Meso-tetra(4-sulfonatophenyl)porphyrin silver/Ag nanoparticles/graphene-phase C₃N₄ with a sandwich-like structure and double-faced active centers via two-step room-temperature photocatalytic synthesis for ractopamine detection

Abstract

Photochemical synthesis under visible light irradiation is a novel approach in the field of green chemistry, and composites with abundant active centers for electrochemical detection are highly attractive. Herein, a meso-tetra(4-sulfonatophenyl)porphyrin silver/Ag nanoparticles/graphene phase C₃N₄ nanosheets (Ag₂TPPS₄/AgNPs/ng-C₃N₄) material with a sandwich-like structure was synthesized using a two-step photocatalytic reaction at room temperature (25 °C). In the first visible light irradiation step and in the presence of a hole capture agent, Ag⁺ ions were photocatalytically reduced onto the surface of ng-C₃N₄ that was used as a photocatalyst. Then, the protons (H⁺) in the core of H₂TPPS₄ were substituted in situ by photo-oxidized Ag⁺ during the second visible light irradiation step and in the presence of an electron capture agent. The electrochemical response of Ag₂TPPS₄ and ng-C₃N₄ to ractopamine (RAC) results in the unique double-faced active centers of Ag₂TPPS₄/AgNPs/ng-C₃N₄, and the cores (AgNPs) are beneficial as bridges for the connection between Ag₂TPPS₄ and ng-C₃N₄ and for high-efficiency electron transfer. Hence, as-synthesized Ag₂TPPS₄/AgNPs/ng-C₃N₄ exhibits high sensitivity (a low detection limit of 5.1 × 10⁻⁸ M, S/N = 3.0), a wide linear range (1 × 10⁻⁷ to 1.2 × 10⁻⁵ M), and long-term stability. Based on the experimental verification of the electrochemical dynamics and electrostatic attraction at the interface between the dual-active-center surface and RAC, the electrochemical mechanism has been clarified. Specifically, in the multi-cycle oxidation of RAC, the blue shift of specific UV-vis peaks also confirms the electrocatalytic oxidation of the two terminal hydroxyl groups of RAC. In brief, Ag₂TPPS₄/AgNPs/ng-C₃N₄ with a sandwich-like structure and double-faced active centers enhances the detection sensitivity and electrocatalytic efficiency towards RAC.

---

Introduction

Both the increasing global demand for green synthesis and worldwide environmental concerns have triggered great attention towards searching for and focusing on novel ways to utilize alternative and renewable synthetic strategies, among which photocatalytic synthesis is known as one of the most abundant and easily accessible methods.¹ Consequently, tremendous efforts have been made to utilize various metallic and organic catalysts capable of exhibiting photo-driven activity and mediating chemical reactions in the production of pharmaceuticals, fine chemicals and advanced materials.^2,3 Thus, diverse composites have been prepared by this protocol, benefiting from the large amount of energy involved and the mild conditions.^4,5 Kim and co-workers utilized UV illumination to excite carbon quantum dots with excellent electron-donating capability, which quickly induced the reduction of metal ions to the corresponding metal nanoparticles.⁶ Carbon-dot-supported silver (CD-Ag nanoparticles) was used to fabricate light-emitting diodes and polymer solar cells based on a solution-processable polymer. Following the early report of Zhou and co-workers, porphyrin-modified ultrathin Zr-MOF nanosheets were constructed to act as an effective photocatalyst for ¹O₂ generation and artemisinin production.⁷ In our previous report, a reduced graphene oxide (rGO)/PbTiO₃ composite was successfully synthesized under UV irradiation and a highly sensitive electrochemisensor to detect pyrrole was fabricated.⁸ Note that photocatalytic synthesis accomplishes photocatalytic redox reactions by producing photogenerated electrons and holes through the use of a photocatalyst under mild conditions, which overcomes the problems of complicated chemical kinetics and saves energy via eliminating the need for high temperatures, which can also protect nanomaterials from damage.⁹

Porphyrins, a class of naturally occurring macrocyclic compounds such as hemoglobin and chlorophyll, are of significance in the metabolism of living organisms.¹⁰ The stable molecular structure of porphyrin is composed of four pyrrole rings linked via methine bridges, leading to aromatic character.¹¹ Porphyrins are electron-rich organic molecules that are characterized by remarkably high extinction coefficients in the visible region.^12,13 They are one of the most promising components of new materials for future electronic devices. For example, meso-tetra(4-sulfonatophenyl)porphyrin (H₂TPPS₄), a water-soluble porphyrin with hydrophilic groups and light-harvesting capacity,¹⁴ is beneficial for use in photocatalytic synthesis to prepare novel materials with specific features.

Graphene-phase C₃N₄ nanosheets (ng-C₃N₄) with a large surface area, which provides abundant reactive sites, and a short bulk diffusion length, which reduces the recombination probability of photo-excited charge carriers, are extremely advantageous for loading with porphyrin and to promote their catalytic performance.^15,16 Nevertheless, the semiconductivity of ng-C₃N₄ hinders electron transfer between ng-C₃N₄ and H₂TPPS₄. Ag nanoparticles (AgNPs) have excellent electrocatalytic abilities and conductivity and are widely loaded onto the surfaces of supports in order to endow the resulting composites with new properties.^17,18 The suitable conduction band (CB) and valence band (VB) (positioned at ca. −1.1 eV and ca. +1.6 eV vs. a normal hydrogen electrode (NHE), respectively^19,20) of ng-C₃N₄ and the divalent vacancy in the center of the porphyrin ring, which is suitable for coordination with a metal, are able to exploit a photocatalytic synthesis strategy in which highly conductive AgNPs are anchored as a “bridge” to achieve connection and effective electron transportation between ng-C₃N₄ and H₂TPPS₄. The sandwich-like structure Ag₂TPPS₄/AgNPs/ng-C₃N₄ would be synthesized via a two-step photocatalytic process under visible light irradiation and at room temperature and be used as a unique double-faced electrocatalyst.

Ractopamine (RAC) is a synthetic beta-adrenergic agonist that can be used to treat congestive heart failure and muscular dystrophy, increase muscle mass, reduce small amounts of fat accumulation, decrease body weight, and enhance neonatal growth.²¹ Moreover, it can increase the daily weight gain of animals, improve feed utilization, and increase the protein content of animals. However, RAC has a stable structure and is not easily decomposed inside animals or humans. Large amounts of RAC residues in meat or dairy products will cause human poisoning, heart palpitations, muscle tremors and other negative effects. Pork containing RAC is now regulated by many countries.²² Subsequently, the monitoring and determination of RAC are extremely significant for food safety and human health. There are various highly selective and sensitive analytical methods for detecting RAC, such as high performance liquid chromatography, mass spectrometry, etc.,^23,24 but their complicated operation, long time and high cost limit their wide application.^25,26 Hence, it is necessary to develop a convenient, low-cost, and highly sensitive RAC detection method.

Here, a Ag₂TPPS₄/AgNPs/ng-C₃N₄ nanocomposite with a novel sandwich-like structure was fabricated via a facile and environmentally friendly two-step photocatalytic synthesis under visible light irradiation and at room temperature (Scheme 1). First, Ag⁺ ions are reduced by photogenerated electrons to form silver nanoparticles on the surface of ng-C₃N₄, which is used as a photocatalyst. Secondly, under the same irradiation, meso-tetra(4-sulfonatophenyl)porphyrin silver (Ag₂TPPS₄) was synthesized via an in situ center-substituted (ISCS) process, i.e., the hydrogen atoms in the core of H₂TPPS₄ were replaced by the unsaturated silver ions on the surface of the AgNPs. Finally, the connection and effective electron transportation resulted in the sandwich structure and double-faced electrocatalytic activity of Ag₂TPPS₄/AgNPs/ng-C₃N₄. The new strategy is expected to simplify the preparation process, reducing the introduction of impurities and enhancing the electron transportation of the hybrid composite. Thereby, Ag₂TPPS₄/AgNPs/ng-C₃N₄ offers excellent conductivity, abundant electrocatalytic active centers, high specific surface area and a stable structure. In virtue of the above benefits, Ag₂TPPS₄/AgNPs/ng-C₃N₄ was applied as an electrochemical sensor for the efficient detection of RAC. The electrochemical analysis results demonstrated that the Ag₂TPPS₄/AgNPs/ng-C₃N₄ exhibited excellent sensitivity, a wide linear range and a low detection limit.

Scheme 1 The photocatalytic synthesis processes of Ag₂NPs/ng-C₃N₄ and Ag₂TPPS₄/AgNPs/ng-C₃N₄.

Experimental

Materials

Meso-tetraphenyl porphyrin (99.95%, 325 mesh) was obtained from URChem (Shanghai, China). RAC was purchased from Sigma-Aldrich (St. Louis, MO, USA). RAC stock solutions (1 mM) were prepared using ultra-pure water in all experiments. The working solutions were diluted from the stock solution using phosphate buffer solution (PBS, 0.1 M, pH 7.2). All other chemicals used were of analytical reagent grade, and were purchased from Beijing Chemical Reagent Company (Beijing, China) and used without further purification. Ultra-pure water was obtained from a Milli-Q plus water purification system (Millipore Co. Ltd., USA) (18 MΩ).

Synthesis of ng-C₃N₄ nanosheets and meso-tetra(4-sulfonatophenyl)porphyrin (H₂TPPS₄)

Melamine was heated at 550 °C for 4 hours with a ramp rate of 5 °C min⁻¹ to produce bulk g-C₃N₄. Then, 400 mg of ground bulk g-C₃N₄ was placed in an open ceramic container and was heated at 500 °C for 2 h with a ramp rate of 5 °C/min to obtain ng-C₃N₄.⁴¹

H₂TPPS₄ was synthesized using the Adleb method from a previous report.²⁷ A mixed solution of 8.5 ml (83.4 mmol) of benzaldehyde and 190 ml of propionic acid was poured into a three-necked flask and refluxed at 136 °C. A mixed solution of 5.6 ml (80.7 mmol) pyrrole and 15 ml propionic acid was slowly added dropwise to the above mixture, which was maintained at a low boil for 30 min. After cooling the mixture to room temperature, 70 ml of anhydrous ethanol was added. The mixture was then filtered and washed with ethanol until the filtered solution was colorless, and the crude product was obtained. The crude product was dissolved in chloroform on a 5 cm silica gel column (200–300 mesh), and all the eluent was collected. 100 ml of methanol was added overnight. The bright purple crystals obtained by extraction were metal-free porphyrins (H₂TPPS₄).

The prepared metal-free porphyrin was sulfonated in chlorosulfonic acid. 2.7 ml of chlorosulfonic acid (98%) was placed in a 100 ml three-necked flask. 1 g of porphyrin powder was dissolved in 50 ml of the solvent 1,2-dichloroethane. The above solution was mixed in the three-necked flask and stirred at 120 °C for 4.5 h. After the stirring was complete and the temperature had spontaneously decreased to below 45 °C, the above solution was diluted with ice water, filtered and washed with distilled water to neutrality, and tested with AgNO₃/HNO₃ (dilute) solution until there was no HCl. The filter cake was dried in a vacuum at 70 °C for 48 h, and the resulting product was subjected to column separation. A basic alumina column was used, and the mobile phase was a mixture of water : methanol : acetone (7 : 2 : 1). The middle pink layer was collected, and the H₂TPPS₄ was obtained later use.

Two-step synthesis of Ag₂TPPS₄/AgNPs/ng-C₃N₄ under visible light irradiation and at room temperature

A two-step photocatalytic process was applied to form the Ag₂TPPS₄/AgNPs/ng-C₃N₄ composite (Scheme 1). 30 mg of the as-synthesized g-C₃N₄ and 0.051 g AgNO₃ were suspended in 20 ml of a mixture of water and isopropanol (3 : 1). Under nitrogen gas protection, the suspension was agitated and irradiated by visible light (300 W xenon arc lamp) for 60 min at room temperature (25 °C). Next, the suspension was centrifuged three times with water to remove Ag⁺ and dispersed again in a mixture of water and isopropanol (3 : 1). Note that isopropanol was used here as a photogenerated-hole capture agent. Subsequently, 20 ml of H₂TPPS₄ (0.22 mM) and p-benzoquinone (BQ, 0.92 mM, as an electron capture agent) were added to the newly prepared AgNPs/ng-C₃N₄ suspension and irradiated within 60 min under visible light in order to synthesize the Ag₂TPPS₄/AgNPs/ng-C₃N₄ suspension. It needs to be emphasized that these synthetic processes were carried out at room temperature (25 °C), and that the as-prepared Ag₂TPPS₄/AgNPs/ng-C₃N₄ was dried in a vacuum at 70 °C for 24 h without further sintering.

Preparation of a RAC electrochemical sensor based on Ag₂TPPS₄/AgNPs/ng-C₃N₄

Prior to the preparation of the sensor, a glassy carbon electrode (GCE) was polished to mirror smoothness using a polishing cloth with 0.3 and 0.05 μm alumina powder, respectively. The GCE was then rinsed with deionized water, and ultrasonicated in deionized water and an ethanol bath, sequentially. 6 μl of Ag₂TPPS₄/AgNPs/ng-C₃N₄ was dropped onto the surface of the working GCE and dried at room temperature. Thus, the modified GCE with a uniform film coating (Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE) for detecting RAC was obtained.

Characterization and electrochemical measurements

Fourier transform infrared spectroscopy (FT-IR) was conducted using a Nicolet Nexus 670 FT-IR spectrophotometer. ¹HNMR was carried using a nuclear magnetic resonance spectrometer (NMR, Avance 400, Switzerland). The suspension in a capped quartz reactor was photo-irradiated with a xenon lamp (Model: PLS-SXE 300C, Perfect Light, China. Optical Filter, UVREF, λ = 200–400 nm). UV-vis measurements were performed using a UV-1800 SHIMADZU spectrophotometer. The fluorescence spectra were obtained at room temperature using a steady/transient fluorescence spectrophotometer (FLS 920, Edinburgh, U.K.). The morphology of the samples and their energy spectra were characterized using field-emission scanning electron microscopy (FE-SEM, Hitachi S4800, Japan). High-resolution transmission electron microscope (HRTEM) images were collected using a Tecnai G2 F20 S-TWIN, 200 kV (FEI Company, USA). Cyclic voltammetric (CV) and differential pulse voltammetry (DPV) plots were obtained using a CHI 660D electrochemical workstation (CH Instrument Company, Shanghai, China). A three electrode cell system was utilized with a modified GCE as the working electrode, Ag/AgCl as the reference electrode, and a platinum wire as the counter electrode. CV and DPV measurements of Ag₂TPPS₄/AgNPs/ng-C₃N₄ were carried out at room temperature (25 °C).

Results and discussion

Structure and morphology of Ag₂TPPS₄/AgNPs/ng-C₃N₄ throughout the two-step photocatalytic synthetic processes

The structure of the ng-C₃N₄ used as the photocatalyst during the first photoreduction is important to the subsequent product in the two-step synthetic process. In order to observe the effect of thermal stripping, the morphology of g-C₃N₄ before and after calcination was analyzed using SEM (Fig. S1†). The bulk g-C₃N₄ before stripping exhibits accumulation and stacking, which is caused by the bulk g-C₃N₄ layers being closely bonded to each other (Fig. S1A†). After the process of thermal stripping (500 °C for 2 h), thin layers of g-C₃N₄ on both sides of bulk g-C₃N₄ are separated by heat under the high-temperature conditions (500 °C, 2 h), inducing thinner g-C₃N₄. Finally, the exfoliated g-C₃N₄ appears as a graphene-like flake, i.e., ng-C₃N₄ (Fig. S1B†). Additionally, the lamellar structure of ng-C₃N₄ after the peeling process remains unchanged, as can be determined from their FT-IR spectra (Fig. S2†). Both the bulk g-C₃N₄ and ng-C₃N₄ samples display the typical vibration peaks of graphitic carbon nitride. The peak at 807 cm⁻¹ is associated with the characteristic breathing mode of triazine units (Fig. S2A†).²⁸ The broad band at approximately 3000 cm⁻¹ corresponds to the stretching modes of terminal –NH₂ or –NH groups (Fig. S2A†). The peaks at 1230, 1300 and 1390 cm⁻¹ are ascribed to the typical stretching modes of aromatic C–N heterocycles, while the peaks at 1530 and 1625 cm⁻¹ are assigned to the double bond C=N stretching vibration (Fig. S2B†). Meanwhile, the XRD pattern for g-C₃N₄ gives low-angle reflection peaks at 27.6° stemming from the g-C₃N₄ produced by stacking between slices (Fig. 1A(a)). The slight shift from 27.6° to 27.7° after calcination indicated that a number of layers of ng-C₃N₄ pieces stacked against each other (Fig. 1B).²⁹ The NMR and UV-vis spectra of the as-prepared H₂TPPS₄ used as a precursor in the second photocatalytic process were measured (Fig. S3 and S4†). The ¹HNMR spectrum (Fig. S3†) of H₂TPPS₄ in D₂O exhibits two broad resonances at 7.08 and 8.69 ppm, which were ascribed to the β pyrrole protons. The two peaks at 7.54 and 8.13 ppm are attributed to the presence of meta and ortho phenyl protons, respectively.³⁰ The behaviour of pure H₂TPPS₄ in a mixed liquid (water : isopropanol = 3 : 1) in the UV-vis spectrum (Fig. S4†) is identified according to its specific nature with a Soret band at 413 nm and Q-band absorption at 515, 552, 579 and 633 nm. Therefore, high-purity H₂TPPS₄ was definitely prepared.

Fig. 1 (A) XRD patterns of (a) bulk g-C₃N₄, (b) ng-C₃N₄, (c) AgNPs/ng-C₃N₄, and (d) Ag₂TPPS₄/AgNPs/ng-C₃N₄, and (B) the XRD patterns from 27–28°.

For the systemic analysis for each product in the two-step photocatalytic process, XRD patterns were employed to characterize the nanostructure of AgNPs/ng-C₃N₄ and Ag₂TPPS₄/AgNPs/ng-C₃N₄, as shown in Fig. 1. In the first photocatalytic process in the presence of the hole capture agent (isopropanol), Ag⁺ was reduced to give AgNPs. In this process, Ag⁺ in AgNO₃ was just reduced by the photogenerated electrons near the photocatalyst, ng-C₃N₄, so the AgNPs were loaded onto the surface of ng-C₃N₄. The XRD pattern of AgNPs/ng-C₃N₄ (Fig. 1A(c)) clearly displays peaks at 38.2°, 44.4°, 64.6° and 77.6°, which are indicative of the (111), (200), (220) and (311) planes of AgNPs (JCPDS-04-0783), respectively, indicating the formation of crystal AgNPs on the surface of ng-C₃N₄. After the AgNPs are loaded on the ng-C₃N₄, the peak of ng-C₃N₄ is shifted from 27.7° to 27.9°, revealing a decreased interlayer spacing and layer stacking of g-C₃N₄. After the second photocatalytic process, due to the rare substitution of the Ag atoms at the surface of the AgNPs, the pattern of the AgNPs is almost unchanged, while the diffracted intensity of ng-C₃N₄ is lower. In addition, the presence of the element Ag in the prepared nanocomposite is also confirmed by EDS (Fig. S5B†). After the second visible light irradiation, the patterns of ng-C₃N₄ and AgNPs remain unchanged due to the very low level of oxidation of the AgNPs to Ag₂TTPS₄.

The Raman spectra of H₂TPPS₄, AgNPs/ng-C₃N₄ and Ag₂TPPS₄/AgNPs/ng-C₃N₄ were measured to confirm the interfacial characteristics of Ag₂TPPS₄/AgNPs/ng-C₃N₄ (Fig. 2). A weak Raman band in the low-frequency region is observed at 312 cm⁻¹, corresponding to the C_αC_mC_α bending coordinate (Fig. 2A). The Raman band of H₂TPPS₄ at 1550 cm⁻¹ is assigned to the vibration ν(C_α/C_m), while the bands at 1494 cm⁻¹ and 1455 cm⁻¹ are ν(C_α/C_m). The interactions between C–C bonds and porphyrin rings results in the 1566 cm⁻¹ vibration peak. The Raman bands in 900–1400 cm⁻¹ region are related to the stretching vibrations of C_αC_β/C_αN bonds. The bands at 1534, 1450, and 1488 cm⁻¹ (Fig. 2C) are ascribed to the ν(C_β/C_β) and ν(C_α/C_m) of H₂TPPS₄, which are slightly changed due to the coplanar conformation of porphyrins as depicted above.³¹ It is worth noting that the peak at 312 cm⁻¹ caused by C_αC_mC_α bending is shifted to 351 cm⁻¹ in the Ag₂TPPS₄/AgNPs/ng-C₃N₄ spectrum. Additionally, the peaks at 1240 and 1329 cm⁻¹ disappear and a new peak at 1340 cm⁻¹ appears, which further confirms that the introduction of Ag⁺ ions in a delocalized state changes the bending mode of the porphyrin rings.³² Characteristic bands at 707 and 1234 cm⁻¹ clearly demonstrate the presence of AgNPs/ng-C₃N₄, indicating the successful connection of Ag₂TPPS₄ and AgNPs/ng-C₃N₄.

Fig. 2 Raman spectra of (A) Ag₂TPPS₄/AgNPs/ng-C₃N₄, (B) H₂TPPS₄, and (C) Ag/ng-C₃N₄.

The morphology and microstructure of ng-C₃N₄, AgNPs/ng-C₃N₄, Ag₂TPPS₄/AgNPs/ng-C₃N₄ were characterized with the HRTEM technique (Fig. 3). The results illustrate that ng-C₃N₄ appears in flakes (Fig. 3A). The AgNPs are inhomogeneously distributed with an average size of 20 nm on the ng-C₃N₄ (Fig. 3B). Then, H₂TPPS₄ was introduced into AgNPs/ng-C₃N₄. Under the second irradiation, Ag₂TPPS₄ was synthesized by means of an ISCS process, i.e., the hydrogen atoms in the core of H₂TPPS₄ are substituted by the unsaturated silver ions on the surface of AgNPs, forming the Ag₂TPPS₄/AgNPs/ng-C₃N₄ composite (Fig. 3C). The morphology of the composite (Ag₂TPPS₄/AgNPs/ng-C₃N₄) has no significant change compared with that of AgNPs/ng-C₃N₄, and the size of the AgNPs of the former is similar to that of the latter. The reason for this is that the Ag₂TPPS₄ molecules are synthesized at the surface of AgNPs. Finally, a sandwich-like structure with AgNPs at the core of ng-C₃N₄ and Ag₂TPPS₄ (Scheme 1) was obtained. The two microcyclic electrochemical catalysts, ng-C₃N₄ and Ag₂TPPS₄, are beneficial to the high-efficient catalytic oxidation towards RAC, as shown later. In addition, the weak regular diffraction spots and diffraction rings with different radii in the SAED pattern (Fig. 3D) are assigned to the AgNP (111) crystal facets and the Ag₂TPPS₄/AgNPs/ng-C₃N₄ composition with a low crystallinity state. Furthermore, elemental mappings reveal that the elemental distributions of Ag are highly uniform (Fig. S9†), demonstrating the successful synthesis of Ag₂TPPS₄ by the ISCS process.

Fig. 3 HRTEM images of (A) ng-C₃N₄, (B) AgNPs/ng-C₃N₄, and (C) Ag₂TPPS₄/AgNPs/ng-C₃N₄, and (D) SAED patterns of Ag₂TPPS₄/AgNPs/ng-C₃N₄.

Photophysicochemical characteristics of ng-C₃N₄, AgNPs/ng-C₃N₄ and Ag₂TPPS₄/AgNPs/ng-C₃N₄

To verify the ISCS process of H₂TPPS₄ on the surface of AgNPs/ng-C₃N₄, a suspension of Ag₂TPPS₄/AgNPs/ng-C₃N₄ before and after visible light irradiation has been characterized using UV-vis spectra (Fig. 4). The initial Soret band for Ag₂TPPS₄/AgNPs/ng-C₃N₄ is recorded at 413 nm, while a new red-shifted Soret band at 434 nm is observed after irradiation, which is related to the molecular flattening of the porphyrin upon interaction with AgNPs/ng-C₃N₄ surfaces.³³ Moreover, a new band for Ag₂TPPS₄/AgNPs/ng-C₃N₄ is observed at 644 nm, accompanied by the disappearance of four free base porphyrin bands at 514, 548, 584 and 638 nm, confirming that the center of H₂TPPS₄ is substitutively coordinated to Ag⁺ ions to assemble on the surface of AgNPs/ng-C₃N₄ and form Ag₂TPPS₄/AgNPs/ng-C₃N₄via the ISCS route.³⁴ In the presence of benzoquinone as an electron capture agent, in the AgNPs/ng-C₃N₄, the photogenic holes partially oxidized the AgNPs (Ag⁰ + photoexcited holes → Ag⁺), inducing the replacement of H⁺ by Ag⁺. Thus, Ag₂TPPS₄ was synthesized in situ at the surface of the AgNPs, which caused the methyl pyridine group of the porphyrin to rotate (a two-dimensional plane conformation) to form a coplanar conformation.³⁵ The deviation from the conformation extends the p-conjugation of the porphyrin molecule and enhances the resonance electron-withdrawing effect of the methylpyridinium groups.³⁶ Hence, the gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) decreased, as illustrated by the consequent red shift (413 nm to 434 nm) in Fig. 4.

Fig. 4 UV-vis spectra of Ag₂TPPS₄/AgNPs/ng-C₃N₄ in BQ solution (A) before and (B) after visible light irradiation. Inset: a structural diagram of Ag₂TPPS₄/AgNPs/ng-C₃N₄.

Fig. 5D shows the photoluminescence (PL) spectra of g-C₃N₄, H₂TPPS₄, Ag-C₃N₄ and Ag₂TPPS₄/AgNPs/ng-C₃N₄. Due to its strong photoluminescence, ng-C₃N₄ exhibits a high peak at 428 nm. The peak at 700 nm is second-order scattering (λ_SOS), which is double its excitation wavelength (λ_ex = 350 nm). Both its photoluminescence and λ_SOS are greatly reduced after the first illumination. Nevertheless, the λ_SOS of Ag₂TPPS₄/AgNPs/ng-C₃N₄ increases after the second irradiation. From the theoretical calculation, the scattering intensity has been calculated to be,

where F is characteristic of the materials, E_m(λ) is the intensity distribution of the scattered light, and E_x(λ) is the excitation spectrum of the scattered light.^37,38 The alteration in the λ_SOS intensity indicates a change in the nature of the composite. Strong fluorescence quenching is detected for both ng-C₃N₄ and H₂TPPS₄ after anchoring AgNPs on ng-C₃N₄ and coupling H₂TPPS₄. The quenching after the first irradiation is caused by the photogenerated electron–hole pairs being effectively injected into the conduction band of the AgNPs through the LUMO of ng-C₃N₄. Moreover, the strong fluorescence quenching for H₂TPPS₄ after addition to AgNPs/ng-C₃N₄ results in the formation of a weak coordination complex, indicating the efficiency of electron transportation between Ag₂TPPS₄ and ng-C₃N₄ through the AgNPs. Thus, AgNPs here play a role as a conductor, accelerating the electrocatalytic process.

Fig. 5 PL spectra of AgNPs/ng-C₃N₄, Ag₂TPPS₄/AgNPs/ng-C₃N₄, H₂TPPS₄, and ng-C₃N₄ in deionized water (the excitation light wavelength is 350 nm).

Electrochemical characterization and electrochemical activity of the modified electrode

Electrochemical impedance spectroscopy (EIS) can be utilized to analyse electrode process kinetics, electric double layers, material properties, etc. The Nyquist plot consists of a semicircle and a line with a 45° inclination for the electrode kinetics control process, and its radius can be used to describe the impedance of the electron transport process. The EIS of the different modified electrodes are shown in Fig. 6A, and an equivalent circuit diagram (R_s: solution resistance, C_d: capacitance, R_ct: charge transfer resistance, and Z_w: Warburg impedance^39–42) is speculated in the inset. As shown in Fig. 6A, the bare GCE (Fig. 6A(a)) exhibits a high R_ct due to its poor electron transfer and weak electrochemical behaviour. The semicircle of ng-C₃N₄/GCE (Fig. 6A(b)) has large radius, demonstrating its high impedance based on its semiconductor characteristics. After being modified with AgNPs/ng-C₃N₄ (Fig. 6A(c)), the R_ct decreases slightly, which could be ascribed to the excellent conductivity of the AgNPs. When Ag₂TPPS₄ is modified on the GCE, the R_ct is also lower than for bare GCE, indicating that Ag₂TPPS₄ (Fig. 6A(d)) can be helpful to improve the conductivity of the sensor. The R_ct of the Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE (Fig. 6A(e)) is much smaller, suggesting the enhanced conductivity and electrochemical properties of Ag₂TPPS₄/AgNPs/ng-C₃N₄. It is supposed that the effect conjugation of the porphyrin structure further improves the efficiency of charge transfer. Cyclic voltammetry (CV) scans were recorded for Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE in blank PBS (pH = 7, 0.1 M) and PBS containing 1 × 10⁻⁶ M RAC (Fig. 6B). No obvious redox peak was observed for Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE in the blank PBS. However, a sensitive oxidation peak appeared at a potential of 0.68 V in PBS containing 1 × 10⁻⁶ M RAC, which was caused by the oxidation of RAC in PBS. It can be seen from the figure that there is no obvious reduction peak, indicating that the electrocatalytic reaction of RAC on the modified electrode is irreversible. To further investigate the electrocatalytic ability of the as-prepared nanocomposite toward RAC, DPV was employed to provide further detail, because it is an advantageous technique for obtaining higher sensitivity by eliminating the non-Faraday currents that occur in CV.⁴³ Compared to the other electrodes in PBS (0.1 M, pH 7.0) containing 1 × 10⁻⁶ M RAC, the response signal of bare GCE to RAC is very faint throughout the scanning interval (Fig. 6C(a)). The ng-C₃N₄/GCE (Fig. 6C(b)) and the mixed H₂TPPS₄+ng-C₃N₄/GCE electrode (Fig. 6C(c)) have a weak response but electrocatalytically oxidize RAC. When AgNPs are modified onto the electrode, the peak current increases and the peak potential decreases (Fig. 6C(d)). In addition, Ag₂TPPS₄/GCE (Fig. 6C(e)) also has a catalytic effect toward RAC, and the peak current is higher than those of curves a–d. For AgNPs/ng-C₃N₄/GCE, the peak current (1.04 μA) increases significantly, indicating that AgNPs/ng-C₃N₄ has a better catalytic oxidation effect towards RAC. Note that when Ag₂TPPS₄/AgNPs/ng-C₃N₄ is modified onto the electrode, the clear increase in the peak current (1.23 μA) and the decrease of peak potential (0.55 V) indicate the reduction of the electrocatalytic activation energy, meaning that the special ring structure of Ag₂TPPS₄ effectively improves the electron transfer ability and overcomes the electrocatalytic barrier. The results reveal that both H₂TPPS₄ + ng-C₃N₄/GCE and AgNPs/ng-C₃N₄/GCE have an electrochemical response to RAC, and that the response of Ag₂TPPS₄/AgNPs/ng-C₃N₄ to RAC is significantly higher. Thereout, it is concluded that Ag₂TPPS₄/AgNPs/ng-C₃N₄ exerts double-faced active centers. Moreover, ng-C₃N₄/GCE, AgNPs/GCE and Ag₂TPPS₄/GCE all have electrocatalytic oxidation effects towards RAC, illustrating that the synergistic effect of the three nanomaterials enhances the electrocatalytic activity and response sensitivity. As a result, Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE possesses excellent electrocatalytic ability and extremely high response sensitivity. Additionally, the high conductivity of the AgNPs in the sandwich-like structure benefits the charge transfer between Ag₂TPPS₄ and ng-C₃N₄, thereby promoting the quick electrochemical response of the electrode to RAC. This is consistent with the fluorescence response characteristics. The potentials and peak current (I_pa) responses of the Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE electrode at different scan rates from 50 to 350 mV s⁻¹ are shown in Fig. 6D. Similarly, with increasing scanning speed (v), the oxidation peak potential shifted positively and I_pa improved. A linear relationship between the scan rate and I_pa can be observed in Fig. 6E. The fitted curve is I_pa = 0.011v + 1.60 (R² = 0.993), which demonstrates that the electrode reaction of RAC on the Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE electrode is controlled by an adsorption process. Simultaneously, the peak potentials and scan rate have a linear relationship with a correlation (R²) of 0.993 (Fig. 6F). According to the Laviron equation (E_p = E₀ + RT ln(RTk⁰/αnF)/αnF − RT ln v/αnF),⁴⁴ the slope of the regression equation is RT/αnF, corresponding to the fact that two-electron transfer is involved in this process (Scheme 2).

Fig. 6 (A) EIS of (a) bare GCE, (b) ng-C₃N₄/GCE, (c) AgNPs/ng-C₃N₄/GCE, (d) Ag₂TPPS₄/GCE, and (e) Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE in 5.0 mM [Fe(CN)₆]^3-/4-; inset: the equivalent circuit. (B) CVs of a Ag₂TPPS₄/AgNPs/ng-C₃N₄-modified GCE in blank solution and in 0.1 M PBS (pH = 7) containing 1 × 10⁻⁶ M RAC. (C) DPV of (a) a bare GCE, (b) ng-C₃N₄/GCE, (c) H₂TPPS₄ + ng-C₃N₄/GCE, (d) AgNPs/GCE, (e) Ag₂TPPS₄/GCE, (f) AgNPs/ng-C₃N₄/GCE, and (g) Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE in PBS (0.1 M, pH 7.0) containing 1 × 10⁻⁶ M RAC. (D) CVs of Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE in PBS (pH 7.0) containing 1 × 10⁻⁶ M RAC at scan rates from 50 to 350 mV s⁻¹. The effects of the scan rate on (E) the oxidation peak currents and (F) the potentials of RAC.

Scheme 2 The electrochemical oxidation mechanism of RAC based on Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE.

To reveal the interfacial features and electrocatalytic mechanism for Ag₂TPPS₄/AgNPs/ng-C₃N₄ and RAC, zeta potential testing was carried out for five samples, as listed in Table S1.† The zeta potential of AgNPs/ng-C₃N₄ is −2.13 mV, indicating that the surface is negatively charged. When AgNPs/ng-C₃N₄ was successfully assembled with Ag₂TPPS₄ to form a sandwich-like structure, the zeta potential of Ag₂TPPS₄/AgNPs/ng-C₃N₄ was significantly more negative (−42.2 mV), indicating that the Ag₂TPPS₄/AgNPs/ng-C₃N₄ has more negative charges on its surface and great potential for electrostatic adsorption towards RAC, which is positively charged (1.26 mV). In addition, the zeta potentials of both AgNPs/ng-C₃N₄ + RAC (7.69 mV) and Ag₂TPPS₄/AgNPs/ng-C₃N₄ + RAC (−15.6 mV) are more positive, confirming that Ag₂TPPS₄/AgNPs/ng-C₃N₄ has double-sided electrostatic interaction. Therefore, the electrostatic attraction benefits the subsequent electrocatalytic process. The results suggest that the adsorption of the sandwich composite towards RAC through electrostatic interactions is conducive to the realization of high-sensitivity detection of RAC. To further investigate the electrocatalytic oxidation mechanism of RAC using Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE, differential pulse voltammetry (DPV) was applied in this work (Fig. S6†). The peak current gradually decreases after cyclic oxidation, confirming the decreased sensitivity toward RAC and the oxidation of RAC. Simultaneously, blue shifts at 223 nm and 280 nm in the UV-vis spectrum are displayed (Fig. S6B†) due to the oxidation of two terminal –OH groups of RAC to weaken the conjugated structure. Furthermore, a new low-intensity absorption band at 313 nm emerges after 250 rounds of cyclic oxidation, which corresponds to the n–π* electronic excitation of an α,β-unsaturated ketone.⁴⁵ The above results demonstrate the consumption and conversion of RAC during the electrocatalytic detection process. Thus, they also prove the existence of double catalytic faces for a sandwich-like structure. Ag₂TPPS₄/AgNPs/ng-C₃N₄ is able to efficiently oxidize two hydroxyl groups and enhance the catalytic rate.

Optimization of the conditions for the modified electrode and the electrochemical detection of RAC

In order to achieve a highly sensitive electrochemical sensor for RAC, the experimental conditions must be optimized. The accumulation potential of the peak current was measured from 0 to 0.4 V using a specific accumulation time (AT = 120 s) (Fig. S7A†), and the best accumulation potential was 0.2 V (Fig. S7B†). The effect of pH on the peak current of the Ag₂TPPS₄/AgNPs/ng-C₃N₄-modified electrode was studied over the range of pH 4 to 9 in PBS solution in the presence of RAC (1 × 10⁻⁶ M) with a potential of +0.2 V (Fig. S8A†). Fig. S8B† depicts a maximum current response at pH = 7. Therefore, for the subsequent experiments, PBS solution at pH 7 was used.

The DPV curves in Fig. 7A show increasing oxidation peak current with increasing RAC concentration at approximately 0.54 V. The calibration data were obtained for RAC solutions under the optimum experimental conditions described in Fig. 7B, and the plot demonstrates linear behaviour in the range of 1 × 10⁻⁷ to 1.2 × 10⁻⁵ mol L⁻¹ (R² = 0.998) with a limit of 5.1 × 10⁻⁸ mol L⁻¹ (S/N = 3). The linear range and the limit of detection of Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE were compared to those reported in the papers listed in Table S2,† which revealed that the Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE electrode has excellent sensitivity. Moreover, the detection sensitivity of the Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE electrode is superior to the CAC (Codex Alimentarius Commission, founded by FAO and WHO) standard (10 ppb).

Fig. 7 (A) DPV plots for Ag₂TPPS₄/AgNPs/mg-C₃N₄/GCE with different amounts of RAC in PBS (0.1 M, pH 7.0) (black: highest concentration; dark purple: lowest concentration). (B) The calibration curve showing the relationship between the RAC concentration and peak current.

Selectivity, reproducibility, and stability of the modified electrode

The selectivity of the Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE was analyzed using different interferents, namely, glucose (Glu), dopamine (DA), ascorbic acid (AA), galactose (Gal), vanillin (Van), fructose (Fru), and acetaminophen (Ace). The relative current response of Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE in 50 μM RAC with the addition of 500 μM of the other chemicals is given in Fig. S10†. Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE shows no remarkable response towards the other interferents, indicating its outstanding selectivity for Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE. The reproducibility of the Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE was investigated by 20 cycles of DPV in PBS (pH = 7, 0.1 M) containing 1 × 10⁻⁶ M RAC. The peak current of oxidation did not decrease remarkably, and the relative standard deviation (RSD) was 3.56%, demonstrating the feasibility of the sensor. The stability of this alternative developed method was also investigated by measuring the current response at a fixed RAC concentration over a long period. The RSD of the stability of oxidation remains 3.7% after storage in air for 7 days, indicating that Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE possesses excellent stability. To examine the actual detection effect of the electrochemical sensor, the Ag₂TPPS₄/AgNPs/ng-C₃N₄/GCE was used to analyse three milk samples purchased from the local market. Recovery experiments were performed using the standard addition method. The value of recovery in milk is in the range from 92 to 114% (Table S3†), revealing that this new method is accurate and feasible.

Conclusions

An environmentally friendly photochemical synthesis method was performed to prepare Ag₂TPPS₄/AgNPs/ng-C₃N₄ with a sandwich-like structure and double-faced active centres. In the first visible light irradiation step and in the presence of a hole capture agent, AgNPs were successfully anchored on the surface of ng-C₃N₄. Ag₂TPPS₄ was formed via an in situ centre-substituted (ISCS) process during the second irradiation step and with the addition of an electron capture agent. The core AgNPs were used as a “conductive bridge” connecting both ng-C₃N₄ and Ag₂TPPS₄ to achieve the formation of a sandwich-like and highly efficient double-faced electrocatalyst. This Ag₂TPPS₄/AgNPs/ng-C₃N₄ with a sandwich-like structure was utilized to construct a double-faced electrocatalytic platform for the nonenzymatic detection of RAC. Ag₂TPPS₄ and ng-C₃N₄ play the role of electrocatalysts, reducing the catalytic barrier and oxidizing the –OH groups of RAC. Based on the experimental results, the electrochemical sensor not only exhibits excellent electrocatalytic activity towards RAC with a broad linear range and low detection limit, but it also provides remarkable long-term stability. This can be ascribed to the high efficiency of charge transport, the double-sided electrocatalytic activity, the high specific surface area, and the synergistic effect of the hybrid composite. The successful application of this electrode demonstrates that modified nanocomposites provide a new platform for designing sensors to determine RAC sensitively.

Author contributions

Xuehua Weng: writing – original draft, formal analysis. Huiling Ye: data curation. Wenqiang Xie: conceptualization, writing – original draft. Meihui Ying: formal analysis. Haibo Pan: supervision, conceptualization, methodology, writing – review & editing. Min Du: project administration.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Natural Science Foundation of China (No. 51872048), Natural Science Foundation of Fujian Province (No. 2016J01222, 2018J01518), and International Cooperation Foundation of Fujian Province (No. 2009I0016).

Notes and references

1. S. Choudhury, J. O. Baeg, N. J. Park and R. K. Yadav, Green Chem., 2014, 16, 4389–4400.
2. M. Fagnoni, D. Dondi, D. Ravelli and A. Albini, Chem. Rev., 2007, 107, 2725–2756.
3. L. C. Jia, G. P. Mane, C. Anand, D. S. Dhawale, Q. M. Ji, K. Ariga and A. Vinu, Chem. Commun., 2012, 48, 9029–9031.
4. Y. Z. Zhao, N. A. Al Abass, R. Malpass-Evans, M. Carta, N. B. McKeown, E. Madrid, P. J. Fletcher and F. Marken, Electrochem. Commun., 2019, 103, 1–6.
5. S. Taylor, M. L. McKee and A. Samokhvalov, J. Photochem. Photobiol., A, 2017, 342, 143–152.
6. H. Choi, S.-J. Ko, Y. Choi, P. Joo, T. Kim, B. R. Lee, J.-W. Jung, H. J. Choi, M. Cha, J.-R. Jeong, I.-W. Hwang, M. H. Song, B.-S. Kim and J. Y. Kim, Nat. Photonics, 2013, 7, 732–738.
7. Y. Wang, L. Feng, J. Pang, J. Li, N. Huang, G. S. Day, L. Cheng, H. F. Drake, Y. Wang, C. Lollar, J. Qin, Z. Gu, T. Lu, S. Yuan and H.-C. Zhou, Adv. Sci., 2019, 6, 1802059.
8. Q. Peng, X. Weng, W. Xie, M. Ying, X. Lin, Y. Dai, Q. Yu, H. Pan, J. Liu and M. Du, J. Colloid Interface Sci., 2020, 560, 502–509.
9. S. Preiß, A. Päpcke, L. Burkhardt, L. Großmann, S. Lochbrunner, M. Bauer, T. Opatz and K. Heinze, Chem.–Eur. J., 2019, 25, 5940–5949.
10. V. Almeida-Marrero, E. van de Winckel, E. Anaya-Plaza, T. Torres and A. de la Escosura, Chem. Soc. Rev., 2018, 47, 7369–7400.
11. M. Biesaga, K. Pyrzynska and M. Trojanowicz, Talanta, 2000, 51, 209–224.
12. J. Otsuki, Coord. Chem. Rev., 2010, 254, 2311–2341.
13. S. Majumder, B. P. Borah and J. Bhuyan, Dalton Trans., 2020, 49, 8419–8432.
14. S. Prathapan, T. E. Johnson and J. S. Lindsey, J. Am. Chem. Soc., 1993, 115, 7519–7520.
15. H. Y. Ji, T. Fei, L. L. Zhang, J. Yan, Y. M. Fan, J. H. Huang, Y. H. Song, Y. H. Man, H. Tang, H. Xu and H. M. Li, Appl. Surf. Sci., 2018, 457, 1142–1150.
16. P. Niu, L. L. Zhang, G. Liu and H. M. Cheng, Adv. Funct. Mater., 2012, 22, 4763–4770.
17. Y. Li, P. P. Zhang, Z. F. Ouyang, M. F. Zhang, Z. J. Lin, J. F. Li, Z. Q. Su and G. Wei, Adv. Funct. Mater., 2016, 26, 2122–2134.
18. N. N. Gao, F. Gao, S. Y. He, Q. H. Zhu, J. F. Huang, H. Tanaka and Q. X. Wang, Anal. Chim. Acta, 2017, 951, 58–67.
19. S. C. Yan, S. B. Lv, Z. S. Li and Z. G. Zou, Dalton Trans., 2010, 39, 1488–1491.
20. X. C. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2009, 8, 76–80.
21. B. L. Dong, S. J. Zhao, H. F. Li, K. Wen, Y. B. Ke, J. Z. Shen, S. X. Zhang, W. M. Shi and Z. H. Wang, Food Chem., 2019, 271, 9–17.
22. L. L. Chu, S. L. Zheng, B. Qu, S. W. Geng and X. J. Kang, Food Chem., 2017, 227, 315–321.
23. C. Li, Y. L. Wu, T. Yang, Y. Zhang and W. G. Huang-Fu, J. Chromatogr. A, 2010, 1217, 7873–7877.
24. F. Karamolegou, M. Dasenaki, V. Belessi, V. Georgakilas and N. Thomaidis, Food Anal. Methods, 2018, 11, 2925–2942.
25. X. B. Wei, Z. R. Zhang, L. F. Zhang and X. H. Xu, J. Mater. Sci., 2019, 54, 2066–2078.
26. Z. Y. Li, C. Q. Liu, V. Sarpong and Z. Y. Gu, Biosens. Bioelectron., 2019, 126, 632–639.
27. M. T. T. S. Srivastava, J. Org. Chem., 1973, 38, 2103.
28. H. Gu, T. S. Zhou and G. Y. Shi, Talanta, 2015, 132, 871–876.
29. X. R. Du, G. J. Zou, Z. H. Wang and X. L. Wang, Nanoscale, 2015, 7, 8701–8706.
30. K. M. Kadish, G. B. Maiya, C. Araullo and R. Guilard, Inorg. Chem., 1989, 28, 2725–2731.
31. J. H. Qin, B. Ma, X. F. Liu, H. L. Lu, X. Y. Dong, S. Q. Zang and H. Hou, Dalton Trans., 2015, 44, 14594–14603.
32. B. Vlckova, P. Smejkal, M. Michl, M. Prochazka, P. Mojzes, F. Lednicky and J. Pfleger, J. Inorg. Biochem., 2000, 79, 295–300.
33. Y. Ishida, D. Masui, T. Shimada, H. Tachibana, H. Inoue and S. Takagi, J. Phys. Chem. C, 2012, 116, 7879–7885.
34. T. J. P. S. S. Cady, Inorg. Chem., 1978, 17, 1501–1507.
35. F. J. Vergeldt, R. B. M. Koehorst, A. Vanhoek and T. J. Schaafsma, J. Phys. Chem., 1995, 99, 4397–4405.
36. J. Malig, A. W. I. Stephenson, P. Wagner, G. G. Wallace, D. L. Officer and D. M. Guldi, Chem. Commun., 2012, 48, 8745–8747.
37. Y. R. Shen, Nature, 1989, 337, 519–525.
38. Z. Ma, Y. Xu, P. Li, D. Cheng, X. Zhu, M. Liu, Y. Zhang, Y. Liu and S. Yao, Anal. Chem., 2021, 93, 3586–3593.
39. S. Zhang, N. Huang, Q. Lu, M. Liu, H. Li, Y. Zhang and S. Yao, Biosens. Bioelectron., 2016, 77, 1078–1085.
40. L. Yang, N. Huang, Q. Lu, M. Liu, H. Li, Y. Zhang and S. Yao, Anal. Chim. Acta, 2016, 903, 69–80.
41. M. Liu, L. Wang, J. Deng, Q. Chen, Y. Li, Y. Zhang, H. Li and S. Yao, Analyst, 2012, 137, 4577–4583.
42. N. Huang, S. Zhang, L. Yang, M. Liu, H. Li, Y. Zhang and S. Yao, ACS Appl. Mater. Interfaces, 2015, 7, 17935–17946.
43. J. J. Jiang and X. Z. Du, Nanoscale, 2014, 6, 11303–11309.
44. X. Fuku, F. Iftikar, E. Hess, E. Iwuoha and P. Baker, Anal. Chim. Acta, 2012, 730, 49–59.
45. F. Li, S. Tang, N. T. Tsona and L. Du, Atmos. Environ., 2020, 237, 117650.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1na00130b

---