Heterojunction-engineered two-dimensional Ti₃C₂–CoFe₂O₄ nanozyme with oxidase-like activity for SERS detection of glutathione in human serum

Abstract

The inherently weak Raman signals from molecules with small scattering cross-sections pose a significant challenge for surface-enhanced Raman scattering (SERS), a technique that is further limited by its reliance on costly precious metal substrates and exogenous labeling strategies. To address these limitations, this study constructs a Ti₃C₂–CoFe₂O₄ heterostructure by anchoring oxidase (OXD)-like CoFe₂O₄ nanoparticles (NPs) on two-dimensional (2D) conductive Ti₃C₂ MXene nanosheets. The resulting interface forms a Mott–Schottky junction, which facilitates rapid charge transfer and synergistically enhances both catalytic and SERS performance. Structurally, the 2D Ti₃C₂ framework provides abundant anchoring sites for the uniform dispersion of CoFe₂O₄ NPs. This effectively prevents particle aggregation and maximizes the exposure of catalytic active sites, thereby enhancing both stability and catalytic activity. Additionally, the Ti₃C₂–CoFe₂O₄ heterojunction effectively suppresses the recombination of charge carriers and promotes the separation of photogenerated charges, generating abundant superoxide anion radicals that oxidize 3,3′,5,5′-tetramethylbenzidine (TMB) for catalytic signal amplification. Therefore, the ingenious combination of nanozymes and SERS technology enables the generation of SERS-active reporters via nanozyme-catalyzed reactions, thus avoiding the need for external labeling modifications. The strategy simultaneously enhances Raman signals through the synergistic effect of photoinduced charge transfer and localized surface plasmon resonance. This Ti₃C₂–CoFe₂O₄ heterojunction exhibits integrated OXD-like activity and SERS enhancement, enabling sensitive glutathione (GSH) detection in human serum samples. Through catalytic oxidation of TMB to oxidized TMB, a distinct Raman peak emerges at 1615 cm⁻¹, with its intensity reduction quantitatively correlating with GSH concentration via competitive reactive oxygen species scavenging. Quantitative analysis demonstrates a linear response range of 0.50–200 μmol L⁻¹ and a detection limit of 0.073 μmol L⁻¹, with serum sample recoveries ranging from 94.7%–115%. This study provides a paradigm for designing non-precious metal nanozyme materials with integrated catalytic and SERS capabilities, demonstrating significant potential for practical applications in clinical diagnostics and biosensing.

---

Introduction

Surface-enhanced Raman scattering (SERS) has become a powerful analytical technique due to its high sensitivity, molecular specificity, and non-destructive characteristics.¹ SERS technology relies on traditional precious metal nanostructures for signal enhancement through localized surface plasmon resonance (LSPR). However, these substrates suffer from aggregation, oxidation, and high costs, limiting their stability and reproducibility in practical applications.^2,3 To circumvent these limitations, non-precious metal substrates have emerged as promising alternatives owing to their excellent chemical stability and cost-effectiveness.⁴ Nevertheless, the SERS enhancement of these substrates mainly stems from the chemical mechanism (CM), specifically charge transfer,⁵ which is often intrinsically weaker than the electromagnetic mechanism (EM) dominant in precious metals. Particularly, the intrinsically weak Raman signals arising from low scattering cross-sections or improper energy-level alignment of target molecules pose a challenge for direct detection with non-precious metal substrates.⁶ Although labeling strategies can be employed to overcome this limitation and achieve the necessary sensitivity,⁷ they inevitably introduce cumbersome procedures and signal nonuniformity.⁸ To address these challenges, the integration of nanozymes with SERS substrates has been proposed to amplify signals through catalytic reactions, enabling an integrated detection and catalysis platform. In such systems, 3,3′,5,5′-tetramethylbenzidine (TMB) is widely used because its oxidized product (oxTMB) yields strong, reproducible Raman bands for indirect detection and has found broad practical applications.^9,10 Such enzymatic enhancement provides a viable SERS-based approach for indirect small-molecule detection. For instance, Li et al. synthesized CeO₂@nanogel/Au nanoparticles (NPs) with enhanced peroxidase (POD)-like activity and established a high-performance SERS platform for the indirect detection of H₂O₂ using its catalytic product as the probe.¹¹ Similarly, Ruan et al. constructed Fe₃O₄@PB@Au@GO_x integrated nanozymes, establishing a colorimetry-SERS dual-mode sensing platform for the indirect detection of glucose and H₂O₂ based on nanozyme-mediated catalytic reactions.¹² Another study reported the design of Co₃O₄@Co–Fe oxide/AuNPs double nanoboxes nanozymes with enhanced POD-like and SERS activities, which enabled the development of a colorimetric-SERS bimodal sensor for the indirect detection of metallothioneins via the radical scavenging effect of the protein.¹³ These established systems primarily leverage POD-like activity. Recently, advances in nanozyme design have significantly broadened the scope of catalytic SERS beyond POD-like systems, focusing on optimizing catalytic specificity and efficiency. For example, dual-active center nanozymes have been engineered to enhance cascade catalysis through oxygen transfer between spatially separated active sites or to dynamically regulate selectivity by mimicking substrate-induced conformational locking effects in natural enzymes.^14,15 Concurrently, catalytic SERS strategies have progressively extended to other enzyme mimics, including oxidase (OXD)-mimicking materials. More recently, this includes AgNPs@metal–organic framework (MOF)-818 nanozymes exhibiting catechol OXD-like activity for selective dopamine analysis in human urine.¹⁶ Among the developed nanozymes, OXD-like nanozymes offer a distinct advantage over POD-like systems by catalyzing substrate oxidation without requiring exogenous H₂O₂, thereby eliminating instability risks and streamlining workflows. However, studies on non-precious metal nanozymes that integrate high OXD-like activity with efficient SERS enhancement remain scarce. To address this limitation, the rational design of non-precious metal nanozymes with high OXD-like activity aims to enhance the detection of small molecules with low Raman scattering cross-sections.

The key factor governing the synergistic enhancement of catalytic activity and SERS signal intensity in non-precious metal nanozymes lies in the efficiency of interfacial charge transfer. In this regard, heterojunction engineering has emerged as an effective strategy to optimize the interface by facilitating efficient charge transfer and increasing the exposure of active sites.¹⁷ Impressively, semiconducting heterojunctions have shown that rational pairing of dissimilar components can synergistically accelerate charge separation and enrich surface carriers, thereby amplifying both enzyme-like and SERS activity.^18,19 For example, the CuO/TiO₂ p–n heterojunction suppresses carrier recombination and promotes interfacial charge transfer, lowering the SERS detection limit of 4-mercaptobenzoic acid to 1 × 10⁻¹⁰ mol L⁻¹.²⁰ In addition, our group demonstrated a MoO_3−x/CuS heterojunction that integrates SERS and nanozyme capabilities, enabling the sensitive detection of the cerebral infarction biomarker S100B.²¹ Further evidence of heterojunction-enhanced catalysis is provided by cobalt ferrite systems. For instance, Guo et al. reported that depositing CoFe₂O₄ on a conjugated porous organic polymer dramatically boosts its OXD-like activity via accelerated electron transfer, enabling ultrasensitive colorimetric detection.²² Therefore, the construction of a heterojunction with excellent OXD-like activity and SERS enhancement capability provides a novel platform for the highly sensitive detection of small molecules.

As two-dimensional (2D) transition metal carbides, MXenes show significant potential for applications in SERS substrates and catalysis due to their high specific surface area, LSPR effects, and exceptional electrical conductivity.^23,24 However, pristine MXene materials (e.g., Ti₃C₂) exhibit limited SERS enhancement capabilities and inherently lack enzyme-like catalytic properties. Fortunately, the functionalization of nanozyme particles with MXene materials provides a viable pathway to enhance their SERS sensitivity through synergistic effects. As reported in the literature, CoFe₂O₄ NPs are magnetic oxides with multiple metal oxidation states and intrinsic enzyme-like activity.^25,26 Nevertheless, CoFe₂O₄ NPs exhibit a pronounced tendency to aggregate due to strong magnetic dipole–dipole interactions, which significantly reduces accessible active sites and compromises the catalytic rate.^27,28 Therefore, the integration of CoFe₂O₄ nanozymes onto the Ti₃C₂ surfaces seems to be a good choice to further improve the catalytic activity and SERS performance of the heterojunction nanozymes.

In this work, we present a 2D Ti₃C₂–CoFe₂O₄ heterojunction with remarkable OXD-like activity and SERS performance. This design achieves a highly dispersed and stable distribution of CoFe₂O₄ NPs on the Ti₃C₂ support, fully exposing catalytic active sites. Furthermore, it capitalizes on the excellent metallic conductivity and LSPR effect of Ti₃C₂ to establish a highly conductive interface with CoFe₂O₄, effectively forming a Mott–Schottky junction that promotes directional charge transfer and separation. Consequently, the interfacial charge-transfer efficiency is significantly enhanced, realizing a synergistic improvement in both nanozyme catalytic activity and SERS performance. Based on the synergistic enhancement of OXD-like activity at the heterojunction interface, TMB is efficiently oxidized into oxTMB, thereby generating a SERS signal. By leveraging glutathione (GSH)-induced signal quenching and exploiting the competitive scavenging of reactive oxygen species (ROS), a highly sensitive detection strategy for human serum GSH is developed. This strategy offers a novel approach for designing non-precious metal nanozymes that integrate enzymatic catalysis with SERS enhancement, paving the way for advanced biomedical sensing applications.

Experimental sections

Reagents and materials

Titanium aluminium carbide (Ti₃AlC₂) was purchased from Rhawn Co., Ltd. Cobalt dichloride (CoCl₂·6H₂O) was purchased from Fu Chen Co., Ltd. 3,3′,5,5′-Tetramethylbenzidine (TMB, ≥99%), iron chloride (FeCl₃·6H₂O), sodium acetate (NaAc), ethylene glycol, ethanol (≥99.8%), hydrogen fluoride (HF, ≥10%, v/v), tetramethylammonium hydroxide (TMAOH, 25 wt%), methanol (≥99.9%), isopropyl alcohol (99.8%), and p-benzoquinone (99%) were purchased from Aladdin Co., Ltd. GSH was purchased from Coolaber Technology Co., Ltd. The reduced GSH detection assay kit (G4305-48 T) was purchased from Servicebio. All the raw materials and reagents were of analytical grade and used directly without further purification.

Synthesis of Ti₃C₂ and Ti₃C₂–CoFe₂O₄

The Ti₃C₂ material was synthesized according to the method described in a previous report.²⁹ In a standard synthesis process, 1.0 g of Ti₃AlC₂ powder was initially immersed in 10.0 mL of 10% (v/v) HF at ambient conditions. This mixture was stirred for 5 minutes, resulting in a suspension that was subsequently centrifuged. The solid product was then washed multiple times with deionized water until the pH of the supernatant reached 5–6 to remove excess HF and byproducts. To exfoliate the cleaned bulk material into a sheet-like configuration, the obtained wet sediment was dispersed in a 25 wt% aqueous solution of TMAOH and the mixture was stirred for 24 hours.

The sample from the TMAOH treatment was collected by centrifugation at 6000 rpm for 15 minutes and then redispersed in deionized water at a weight ratio of 1 : 300 (sample : water). By manually shaking this solution for 10 minutes, colloids containing uniformly sized Ti₃C₂ sheets terminated with Al(OH)₄⁻ were formed. To remove unexfoliated particles, the colloid was centrifuged at 3500 rpm for 15 minutes. The supernatant was collected and dried under vacuum at 60 °C for 12 hours to obtain the final product of delaminated Ti₃C₂.

The preparation of Ti₃C₂–CoFe₂O₄ composites were as follows. The Ti₃C₂–CoFe₂O₄ composites were synthesized with controlled mass ratios by systematically varying the Ti₃C₂ MXene loading while maintaining identical precursor inputs for CoFe₂O₄ formation. Specifically, 5 mg of Ti₃C₂ nanosheets dispersed in 30 mL ethylene glycol with 2 mmol FeCl₃·6H₂O and 1 mmol CoCl₂·6H₂O yielded the 1 : 1 Ti₃C₂–CoFe₂O₄ composite. Proportional increases to 10 mg and 15 mg of Ti₃C₂ under identical precursor conditions and solvent volume produced the 1 : 2 and 1 : 3 Ti₃C₂–CoFe₂O₄ composites, respectively.

All mixtures underwent ultrasonication followed by hydrothermal treatment at 180 °C for 24 hours to crystallize the heterostructures. After being allowed to cool naturally to room temperature, the resulting black precipitate was collected by filtration and washed sequentially with absolute ethanol and deionized water several times. The final product was dried in a vacuum oven at 60 °C for 12 hours prior to use. The nanocomposite preparation procedure is illustrated in Scheme S1. Pure CoFe₂O₄ was prepared following the same procedure but without the addition of Ti₃C₂.

Characterization

Scanning electron microscopy (SEM, Bruker Gemini500 and SU8010) and transmission electron microscopy (TEM, JEM-2010 HR) were employed to characterize the morphology of the synthesized samples. Powder X-ray diffraction (XRD) measurements were performed on a SmartLab X-ray Powder Diffractometer (Rigaku, Japan). X-ray photoelectron spectroscopy (XPS) was carried out using a Thermo Scientific Nexsa to determine the compositional and valence state information of the prepared samples. The contents of Fe and Co in the prepared Ti₃C₂–CoFe₂O₄ samples were determined by inductively coupled plasma optical emission spectrometry (ICP-OES) with an Agilent 720ES ICP-OES spectrometer. Ultraviolet–visible-near-infrared (UV-vis-NIR) and ultraviolet-visible (UV-vis) absorbance spectra were measured by a Shimadzu UV-3600 spectrophotometer (Shimadzu, Japan). Zeta potential (ζ) was measured to determine the surface charge of the materials (EliteSizer). The photoluminescence spectra were recorded using a fluorescence spectrophotometer (Shimadzu, Japan) at an excitation wavelength of 244 nm. Diffuse reflectance spectra were recorded using UV-vis diffuse reflectance spectroscopy (UV-vis DRS, PerkinElmer Lambda 950) in the range of 200–800 nm. Mott–Schottky (M–S) measurements were performed in 0.1 mol L⁻¹ Na₂SO₄, while electrochemical impedance spectroscopy (EIS) was conducted in 5 mmol L⁻¹ K₃[Fe(CN)₆]/K₄[Fe(CN)₆] and 0.1 mol L⁻¹ KCl, using an electrochemical workstation (CHI-660E, Shanghai, China) with a standard three-electrode system comprising a glassy carbon working electrode, a platinum counter electrode and an Ag/AgCl reference electrode. Raman measurements were performed on liquid samples using an RM5 microscopy Raman spectrometer at an excitation laser wavelength of 633 nm. A 10× objective lens with an accumulation time of 10 seconds was used for SERS signal collection.

Evaluation of oxidase-like property of Ti₃C₂–CoFe₂O₄

The verification of the Ti₃C₂–CoFe₂O₄ OXD-like activity was as follows: 100.0 μL 5 mmol L⁻¹ TMB solution and the prepared Ti₃C₂–CoFe₂O₄ dispersion (100.0 μL, 1 mg mL⁻¹) were mixed with the buffer (HAc–NaAc, 800.0 μL, 0.1 mol L⁻¹, pH = 3.0). After 10 minutes, the spectrum was measured using UV-vis spectroscopy.

SERS measurements of GSH

GSH solutions (50.0 μL) at various concentrations, TMB (5 mmol L⁻¹, 20.0 μL), Ti₃C₂–CoFe₂O₄ (50.0 μL, 1 mg mL⁻¹), and buffer (0.1 mol L⁻¹ HAc–NaAc, pH = 3.0) were mixed to a final volume of 200 μL. After incubation for 10 minutes at 35 °C, the SERS spectrum was measured. The analytical signal was calculated as the difference in SERS intensity between the samples with and without GSH. All experiments were performed in triplicate.

Analysis of serum samples

Serum samples were provided by the Affiliated Cancer Hospital of Sun Yat-sen University. The analytical procedure was as follows: the samples were collected and diluted 20-fold with a pH 3.0 buffer solution before proceeding with SERS analysis. The diluted serum samples (50.0 μL) were mixed with TMB (5 mmol L⁻¹, 20.0 μL), Ti₃C₂–CoFe₂O₄ (50.0 μL, 1 mg mL⁻¹), and buffer (0.1 M HAc–NaAc, pH = 3.0) to obtain a total volume of 200 μL. After incubation for 10 minutes at 35 °C, the SERS spectrum was measured. The SERS signals were recorded on a spectrometer under 633 nm laser excitation with an accumulation time of 10 seconds. For comparison, the GSH levels in human serum were also determined by the standard method using a Reduced GSH Detection Assay Kit (Servicebio, G4305-48 T), which is based on Ellman's method.

Results and discussion

Characterization of Ti₃C₂–CoFe₂O₄

The morphology and microstructure of the as-prepared Ti₃C₂–CoFe₂O₄ composite (with a 1 : 2 mass ratio, used hereafter unless otherwise stated) were investigated by SEM and TEM. After TMAOH etching, Ti₃C₂ exhibited ultrathin lamellar nanostructures (Fig. 1A and S1A, B), which provided abundant active sites for probe molecules and thereby significantly enhanced the SERS detection sensitivity. Bare CoFe₂O₄ formed spherical NPs with an average size of approximately 100 nm (Fig. 1B and S1C), but exhibited severe agglomeration due to magnetic dipole–dipole interactions. In contrast, after anchoring CoFe₂O₄ onto Ti₃C₂ nanosheets, magnetic agglomeration was significantly suppressed while the particulate morphology was retained (Fig. 1C and S1D). Moreover, EDS elemental mapping (Fig. 1E–J) demonstrated the homogeneous distribution of Co, Fe, O, C, and Ti, confirming the successful construction of the composite heterostructure. Quantitative analysis by ICP-OES indicated that the mass fractions of Fe and Co were 36.43% and 17.34%, respectively, corresponding to a molar ratio of approximately 2.22 : 1. This ratio is consistent with the theoretical stoichiometric ratio of CoFe₂O₄, and the loading content of CoFe₂O₄ was determined to be 72.79 wt%. HRTEM imaging (Fig. 1D) showed the clear deposition of CoFe₂O₄ on Ti₃C₂. The measured interplanar spacing of 0.218 nm corresponded to the Ti₃C₂ (105) plane.^30,31 The spacing of 0.257 nm was attributed to the CoFe₂O₄ (311) plane. The XRD patterns of the CoFe₂O₄ and Ti₃C₂–CoFe₂O₄ composites (Fig. S2) were consistent with the standard CoFe₂O₄ structure (JCPDS no. 22-1086), and the prominent diffraction peaks corresponded to the (220), (311), (400), (511), and (440) planes.³² These collective findings confirmed the formation of heterojunctions between the two components, which facilitated faster charge transport.

Fig. 1 TEM images of (A) Ti₃C₂, (B) CoFe₂O₄, (C) Ti₃C₂–CoFe₂O₄. (D) HRTEM images of Ti₃C₂–CoFe₂O₄ (inset: SAED pattern of Ti₃C₂–CoFe₂O₄). (E) SEM of Ti₃C₂–CoFe₂O₄ and EDS mapping of (F) Co, (G) Fe, (H) O, (I) Ti and (J) C of Ti₃C₂–CoFe₂O₄.

XPS is an important technology for determining the binding energy of elements, which can be used to estimate the charge transfer behavior of heterostructures.³³ Fig. 2 presented the XPS spectra of CoFe₂O₄ and Ti₃C₂–CoFe₂O₄. Fig. 2A–D collectively revealed the elemental composition (C, O, Ti, Co, Fe) and chemical bonding states (C–Ti, C–C, C–F, Ti–O) in the Ti₃C₂–CoFe₂O₄ composite.^34–37 Moreover, the XPS analysis provided critical insights into the chemical states and interfacial interactions in the Ti₃C₂–CoFe₂O₄ composite. For the pure CoFe₂O₄ sample, multiple peaks were detected in the Co 2p spectrum (Fig. 2E), reflecting the characteristics of Co species in different lattice sites. The peaks at 780.2/795.6 eV corresponded to Co²⁺ 2p_3/2/2p_1/2 in octahedral sites, while those at 781.8/797.6 eV corresponded to Co²⁺ 2p_3/2/2p_1/2 in tetrahedral sites. The Fe 2p spectrum (Fig. 2F) also revealed Fe species in different coordination environments. The peaks at 710.4/723.8 eV corresponded to Fe³⁺ 2p_3/2/2p_1/2 in octahedral sites, and those at 712.3/726.1 eV corresponded to Fe³⁺ 2p_3/2/2p_1/2 in tetrahedral sites.³⁷ In contrast, the Ti₃C₂–CoFe₂O₄ composite material displayed slight but consistent binding energy downshifts across O 1s, Co 2p, and Fe 2p spectra compared to pure CoFe₂O₄. This negative shift indicated the electron transfer from Ti₃C₂ to CoFe₂O₄, driven by Fermi level equilibration at the heterojunction interface. The alignment of energy levels facilitated thermodynamic equilibrium, confirming the formation of a Mott–Schottky junction at the Ti₃C₂–CoFe₂O₄ interface. This electronic interaction underpinned the enhanced charge separation observed in catalytic performance.

Fig. 2 XPS analyses of CoFe₂O₄ and Ti₃C₂–CoFe₂O₄: (A) full-scan XPS spectra; (B) C 1s; (C) O 1s; (D) Ti 2p; (E) Co 2p; (F) Fe 2p.

The oxidase-like property of Ti₃C₂–CoFe₂O₄

Following characterization, the OXD-like activity of Ti₃C₂–CoFe₂O₄ composites were assessed by UV-vis and SERS monitoring of TMB oxidation. Ti₃C₂–CoFe₂O₄ composites with different ratios were tested for their ability to catalyze TMB oxidation, with SERS signals at 1615 cm⁻¹ (assigned to ring stretching and C–H bending modes of the single-electron oxidation product)³⁸ and UV absorption at 652 nm monitored. The 1 : 2 Ti₃C₂–CoFe₂O₄ showed the most intense Raman peak at 1615 cm⁻¹ and the highest absorbance at 652 nm (Fig. 3A and B). Meanwhile, the absorbance of the Ti₃C₂–CoFe₂O₄ heterojunction with TMB increased in a time-dependent manner until TMB was completely oxidized in 10 minutes (Fig. S3). Such excellent catalytic performance was not observed for the simply physically blended Ti₃C₂/CoFe₂O₄, which only exhibited weak and broad oxTMB absorption peaks in the reaction system (Fig. S4). This superior performance may stem from charge transfer between Ti₃C₂ and CoFe₂O₄. Zeta potential characterization (Fig. 3C) revealed that a more negative surface charge was exhibited by the Ti₃C₂–CoFe₂O₄ composite compared to pure CoFe₂O₄ following incorporation of negatively charged Ti₃C₂. This enhanced negative charge facilitated electrostatic interactions with positively charged TMB, thereby improving OXD-like activity. Subsequent optimization of catalytic performance was conducted through systematic variation of pH, temperature, and component concentrations (Fig. S5). Optimal reaction conditions were identified as pH 3.0, 35 °C, 0.25 mg mL⁻¹ composite concentration, and 0.5 mmol L⁻¹ TMB concentration.

Fig. 3 (A) SERS spectra of TMB with CoFe₂O₄ and Ti₃C₂–CoFe₂O₄ composites at different mass ratios. (B) UV-vis absorbance at 652 nm for TMB, CoFe₂O₄ with TMB, and TMB with Ti₃C₂–CoFe₂O₄ composites at different mass ratios. (C) Zeta potential spectra of Ti₃C₂, CoFe₂O₄ and Ti₃C₂–CoFe₂O₄. (D) Absorption spectra of TMB with Ti₃C₂–CoFe₂O₄ system under different gaseous conditions (air, N₂). (E) Absorbance of TMB at 652 nm in the Ti₃C₂–CoFe₂O₄ using different scavengers. (F) Michaelis–Menten equation of Ti₃C₂–CoFe₂O₄ material at different concentrations of TMB (inset: Lineweaver–Burk plots of the velocity versus varying concentrations of TMB).

Additionally, to explore the catalytic mechanism of the Ti₃C₂–CoFe₂O₄ material, the effect of oxygen on its OXD-like activity was investigated. As shown in Fig. 3D, the absorbance peak at 652 nm decreased significantly when N₂ was introduced into the reaction system, suggesting that O₂ was essential for the OXD-like activity of Ti₃C₂–CoFe₂O₄. Furthermore, different radical scavengers were used to identify the radicals produced during the catalytic reaction of Ti₃C₂–CoFe₂O₄. The addition of p-benzoquinone (PBQ), a superoxide anion radical (O₂˙⁻) scavenger, caused a significant decrease in the absorbance at 652 nm, indicating strong inhibition of the catalytic activity of Ti₃C₂–CoFe₂O₄. In contrast, methanol (MeOH), a hydroxyl radical scavenger, and L-histidine (L-His), a singlet oxygen scavenger, caused only slight decreases in the absorbance when added to the system. These results confirmed that O₂˙⁻ played a primary role in the catalytic system and demonstrated that Ti₃C₂–CoFe₂O₄ generates O₂˙⁻ by catalyzing O₂ (Fig. 3E).

The steady-state kinetics of the Ti₃C₂–CoFe₂O₄ were calculated using the Michaelis–Menten equation (V = V_max × [S]/(K_m + [S])) and the Lineweaver–Burk equation (1/V = K_m/V_max × [1/S] + 1/V_max), based on changes in TMB concentration while keeping other concentrations constant.³⁹ This further confirmed the catalytic mechanism, where V_max represents the maximum reaction rate, [S] is the substrate concentration, and K_m is the Michaelis–Menten constant. Fig. 3F illustrates the relationship between reaction rate and substrate concentration, and the inset in Fig. 3F shows the linear relationship obtained by double reciprocal transformation of the Michaelis–Menten equation. Using the Michaelis–Menten calculation method, the K_m value for TMB and the maximum reaction rate V_max of the system were determined. The Ti₃C₂–CoFe₂O₄ nanozyme has the smallest K_m value for TMB (Table S1), indicating the highest affinity and best catalytic performance among the materials compared.

Synergistic mechanism for Ti₃C₂–CoFe₂O₄ oxidase-like catalysis and SERS enhancement

The synergistic enhancement mechanism of OXD-like activity and SERS signals for the Ti₃C₂–CoFe₂O₄ heterojunction was explored by measuring the UV-vis DRS spectra of CoFe₂O₄ and Ti₃C₂–CoFe₂O₄, as well as the Mott–Schottky plot of CoFe₂O₄ (Fig. S6 and S7). Ti₃C₂–CoFe₂O₄ exhibits a narrowed bandgap compared to the pristine CoFe₂O₄ (Fig. S6), which enhances visible light absorption and facilitates the generation of photogenerated electron–hole pairs for efficient charge separation. To evaluate the contribution of light, the catalytic oxidation of TMB was monitored by UV-vis absorption at 652 nm under dark and ambient light conditions. The results showed accelerated oxTMB formation under ambient light (Fig. S8), supporting a photo-assisted charge transfer process. Based on these findings, Fig. 4A depicts a plausible mechanism for this synergistic enhancement. The SERS enhancement of the heterojunction involves both interfacial charge transfer through the Mott–Schottky junction and electromagnetic contribution from Ti₃C₂ LSPR. When metallic Ti₃C₂ (Fermi level E_f = −0.04 V vs. normal hydrogen electrode (NHE)⁴⁰) contacts the p-type semiconductor CoFe₂O₄, the E_f of CoFe₂O₄ is well-documented to be closely aligned with its valence band (VB).⁴¹ Driven by the Fermi level gradient from higher to lower energy, spontaneous electron transfer occurs from Ti₃C₂ to CoFe₂O₄ (as corroborated by the XPS results in Fig. 2) until the system reaches a unified Fermi level at equilibrium. This charge transfer process creates positive charge centers on the Ti₃C₂ side and negative charge centers on the CoFe₂O₄ side at the interface, thereby establishing a built-in electric field pointing from Ti₃C₂ to CoFe₂O₄. Interfacial electronic coupling between Ti₃C₂ and CoFe₂O₄ establishes a rectifying contact by forming a Mott–Schottky junction, which generates abundant ROS-mediated oxidation sites.^42,43 Under light irradiation, electrons in the VB of CoFe₂O₄ are excited to the conduction band (CB), generating electron–hole pairs. The built-in field facilitates the rapid transfer of these photogenerated electrons from the CB of CoFe₂O₄ to Ti₃C₂, where Ti₃C₂ acts as an electron reservoir to stabilize charge separation and suppress recombination. Concurrently, the LSPR effect of Ti₃C₂ under laser excitation generates hot electrons.⁴⁴ UV-vis-NIR spectra confirm the LSPR response of Ti₃C₂ and its retention in the heterojunction, as evidenced by the NIR peak shift and broadening after coupling with CoFe₂O₄ (Fig. S9). These hot electrons can inject into the CB of CoFe₂O₄ under photoexcitation via Landau damping. This bidirectional electron transfer, coupled with the built-in field-driven carrier migration, effectively suppresses charge recombination and achieves efficient separation of photogenerated electrons and holes. It thereby lays the critical foundation for the two core functionalities of the heterojunction, which are enhanced OXD-like activity and amplified SERS enhancement. The OXD-like activity is realized through a ROS-mediated catalytic oxidation pathway. Specifically, electrons accumulated on Ti₃C₂ reduce dissolved oxygen to generate O₂˙⁻, and these radicals further convert low Raman-active TMB into high Raman-active oxTMB.^45,46 The SERS enhancement relies on a synergistic effect of CM and EM. The CM originates from the photoinduced charge transfer (PICT) between CoFe₂O₄ and adsorbed TMB, where holes in the VB of CoFe₂O₄ are transferred to TMB molecules. This charge transfer process not only increases the molecular polarizability of TMB and its oxidized products but also promotes the formation of TMB˙⁺ radicals and oxTMB, which are species with larger Raman cross-sections, thereby significantly boosting the CM for SERS.⁴⁷ Meanwhile, the localized electromagnetic field generated by the LSPR effect of Ti₃C₂ under light excitation affords an auxiliary contribution to the EM and further amplifies the Raman signal of oxTMB.⁴⁸ Notably, Ti₃C₂–CoFe₂O₄ establishes a quantifiable signal transduction pathway by catalyzing the conversion of O₂ to O₂˙⁻ for TMB oxidation. GSH competes with TMB for O₂˙⁻ consumption, which in turn reduces the production of oxTMB.⁴⁹ Based on this mechanism, a SERS signal attenuation model is constructed to achieve sensitive detection of GSH. Collectively, the optimized interfacial band alignment and electronic coupling in the Ti₃C₂–CoFe₂O₄ heterojunction enable efficient charge separation and utilization. These features synergistically boost OXD-like catalytic activity and amplify SERS enhancement through oxidation mediated by ROS, chemical enhancement driven by PICT, and electromagnetic enhancement derived from LSPR.

Fig. 4 (A) Synergistic enhancement of OXD-like activity and SERS signals by Ti₃C₂–CoFe₂O₄ heterojunction. (B) PL spectra of CoFe₂O₄ and Ti₃C₂–CoFe₂O₄ composites. (C) EIS Nyquist plots of Ti₃C₂, CoFe₂O₄ and Ti₃C₂–CoFe₂O₄ composites.

The spatial charge separation efficiency was examined by the photoluminescence (PL) technique. As shown in Fig. 4B, the pristine CoFe₂O₄ presented a broad and intense emission peak centered around 368 nm, originating from the band-gap recombination of photogenerated electrons and holes. The incorporation of Ti₃C₂ resulted in a significant reduction in photoluminescence intensity from the composite, demonstrating effective suppression of charge recombination and enhanced charge separation efficiency. To gain deeper insight into charge carrier dynamics, electrochemical impedance spectroscopy (EIS) analysis was performed. Results showed the Ti₃C₂–CoFe₂O₄ composite exhibited a smaller Nyquist plot diameter than CoFe₂O₄, indicating lower impedance and charge transfer resistance (Fig. 4C). This reflected the high conductivity of Ti₃C₂ and the faster interfacial charge-transfer kinetics of the formed heterojunction.

Quantitative detection of GSH

The Ti₃C₂–CoFe₂O₄ nanocomposite exhibits both OXD-like activity and SERS enhancement, enabling sensitive GSH detection through a redox-mediated signal transduction mechanism. As an OXD-like nanozyme, the Ti₃C₂–CoFe₂O₄ catalyzes TMB oxidation to blue oxTMB without H₂O₂, generating strong SERS signals. Upon GSH introduction, the antioxidant properties of GSH scavenge O₂˙⁻ and reduce oxTMB back to TMB, leading to decreased SERS intensity. This strategy enables sensitive and selective GSH detection by combining catalytic signal amplification with SERS, offering clinical potential for monitoring oxidative stress biomarkers.

A SERS strategy for the quantitative analysis of GSH was successfully established under optimal experimental conditions. As depicted in Fig. 5A, the SERS intensity at 1615 cm⁻¹ decreased as the GSH concentration increased. The standard curve was plotted based on the relationship between the 1615 cm⁻¹ intensity and the GSH concentration (Fig. 5B). The regression equation was ΔI = 22 772.99 log C_[GSH] + 28 412.67 with a linear correlation coefficient (R) of 0.9949. The limit of detection was 0.073 μmol L⁻¹, and the linear range extended from 0.50 to 200 μmol L⁻¹. Compared to other GSH detection methods (Table S2), this SERS strategy shows higher sensitivity and a broader linear range.

Fig. 5 (A) SERS spectra of GSH with different concentrations. (B) The linear relationship between SERS signal changes at 1615 cm⁻¹ and the concentrations of GSH. ΔI = I₀ − I_i, I₀: the initial SERS intensity at 1615 cm⁻¹ of oxTMB without GSH; I_i: the SERS intensity at 1615 cm⁻¹ of oxTMB with various levels of GSH. (C) The selectivity investigation with various substances, including GSH (10.0 µmol L⁻¹), UA (100 µmol L⁻¹), ascorbic acid (100.0 µmol L⁻¹), glucose (5.00 mmol L⁻¹), glutamine (100 µmol L⁻¹), L-cysteine (2.00 µmol L⁻¹), L-glycine (100 µmol L⁻¹), L-glutamic acid (100 µmol L⁻¹), and valine (100 µmol L⁻¹). (D) The influence of potential interferent substances on the SERS detection method for GSH. The GSH concentration was 10.0 μmol L⁻¹, while glucose, fructose, urea, Na⁺, Mg²⁺, K⁺, and Cl⁻ were each at 5 mmol L⁻¹.

To evaluate the selectivity and interference resistance of the proposed SERS detection method for GSH, the method was exposed to several potential human serum interfering substances. The selectivity experiment results are shown in Fig. 5C. A significant SERS signal reduction was observed for GSH, while other interfering substances exhibited almost no SERS signal reduction. Additionally, monitoring results of key interfering substances are presented in Fig. 5D. When the concentrations of these substances were 500 times those of GSH, they had almost no effect on the SERS intensity. Notably, the proposed SERS detection system operates optimally at pH 3.0, which is lower than the typical pH range (4.0–5.0) suitable for peroxidases. Consistent with previous research,⁵⁰ such strongly acidic conditions (pH < 3.5) can enhance the stability of oxidation-related species while mitigating undesirable side reactions and competitive adsorption from interfering substances, thereby further contributing to the excellent interference resistance. These results indicate that the developed method shows excellent sensitivity, selectivity, and interference resistance, enabling reliable GSH detection in human serum. The stability and repeatability of Ti₃C₂–CoFe₂O₄ were also tested due to the critical role in practical application. Time-dependent SERS measurements of TMB on the Ti₃C₂–CoFe₂O₄ composite over 40 minutes gave a relative standard deviation (RSD) of 1.6% (Fig. S10), verifying excellent temporal SERS signal stability. As shown in Fig. S11, the corresponding SERS intensity of the peak at 1615 cm⁻¹ was collected, indicating that Ti₃C₂–CoFe₂O₄ still retained high SERS activity and OXD-like activity after being stored at 4 °C for 35 days with an RSD of 3.6%. As manifested in Fig. S12, the SERS signals of 10 different batches of Ti₃C₂–CoFe₂O₄ substrates were detected. It can be observed that the intensity of the peak was essentially stable with the RSD of 1.5%. This analytical strategy exhibited excellent stability and batch reproducibility, thus laying a solid foundation for the subsequent detection of GSH in clinical samples.

Clinical serum sample analysis

To validate the practical application of the proposed SERS detection method for GSH, human serum samples were analyzed that had been collected from hepatocellular carcinoma patients and myocarditis patients. The results were compared with those from the Ellman's method. As shown in Table S3, the GSH levels in 10 human serum samples detected by this method were consistent with the values provided by the standard method but with more convenient operation and lower cost. The relative error between the two methods was within ±5.5%. Meanwhile, the RSD was below 9.7%, indicating its accuracy and reliability in detecting GSH in real serum samples. Additionally, a recovery test was performed by spiking different GSH concentration gradients into three selected serum samples. The recoveries ranged from 94.7% to 115% with RSD less than 7.6% (Table 1). These results suggest that the method has broad application prospects for detecting GSH in human serum.

Table 1 Detection of GSH concentration in human serum samples

Samples	SERS method (μmol L^−1)	Added (μmol L^−1)	Found (μmol L^−1)	Recovery (%)	RSD (%, n = 3)
Serum 1	35.5	10.0	46.7	112	5.0
		30.0	70.0	115	1.6
		50.0	85.1	99.2	4.0
Serum 2	14.6	10.0	25.0	104	0.7
		20.0	34.1	97.5	5.4
		30.0	43.0	94.7	7.6
Serum 3	25.4	20.0	46.7	111	5.0
		30.0	54.6	100	2.7
		50.0	73.1	97.2	3.7

---

Conclusions

In summary, a novel 2D Ti₃C₂–CoFe₂O₄ nanozyme was constructed through heterojunction engineering, successfully integrating high OXD-like activity and intrinsic SERS enhancement into a single non-precious metal platform. This design overcomes the challenge of coupling efficient nanozyme catalysis with sensitive, label-free detection in complex biological matrices. The key innovation lies in the markedly enhanced interfacial charge transfer, which under photoexcitation promotes charge separation and hot electron injection. This process leads to a significant enhancement in ROS generation and a strong SERS signal from the catalytic product, synergistically amplified by PICT and LSPR of the heterojunction. This mechanism enabled the development of a label-free sensing platform for GSH detection in human serum, exhibiting ultrahigh sensitivity, a wide linear range, excellent anti-interference capability, and reliable accuracy validated against the Ellman's method. The heterojunction engineering strategy presented herein provides a versatile and generalizable paradigm for designing next-generation multifunctional nanozymes. Capitalizing on the excellent anti-interference capability, this platform holds great potential for multiplexed biomarker detection in complex biological fluids such as saliva and urine. Furthermore, through rational modulation of heterojunction components, this strategy can be extended to monitor diverse redox-active biomolecules, offering a powerful tool for precision diagnostics and rapid clinical analysis.

Author contributions

Huiqi Zhu: conceptualization, investigation, writing – original draft. Ying Chen: data curation. Weiqing Yang: visualization. Zunxiang Zeng: investigation. Yuling Hu: conceptualization, writing – review & editing, resources, and supervision. Ji Zhang: resources, project administration.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data are available from the corresponding author on reasonable request.

Supplementary information (SI) is available. Supplementary material: Additional information on steady-kinetic analysis; band edge position determination for CoFe₂O₄; SEM and spectral characterization; optimization of OXD-like catalytic conditions; SERS stability and reproducibility tests; kinetic parameter comparison; GSH detection method evaluation; clinical serum sample analysis. See DOI: https://doi.org/10.1039/d6an00055j.

Acknowledgements

This research was funded by the National Natural Science Foundation of China (No. 22574176, 22074162).

References

1. S. Jin, D. Zhang, B. Yang, S. Guo, L. Chen and Y. M. Jung, Analyst, 2024, 149, 11–28.
2. T. Endo, H. Yamada and K. Yamada, Micromachines, 2020, 11, 936.
3. Y. Qiu, C. Kuang, X. Liu and L. Tang, Sensors, 2022, 22, 4889.
4. L. Chen, H. Liu, J. Gao, J. Wang, Z. Jin, M. Lv and S. Yan, Nanomaterials, 2024, 14, 1654.
5. X. Wang, E. Zhang, H. Shi, Y. Tao and X. Ren, Analyst, 2022, 147, 1257–1272.
6. D. P. Dos Santos, M. M. Sena, M. R. Almeida, I. O. Mazali, A. C. Olivieri and J. E. L. Villa, Anal. Bioanal. Chem., 2023, 415, 3945–3966.
7. I. B. Becerril-Castro, I. Calderon, J. Ockova, M. Liebel, N. F. van Hulst, V. Giannini and R. A. Alvarez-Puebla, ACS Appl. Mater. Interfaces, 2022, 14, 57165–57170.
8. L. Li, X. Cao, T. Zhang, Q. Wu, P. Xiang, C. Shen, L. Zou and Q. Li, Foods, 2022, 11, 2165.
9. Y. Wang, S. Zhang, Q. Yin, J. Wei, J. Liu and L. Wu, Trends Anal. Chem., 2026, 196, 118632.
10. Z. Li, M. Shen, F. Meng, Y. Zhang, W. Duan, C. Hou and M. Zhang, Small, 2025, 21, 2501976.
11. Y. Li, Y. Zhang, H. Jiang, M. Qi, X. Zhang, B. Zhu and L. Han, Microchem. J., 2023, 195, 109467.
12. S. Ruan, S. Chen, Y. Tu, Y. Liu, R. You, Q. Lin and Y. Lu, Microchem. J., 2024, 205, 111377.
13. Y. Long, J. Li, F. Shi, Y. Xue, Y. Wang, Z. Yang and J. Li, Anal. Chem., 2025, 97, 22279–22287.
14. H. Shen, S. Chen, S. Mo, H. Huang, H. Liang, J. Zhang, Z. Xu, W. Liu and Y. Liu, Adv. Funct. Mater., 2025, 35, 2418360.
15. H. Shen, H. Huang, B. Shu, P. Zhang, Z. Xu, W. Liu, S. Chen and Y. Liu, Adv. Mater., 2025, e21184.
16. S. Liu, G. Li and Y. Hu, Talanta, 2026, 298, 128907.
17. D. Wu, Q. Liang, H. Si, X. Yan, H. Huang, Z. Li and Z. Kang, J. Mater. Chem. A, 2022, 10, 24519–24528.
18. L. Jiang, R. Hu, A. Wang, L. Mei and J. Feng, Sens. Actuators, B, 2023, 382, 133491.
19. H. Zhang, Y. Tang, W. Wang, D. Yu, L. Yang, X. Jiang, W. Song and B. Zhao, Food Chem., 2024, 431, 137163.
20. D. Yu, L. Xu, H. Zhang, J. Li, W. Wang, L. Yang, X. Jiang and B. Zhao, Chin. Chem. Lett., 2023, 34, 107771.
21. Y. Chen, J. Zhang, Y. Li, G. Li and Y. Hu, Anal. Chem., 2024, 96, 17711–17719.
22. X. Guo, F. Yang, L. Jing, J. Li, Y. Li, R. Ding, B. Duan and X. Zhang, J. Hazard. Mater., 2022, 431, 128621.
23. Y. Gogotsi and B. Anasori, ACS Nano, 2019, 13, 8491–8494.
24. Z. Said, A. K. Pandey, A. K. Tiwari, B. Kalidasan, F. Jamil, A. K. Thakur, V. V. Tyagi, A. Sarı and H. M. Ali, Prog. Energy Combust. Sci., 2024, 104, 101162.
25. Y. Han, Z. Zhang and L. Guo, Catal. Surv. Asia, 2020, 24, 166–177.
26. F. He, W. Li, F. Zhao, X. Zhu, Q. Liu, Z. Liu, X. Zhang and X. Zhang, Microchem. J., 2020, 158, 105264.
27. Y. Zhao, K. Xia, Z. Zhang, Z. Zhu, Y. Guo and Z. Qu, Nanomaterials, 2019, 9, 455.
28. D. Zhang, J. Cheng, J. Chai, J. Deng, R. Ren, Y. Su, H. Wang, C. Ma, C. S. Lee, W. Zhang, G. P. Zheng and M. Cao, J. Alloys Compd., 2018, 740, 1067–1076.
29. J. Xuan, Z. Wang, Y. Chen, D. Liang, L. Cheng, X. Yang, Z. Liu, R. Ma, T. Sasaki and F. Geng, Angew. Chem., Int. Ed., 2016, 55, 14569–14574.
30. L. Sun, J. Xie, L. Zhang, R. Jiang, J. Wu, L. Fan, R. Shao, Z. Chen and Z. Jin, FlatChem, 2020, 20, 100152.
31. X. Zhang, J. Nie, F. Rao, H. Liu, Y. Wang, D. Qu, W. Wu, P. Zhong and G. Zhu, Ceram. Int., 2021, 47, 31302–31310.
32. S. B. Das, R. K. Singh, V. Kumar, N. Murali and S. Betal, Bull. Mater. Sci., 2024, 47, 171.
33. K. D. Jayeola, D. S. Sipuka, T. I. Sebokolodi, O. V. Nkwachukwu, C. Muzenda, B. A. Koiki, J. O. Babalola, M. Zhou and O. A. Arotiba, Chem. Eng. J., 2024, 479, 147482.
34. B. Li, S. Liu, C. Lai, G. Zeng, M. Zhang, M. Zhou, D. Huang, L. Qin, X. Liu, Z. Li, N. An, F. Xu, H. Yi, Y. Zhang and L. Chen, Appl. Catal., B, 2020, 266, 118650.
35. D. Liu, C. Li, J. Ge, C. Zhao, Q. Zhao, F. Zhang, T. Ni and W. Wu, Appl. Surf. Sci., 2022, 579, 152180.
36. C. Yuan, Z. He, Q. Chen, X. Wang, C. Zhai and M. Zhu, Appl. Surf. Sci., 2021, 539, 148241.
37. W. Shi, X. Sun, M. Xu, S. Wang, G. Liu and H. Yang, J. Water Process Eng., 2024, 67, 106229.
38. J. Wan, Q. Liu, P. Tang, Y. Ji, W. Zhong, W. Cheng, X. Xing, X. Lu and L. Zhong, Analyst, 2023, 148, 869–875.
39. Y. Chen, J. Zhang, J. Li, Y. Hu, K. Ge, G. Li and S. Liu, Anal. Chem., 2024, 96, 2998–3007.
40. J. Low, L. Zhang, T. Tong, B. Shen and J. Yu, J. Catal., 2018, 361, 255–266.
41. W. Miao, W. He, L. Shen, Y. Li, Z. Fang, Z. Yang and G. Kai, J. Environ. Chem. Eng., 2024, 12, 114202.
42. Y. Yang, Z. Zeng, G. Zeng, D. Huang, R. Xiao, C. Zhang, C. Zhou, W. Xiong, W. Wang, M. Cheng, W. Xue, H. Guo, X. Tang and D. He, Appl. Catal., B, 2019, 258, 117956.
43. K. Huang, C. Li, H. Li, G. Ren, L. Wang, W. Wang and X. Meng, ACS Appl. Nano Mater., 2020, 3, 9581–9603.
44. Y. Tang, R. Wang, W. Wang, Y. Lu, X. Jiang, B. Zhao and L. Yang, ACS Nano, 2025, 19, 39384–39397.
45. P. Jiang, L. Zhang, X. Liu, C. Ye, P. Zhu, T. Tan, D. Wang and Y. Wang, Nat. Commun., 2024, 15, 1010.
46. Y. Tan, H. Jiang, B. Wang and X. Zhang, New J. Chem., 2021, 45, 19593–19604.
47. J. Wu, L. Qu, Z. Li, L. Zhao, Y. Sun and R. Yang, Food Chem., 2023, 427, 136672.
48. F. Wang, S. Li, J. Liang, Y. Wang, H. Song, J. Yang, X. Zou and C. Li, J. Hazard. Mater., 2023, 458, 131954.
49. D. Yang, J. Liu, W. Hu, Y. Xiao, H. Chen, Y. Long and H. Zheng, Sens. Actuators, B, 2023, 393, 134170.
50. S. Wei, C. Zhang and S. Zhang, Adv. Funct. Mater., 2025, e14080.

---