CVD-grown MXene–CNT nanocomposite-assisted electrochemical immunosensors for label-free and ultra-sensitive monitoring of thyroid-stimulating hormone in artificial serum

Abstract

Thyroid-stimulating hormone (TSH) is a critical clinical biomarker for evaluating thyroid function and is essential in diagnosing and monitoring related hormonal disorders. However, the development of point-of-care diagnostic devices for the early detection of TSH-associated diseases remains challenging, particularly due to constraints in measurement methodologies in resource-limited settings. This research describes an effective electrochemical immunosensor developed from a novel seamless three-dimensional (3D) Ti₃C₂T_x (MXene)–carbon nanotube (CNT) nanocomposite, specifically engineered for the ultrasensitive and label-free detection of low TSH concentrations. The MXene–CNT nanocomposite was synthesized using the chemical vapor deposition (CVD) method, and its physical characteristics and electrochemical sensing capabilities were examined. The unique synergetic properties of the MXene–CNT nanocomposite enhance antibody immobilization, expand the sensor's surface area (0.119 cm²), provide additional binding sites, and improve electrical conductivity, leading to superior sensitivity and stability of the biosensor. With a label-free format, the redox mediator signal decreased significantly in response to varying concentrations of TSH, attributable to the insulating immunocomplexes formed between anti-TSH and TSH on the MXene–CNT-modified disposable electrode. The method achieved a dynamic linear range of 0.1–10 000 pg mL⁻¹, with a detection limit of 0.03 pg mL⁻¹, showcasing high selectivity against interferents. Importantly, this study presents the first successful electrochemical detection of TSH in artificial serum using the CVD-grown MXene–CNT nanocomposite immunosensor, underscoring its potential for real-sample analysis in complex biological environments. The current study opens new avenues for leveraging the CVD-grown MXene–CNT nanocomposites in designing and developing molecular diagnosis-based electrochemical and related biosensors.

---

1. Introduction

The ligand–receptor interaction is crucial for biological functions, immune responses, and signal transductions, significantly impacting biosensing technology. In healthcare, there is a trend toward diagnosing and treating infectious diseases in their early stages. Hormonal imbalances can lead to serious illnesses such as cancer, adenoma, cardiovascular disease, osteoporosis, and hyperplasia.¹ Thyroid disorders affect nearly 200 million people globally, with 60% unaware of their condition.² Many thyroid diseases are hormone-related; thyroid-stimulating hormone (TSH) is a key biomarker for thyroid function and illness diagnosis. It promotes the production of thyroid hormones and their release into the serum.^3,4 Normal adult TSH levels range from 0.4 to 5.0 μIU per mL (48–600 pg mL⁻¹).⁵ Significant fluctuations in thyroid hormones can cause symptoms of hyperthyroidism or hypothyroidism. A low TSH level may indicate subclinical hyperthyroidism.⁶ Without intervention, hyperthyroidism can lead to complications, such as decreased bone density, cardiac dysfunction, and so forth.⁷ Elevated TSH levels can cause hypothyroidism, chronic lymphocytic thyroiditis, and pituitary tumors.⁸ Thus, systematic clinical assessments for thyroid hormone analysis benefit from easy and affordable TSH monitoring.

Recently, various TSH monitoring tools, such as radioimmunoassay (RIA),⁹ chemiluminescent immunoassay (CLIA),¹⁰ liquid chromatography,¹¹ and liquid chromatography-mass spectrometry (LC-MS),¹² have been developed. However, these systems face obstacles such as a lack of miniaturization, including high costs and lengthy processes. Developing simple, rapid, and sensitive handheld electrochemical devices for TSH testing remains a significant challenge. Thanks to innovative electrochemical biosensing devices such as micro- and nano-fabrication, lab-on-a-chip development is driven by advanced biomolecular electronics that facilitate rapid testing, integration, and automation.¹³ Among immunosensors, electrochemical immunosensors are powerful screening tools that are highly sensitive, specific, affordable, convenient, and efficient.¹⁴ They are useful for food quality assessment, environmental control, clinical analyses, and medical diagnostics.¹⁵ Antibody–antigen detection modifies the transducer's surface characteristics (current or voltage) without reporter labels, such as radioactive elements, fluorescent dyes, aptamers, or enzymes. This approach holds promise in electrochemical immunosensing due to its straightforward and cost-effective diagnostic process. A key aspect of assembling electrochemical immunosensors is immobilizing antibodies or antigens as bioreceptors. Various methods, including physical or chemical adsorption, have been proposed for binding them to diagnostic devices.¹⁶ Recently, researchers have enhanced detection signals using advanced nanomaterials in electrochemical immunosensors to identify protein indicators from clinical samples.^17,18 By employing nanomaterials with improved conductivity, biocompatibility, stability, and a large surface area, electrochemical immunosensor devices achieve superior results in clinical analyses. This technique boosts detection sensitivity and monitoring of biological indicators, aiding in the early identification of clinical biomarkers.^18–20 Tracking real-world samples is crucial for timely medical interventions. Numerous materials, such as mesoporous silica materials,²¹ conducting carbon polymers,²² graphene composites,²³ carbon nanotube (CNT) composites,²⁴ Au–Ag nanoparticles,²⁵ and molybdenum disulfide (MoS₂) composites,²⁶ have been utilized as transducer components in developing immunosensors or immunoassay techniques.

MXenes are a newly discovered class of two-dimensional (2D) transition metal carbides and nitrides that have seen rising demand since their first production in 2011 by etching a precursor M_n+1AX_n phase. In this context, M represents a d-block transition metal, A indicates a group IIIA–VIA element, and X denotes carbon and/or nitrogen.²⁷ Selective etching can yield structured MXene (M_n+1AX_nT_x), with surface functional groups (–F, –O, and –OH) represented by T_x. Approximately 100 MXene compounds have been theoretically recognized, primarily based on experimental findings.²⁸ MXenes are compelling materials for transduction due to their large active surface area, layered structure, chemically modifiable nature, tailored electrical properties, high hydrophilicity, stability, and biocompatibility. They show significant potential in electrochemistry, including electrocatalysis, batteries, supercapacitors, sensors, and biosensors.^29,30 Recent studies have focused on electrochemical biosensors for detecting clinical biomarkers.³¹ MXene surfaces may negatively affect the contact interface due to electron deficiency at the flake surface, leading to limited selectivity, reduced conductivity, and low carrier mobility. Restacking and reduction related to electro-active sensitive areas can occur due to weak van der Waals forces.³² Additionally, modifying sensor surfaces is essential to enhance the MXene stability in biosensing applications, as they suffer from oxidation in solutions and aggregation in biological environments.^32,33 Despite their excellent conductivity, versatile surface chemistry, and abundant active sites, MXenes continue to encounter several issues that limit their practical use in electrochemical biosensing. Major challenges include nanosheet restacking, which decreases accessible surface area, surface oxidation, reduced conductivity and stability, poor stability in aqueous and biological media, and nonspecific adsorption that affects selectivity. To address these problems, researchers suggest approaches such as hybridizing with conductive nanomaterials like graphene and CNTs to prevent restacking and enhance charge transfer; functionalizing surfaces with polymers, biomolecules, or heteroatom dopants to improve sensitivity and selectivity; engineering heterostructures and adjusting interlayer spacing to allow better ion access; and applying antifouling coatings to increase stability and biocompatibility. Recent advances demonstrate that integrating MXenes into hybrid or nanostructured architectures can significantly boost sensitivity, reproducibility, and detection limits, as shown by recent reports.^34–37 These approaches offer promising ways to overcome the inherent limitations of MXenes and develop highly sensitive, reliable, next-generation MXene-based biosensing platforms. While various techniques aim to regulate the binding position of antibodies on MXene surfaces, achieving overall directional control remains challenging.

Inspired by this perspective, MXene has emerged as a promising material, with several scientists expressing enthusiasm for MXene-based electrochemical transduction devices. This growing interest is primarily attributed to MXene materials, particularly Ti₃C₂T_x, which show high electrical conductivity and versatile surface chemistry, making them ideal for electrochemical composites.³⁸ Adding conductive fillers like CNTs helps prevent MXene restacking and further improves electrical performance. Silver nanoparticles (AgNPs) can also enhance conductivity and provide antibacterial properties, which are beneficial for wearable and biomedical applications.^39,40 MXene facilitates the production of nanosheets and carbon materials that enhance the mechanical strength and electro-catalytic and conductive properties.⁴¹ Integrating MXene with CNT-derived nanocomposites significantly increases the number of active sites and surface area, thereby enhancing electroanalytical detection capabilities.^42,43 MXene–CNT devices enhance signal intensity, electrocatalytic activity, and stability, which boosts biosensor sensitivity for diagnosis.^44,45 Additionally, MXene-based functional materials demonstrate that –F, –O, and –OH groups on MXene promote adhesion to synthetic and natural surfaces (hydroxyl) through hydrogen bonding and electrostatic interactions, forming linkages with Ti.⁴⁶ Recent reports have described combined MXene–CNT-based composite materials through physical mixing to create electrochemical sensors, DNA sensors, and immunosensors.^47–49 However, interaction between MXene and CNT composite materials can lead to limited electrical conductivity due to nonohmic contact, and structural flaws. In contrast, the MXene–CNT nanocomposite produced via CVD shows increased electrical conductivity, better interaction sites, and reduced contact resistance. The potential use of covalent surface modification antibody–antigen binding on the seamless 3D MXene–CNT nanocomposite in label-free electrochemical immunosensors has not been addressed.

In this study, we propose a novel electrochemical immunosensor based on a 3D seamless MXene–CNT nanocomposite, produced through the CVD growth of CNT on MXene, for the sensitive and rapid detection of trace levels of TSH. The conductive MXene promotes electron transfer, while CNTs enhance ion migration by preventing MXene re-stacking and toxic effects, thereby increasing adaptability to biological settings. The synergistic effect of the MXene–CNT nanocomposite boosts biosensor sensitivity by amplifying the electronically transmitted signal through improved antibody affinity, showcasing its potential as a transformative electronic tool for TSH antigen monitoring. This research aims to create a reliable framework for screening interconnected 3D conductive transducer surfaces to address the limitations of portable diagnostic techniques. It emerges as a strong candidate for early diagnosis of trace-level TSH through efficient label-free electrochemical immunosensing in artificial blood serum samples, marking a first in the field.

2. Experimental

2.1 Chemicals, reagents, and instrumentation

Chemicals, reagents, and instrumentation details are provided in the SI.

2.2 MXene etching and delamination

Ti₃C₂T_x (MXene) was prepared following our earlier reports.⁵⁰ In a standard process, bulk powders of the MAX (Ti₃AlC₂) precursor were gradually added to a 48% hydrofluoric acid (HF) solution. The reaction mixture was then placed in a polypropylene beaker with a Teflon stirring magnet in the exhaust hood and constantly stirred for 30 hours at ambient temperature. The solution was rinsed with deionized water (DI) using a centrifuge set to 4000 rpm for 20 minutes, aiming for a pH of about 7. DMSO solutions were utilized to enlarge and delaminate the obtained (HF) etching mixed solutions as hazardous chemical compounds intercalated into the as-synthesized MXene. Finally, the DMSO mixture containing MXene substances was centrifuged again with DI water to remove residual DMSO. The resulting product of MXene was dried for 12 hours at 100 °C.

2.3 MXene–CNT nanocomposite through the CVD method

The MXene–CNT nanocomposite was synthesized based on our previous literature report.⁵⁰ This nanocomposite was grown using the CVD method. MXene powders were produced and employed as a catalyst. A quartz boat was loaded with 0.2 g of MXene catalyst, which was then inserted and spread out at the center of a tube-shaped CVD furnace. Subsequently, CNTs were grown directly using MXene catalyst powder. The MXene–CNT nanocomposite was created using argon (Ar) gas in an inert atmosphere, and before growth, the catalyst (MXene) activation was conducted with hydrogen gas (H₂), while maintaining mass flow rates of Ar (150 sccm) and H₂ (150 sccm), respectively (sccm – standard cubic centimeters per minute). The interval was set at 30 minutes, and the temperature was increased to 700 °C at 5 °C per minute. Once the furnace reached room temperature at 700 °C, acetylene (C₂H₂) was introduced as a carbon source with a predetermined reaction time and a mass flow of 40 sccm. Following this process, the MXene–CNT nanocomposite was produced in powdered form and allowed to cool to room temperature in the CVD furnace before being dried in an oven set to 70 °C for 12 hours. CNTs were directly grown on MXene to form the nanowires and/or nanocomposite, and the resultant product, the MXene–CNT nanocomposite, was treated with a 1 : 1 (v/v) mixture of HNO₃ and H₂SO₄ to activate their surface, as described in our earlier procedure.^51,52 The treated MXene–CNT nanocomposite was evaluated using physical and electrochemical characterization techniques. Fig. S1 shows a schematic illustration of the MXene–CNT nanocomposite synthesized using the CVD method.

2.4 Nanocomposite immunosensor construction

An in-house screen-printing procedure was employed to fabricate the disposable SPGEs. To construct the immunosensors, a 5 μL aliquot of the MXene–CNT nanocomposite (optimized at 2 mg mL⁻¹) was drop-cast onto the surface of the SPGE and allowed to dry in air. Next, 5-μL of NHS (400 mM) and EDC (300 mM) in a 1 : 1 ratio were applied for covalent surface chemistry to the MXene–CNT nanocomposite modified SPGE and left for 1 hour at room temperature (RT) to facilitate the activation process.⁴⁴ Following this, the EDC–NHS/MXene–CNT/SPGE was treated with 5-μL of anti-TSH (optimized at 10 μg mL⁻¹, sourced from 7.4 pH PBS) and incubated for 1 hour at RT, with a careful rinse using PBS (pH 7.4) afterward. After electrode modification, anti-TSH antibodies were covalently immobilized using EDC–NHS surface chemistry, forming stable amide bonds with carboxyl groups on the MXene–CNT surface to enhance the sensor stability. Subsequently, the remaining active sites on the anti-TSH/EDC–NHS/MXene–CNT/SPGE were blocked for 1 hour at RT using 1% BSA (prepared in 7.4 pH PBS) to prevent non-specific reactions, followed by another rinse with PBS. Finally, the BSA/anti-TSH/EDC–NHS/MXene–CNT/SPGE was treated with various concentrations of 5 μL TSH antigen, incubated for 45 minutes at RT, followed by careful rinsing with PBS, after which the electrochemical responses were measured. Schematic illustration of Ti₃C₂T_x MXene synthesis via selective etching of Al from the Ti₃AlC₂ MAX phase, along with the CVD-based synthesis of the MXene–CNT nanocomposite (A), and the fabrication process of the TSH electrochemical immunosensors is illustrated (B) in Scheme 1.

Scheme 1 Schematic illustration of Ti₃C₂T_x MXene synthesis by etching Al from Ti₃AlC₂ and CVD growth of the MXene–CNT nanocomposite (A), and stepwise construction of the electrochemical immunosensor for TSH detection (B).

2.5 Sample preparation process

A 10 μg sample of RhTSH was dissolved in 100 μL of Milli-Q water to prepare a standard stock solution containing 100 μg mL⁻¹ concentration. This solution was divided into 10 aliquots of 10 μL each and stored at −20 °C until needed. Similarly, a standard stock solution of the anti-TSH antibody (Ab) was prepared by dissolving it in 200 μL of Milli-Q water, resulting in a final concentration of 0.5 mg mL⁻¹. The solution was also divided into 10 separate aliquots, 10 μL each, and stored at −20 °C for future use. Before testing, the stock solution was diluted with PBS to create the working solution of RhTSH containing RhTSH at the desired concentrations.

3. Results and discussion

3.1 Surface characteristics of the modified MXene interfaces

Morphological and structural changes of MXene and MXene–CNT surfaces were analyzed using FE-SEM and TEM (Fig. 1). The FE-SEM results showed that the Ti₃AlC₂ structure has an accordion-like morphology with increased lamellar spacing in Ti₃C₂T_x MXene after selectively etching Al layers. This enhances contact interfaces and creates conduits that are essential for ion migration (Fig. 1A).⁵³ The expanded structure indicates that MXene's stratified foundation allows stronger attachment for additional materials, enhancing the functionality of the MXene–CNT nanocomposite for biomolecular diagnostics. Post-etching, MXene is used in the CVD process for consistent growth of CNT nanowires on the edges of the MXene layers, forming a highly structured 3D conductive network of the MXene–CNT nanocomposite (Fig. 1B and C). These structural changes provide rich active sites, increased surface area, and electrical conductivity, developing a favorable interface for antibody binding. The covalent attachment of TSH antibodies to the MXene–CNT surface was confirmed by the observation of a stacked layer and tubular-shaped structures (Fig. 1D). Evaluations showed that the MXene–CNT nanocomposite effectively immobilized antibodies on the transducer surface, aiding further electrochemical analysis. Additionally, thorough TEM evaluations of MXene and the MXene–CNT nanocomposite were performed (Fig. 1E–I). TEM analysis confirmed that high-quality monolayer MXene nanosheets were synthesized post-etching, characterized by an ultra-thin, sheet-like morphology (Fig. 1E and F). The CVD process yielded neat interconnected monolayer and tubular structures, indicating effective interactions (Fig. 1G–I). The monolayer structure and 3D conductive network of the MXene–CNT nanocomposite provide an advantage for substantial CNT bundle growth on MXene, increasing electro-active sites and large surface area for rapid electron transfer in electrochemistry applications.

Fig. 1 FE-SEM images of MXene (A), MXene–CNT (B), MXene–CNT at low magnification (C), and MXene–CNT/Ab (D). TEM images of MXene (E), MXene at low magnification (F), MXene–CNT (G), and MXene–CNT at low magnification (H and I).

The XRD patterns of MXene and the MXene–CNT nanocomposite are shown in Fig. SI-2, illustrating the crystalline structural changes caused by CNT integration. The pristine MXene exhibits a basal (002) reflection at 8.93° (d ≈ 9.89 Å), signifying expanded interlayer spacing and partial nanosheet exfoliation.⁵⁴ Higher-order reflections at 16.25° (004) and 18.40° (006) confirm layered stacking, while peaks at 35.20° (103), 41.16° (105), and 76.28° (110/112) demonstrate MXene's crystallinity and structural integrity. During the in situ CVD growth of CNTs on MXene, the nanocomposite displays a shifted basal (002) reflection at 7.44° (d ≈ 11.88 Å), slightly higher than the ∼7.2–7.3° reported by Li et al., indicating further nanosheet delamination and subtle interlayer expansion driven by strong MXene–CNT interactions. Distinct graphitic CNT (002) peaks at 25.37° and 26.51° confirm successful CNT formation and hybridization, consistent with recent reports.^55–58 The higher-order 35.20° (103) reflection shows significantly increased intensity compared to pristine MXene, reflecting improved crystallinity. In contrast, the 76.28° (110/112) reflection exhibits a slight decrease, indicating minor structural adjustments at the highest crystallographic planes. These results demonstrate that the MXene–CNT nanocomposite preserves MXene's fundamental crystalline structure while enabling modifications that can boost the electrical conductivity and open avenues for high-performance electrochemical applications.

The FTIR spectrum verified the surface functionalization of pristine MXene, MXene–CNT, and MXene–CNT/Ab surfaces (Fig. 2A). The spectra showed characteristic absorption bands for each surface. For pristine MXene, the peaks at 3300 cm⁻¹ (–OH stretching) and 1638 cm⁻¹ (C=O bending) indicated the presence of hydroxyl and water groups.⁵⁹ A strong peak at 592 cm⁻¹ was related to Ti–O deformation.⁶⁰ In the MXene–CNT spectra, new peaks indicated chemical bonds that enhance the functionality and stability, with peaks at 1074 cm⁻¹, 1455 cm⁻¹, and 2930 cm⁻¹ corresponding to C–O and C–H vibrations, and a peak at 2106 cm⁻¹ (C≡C) indicating alkyne groups.⁵⁰ The C=O peak at 1725 cm⁻¹ confirmed the carboxylic modification, which is absent in pristine MXene.⁶¹ After covalent antibody immobilization via EDC/NHS surface chemistry, distinct FTIR peaks at 3465 cm⁻¹, 1223 cm⁻¹, 1712 cm⁻¹, 1635 cm⁻¹, and 1561 cm⁻¹ arise due to N–H, C–O, C=O, amide I, and amide II groups, confirming the presence of TSH antibody.⁵⁹ These results verify MXene–CNT surface antibody immobilization, showing strong potential for biosensing applications.

Fig. 2 FTIR spectra of MXene, MXene–CNT, and MXene–CNT/Ab (A). Raman spectra of CNT, MXene, and MXene–CNT (B).

The Raman study provided evidence of MXene, MXene–CNT, and CNT structure and phase changes (Fig. 2B). The Raman spectrum from 100 to 2000 cm⁻¹ showed characteristic peaks of MXene: 156 cm⁻¹, 204 cm⁻¹, 396 cm⁻¹, 515 cm⁻¹, 624 cm⁻¹, 708 cm⁻¹, 1342 cm⁻¹, and 1564 cm⁻¹, correlating with MXene characteristics.⁶² The CNT grown on MXene and the nanocomposite exhibited similar peak patterns, compared to pure CNT's indistinct peaks, revealing the presence of Ti₃C₂T_x. The MXene–CNT and pure CNT had broad D band (1340 cm⁻¹) and G band (1590 cm⁻¹) peaks, indicating lattice defects and sp² hybridization.^50,63 The intensity ratios (I_D/I_G) from the Raman spectra were 0.69 for CNT, 0.49 for MXene, and 0.79 for MXene–CNT, indicating the crystalline carbon lattice structure with minimal defects and high graphitization. This affirms efficient MXene–CNT nanocomposite production, offering active binding sites that make it a promising platform for electrochemical applications.

The XPS characterization examined the surface chemical compositions and element valence states of CNT, MXene, MXene–CNT, and MXene–CNT/Ab. Fig. 3A displays Ti 2p, F 1s, C 1s, O 1s, and N 1s in the full XPS spectrum of CNT, MXene, and MXene modifications. The N 1s spectra show a strong peak at 400.74 eV, indicating that TSH antibodies bind to the MXene–CNT (Fig. 3A). The intensity of this peak decreased in the CNT, MXene, and MXene–CNT spectra (Fig. 3A). TSH antibodies activate via covalent surface chemistry and bind to carboxyl groups on MXene–CNT, ensuring successful attachment.⁶⁴ The XPS core-level spectra of MXene–CNT revealed an additional Ti peak and increased O intensity compared to CNT (Fig. 3A), confirming a strong interconnection between MXene and CNT.⁴³ The high-resolution C 1s spectra of MXene (Fig. 3B) revealed four deconvoluted peaks: Ti–C (284.07 eV), C=C (284.65 eV), C–OH (286.02 eV), and C=O (288.52 eV). The MXene C=C peak (284.65 eV) remained constant, though its intensity was lower than that on MXene–CNT. The peaks for Ti–C, C–OH, and C=O shift to 283.79 eV, 285.89 eV, and 287.35 eV,⁶¹ respectively; the intensity of Ti–C (283.79 eV) diminished for the MXene–CNT, indicating CNT surface growth on MXene. The XPS spectra of MXene–CNT and MXene exhibit additional Ti–C peaks (283.79 eV and 284.07 eV), with shifted intensity compared to CNT, validating the interaction between MXene and CNT.⁴³

Fig. 3 XPS survey spectra of MXene-based stepwise modified surfaces (A). The C 1s (B) and O 1s (C) spectra of CNT, MXene, and MXene–CNT modified surfaces. The Ti 2p spectra of MXene and MXene–CNT modified surfaces (D).

Fig. 3C illustrates high-resolution XPS peak fitting for MXene in the O 1s region, with peaks at 530.77 eV, 531.95 eV, and 533.27 eV associated with C–Ti–O, OH/Ox, and C–O.⁶⁵ The intensities of the C–Ti–O and C–O peaks are higher in MXene than in MXene–CNT; C–O displays greater intensity in MXene than in CNT. After the CVD process, the OH/Ox peak characteristic of MXene–CNT disappeared when compared to MXene. The O 1s spectrum indicates that MXene–CNT has a weaker C–Ti–O signal, with intensity shifting from 531.59 eV to 530.77 eV compared to MXene, with no C–Ti–O peak in CNT due to the consistent CVD growth of the MXene–CNT nanocomposite.

Fig. 3D presents the Ti 2p XPS analysis with four predominant peaks at 458.74 eV, 459.42 eV, 464.35 eV, and 465.32 eV attributed to Ti(II), Ti–O, Ti 2p_1/2, and Ti(III) in MXene–CNT and MXene. The MXene–CNT spectrum reveals two new peaks at 454.5 eV and 462.47 eV for Ti–C and Ti–O, with slight displacements compared to MXene. Additionally, peaks for Ti–O are observed at 459.46 eV and 462.47 eV.⁶⁶ The ensemble of Ti 2p and C 1s peaks of Ti–C indicates MXene's involvement. This finding illustrates that the CVD process produced uniform CNT growth on the MXene interlayer, resulting in robust interfacial bonding and an enriched nanocomposite, enhancing properties that facilitate electron migration at the interface.⁶⁷

Based on these analyses, the structural integrity, chemical stability, and biocompatibility of the MXene–CNT nanocomposite were ensured by HF-etched MXene, which served as the template for direct CNT growth via CVD. This method improved electrical conductivity and minimized oxidation by stabilizing the MXene sheets within the CNT network. After CNT growth, the composite was treated with HNO₃/H₂SO₄ to introduce surface groups (–OH, –COOH) for biomolecule immobilization, which was confirmed by FTIR analysis. The thorough washing (to neutral pH) and acid treatment eliminated residual HF or fluoride ions. The structural and chemical integrity were verified through XRD, XPS, and Raman analysis, demonstrating preserved MXene crystallinity, characteristic surface terminations (–O, –OH, –F), and no residual etchant traces. Control measurements showed consistent hormone sensing, indicating that oxidation and residual contaminants negatively impact the performance of the MXene–CNT nanocomposite.

3.2 Electrochemical properties of modified MXene interfaces

The conductive electrodes were characterized to evaluate changes in sensing interface properties using CV and EIS analyses. Fig. 4A illustrates that various modified electrodes were analyzed using CV in a 0.1 M KCl solution containing 5 mM [Fe(CN)₆]^3−/4− at a scan rate of 50 mV s⁻¹. The MXene–CNT modified SPGE showed a redox peak current of 58.51 μA, compared to 41.46 μA for the MXene-modified SPGE and 34.61 μA for the bare SPGE, attributed to its improved electrical conductivity and large surface area, which facilitates efficient electron transport. After continuous loading, the peak current decreased to 47.57 μA for anti-TSH and 35.05 μA for target TSH on the MXene–CNT modified SPGE, a result of limited electron migration caused by biomolecule attachment, which reduces conductivity and hinders charge transport at the interface. The CVD-grown MXene–CNT nanocomposite modified SPGE exhibited current responses 1.4-fold and 1.7-fold higher than the MXene and bare SPGE, respectively, due to its abundant active sites that enhance rapid electron transfer in the redox reaction with [Fe(CN)₆]^3−/4−. Each electrode's electrochemically active surface area was determined using the Randles–Ševčík equation,⁶⁸ as presented in eqn (1).

i_p = 2.69 × 10⁵n^3/2ACD^1/2ν^1/2 (1)

Fig. 4 CV (A) and EIS (B) responses of stepwise-modified surfaces: SPGE, MXene/SPGE, MXene–CNT/SPGE, BSA/anti-TSH/EDC–NHS/MXene–CNT/SPGE, and after TSH binding, measured in 5 mM [Fe(CN)₆]^3−/4− with 0.1 M KCl. CV scan rate: 50 mV s⁻¹; EIS frequency range: 0.1 Hz to 10 000 Hz at 5 mV amplitude. Bar chart of CV peak current and EIS R_ct value confirms successful surface assembly and TSH detection (C). CV of the MXene–CNT/SPGE at scan rates of 10–100 mV s⁻¹ (D). Linear plot of redox peak current variation vs. scan rate (E). The DPV response of the MXene–CNT sensor before and after TSH binding shows reduced current due to antibody–antigen interaction (F).

In this equation, i_p represents the peak current, n is the number of electrons exchanged, A denotes the electrode's active surface area, D is the diffusion coefficient of the redox species, C refers to the concentration of [Fe(CN)₆]^3−/4−, and ν is the scan rate. The electro-active surface areas were estimated to be 0.119 cm² for MXene–CNT/SPGE, 0.085 cm² for MXene/SPGE, and 0.071 cm² for bare SPGE, confirming the CVD growth of MXene–CNT as an effective immunosensor interface.

Fig. 4B shows that the EIS characterization of the modified electrodes was performed from 0.1 Hz to 10 000 Hz, with a 5 mV amplitude. The charge transfer resistance (R_ct) correlates with the semicircle diameter in the Nyquist plot, reflecting electron transfer kinetics at the interface.⁵¹ The MXene–CNT modified SPGE had an R_ct of 10.65 kΩ, lower than 12.25 kΩ for the MXene-modified SPGE and 13.71 kΩ for the bare SPGE, indicating rapid electron transfer and increased active sites. Following the assembly of anti-TSH (15.74 kΩ) and target TSH (18.95 kΩ) on the MXene–CNT modified SPGE, the semicircle diameters increased, suggesting higher resistivity for interfacial electron transfer. The R_ct values were significantly lower for the MXene–CNT nanocomposite compared to other SPGE types, highlighting its potential as a transducer surface.

The heterogeneous electron transfer rate constant (K_et) was calculated from the relationship between R_ct and K_et using eqn (2) for each modified electrode.⁶⁹

K_et was calculated using the geometrical electrode area and [Fe(CN)₆]^3−/4− concentration, yielding 2.64, 2.95, 3.40, 2.30, and 1.90 × 10⁻⁴ cm s⁻¹ for bare, MXene-, MXene–CNT-, BSA/anti-TSH/EDC–NHS/MXene–CNT-, and TSH/BSA/anti-TSH/EDC–NHS/MXene–CNT-modified SPGE, respectively. Fig. 4C further validates these findings with a bar chart quantifying changes in CV current and EIS R_ct during modification. An increasing R_ct and decreasing currents after the binding of anti-TSH and target TSH confirm effective sensor assembly and selective TSH recognition via antibody–antigen interaction. These findings across CV and EIS studies validate the successful construction of the TSH electrochemical immunosensor using the MXene–CNT nanocomposite.⁶⁹

Fig. 4D presents CV curves of the MXene–CNT nanocomposite recorded at scan rates from 10 to 100 mV s⁻¹ in 5 mM [Fe(CN)₆]^3−/4− with 0.1 M KCl. The increasing redox peak current with scan rate indicates enhanced electrochemical activity. A linear relationship between the peak current and scan rate (Fig. 4E), with i_pa = 0.5642x + 22.667 (R² = 0.981) and i_pc = −5073x − 24.8 (R² = 0.978), suggests an adsorption-controlled process. Peak shifts confirm surface-controlled electron transfer with minimal diffusion, indicating that the MXene–CNT surface is well-suited for antibody–antigen binding.

3.3 Electrochemical measurement of TSH using the established platform

The performance of the MXene–CNT-based immunosensor was evaluated using DPV with and without the target TSH (Fig. 4F). A significant change in peak current was observed with MXene–CNT (169 μA), without TSH (119 μA), and with TSH (75 μA). This decline results from the binding of the antibody–antigen complex, which obstructs electron transfer at the electrode surface, confirming successful TSH binding and highlighting the sensor's specificity for this hormone. In this mechanism, as TSH binds to anti-TSH antibodies on the MXene–CNT surface, it results in reduced electrode availability for redox reactions, leading to decreased current. This response can be quantitatively analyzed for sensitive TSH detection. The design of the electrochemical interface and MXene–CNT nanocomposite enhances sensor sensitivity and stability, allowing reliable measurements of various TSH concentrations and effective monitoring of TSH levels for diagnosing endocrine disorders.

3.4 Optimization of the nanocomposite-derived immunosensor

The experimental conditions are optimized for better electrochemical sensing performance. DPV analysis explored various parameters to enhance the sensitivity and selectivity of immunosensors by evaluating variables such as MXene–CNT concentration, TSH antibody levels, immobilization time, and antigen incubation time. MXene–CNT concentrations are crucial for developing immunosensors, as their electronic properties significantly improve the surface area of the screen-printed graphene electrodes. To assess the effect of MXene–CNT amounts on the TSH immunosensor's electrical signal, different concentrations (0.1 to 10 mg mL⁻¹) were tested (Fig. 5A). The peak δ current (ΔI) increased with MXene–CNT concentration; the signal response (ΔI) of the [Fe(CN)₆]^3−/4− redox probe peaked at 2.0 mg mL⁻¹ before declining with higher amounts. This trend suggests an optimally regulated MXene–CNT layer thickness on the electrode, which affects the redox probe's electron mobility. Thus, a concentration of 2.0 mg mL⁻¹ was chosen for the immunosensor to maximize sensitivity in TSH monitoring.

Fig. 5 DPV signals captured during the optimization of the MXene–CNT concentration (A), anti-TSH quantity (B), anti-TSH binding time (C), and TSH antigen incubation time (D).

Antibody concentration significantly impacts analyte detection. TSH antibody concentrations ranged from 1 to 20 μg mL⁻¹ and were loaded onto EDC–NHS/MXene–CNT/SPGE. The change in signal response (ΔI) was monitored at different TSH antibody concentrations before and after conjugating with 10 μg mL⁻¹ TSH. As shown in Fig. 5B, the signal change (ΔI) slightly increased from 1 to 10 μg mL⁻¹ and then decreased from 10 to 20 μg mL⁻¹ due to limited binding between the TSH antibody and the active site of the interface material. Therefore, we chose 10 μg mL⁻¹ as the optimal TSH antibody concentration for further studies.

An immobilization time is essential for creating a recognition surface to bind the target protein. A very short time results in limited sensing efficiency as the target protein fails to attach effectively. Conversely, extending the antibody immobilization period can improve analytical detection performance. It is vital to identify the optimal time for antibody attachment to validate screening capabilities. The effective recognition time was explored for 10 to 100 minutes, as shown in Fig. 5C. The signal response (ΔI) increased significantly from 10 to 60 minutes, then declined after 60 minutes of TSH antibody immobilization. The findings indicated that 60 minutes is optimal for antibody immobilization on EDC–NHS/MXene–CNT/SPGE to detect the TSH antigen, yielding the maximum current signal change for a quick evaluation.

The incubation time of the TSH antigen is a key variable influencing antibody–antigen binding effectiveness and the outcomes of the electrochemical immunosensors. To determine the optimal incubation time, the BSA/anti-TSH/EDC–NHS/MXene–CNT/SPGE was immersed in the TSH antigen for various intervals. Fig. 5D shows that as the incubation time ranged from 15 to 90 minutes, the change in value (ΔI) increased rapidly, peaking at 45 minutes. However, effective binding does not occur beyond 45 minutes due to hindered diffusion and transport of specific antibody binding sites to the antigen. Thus, 45 minutes was established as optimal for TSH antigen detection to ensure monitoring efficiency of real-world samples.

3.5 Immunosensor performances and calibration for TSH

The analytical performance of the TSH immunosensor was evaluated using DPV after electrochemical characterization. Fig. 6A illustrates the label-free electrochemical sensing mechanism of the MXene–CNT-based immunosensor for TSH detection. Electrochemical detection was conducted under optimized DPV conditions on a disposable, cost-effective antibody-functionalized surface. The sensor tested various TSH antigen concentrations in 0.1 M KCl with 5 mM [Fe(CN)₆]^3−/4−, and the ΔI responses were used to create a calibration curve (Fig. 6B and C). The immunosensor was incubated with the TSH antigen for 45 minutes to facilitate antibody–antigen interaction. This incubation formed an insulating barrier between antibodies and antigens, increasing the EIS of the electrode (Fig. 4B) and hindering electron transmission. Fig. 6B shows a DPV curve demonstrating a decrease in the electron transfer rate and current signal. Higher TSH antigen concentrations correlated with a decline in the DPV oxidation current response. The DPV signal exhibited a dynamic linear range of 0.1–10 000 pg mL⁻¹ (R² = 0.990) and a detection limit of 0.03 pg mL⁻¹. A calibration plot of ΔI against the logarithm of TSH concentration is shown in Fig. 6C. This indicates that the interface material is highly sensitive and can effectively monitor TSH in artificial serum samples. The developed immunosensor outperformed previously studied techniques, as shown in Table 1, and is advantageous for early TSH antigen diagnosis due to its low detection limit. These features enhance the analytical sensing capability, supported by a large effective surface area and robustness of MXene–CNT, which facilitates electron transport due to excellent conductivity. Additionally, the synergistic effect of the MXene–CNT nanocomposite provides abundant active sites for immobilizing antibodies, improving the effectiveness of immunosensor detection.

Fig. 6 Schematic of the label-free MXene–CNT-based electrochemical TSH sensing mechanism (A). DPV signal response of the immunosensor at various concentrations of TSH (0.1–10 000 pg mL⁻¹) (B). Calibration plot of the signal change (ΔI) vs. log concentration of TSH (C). The influence of interference on the immunosensor for 50 pg mL⁻¹ against other biomolecules (D). The stability study of the immunosensor over 35 days (E). The reproducibility test was conducted with seven immunosensors for TSH detection (F).

Table 1 Comparison of other analytical performances for TSH detection

Detection techniques	Sensor material	Linear range (pg mL^−1)	LOD (pg mL^−1)	Ref.
Electrochemical	Nanogold-ionic liquid carbon paste electrode	200.0–90 000	100.0	3
Electrochemical	Nanogold-functionalized magnetic beads	0.168–3360	0.84	5
ELISA	Microfluidics and thermal lens detection	16.8–1680	5.04	77
Chemiluminescence	Biotin–streptavidin	16.8–6400	1.68	78
Chemiluminescence	Polymer lab-on-a-chip	19.2–9240	319.2	79
ELISA
Electrochemical	Non-labelled immunosensor	117.6–588	5.712	80
Electrochemical	MXene–CNT/SPGE	0.10–10 000	0.03	This work

---

The developed immunosensor demonstrated ultrasensitive detection of TSH, with DPV measurements spanning a wide dynamic range (0.1–10 000 pg mL⁻¹). Fitting the data to the Langmuir isotherm yielded a dissociation constant (K_D ≈ 54 pg mL⁻¹, 1.8–1.9 pM), representing the point at which half of the antibody binding sites are occupied. This low picomolar K_D indicates strong antibody–antigen interactions and highlights the sensor's high sensitivity. Concurrently, the antibody surface coverage on the MXene–CNT modified screen-printed graphene electrode was estimated from the applied antibody solution (10 μg mL⁻¹, 5 μL) and the electrode area (0.07 cm²). Assuming full immobilization, the surface coverage was approximately 4.8 pmol cm⁻², consistent with the reported values for IgG on electrode surfaces.^70–72 The ideal antibody loading and low picomolar K_D demonstrate that the MXene–CNT interface supports strong binding interactions and enables reliable assessment of antibody–antigen affinity, significantly enhancing the ultrasensitive performance of the immunosensor, consistent with other advanced nanomaterial-based biosensors.^73,74

3.6 Selectivity, stability, reproducibility, and repeatability studies of the immunosensor

Confirming the performance of the developed immunosensor for selective TSH screening in the presence of interfering agents is a primary challenge, involving DPV analysis and various biological indicators such as troponin (Tr), albumin (Al), bilirubin (Bi), BSA, interleukin-6 (IL-6), and mixed compounds (Fig. 6D). The immunosensor exhibited a TSH signal response (ΔI) that was 100 times greater than that of other potential interferents, with negligible variations. The results indicate that the transducer surfaces demonstrate high specificity for the TSH protein, making them ideal for the selective detection of TSH in complex media.

The developed immunosensor, based on BSA/anti-TSH/EDC–NHS/MXene–CNT/SPGE, was tested for storage stability over 35 days, demonstrating excellent performance. As illustrated in Fig. 6E, the DPV signal was recorded on seven different days, maintaining current values at 99, 98, 89, 88, and 81% of the original at 7, 15, 21, 28, and 35 days, respectively (with %RSD below 5%). This impressive performance is attributed to the strong covalent linkage between the antibody and the nanocomposite surface. Utilizing EDC/NHS surface chemistry for covalent immobilization improved sensor stability by forming amide bonds between carboxyl and amine groups, thereby minimizing desorption. The CVD-grown MXene–CNT architecture provided a stable, 3D conductive interface for reliable TSH detection. The system exhibited advantageous diagnostic characteristics, including a broad dynamic range, low detection limit, and high sensitivity and selectivity for determining a key thyroid function biomarker. This framework is suitable for point-of-care testing due to its straightforward technique, involving a CVD-grown MXene–CNT nanocomposite and an accessible homemade SPGE.

The reproducibility and repeatability were systematically evaluated through multi-electrode, inter-day, and intra-day assessments. Multi-electrode tests with seven independent electrodes for each concentration at 100, 500, and 1000 pg mL⁻¹ yielded low RSDs (1.59–2.73%), with all t-calculated values (0.003) below the t-critical value (2.447), confirming excellent reproducibility (Fig. 6F and Table SI-1). Inter-day evaluation over six consecutive days involved one replicate per day, with the obtained average concentrations between 99.18 and 101.72 pg mL⁻¹, RSD = 1.07%, and a t-calculated value (0.250) below the t-critical value (2.571), indicating consistent and high performance (Table SI-2). Intra-day precision was performed on a single day with five replicates per measurement of 100 pg mL⁻¹ at two distinct times (Table SI-3). The mean recovery concentration ranges from 99.99 to 100.05 pg mL⁻¹, RSDs of 0.160–0.300%, and a t-calculated value of 0.150, which is below t-critical (2.776), confirming high precision. All Student's t-calculated values at 95% confidence level (α = 0.05) were below t-critical, indicating no significant difference between measured and nominal concentrations. These findings demonstrate the immunosensor's high accuracy, reproducibility, and robustness, making it suitable for reliable biosensing, consistent with earlier studies.^75,76

3.7 Monitoring in an artificial serum matrix

A real-time evaluation introduced the TSH biomarker into an artificial serum sample to assess the feasibility of a new biomolecular sensor for rapid TSH diagnosis. Artificial serum created a controlled environment that simulates the key properties of human serum, thereby avoiding variability associated with clinical samples. This is a standard step in early biosensor development before clinical validation. We measured TSH concentrations at 0.5, 1, and 3 pg mL⁻¹ in artificial serum, using a standard addition method to evaluate TSH levels in biological fluids. The serum sample was incubated on a BSA/anti-TSH/EDC–NHS/MXene–CNT/SPGE for 45 minutes at 4 °C and rinsed with PBS to remove excess protein. Recovery trials measured the immunosensor's accuracy, as shown in Table 2. The developed sensor offers high analytical reliability and was further evaluated through recovery, precision, and t-test results (Table SI-4). The %recovery values were within the acceptable range (98.3–108.0%), indicating good accuracy and reliable detection of the target analyte in spiked samples. A statistical comparison using the Student's t-test confirmed no significant differences between the spiked and detected concentrations at the 95% confidence level (α = 0.05), reinforcing the method's accuracy. Additionally, the low %RSD values (<5.2%) demonstrate excellent reproducibility, aligning with standard analytical performance criteria reported in the literature.^81,82 These results confirm that the MXene-based sensor provides accurate, precise, and reproducible measurements, making it suitable for practical biosensing applications.

Table 2 Determination of TSH in artificial serum analysis using the constructed immunosensor (n = 5)

Samples	Spiked (pg mL^−1)	Found (pg mL^−1)	Recovery (%)	RSD (%)
1	0.50	0.54	108.0	1.0
2	1.00	0.99	99.0	2.5
3	3.00	2.95	98.3	5.1

---

Real-time analysis confirmed the sensor's reliability and effectiveness for monitoring TSH levels in artificial serum. TSH concentrations were validated by spiking with known standard amounts and confirming accuracy using a commercial TSH (Catalog number: 4610-TH), with measured values consistent with expected concentrations. Analyzing real samples can pose challenges such as nonspecific binding and matrix interference, but the CVD-grown MXene–CNT nanocomposite improves surface conductivity and mitigates matrix effects by providing a high surface area, abundant electro-active sites, and antifouling properties. Similar effects have been reported for the solution-processed MXene–CNT system.⁴⁴ This engineered interface enhances antibody immobilization and minimizes signal instability, addressing common limitations in complex biological media. Further optimization, including controlled surface chemistry and mild sample pre-treatment, could enhance clinical performance.

4. Conclusions

This study develops a new electrochemical immunosensing framework for label-free, ultrasensitive detection of the thyroid function biomarker TSH using the MXene–CNT nanocomposite. The engineered nanocomposite was prepared with CVD techniques, integrating large CNT bundles that grew on catalyst-like MXene. These synergistic effects create more active sites, increase surface area, and enhance electrical conductivity, facilitating rapid electron transfer at an optimal interface through covalent surface chemistry, enabling the binding of anti-TSH to the MXene–CNT modified SPGE via an EDC/NHS system. Under optimal conditions, the electrochemical immunosensors exhibit a dynamic linear concentration range from 0.1–10 000 pg mL⁻¹, with a detection limit of 0.03 pg mL⁻¹ and greater selectivity by eliminating potential interfering compounds. The TSH immunosensor also demonstrates high selectivity, stability, and reproducibility. Importantly, this immunosensor offers a promising approach for early TSH diagnosis and effectively measures TSH in artificial serum samples. Furthermore, the immunosensors are user-friendly, rapid, and cost-effective, suitable for integration with wearable and flexible electronic sensors for TSH monitoring in artificial and human blood serum.

Conflicts of interest

The authors declare no conflicts of interest.

Data availability

All relevant data generated or analyzed during this study are included in this published article. Additional information is available from the corresponding author upon reasonable request.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d5tb01632k.

Acknowledgements

This research was supported by the Second Century Fund (C2F), Chulalongkorn University. Vijayaraj Kathiresan gratefully acknowledges the C2F for providing a postdoctoral fellowship.

References

1. O. S. Kwon, S. R. Ahn, S. J. Park, H. S. Song, S. H. Lee, J. S. Lee, J.-Y. Hong, J. S. Lee, S. A. You, H. Yoon, T. H. Park and J. Jang, ACS Nano, 2012, 6, 5549–5558.
2. A. S. Alzahrani, M. Al Mourad, K. Hafez, A. M. Almaghamsy, F. A. Alamri, N. R. Al Juhani, A. S. Alhazmi, M. Y. Saeedi, S. Alsefri, M. D. A. Alzahrani, N. Al Ali, W. I. Hussein, M. Ismail, A. Adel, H. El Bahtimy and E. Abdelhamid, Adv. Ther., 2020, 37, 3097–3111.
3. H. Beitollahi, S. G. Ivari and M. Torkzadeh-Mahani, Biosens. Bioelectron., 2018, 110, 97–102.
4. Y. Liu, Q. Zhang, H. Wang, Y. Yuan, Y. Chai and R. Yuan, Biosens. Bioelectron., 2015, 71, 164–170.
5. B. Zhang, D. Tang, B. Liu, Y. Cui, H. Chen and G. Chen, Anal. Chim. Acta, 2012, 711, 17–23.
6. D. S. Cooper and B. Biondi, Lancet, 2012, 379, 1142–1154.
7. M. I. Surks, E. Ortiz, G. H. Daniels, C. T. Sawin, N. F. Col, R. H. Cobin, J. A. Franklyn, J. M. Hershman, K. D. Burman, M. A. Denke, C. Gorman, R. S. Cooper and N. J. Weissman, JAMA, 2004, 291, 228–238.
8. Rajesh, K. Kumar, S. K. Mishra, P. Dwivedi and G. Sumana, TrAC, Trends Anal. Chem., 2019, 118, 666–676.
9. K. Okuzawa, Y. Kazeto, S. Uji, T. Yamaguchi, H. Tanaka, M. Nyuji and K. Gen, Gen. Comp. Endocrinol., 2016, 239, 4–12.
10. C. Shim, R. Chong and J. H. Lee, Talanta, 2017, 171, 229–235.
11. R. Wang, Z.-P. Jia, X.-L. Hu, L.-T. Xu, Y.-M. Li and L.-R. Chen, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2003, 785, 353–359.
12. T. Kunisue, J. W. Fisher and K. Kannan, Anal. Chem., 2011, 83, 417–424.
13. F. Sassa, G. C. Biswas and H. Suzuki, Lab Chip, 2020, 20, 1358–1389.
14. S. E. Zahra Jawad, D. Hussain, M. Najam-ul-Haq and B. Fatima, TrAC, Trends Anal. Chem., 2024, 181, 117993.
15. F. S. Felix and L. Angnes, Biosens. Bioelectron., 2018, 102, 470–478.
16. S. C. Gopinath, T. H. Tang, M. Citartan, Y. Chen and T. Lakshmipriya, Biosens. Bioelectron., 2014, 57, 292–302.
17. N. Razmi and M. Hasanzadeh, TrAC, Trends Anal. Chem., 2018, 108, 1–12.
18. A. Tripathi and J. Bonilla-Cruz, ACS App. Nano Mater., 2023, 6, 5042–5074.
19. H. Bhardwaj, Archana, A. Noumani, J. K. Himanshu, S. Chakravorty and P. R. Solanki, Mater. Adv., 2024, 5, 475–503.
20. L. Guo, Y. Zhao, Q. Huang, J. Huang, Y. Tao, J. Chen, H. Y. Li and H. Liu, Microsyst. Nanoeng., 2024, 10, 65.
21. M. Hasanzadeh, N. Shadjou, E. Omidinia, M. Eskandani and M. de la Guardia, TrAC, Trends Anal. Chem., 2013, 45, 93–106.
22. R. Han, Y. Li, M. Chen, W. Li, C. Ding and X. Luo, Anal. Chem., 2022, 94, 2204–2211.
23. K. Vijayaraj, J. H. Lee, H. S. Kim and S.-C. Chang, J. Food Hyg. Saf., 2017, 32, 83–88.
24. S. Er Zeybekler and D. Odaci, ACS Omega, 2023, 8, 5776–5786.
25. L. S. Wang, Y. J. Guo, Y. H. Li, Y. S. Zhao, Q. Wei and Z. F. Gao, Anal. Methods, 2024, 16, 6802–6809.
26. P. Yaiwong, J. Jakmunee, D. Pimalai, K. Ounnunkad and S. Bamrungsap, Colloids Surf., B, 2024, 243, 114124.
27. M. Naguib, M. Kurtoglu, V. Presser, J. Lu, J. Niu, M. Heon, L. Hultman, Y. Gogotsi and M. W. Barsoum, Adv. Mater., 2011, 23, 4248–4253.
28. C. C. Hsiao, J. Kasten, D. Johnson, B. Ngozichukwu, R. M. S. Yoo, S. Lee, A. Erdemir and A. Djire, ACS Nano, 2024, 18, 7180–7191.
29. B. Anasori and Y. Gogotsi, 2D Metal Carbides and Nitrides (MXenes), Springer, 2019, vol. XVII, p. 534.
30. K. Khan, A. K. Tareen, M. Iqbal, Z. Ye, Z. Xie, A. Mahmood, N. Mahmood and H. Zhang, Small, 2023, 19, 2206147.
31. D. Khorsandi, J.-W. Yang, Z. Ülker, K. Bayraktaroğlu, A. Zarepour, S. Iravani and A. Khosravi, Microchem. J., 2024, 197, 109874.
32. Sweety and D. Kumar, J. Colloid Interface Sci., 2023, 652, 549–556.
33. Y. Sun, P. Li, Y. Zhu, X. Zhu, Y. Zhang, M. Liu and Y. Liu, Biosens. Bioelectron., 2021, 194, 113600.
34. S. M. Khoshfetrat, J. Electroanal. Chem., 2025, 992, 119215.
35. S. M. Khoshfetrat, M. Motahari and S. Mirsian, Sci. Rep., 2025, 15, 6804.
36. S. M. Khoshfetrat, M. Nabavi, S. Mamivand, Z. Wang, Z. Wang and M. Hosseini, Mikrochim. Acta, 2025, 192, 113.
37. F. Momeni, S. M. Khoshfetrat, H. Bagheri and K. Zarei, Biosens. Bioelectron., 2024, 250, 116078.
38. A. Lipatov, S. Bagheri and A. Sinitskii, ACS Mater. Lett., 2023, 6, 298–307.
39. F. Mokhtari, R. W. Symes, Ž. Simon, B. Dharmasiri, L. C. Henderson, M. W. Joosten and R. J. Varley, J. Mater. Chem. A, 2025, 13, 6482–6492.
40. C. T. Pan, K. Dutt, A. Kumar, R. Kumar, C. H. Chuang, Y. T. Lo, Z. H. Wen, C. S. Wang and S. W. Kuo, Int. J. Bioprint., 2023, 9, 647.
41. B. Muthukutty, P. S. Kumar, A. K. Vivekanandan, M. Sivakumar, S. Lee and D. Lee, Chemosphere, 2024, 355, 141838.
42. A. Aris, W. T. Wahyuni, B. R. Putra, A. Hermawan, F. A. A. Nugroho, Z. W. Seh and M. Khalil, Nanoscale, 2025, 17, 2554–2566.
43. L. Lei, J. Zhang, W. Zhang, J. Hao and K. Wu, Biosens. Bioelectron., 2024, 263, 116609.
44. L. Tian, M. Jiang, M. Su, X. Cao, Q. Jiang, Q. Liu and C. Yu, Microchem. J., 2023, 185, 108172.
45. M. Yang, L. Wang, C. Xie, H. Lu, J. Wang, Y. Li, H. Li, J. Yang, T. Zhang and S. Liu, Talanta, 2025, 281, 126893.
46. B. Dharmasiri, K. A. S. Usman, S. A. Qin, J. M. Razal, N. T. Tran, P. Coia, T. Harte and L. C. Henderson, Chem. Eng. J., 2023, 476, 146739.
47. M. Gao, Y. Xie, W. Yang and L. Lu, Int. J. Electrochem. Sci., 2020, 15, 4619–4630.
48. S. Rajaie, M. Nasiri, A. Pasdar, M. Rezayi, M. Khazaei and B. Hatamluyi, Microchem. J., 2023, 193, 109016.
49. G. Zhang, H. Zhang, Y. Fu, P. Xia, C. Chen, Y. Wei, S. Qu and S. Feng, Mikrochim. Acta, 2024, 191, 626.
50. V. Thirumal, R. Yuvakkumar, P. S. Kumar, G. Ravi, S. P. Keerthana and D. Velauthapillai, Chemosphere, 2022, 286, 131733.
51. V. Kathiresan, D. Thirumalai, T. Rajarathinam, M. Yeom, J. Lee, S. Kim, J.-H. Yoon and S.-C. Chang, J. Anal. Sci. Technol., 2021, 12, 5.
52. S. Woo, Y.-R. Kim, T. D. Chung, Y. Piao and H. Kim, Electrochim. Acta, 2012, 59, 509–514.
53. P. Yu, G. Cao, S. Yi, X. Zhang, C. Li, X. Sun, K. Wang and Y. Ma, Nanoscale, 2018, 10, 5906–5913.
54. Y. Zhao, J. Wang, D. Yang, Z. Du, X. Zhi, R. Yu, Z. Guo, C. Tang and Y. Fang, Carbon, 2025, 232, 119825.
55. S. Adomaviciute-Grabusove, A. Popov, S. Ramanavicius, V. Sablinskas, K. Shevchuk, O. Gogotsi, I. Baginskiy, Y. Gogotsi and A. Ramanavicius, ACS Nano, 2024, 18, 13184–13195.
56. J.-J. Zhu, A. Hemesh, J. J. Biendicho, L. Martinez-Soria, D. Rueda-Garcia, J. R. Morante, B. Ballesteros and P. Gomez-Romero, J. Colloid Interface Sci., 2022, 623, 947–961.
57. Y. Yue, Y. Wang, X. Xu, C. Wang, Z. Yao and D. Liu, Ceram. Int., 2022, 48, 6338–6346.
58. X. Li, X. Yin, M. Han, C. Song, H. Xu, Z. Hou, L. Zhang and L. Cheng, J. Mater. Chem. C, 2017, 5, 4068–4074.
59. C. G. Sanjayan, L. Gokavi, C. H. Ravikumar and R. G. Balarkishna, Biosens. Bioelectron., 2025, 271, 117028.
60. Z. Yang and J. Wang, Langmuir, 2023, 39, 4179–4189.
61. Y.-S. Kim, J.-H. Cho, S. G. Ansari, H.-I. Kim, M. A. Dar, H.-K. Seo, G.-S. Kim, D.-S. Lee, G. Khang and H.-S. Shin, Synth. Met., 2006, 156, 938–943.
62. A. Sarycheva and Y. Gogotsi, Chem. Mater., 2020, 32, 3480–3488.
63. R. A. DiLeo, B. J. Landi and R. P. Raffaelle, J. Appl. Phys., 2007, 101, 064307.
64. K. Vijayaraj, S. W. Hong, S.-H. Jin, S.-C. Chang and D.-S. Park, Anal. Methods, 2016, 8, 6974–6981.
65. M. Han, X. Yin, H. Wu, Z. Hou, C. Song, X. Li, L. Zhang and L. Cheng, ACS Appl. Mater. Interfaces, 2016, 8, 21011–21019.
66. S. Choudhury, S. Zafar, D. Deepak, A. Panghal, B. Lochab and S. S. Roy, J. Mater. Chem. B, 2024, 13, 274–287.
67. W. Ding, P. Liu, Z. Bai, Y. Wang, G. Liu, Q. Jiang, F. Jiang, P. Liu, C. Liu and J. Xu, Adv. Mater. Interfaces, 2020, 7, 2001340.
68. V. Kathiresan, T. Rajarathinam, S. Lee, S. Kim, J. Lee, D. Thirumalai and S. C. Chang, Sensors, 2020, 20, 6083.
69. S. Jampasa, P. Lae-Ngee, K. Patarakul, N. Ngamrojanavanich, O. Chailapakul and N. Rodthongkum, Biosens. Bioelectron., 2019, 142, 111539.
70. S. Mehdi Khoshfetrat and M. A. Mehrgardi, ChemElectroChem, 2014, 1, 779–786.
71. S. M. Khoshfetrat, P. Seyed Dorraji, M. Shayan, F. Khatami and K. Omidfar, Anal. Chem., 2022, 94, 8005–8013.
72. S. M. Khoshfetrat, H. Khoshsafar, A. Afkhami, M. A. Mehrgardi and H. Bagheri, Anal. Chem., 2019, 91, 6383–6390.
73. S. M. Khoshfetrat, P. S. Dorraji, L. Fotouhi, M. Hosseini, F. Khatami, H. R. Moazami and K. Omidfar, Sens. Actuators, B, 2022, 364, 131895.
74. R. El-Shaheny, L. A. Al-Khateeb and M. El-Maghrabey, Sens. Actuators, B, 2021, 348, 130658.
75. B. Wang, S. M. Khoshfetrat and H. Mohamadimanesh, Microchem. J., 2024, 207, 111796.
76. S. Mehdi Khoshfetrat, M. Moradi, H. Zhaleh and M. Hosseini, Microchem. J., 2024, 205, 111382.
77. T. Ohashi, O. Fukahori, H. Tazawa, A. Harano, T. Ebata, K. Mawatari and T. Kitamori, 14th International Conference on Miniaturized Systems for Chemistry and Life Sciences (MicroTAS 2010), 2010, vol. 2, pp. 821–823.
78. Z. Lin, X. Wang, Z. J. Li, S. Q. Ren, G. N. Chen, X. T. Ying and J. M. Lin, Talanta, 2008, 75, 965–972.
79. W. Jung, J. Han, J. Kai, J. Y. Lim, D. Sul and C. H. Ahn, Lab Chip, 2013, 13, 4653–4662.
80. E. Asav, Turk. J. Chem., 2021, 45, 819–834.
81. S. M. Khoshfetrat, S. Mamivand and G. B. Darband, Mikrochim. Acta, 2024, 191, 546.
82. S. M. Khoshfetrat, K. Fasihi, F. Moradnia, H. Kamil Zaidan and E. Sanchooli, Anal. Chim. Acta, 2023, 1252, 341073.

---