepared for the first time. Characterization of the physico-chemical performance of the prepared materials was conducted. In addition, the charge transfer and photodegradation mechanism of TC nanocomposites were discussed.

Para-phtalic acid (PTA, ≥99%), pyrrole (C₄H₅N, AR), methyl 4-formylbenzoate (C₉H₈O₃, ≥98%), formic acid (CH₂O₂, ≥85%), potassium hydroxide (KOH, 90%), potassium dichromate (≥99.8%), and sodium hydroxide (NaOH, ≥96.0%) were purchased from Mackiln Chemical Co., Ltd Propionic acid (C₃H₆O₂, ≥99.5%), tetrahydrofuran (C₄H₈O, ≥99.5%), methanol (CH₄O, ≥99.5%), copper(II) chloride dihydrate (CuCl₂·2H₂O, ≥99.0%), isopropyl alcohol (IPA, ≥99.5%), N,N-dimethylformamide (DMF, ≥99.5%), para-benzoquinone (p-BQ, ≥99.0%), nickel(II) nitrate hexahydrate (Ni(NO₃)₂·6H₂O, ≥98.0%), and ethylenediaminetetraacetic acid disodium (EDTA-2Na, ≥98.0%) were purchased from Shanghai Test. HPLC-methanol (≥99.999%) were bought from Kermel. Hydrochloric acid (HCl, 36.0–38.0%) and diethyl ether anhydrous ((CH₃CH₂)₂O, ≥99.6%) were purchased from Yantai Yuandong Chemicals Co., Ltd.

The method used to prepare Ni-MOL can be found in literature (Yang et al. 2021). PTA (0.166 g) was dissolved into 5 mL DMF solution with stirring for 10 min to form solution A. Ni(NO₃)₂⋅6H₂O (0.436 g) was added into 10 mL DMF solution with stirring for 10 min to form solution B. Subsequently, solution A was added drop-wise into solution B. Afterwards, 2 mL water containing NaOH (3.2 mg, 0.08 mmol) was slowly added into the mixed solution and stirring for 1 h. Thereafter, the mixture was transferred into Teflon-lined stainless-steel autoclave and heated at 120 °C for 10 h. The resulting precipitate was collected by filtration and washed with DMF and alcohol for several times. The collected sample was dried at 50 °C for 12 h in an oven to obtain the final product.

Pyrrole (3.0 g, 0.043 mol) and methyl 4-formylbenzoate (6.9 g, 0.042 mol) were placed in a three-necked flask. The mixtures were completely dissolved by 100 mL propionic acid. Subsequently, the mixture was placed in an oil bath at 160 °C for 12 h, followed by cooling down. The solid material was collected after filtration and washed with ethanol, ether and tetrahydrofuran. After drying, a purple substance was obtained.

Tetracarboxymethyl porphyrin (0.99 g) was dissolved in 30 mL tetrahydrofuran and 30 mL methanol. Then, 30 mL potassium hydroxide solution (containing 6.82 g potassium hydroxide) was added into the mixed solution, followed by being refluxed in an oil bath at 80 °C for 12 h. The solids were acidified with 1M acid solution, filtered and washed with water afterwards. The final product (TCPP) was dried in an oven.

TCPP was impregnated in 0.5 mol/L CuCl₂•2H₂O solution and sonicated for 30 min. Later, it was settled down for 24 h. After soaking, 1 g of the soaked TCPP was calcined at 250 °C for 2 h with the rate of 5 °C/min in N₂. Later, the material and NaOH solution (100 ml, 0.15 mol/L) were added into a Teflon autoclave at 80 °C for 24 h. Following alkali etching, the CD(Cu) suspension was centrifuged at 10,000 rpm/min for 5 min. Subsequently, Ni-MOL (0.2 g) was added into 50 mL of supernatant and hydrothermal reaction was carried out at 90 °C for 3 h. Then, the precipitate was washed with ultrapure water.

The FT-IR spectra of Ni-MOL, CD, CD(Cu), CD-Ni-MOL and CD(Cu)-Ni-MOL are presented in Figure 2(b). Peaks sited at 3,433 cm⁻¹ were accorded to the stretching vibrations of O-H of samples. For Ni-MOL, a series of peaks located at 1,446–1,680 cm⁻¹ were related to the feature peaks of benzene rings. A peak appearing at 823 cm⁻¹ belonged to the feature peak of Ni-O (Yang et al. 2021). A peak appearing at 1,750 cm⁻¹ was ascribed to the carbonyl group of CD and CD(Cu). The characteristic peak of C-N of CDs and CD(Cu) was located at 1,350 cm⁻¹. Compared with CD, the new peak presented at 1,004 cm⁻¹ was ascribed to Cu-O in CD(Cu). All characteristic peaks of CD and Ni-MOL were detected in the CD-Ni-MOL and CD(Cu)-Ni-MOL, proving the successful combination of CD and Ni-MOL. The absorption peak of C = O of CDs was blue shifted to 1,700 cm⁻¹ due to the influence of the Ni as the electron withdrawing group and the combination with Ni-MOL molecular bond, which was covered by the characteristic peaks of benzene ring.

Figure 5(b) shows the photoluminescence (PL) spectra of Ni-MOL, CD, CD(Cu), CD-Ni-MOL and CD(Cu)-Ni-MOL. The fluorescence intensities were different at 720 nm for all samples. In general, a higher intensity of PL emission resulted in a higher rate of photogenerated electron/hole pair complexation. CD(Cu)-Ni-MOL emission peak possessed the lowest peak intensity among all the samples, which attributed to the valid charge transfer between Ni-MOL and CD particles. Thus, it decreased the carrier complication process. In addition, CD(Cu) had lower PL emission intensity than CD. It showed that the doping of Cu²⁺ could significantly reduce the compounding rate of photogenerated carriers by trapping electrons, resulting in an improved photocatalytic activity. It is worthy to note that the PL intensity of Ni-MOL is lower than that of CD(Cu)-Ni-MOL. This is because the energy band gap of Ni-MOL is 3.536 eV, which can only absorb ultraviolet light and does not produce electron transition under visible light irradiation, so PL has a low intensity. In addition, there are certain intrinsic relationships between the PL spectrum and photocatalytic activity of a semiconductor material according to the mechanisms of PL and photocatalysis. However, the inherent relationships are very complicated, and can properly be disclosed only on the basis of PL attributes, closely related to dopant species. For excitonic PL signals, the lower the PL intensity, the higher the separation rate of photo-induced charges and, possibly, the higher the photocatalytic activity (Zhang et al. 2000). The stronger the PL signal, the higher the content of surface oxygen vacancies and defects and possibly the higher the photocatalytic activity. Compared with Ni-MOL, the doped Cu in CDs and the loaded CD(Cu) in Ni-MOL could form more oxygen vacancies and defects. Therefore, the PL intensity of Ni-MOL was lower than that of CD(Cu)-Ni-MOL, while the photocatalytic performance of CD(Cu)-Ni-MOL was higher than that of Ni-MOL.

Figure S1 shows the effect of initial concentration on removal rate. As shown in Figure S1, it can be found that the value of C/C₀ decreases with the increases of initial concentration. However, 83% removal capacity can still be maintained at the initial concentration 125 mg/L under visible light condition, indicating that CD(Cu)-Ni-MOL has excellent catalytic performance. Compared with visible light condition, the degradation performance of CD(Cu)-Ni-MOL decreased significantly under ultraviolet light conditions. It is based on the fact that the band gap of CD(Cu)-Ni-MOL is 2.071 eV (Figure 5(a) inset). 2.071 eV is corresponded to the absorption of visible light. Therefore, no catalytic degradation occurs under UV light. The existence of adsorption phenomenon leads to the decrease of C/C₀ value with increases of initial concentration. In order to further investigate the adsorption progress, the adsorption equilibrium analysis was performed. The adsorption of TC onto CD(Cu)-Ni-MOL by different initial TC concentrations (25, 50, 75, 100, 125 and 150 mg/L) were shown in Figure S2. It is clear that the adsorption capacity of TC onto CD(Cu)-Ni-MOL was increased with the increase of initial TC concentrations. The remove rate at the initial TC concentration of 150 mg/L has reached 30.2%. The adsorption equilibrium was often described by adsorption isotherms. Freundlich isotherm models are the two most frequently-used models used to describe the adsorption isotherm process. Figure S3 shows the adsorption isotherms of TC with corresponding Langmuir and Freundlich plots. The fitted parameters of both Freundlich and Langmuir isotherms are shown in Table S1. The correlation coefficient of Freundlich model was higher than that of Langmuir model, indicating that isotherms results were better described by Freundlich model. In another word, the adsorption process was non-uniform and multilayer.

Figure S7 shows the effect of pH of degradation process. From Figure S7, it can be found that the removal rate reaches the maximum (62.4%) at pH = 8 under the dark condition, indicating that adsorption plays an important role in the degradation and removal of TC. Similarly, the maximum removal rate has reached to 98.3% at pH = 8 under the visible light. Compared with dark environment, the change of pH has less influence on removal rate under visible light. According to the adsorption-degradation theory, the initial stage of degradation removal is adsorption, and then the degradation stage. In the adsorption process, it is greatly affected by pH change, while in the degradation stage, it is fairly affected by pH change.

In order to further confirm the role of active free radical in the process of photocatalysis, electron spin resonance (ESR) technology was used to investigate the role of •OH and in the photocatalytic degradation of TC. As shown in Figure S8, the signals of DMPO- could be clearly observed. Compared with DMPO-⁠, there was no signal for DMPO-•OH, suggesting that superoxide radicals indeed played a major role in the photocatalytic reaction. The results of ESR further verified the results of trapping experiments.

The bandgap energy of Ni-MOL was calculated as 3.536 eV as shown in Taucplot (Figure 6 inset). We plotted and analyzed the Mott-Schottky curve (Figure S9) with the purpose of identifying the semiconductor type and the flat charged potential of Ni-MOL. Ni-MOL was an n-type semiconductor and the flat-band potential was 0.063 eV (vs. Ag/AgCl). Therefore, the valence band potential of Ni-MOL could be identified as 3.257 V vs. NHE.

The stability of photocatalysts is a decisive factor in practical applications. Therefore, the stability of CD(Cu)-Ni-MOL was assessed, which showed the best degradation performance of TC with cyclic experiments in this study. Figure S11 shows that the decrease of the photo-catalytic activity of CD(Cu)-Ni-MOL in five cyclic experiments was tiny, suggesting that CD(Cu)-Ni-MOL could maintain a stable degradation activity in long-term use.

The HPLC-MS was applied to identify the intermediate products and to explore the degradation pathway (Figure S12). Interestingly, toroidal structures of tetracyclines were related to several readily ionizable functional and electron-rich groups (e.g. dimethylphenol groups and amino groups) that were susceptible to attack by reactive agents (Shen et al. 2022).

In pathway I, a double bond on P1 was oxidised by to produce a ketone group and a new hydroxyl group to give P2 (m/z = 477). Subsequently, compound P3 (m/z = 449) was obtained by demethylation as a result of attacking the position of the tertiary amine on the ring. Afterwards, It was possible to convert P3 to P4 (m/z = 393) by removing another methyl group and losing the amide group. It is mainly through open-loop and oxidation processes that the formation of P5 (m/z = 335) and P6 (m/z = 274) could be determined. The intermediate P6 was further degraded to small molecules (m/z = 102).

In pathway II, TC molecules were firstly converted to P7 (m/z = 427), which was ascribed to the dehydration products under the attack of h⁺. With the progression of the reaction, intermediates P8–P10 (with m/z = 301, 291, 207, and 149, respectively) were formed following terminal oxidative and ring-opening reactions triggered by •OH.

In this work, a novel 2D layered photocatalyst was successfully prepared by a one-step hydrothermal method. It exhibited excellent photocatalytic degradation ability of TC under visible light irradiation, with degradation rate reaching approximately 93.5% within 60 min. The results of UV-DRS, PL, transient photocurrent and EIS suggested that CD(Cu)-Ni-MOL could efficiently inhibit the recombination of photogenerated charge. In addition, CD could convert low-energy visible light to high-energy visible light through upconversion effect and the addition of Cu resulted in the change of carrier transport path and the formation of Cu⁰, which effectively improved the photocatalytic performance of Ni-MOL. h⁺ played a significant role and •O₂^– also played a role in the photo-degradation process. The TC degradation process and decomposition pathways were proposed based on HPLC-MS.

This work was financially supported by the Open Research Fund Program of Key Laboratory of Cleaner Production and Integrated Resource Utilization of China National Light Industry (CP2021YB10), Integration of Science, Education and Industry Innovation Pilot Project of Qilu University of Technology (Shandong Academy of Sciences) (2020KJC-ZD12), Cooperation Foundation for the Youth Scholars of Qilu University of Technology (Shandong Academy of Sciences) (2018BSHZ0022) and Natural Science Foundation of Shandong Province (ZR2017LEE024).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.