Hydrogen peroxide (H₂O₂) is widely used as antimicrobial, oxidizing, reducing and bleaching agents in many fields including pharmaceutical, medical, textile, paper, and food processing^[1]. The United States Food and Drug Administration (USFDA) has affirmed the Generally Recognized As Safe (GRAS) status of H₂O₂ for use in food with a maximum permitted concentration in specified foods and residual must be removed by appropriate processing^[2]. Excessive amount of H₂O₂ has been reported to have a destructive impact on central nervous system of human body and can result in oxidative stress which is associated with many diseases including neurodegenerative disorders, diabetes, atherosclerosis and cancers^[3,4]. Therefore, monitoring H₂O₂ residual in food is of practical significance to both academic and industry. To date, a variety of techniques including fluorometry^[5], spectrophotometry^[6,7] and electrochemistry^[8,9] have been developed for detection and quantification of H₂O₂.

Electrochemical biosensing technique has generated much interest due to its advantages of simple instrumentation, easy miniaturization, high sensitivity and selectivity, as well as rapid response^[10]. At present, very few electrochemical biosensors reached practical application and commercialization mainly due to its inconsistent operational stability^[11]. The sensitivity, selectivity and operational stability of electrochemical biosensors are strongly dependent on structure and properties of electrode materials and enzyme immobilization matrixes^[1,12-13].

Two-dimensional (2D) transition metal carbides, nitrides and carbonitrides (MXene) are produced by etching layers of sp elements (specifically groups 13 and 14) from their corresponding three-dimensional (3D) MAX phases which correspond to the general formula M_n+1AX_n (n=1, 2, 3) where M represents early d-block transition metals (Ti, Sc, V, Cr, Ta, Nb, Zr, Mo, Hf), A represents main group sp elements and X is either C or N atom^[14,15]. MXenes have generated a lot of interest due to their hydrophilic surfaces, good structural and chemical stabilities, excellent electrical conductivities, and environment- friendly characteristics^[16,17]. As MXene surfaces can be used for easy immobilization of enzymes/protein to achieve accelerated reaction kinetics, low detection limits, high sensitivity and selectivity. Sit is suitable for use as highly sensitive and selective detection platform for biosensing applications^{[18,19,20,21]}. Understanding of the sensitivity, selectivity and long term operational stability of MXene electrochemical biosensors are important for application of MXene biosensors for various purposes.

Present study aims to fabricate a horse radish peroxidase@MXene electrochemical biosensor for detection of H₂O₂ in food. HRP, a heme-containing enzyme, has been widely used to catalyze oxidation of a wide variety of substrates including hydrogen peroxide^[22,23]. MXene with vertical junction structure in which MXene sheets are perpendicular to the plane of graphite has been demonstrated to have good electromagnetic absorption properties^[24]. We proposed that this vertical junction structure will improve HRP immobilization and demonstrate good electron transfer properties which made it a suitable enzyme immobilization matrix for fabrication of H₂O₂ electrochemical biosensor. We synthesized and characterized MXene using X-ray powder diffraction (XRD), Fourier transform infared spectroscopy (FT-IR) and scanning electron microscopy (SEM). MXene was then used as HRP immobilization matrixes to fabricate HRP@MXene/ chitosan/GCE biosensor. Electrochemical behavior of the fabricated HRP@MXene/chitosan/GCE biosensor was investigated and optimized using cyclic voltammetry (CV) and different pulse voltammetry (DPV). Amperometric method was used to detect concentration of H₂O₂ in real food samples. Selectivity and storage stability of the HRP@MXene/chitosan/GCE were also elucidated.

Horseradish peroxide (HRP, activity units•mg^-1) was purchased from Sigma Aldrich. Natural flake graphite (48 μm), Ti powders (48 μm, purity of 99.9%), Al powders (48 μm, purity of 99.9%), hydrogen peroxide solution (30wt%), hydroquinone (HQ), chitosan (deacetylation 95%), potassium chloride, acetic acid were obtained from Aladdin, China. Other reagents including NaCl, KCl, sodium hydroxide (NaOH), K₃[Fe(CN)₆], K₄[Fe(CN)₆]• 3H₂O were obtained from Sinoreagent, China. 0.1 mol•L^-1 phosphate buffer solutions (PBS, pH 7.0) comprising NaH₂PO₄ and Na₂HPO₄ were used as electrolyte. All aqueous solutions were freshly prepared with ultra-pure water (18 MW•cm).

G(graphite)/TiC/Ti₃AlC₂ were fabricated according to a previously reported method with slight modifications^[24]. Graphite powder (48 μm), Ti powders (48 μm, purity of 99.9%), Al powders (48 μm, purity of 99.9%), NaCl and KCl were mixed at a molar ratio of 4 : 4 : 1 : 10 : 10, placed in alumina crucible and packaged in a tube furnace. The tube furnace was heated to 800 ℃ at a heating rate of 4 ℃•min^-1 under argon protection and kept for 300 min. Following that, the mixtures were heated to 1100 ℃ at a heating rate of 4 ℃•min^-1, kept for 180 min and finally cooled to room temperature at a cooling rate of 4 ℃•min^-1. The resulting product (G/TiC/Ti₃AlC₂) was washed by deionized water to remove salts and dried at 80 ℃. G/TiC/Ti₃C₂ were obtained by etching process using HF to remove the Al atoms.

MXene was characterized using XRD, FT-IR and SEM. XRD analysis were conducted at room temperature using Bruker D8 Discover XRD (Cu radiation, λ=0.l540596 nm) over the 2θ range of 5°~70° at room temperature. FT-IR spectra was obtained in the range of 500 to 4000 cm^-1 by using a Fourier-transform infrared (FT-IR) spectroscopy (Nicolet 6700, Thermo, USA).

The microstructures of the powders were examined by a field emission scanning electron microscopy (FEI Quanta FEG 250) equipped with an EDS system and a TEM instrument (FEI Tecnai F20).

Fabrication of the HRP@MXene electrochemical biosensor is illustrated in Fig. 1. Glassy carbon electrodes (GCE, 3 mm) was firstly polished using Al₂O₃ (1.0, 0.3, 0.05 μm), cleaned by ethanol and water for three times, and finally dried under gentle N₂ stream. Ten microliter of HRP solution [10 mg•mL^-1, PBS (0.1 mol•L^-1, pH 6.0)] and 20 μL of MXene aqueous solution (5 mg•mL^-1) were mixed and shaked at 200 r•min^-1 for 10 h at low temperature.

Following that, 10 μL chitosan solution (6 mg•mL^-1, adjusted to pH 6.0 by 10 mg•mL^-1 NaOH) was added to the mixture and vibrated for 3 min. Chitosan solution has been previously reported to be positively charged and have good electrical conductivity at pH 6.0 due to the protonation of amino groups^[8,25]. As MXenes synthesized in present study is negatively charged due to the abundance of hydroxyl or fluoride groups, it could be well adhered in chitosan solution via Coulomb effect and formed a unique film on the surface of GCE^[26]. 5 μL of the resultant HRP@MXene/chitosan was dropwisely casted onto the surface of a freshly polished GCE. The prepared electrodes (HRP@MXene/chitosan/GCE) were dried and stored in 0.05 mol•L^-1 PBS (pH 7.5) in a refrigerator (4 ℃) prior to usage.

All electrochemical experiments were carried out using CHI760E electrochemical workstation (Chenhua, Shanghai) with GCE as working electrode, platinum wire as counter electrode and saturated calomel electrode (SCE) as reference electrode. The electrochemical impedance spectroscopy (EIS) and cyclic voltammograms (CVs) of electrodes fabricated using chitosan of different pH was conducted in N₂-saturated 0.1 mol•L^-1 KCl solution containing 5.0 mmol•L^-1 Fe(CN)₆^3-/4- at open circuit potential in the frequency range from 0.1 Hz to 10⁵ Hz with the amplitude 5 mV. The EIS data were analyzed using ZVIEW software.

CVs were carried out in N₂-saturated 0.1 mol•L^-1 PBS (pH 7.5) in the presence of 2.0 mmol•L^-1 H₂O₂ and 1 mmol•L^-1 HQ (dissolved in methanol) at a scanning rate of 50 mV•s^-1. Differential pulse voltammetry (DPV) was performed in N₂-saturated 0.1 mol•L^-1 PBS (pH 7.5) containing 2 mmol•L^-1 H₂O₂ and 1 mmol•L^-1 HQ (dissolved in methanol) with amplitude of 5 mV and pulse width of 0.2 s after five times of CV at a scanning rate of 50 mV•s^-1 ranging from 0.8 V to -0.8 V. The effects of electrolyte PBS buffer pH (5.5 to 8) and the concentration of MXene were evaluated and optimized in terms of CV and DPV signal.

Amperometric current-time curves for H₂O₂ were carried out to construct a calibration curve of current response at different H₂O₂ concentration. Measurements were performed in 10 mL of stirring 0.1 mol•L^-1 PBS (pH 7.5) in the presence of 1 mmol•L^-1 HQ with successive addition of H₂O₂ at room temperature under an applied peak potential value of -0.1 V. LOD was determined according to the following equation:

whereby SD refers to the standard deviation of the control measurement, and K refers to slope of the calibration curve.

Milk and dried scallop were chosen as model of liquid and solid food. Milk sample was used directly for H₂O₂ detection. Dried scallop was pre-treated according to the following procedure to extract H₂O₂ residual. Briefly, 2 g of dried scallop was immersed in 5 mL of H₂O₂ aqueous solution (3%) for 1 h. Following that, the scallop was immersed in 5 mL of water for 0.5 h to extract H₂O₂ residue. H₂O₂ concentration in spiked dried scallop test solution and milk solution (12.5, 50 and 125 μmol•L^-1 H₂O₂) were detected using the amperometric current- time curves for H₂O₂. Recovery of the HRP@MXene/ Chitosan/GCE was calculated.

Selectivity of the fabricated HRP@MXene/chitosan/ GCE biosensor was evaluated using potentially interfering substances including uric acid, glucose and ascorbic acid [100 μmol•L^-1 in 0.1 mol•L^-1 PBS (pH 7.5)].

Storage stability of the HRP@MXene/GCE was evaluated by monitoring reduction peak in CVs in 0.1 mol•L^-1 PBS with 1 mmol•L^-1 HQ and 2 mmol•L^-1 H₂O₂ during electrodes storage in 0.05 mol•L^-1 PBS at 4 ℃.

XRD patterns of the synthesized MXene (G/TiC/Ti₃C₂) and G/TiC/Ti₃AlC₂ are showed in Fig. 2(A). G/TiC/Ti₃C₂ demonstrates a dominant phase of graphite (peak at ~26°) and TiC (peaks at 35.9°, 41.8°). This is in agreement with previously reported finding^[24]. In addition, after HF etching, the peak at 39° corresponded to the (104) plane of Ti₃AlC₂ disappears compared with the XRD pattern of Ti₃AlC₂ which indicates the elimination of Al during the G/TiC/Ti₃C₂ syntheses process.

As shown in Fig 2(B), FT-IR spectra of MXene do not display any absorption peaks from 3800 to 400 cm^-1. Meanwhile, HRP demonstrates characteristic peaks at 2961, 1647, 1541, and 1080 cm^-1. The amide I band (1700-1600 cm^-1) can be assigned to the α-helical conformation of the HRP; meanwhile, the amide II band can be assigned to the β-sheet structure of the HRP^[3]. Following immobilization of HRP onto the two dimensional MXene nanosheets, the major bands of HRP can be observed on the FT-IR spectra of HRP@MXene indicating successful immobilization process without any conformational change in the secondary structure of HRP.

SEM analysis shows a two dimensional multilayered structured of Ti₃C₂ (<1 μm) standing perpendicular to the plane of G/TiC forming interfacial junctions (Fig. 1(C)). The multilayer Ti₃C₂ also demonstrated typical MXene morphology of two-dimension structure (Fig. 1(D)). This two-dimensional multilayered interfacial junctions structure provides a large specific surface area for efficient enzyme immobilization/entrapment.

Chitosan, a natural film-forming agent, is commonly used in fabrication of enzyme electrodes. It is positively charged at pH<6.3 due to protonation of amino groups^{[8, 27]}. At pH>6.3, chitosan demonstrated decreased solubility in aqueous solution with the decline of adhesion. Fig. S1(A) shows the effects of pH of chitosan solution on charge transfer resistance (R_ct) of chitosan/GCE electrodes. R_ct was found to slightly increase with pH increasing from pH 5.0 to 6.0. However, a dramatic increase in R_ct from 0.347 kΩ to 1.304 kΩ can be observed as pH of the chitosan solution increased from 6.0 to 6.5 and reached 4.663 kΩ at pH 7.0.

In addition, according to Fig. S1(B), redox peaks current decreased with increased pH, and peak separation (DEp) became bigger when pH from 6.0 to 7.0. The increasing R_ct reflected the degressive electrical conductivity of chitosan because of protonation of amino groups, and the increasing DEp indicated the declined ability of electronic transfer. Considering the film-forming and electrical conductivity of chitosan, in addition, HRP was reported to be most active at nearly neutral^[28,29], chitosan solution at pH 6.0 was used in the fabrication of HRP@MXene/ chitosan/GCE biosensor. Fig. 3(A) shows the Nyquist plots of chitosan/GCE, MXene/chitosan/GCE and HRP@MXene/ chitosan/GCE. All three electrodes demonstrated the electron transfer- limited process in the high frequency area. Chitosan/GCE electrode had an R_ct value of 174.40 Ω. Incorporation of MXene onto the chitosan/GCE matrix resulted in a decreased R_ct value of MXene/chitosan/GCE to 52.88 Ω indicating good electron transfer property of MXene from the redox probe of [Fe(CN)₆]^3-/4-. Nevertheless, immobilization of HRP onto the MXene/chitosan/GCE matrix increase of the R_ct value of HRP@MXene/chitosan/ GCE to 542.60 Ω. Increased of the R_ct value is mainly caused by steric hindrance, electrostatic interactions and partial blockage of interfacial electrons by enzyme molecules which has poor conductivity^[10]. Cyclic voltammetry(CV) for the different electrodes were carried out in 5.0 mmol•L^-1 Fe[(CN)₆]^3-/4- and 0.1 mol•L^-1 KCl (Fig. 3(B)).

In comparison with Chitosan/GCE (curve a), MXene/ chitosan/GCE (curve b) demonstrated an increase in current response and similar ΔEp value (differences between anodic and cathodic peaks potential) indicating MXene is an excellent electric conducting material. Meanwhile, HRP@MXene/chitosan/GCE (curve c) demonstrated a decrease in current response and an increase in ΔEp value indicating HRP hindered the electron conductivity.

Fig. 4 shows the CV of chitosan/GCE, MXene/chitosan/GCE, HRP@chitosan/GCE, and HRP@MXene/chitosan/ GCE electrodes obtained in 0.1 mol•L^-1 N₂-saturated PBS (pH 7.5) containing 1 mmol•L^-1 HQ and 2 mmol•L^-1 H₂O₂. Chitosan/GCE electrode demonstrated a pair of well-defined redox peaks with potentials at about 0.14 and -0.07 V which is characteristic of redox process of HQ and H₂O₂^[30]. In comparison with the signal obtained from chitosan/GCE, modification of the GCE with MXene/chitosan resulted in signal enhanced of the redox peaks. Following HRP immobilization, both HRP@chitosan/ GCE and HRP@MXene/chitosan/GCE demonstrated further enhanced reduction peak with HRP@MXene/chitosan/ GCE showing highest increase in reduction peak’s current (52 μA). Increase in the reduction peak current can be attributed to reduction process of H₂O₂ catalyzed by HRP at its reducing state (HRP_RED) (Fig. 1). During this reduction process, the redox centre of HRP_RED turned into its oxidizing state (HRP_OX). HRP_OX were then regenerated into HRP_RED with the aid of HQ which was oxidized to form benzoquinone. Finally, benzoquinone exchanged electrons with the electrode to electrochemically produced HQ. The redox processes of H₂O₂ and hydroquinone were in agreement with those previously reported findings^[30]. The aforementioned findings showed HRP@MXene/ chitosan/GCE biosensor can be used for electrochemical biosensing of H₂O₂ and MXene provided a favorable microenvironment to retain the bioactivity of HRP.

Fig. S2(A) shows the CV of HRP@MXene/chitosan/GCE obtained in 0.1 mol•L^-1 N₂-saturated PBS (pH 7.5) containing 1 mmol•L^-1 HQ and 2 mmol•L^-1 H₂O₂ at various scan rates. The redox peaks of HRP@MXene/chitosan/GCE increased linearly versus the square root of scanning rates from 20 to 500 mV•s^-1 (Fig. S2(B)). The electrochemical behaviors were in accordance with a diffusion-controlled process occurring at the surface of the biosensor^[31]. Similar results for different electrodes with mediator were also reported^{[28, 32]}.

Based on aforementioned findings, PBS buffer’s pH of 7.5 and MXene concentration of 5 mg•mL^-1 were used for fabrication of HRP@MXene/chitosan/GCE in the subsequent analysis.

Electrochemical biosensing of H₂O₂ by HRP@MXene/ chitosan/GCE was optimized in terms of electrolyte PBS buffer’s pH (pH 5.5-8.0) and concentration of MXene (0.5-10 mg•mL^-1). The pH value of the electrolyte is important for the performance of enzyme electrode as HRP activity is greatly affected by pH. Fig. S3(A) shows that the peak current of HRP@MXene/chitosan/GCE increased from pH 5.5 and reached maximum at pH 7.5. The value of pH was chosen for further study and was also in agreement with previous observations reported^[33]. Fig. S3(B) shows the peak currents of cyclic voltammograms of HRP@MXene/chitosan/GCE fabricated with different concentration of MXene. Peak current of HRP@MXene/chitosan/GCE was the highest at 5 mg•mL^-1 MXene (MXene: HRP ratio IS 1 : 1). At this concentration of MXene, HRP was fully immobilized on the surface of MXene and the biosensor demonstrated most effective performance. In terms of DPV responses (Fig. S3(C)), negative shifts in peak potentials can be observed with increased pH value. This indcated that H⁺ participated in the HRP catalyzed H₂O₂ reduction reaction to produce water. Peak potential was also affected by concentration of MXene with negative shift in peak potential and highest peak current can be observed at 5 mg•mL^-1 MXene (Fig. S3(D)).

The current-time curve which is a potential-controlled electrochemical analysis method was used to build a calibration curve of amperometric response at a series of H₂O₂ concentration. Fig. 5(A) shows the amperometric response of HRP@MXene/chitosan/GCE following successive additions of H₂O₂ to PBS buffer (Potential = -0.1 V). The corresponding calibration curves of HRP@MXene/ chitosan/GCE biosensor were presented in Fig. 5(B), which was linear at two concentration ranges (5~190 and 190~1650 μmol•L^-1 H₂O₂) with a linear regression equation of Y=0.02644X+0.55914(R²=0.999) and Y=0.01959X +1.84114 (R²=0.996). Moreover, the fabricated biosensor also showed very low detection limit of 0.74 μmol•L^-1. A comparison of linear range and detection limit for H₂O₂ with other H₂O₂ sensors reported in literature are summarized in Table S1. The data demonstrated that both the linear range and detection limit for H₂O₂ are comparable or even better than those detected using sensors recently reported. The excellent biosensing performance of HRP@MXene/chitosan/GCE can be ascribed to the unique vertical junction structure of the two dimensional MXene nanosheets which provided a suitable matrix for HRP immobilization and also platform for H₂O₂ and HQ redox reactions.

Present work used dried scallop and milk as representative of solid and liquid food system to explore the application of HRP@MXene/chitosan/GCE biosensor in detection of H₂O₂ in food samples. Fig. 5(C, D) shows the amperometric response of HRP@MXene/chitosan/GCE following additions of solutions extracted from milk and dried scallop with different concentration of H₂O₂. The curves show HRP@MXene/chitosan/GCE is a rapid and sensitive method to detect H₂O₂ at different concentraitons. The recovery of H₂O₂ in food samples at different concentrations ranged from (90.24±6.97)% to (109.20± 3.33)% (Table 1). The results indicated that the fabricated biosensor is a reliable tool for detection of residual H₂O₂ in food samples.

The anti-interference performance of HRP@MXene/ chitosan/GCE biosensor was evaluated by detecting 100 μmol•L^-1 H₂O₂ in the presence of the same concentration of ascorbic acid, glucose and uric acid as interfering substances. As shown in Fig. S4(A), there were no noticeable amperometric responses from glucose and uric acid. However, amperometric responses can be detected by ascorbic acid (34% H₂O₂) indicating ascorbic acid has the capability to participate in the redox process of HQ and H₂O₂; hence, interfering with the measurement of H₂O₂.

HRP@MXene/chitosan/GCE demonstrated good storage and operational stability. When stored in 0.05 mol•L^-1 PBS (pH 7.5) at 4 ℃, HRP@MXene/chitosan/GCE was able to retain 84.8% of its initial response to H₂O₂ after a period of 10 d (Fig. S4(B)). This indicated that the vertical junction structure of the MXene (Graphite/ TiC/Ti₃C₂) were able to act as an effective and stable platform for entrapment enzyme HRP.

In summary, we have explored a new type of supporting material for immobilizing HRP and fabricated an electrochemical H₂O₂ biosensor for in situ detection of H₂O₂ in food products. The synthesized MXene exhibited large specific area, biocompatibility, excellent electronic conductivity, and good dispersion in aqueous phase. HRP enzymes molecules immobilized on MXene/chitosan/GCE electrode showed good electrochemical behaviors and electrocatalytic activity toward reduction of H₂O₂. The fabricated HRP@MXene/chitosan/GCE biosensor exhibited a wide linear range from 5 μmol•L^-1 to 1.650 mmol•L^-1 and a low detection limit of 0.74 μmol•L^-1 with long-term stability, good reproducibility and high selectivity. The fabricated biosensor has also been successfully employed for detection of trace level of H₂O₂ in real food products (both solid and liquid food). The study provides a good concept for construction of electrochemical H₂O₂ biosensor based on MXene.

Supporting materials

Supporting materials related to this article can be found at https://doi.org/10.15541/jim20190139

Enzyme-MXene Nanosheets: Fabrication and Application in Electrochemical Detection of H₂O₂

MA Bao-Kai^1,2,3, LI Mian³, CHEONG Ling-Zhi², WENG Xin-Chu¹, SHEN Cai³, HUANG Qing³

(1. School of Life and Sciences, Shanghai University , Shanghai, 200444, China; 2. Department of Food Science and Engineering, College of Food and Pharmaceutical Sciences, Ningbo University, Ningbo 315211,China; 3. Institute of Materials Technology & Engineering, Chinese Academy of Sciences, Ningbo, 315201, China)

* HRP: Horseradish Peroxidase; CTAB: cetyltrimethylammonium bromide; GO: graphene oxide; TB: Toluidine blue; CCB: ceramic composite biosensor; BMIM·BF4: 1-butyl-3-methylimidazolium tetrafluoroborat; SWCNTs: Single-walled carbon nanotubes; PGN: porous grapheme; Hb: hemoglobin; naf: nafion.

References:

[1]YANG S, DING S, LI L, et al. One-step preparation of direct electrochemistry HRP biosensor via electrodeposition. Journal of The Electrochemical Society, 2017, 164(13): B710-B714.

[1]WANG Y, ZHAO K J, ZHANG Z Q, et al. Simple approach to fabricate a highly sensitive H₂O₂ biosensor by one-step of graphene oxide and horseradish peroxidase co-immobilized glassy carbon electrode. International Journal of Electrochemical Science, 2018, 13(3): 2921-2933.

[2]THENMOZHI K, NARAYANAN S S. Horseradish peroxidase and toluidine blue covalently immobilized leak-free Sol-Gel composite biosensor for hydrogen peroxide. Materials Science & Engineering C, Materials for Biological Applications, 2017, 70(Pt 1): 223-230.

[3]REN Q Q, WU J, ZHANG W C, et al. Real-time in vitro detection of cellular H₂O₂ under camptothecin stress using horseradish peroxidase, ionic liquid, and carbon nanotube-modified carbon fiber ultramicroelectrode. Sensors and Actuators B: Chemical, 2017, 245: 615-621.

[4]LIU Y, LIU X, GUO Z, et al. Horseradish peroxidase supported on porous graphene as a novel sensing platform for detection of hydrogen peroxide in living cells sensitively. Biosensors & Bioelectronics, 2017, 87: 101-107.

[5]ZHENG J, DIAO J, JIN Y, et al. An inkjet printed Ti₃C₂-GO electrode for the electrochemical sensing of hydrogen peroxide. Journal of The Electrochemical Society, 2018, 165(5): B227-B231.

[6]LORENCOVA L, BERTOK T, DOSEKOVA E, et al. Electrochemical performance of Ti₃C₂T_x MXene in aqueous media: towards ultrasensitive H₂O₂ sensing. Electrochimica Acta, 2017, 235: 471-479.

[7]WANG F, YANG C, DUAN C, et al. An organ-like titanium carbide material (MXene) with multilayer structure encapsulating hemoglobin for a mediator- free biosensor. Journal of The Electrochemical Society, 2014, 162(1): B16-B21.

[8]WANG F, YANG C, DUAN M, et al. TiO₂ nanoparticle modified organ-like Ti₃C₂ MXene nanocomposite encapsulating hemoglobin for a mediator-free biosensor with excellent performances. Biosensors and Bioelectronics, 2015, 74: 1022-1028.