Heterostructures of CdSe Quantum Dots and g-C₃N₄ Applied as Electrochemiluminescent Probes for the Detection of Hydrogen Peroxide in Human Serum

Abstract

In this work, we developed a highly sensitive and reproducible electrochemiluminescent sensor based on a heterostructure of cadmium selenide quantum dots capped with 3-mercaptopropionic acid (MPA) + 3-morpholinoethanesulfonic acid (MES) (QDs CdSe) and carbon nitride nanosheets (g-C₃N₄) for the detection of H₂O₂ in lyophilized serum samples. To enhance the sensor sensitivity, g-C₃N₄ nanosheets were utilized as a platform to immobilize the QDs CdSe. An exhaustive characterization of the heterostructure was conducted, elucidating the interaction mechanism between QDs CdSe and g-C₃N₄. It was revealed that g-C₃N₄ acts as a hole (h⁺) donor, while QDs CdSe act as energy acceptors in a resonance energy transfer process, with the electrochemiluminescence emission originating from the QDs CdSe. The electrochemiluminescence intensity decreases in the presence of H₂O₂ due to the deactivation of the excited states of the QDs CdSe. This electrochemiluminescent sensor demonstrates exceptional performance for detecting H₂O₂ in aqueous systems, achieving a remarkably low limit of detection (LOD) of 1.81 nM, which is more sensitive than most reported sensors to detect H₂O₂. The applicability of the sensor was successfully tested where sub-µM levels of H₂O₂ were accurately quantified. These results highlight the potential of this electrochemiluminescent sensor as a reliable and pre-treatment-free tool for H₂O₂ detection in biochemical studies and human health applications.

1. Introduction

Hydrogen peroxide (H₂O₂) is a compound that remains a topic of interest across a wide variety of scientific disciplines. H₂O₂ is a critical biomarker for multiple pathological conditions, including cardiovascular diseases, cancer, diabetes, Parkinson’s disease, and neurodegenerative disorders [1]. Its detection and quantification are therefore essential for diagnostic and therapeutic applications in human health. At the molecular level, H₂O₂ decomposes into hydroxyl radicals (OH·), highly reactive species that induce oxidative damage to cellular components, leading to cell death. When present at supraphysiological concentrations, H₂O₂ drives excessive generation of reactive oxygen species (ROS), resulting in systemic oxidative stress. This stress exhibits a dual role in cellular pathophysiology: while it can trigger apoptosis (a regulated cell death process), it paradoxically promotes disease progression through mechanisms such as aberrant cell proliferation, angiogenesis, and metastasis [2]. The concentration-dependent effects of H₂O₂ underscore its complex role as both a signalling molecule and a cytotoxic agent in human health and disease. Therefore, the development of new methods for detecting H₂O₂ at both micromolar (µM)- and nanomolar (nM)-level concentrations in human serum sample is of considerable importance. Consequently, a wide variety of methods for the detection of H₂O₂ are currently being developed [3]. The majority of methodologies developed for the detection of H₂O₂ are based on the use of enzymes, such as horseradish peroxidase [4]; nevertheless, their enzymatic activity is dependent upon environmental factors, including the medium (pH, ionic strength, solvent polarity, etc.), the temperature and the concentration of the enzyme substrate. Today, new enzyme-free methodologies are currently being developed with the goal of replacing biosensors for the quantification of H₂O₂, where the most prevalent of these are based on colorimetric sensors [5,6], as well as electrochemical sensors [7,8]. In recent years, the use of systems based on nanostructured materials, including semiconductors [9,10,11,12] and quantum dots [13,14,15], has facilitated the advancement of novel sensor technologies for the detection and quantification of H₂O₂. In this context, the use of nanomaterials with well-defined optical and electrochemical properties has enabled the development of electrochemiluminescent sensors. Electrochemiluminescence (ECL) has gained interest in diverse fields of analytical chemistry. The key advantage of ECL lies in its synergistic integration of electrochemical and photoluminescent methodologies. The ECL mechanism generates excited states through electrochemically produced radical intermediates, eliminating the requirement for external optical excitation. This unique feature (i) minimizes background interference from scattered light, (ii) enhances signal-to-noise ratios, and (iii) provides superior spatial control compared to conventional luminescence techniques. Also, ECL includes low background emission, high temporal resolution and high sensitivity [16].

The introduction of probes composed of nanomaterials with electrochemiluminescent properties has emerged as a highly valuable alternative for H₂O₂ detection. This is due to the ability to tune the optical and electrochemical properties of these probes during their synthesis, enabling optimized performance in sensing applications [17].

Graphitic carbon nitride (g-C₃N₄) is a metal-free semiconductor polymeric material with a two-dimensional structure [18]. It exhibits high chemical stability, biocompatibility and structural defects that facilitate functionalization. These properties make g-C₃N₄ an ideal material for use in the development of sensors and biosensors [19,20]. Due to its appropriate bandgap (E_g), between 2.6 and 2.8 eV, g-C₃N₄ can be used as a platform for electrochemiluminescent sensors in the presence of co-reactants [21,22,23,24,25]. The planar structure of g-C₃N₄ allows for the loading of nanoparticles, facilitating the formation of stable heterostructures, with both photocatalytic and electrocatalytic properties. These characteristics make g-C₃N₄ an effective and versatile probe for a wide variety of applications [26,27]. With respect to the semiconductor material, one of the main challenges is the limited ability to separate generated electron–hole pairs (excitons) when they are excited electrochemically and/or by light, where a rapid recombination between electrons and holes is produced due to the electrostatic attraction between them, decreasing the material’s lifetime and, therefore, the efficiency of the system [28,29]. In the case of g-C₃N₄, one effective strategy to overcome this limitation is the formation of heterojunctions by coupling g-C₃N₄ with nanoparticles such as quantum dots of semiconducting metals [30]. These heterostructures are presented as excellent electrochemiluminescent probes. In some cases, there is often a coupling between the valence and conduction bands of g-C₃N₄ and nanomaterials, characterized by the overlap of the emission band of g-C₃N₄ with the absorption band of the quantum dots. In this case, if the distance between them is very close (<10 nm), it is possible for a resonance energy transfer (RET) to be successfully established, enhancing the electrogenerated light emission processes [31,32,33,34,35,36]. This strategy allows the synthesis of hybrid nanomaterials, which can be employed in the development of more sensitive sensors based on ECL. In this respect, QDs CdSe are a valuable alternative to coupling with g-C₃N₄ due to their fluorescent and electrochemical properties [37]. QDs CdSe also exhibit a high rate of recombination of photogenerated electrons and holes. Nevertheless, heterostructures comprising g-C₃N₄ and QDs CdSe have the potential to generate new charge separation dynamics. The incorporation of an electric field at the heterojunction interface facilitates the spatial separation of hole–electron pairs, thereby reducing their recombination rate and enhancing the efficiency of the ECL [38]. One of the great advantages of using QDs CdSe is that they can be obtained in a simple way, with a controlled size [39]. The E_g can be adjusted and their optical properties are tunable according to the need for their application. Furthermore, the appropriate selection of capping molecules for QDs CdSe provides protection and enhances their optical properties, facilitating the ECL process, as previously reported [39].

In this work, we report the development of a new electrochemiluminescent sensor for the detection and quantification of H₂O₂ in lyophilized serum samples, without any pre-treatment, based on a heterostructure formed by g-C₃N₄ and QDs CdSe (QDs CdSe@g-C₃N₄). The electrochemiluminescent sensor demonstrated a high level of sensitivity and selectivity for detecting H₂O₂ in reconstituted serum samples.

2. Materials and Methods

2.1. Reagents and Materials

All reagents used in this work were of analytical grade and were used as received. For more details, they are described in Supplementary Materials (Section S1).

2.2. Instrumentation

Absorption spectra were obtained by using a Hewlett-Packard 8452A diode array spectrophotometer (Agilent, LA, USA). Fluorescence emission measurements were performed by using a FluoromaxTM spectrofluorometer (Spex Industries Inc., Edison, NY, USA) and a FluoroMax4^® (Horiba Scientific, Kyoto, Japan). The Fourier transform infrared (FT-IR) spectra were obtained using a Bruker Tensor 27 FT-IR (Bruker Optics, Ettlingen, Germany). The FT-IR spectra were acquired by transmission on a KBr cell containing a drop of the solution, where the solvent was previously evaporated by using a sodium lamp. The analysis of the luminescence decay of the nanomaterials was performed with an Edinburgh Instruments OB 900 Time Correlated Single-Photon Counting (TCSPC) fluorometer (Edinburgh Instruments Ltd., Livingston, Scotland).

Electrochemical measurements were performed using a PS4 No. 10 potentiostat (PalmSens, Houten, The Netherlands), coupled to a PC with incorporated software (PSTrace 5.9, Houten, The Netherlands). The electrochemiluminescence measurements were conducted by coupling the spectrofluorometer with the potentiostat as described in the Supplementary Materials (Section S2. Equipment for ECL Measurement). The working electrode used was an unpolished float glass, 7 × 50 × 0.7 mm, SiO₂ passivate/indium tin oxide-coated, with a resistance of 8–12 Ω. Prior to use, the electrodes were cleaned in an ultrasonic bath with 99.8% ethanol, HPLC-grade (Sintorgan, Provincia de Buenos Aires, Argentina), then dried in an oven at 40 °C. A platinum wire and a Ag/AgCl reference electrode (3 mol L⁻¹ NaCl) (CHI Instruments, Austin, TX, USA) were used as auxiliary (CE) and reference (RE) electrodes, respectively.

2.3. Synthesis of g-C₃N₄

Synthesis of bulk g-C₃N₄ was carried out in accordance with well-established literature protocols [40]. Briefly, 7.5 g of urea was placed in a covered porcelain chamber and calcined in an oven. A temperature ramp was applied at a rate of increase of 10 °C min⁻¹, reaching a maximum of 550 °C, and this temperature was maintained constant for 2 h. Then, the oven was switched off and allowed to cool to room temperature. In this process, the structural units of tri-S-triazine/heptazine heterocyclic compounds were formed [41].

Then, to separate the g-C₃N₄ sheets, 10 mL of H₂SO₄ concentrate was added to 500 mg of bulk g-C₃N₄ and mixed at 1500 rpm for 60 min [42]. Next, 20 mL of H₂O was added and sonicated for 2 h at 40 °C, and the solution changed colour from yellow to white. This process resulted in the formation of smaller and separated sheets (see Section 3). The purification of exfoliated g-C₃N₄ was carried out through centrifugation and water washing steps until a pH-neutral solution of g-C₃N₄ was achieved. Then, the g-C₃N₄ was dried in an oven at 60 °C for 12 h, resulting in exfoliated g-C₃N₄ powder.

2.4. Synthesis of Heterostructures of g-C₃N₄ and QDs CdSe

The synthesis of heterostructures of QDs CdSe and g-C₃N₄ (QDs CdSe@g-C₃N₄) was carried out by an in situ growth process of QDs CdSe on g-C₃N₄ sheets, as illustrated in Scheme 1. To 5 mL of an aqueous Cd²⁺ solution (7 mmol L⁻¹), MPA and MES were added simultaneously as stabilizers, maintaining a Cd²⁺/stabilizer molar ratio of 1:10. The pH was adjusted to 10 using NaOH, and the solution was degassed for 30 min under argon bubbling with constant stirring at 300 rpm. The dual role of MPA and MES was to stabilize the quantum dots and control their growth. Next, 200 µL of 2.5 mg mL⁻¹ of aqueous dispersion of exfoliated g-C₃N₄ was added. Subsequently, a Se²⁻ aqueous solution was rapidly introduced, producing a change of colour to yellow/orange. As previously reported, to maximize the electrochemiluminescence of QDs CdSe, a molar ratio of Cd²⁺ to Se²⁻ of 2:1 was used. The MPA:MES molar ratio (30:70) was selected based on our prior work, which demonstrated its superior enhancement of ECL intensity [39]. The product obtained was purified through several steps of precipitation by changing the polarity of the solvent (using ethanol), centrifugation and washing. Finally, QDs CdSe@g-C₃N₄ were dispersed in water.

3. Results

3.1. Spectroscopic Characterization of QDs CdSe@g-C₃N₄

FTIR was performed on g-C₃N₄, QDs CdSe, and QDs CdSe@g-C₃N₄, as shown in Figure 1. Exfoliated g-C₃N₄ (green line) shows typical FTIR spectra as reported in the literature, with a peak at 809 cm⁻¹ due to the presence of triazine rings. Bands at 1572 cm⁻¹ and 1632 cm⁻¹ were attributed to C=N stretching, while the bands at 1235 cm⁻¹, 1327 cm⁻¹, and 1405 cm⁻¹ corresponded to aromatic C-N stretching [43]. Peaks between 3000 cm⁻¹ and 3600 cm⁻¹ were due to N-H groups and adsorbed O-H groups, respectively. The FTIR of QDs CdSe stabilized with MPA:MES (orange line) showed peaks in the range of 770 cm⁻¹ to 796 cm⁻¹ attributed to the stretching vibrational mode of the Cd-Se bond [44], and peaks at 1492 cm⁻¹ and 1525 cm⁻¹ corresponded to C=O and C-O, respectively, due to MPA [39]. In the case of QDs CdSe@g-C₃N₄ (black line), the characteristic bands of g-C₃N₄ were smaller in comparison to bands of QDs CdSe. However, peaks located between 1572 cm⁻¹ and 1632 cm⁻¹ were attributed to the stretching of the C=N bonds. In addition, principal peaks of QDs CdSe were also observed, demonstrating the presence of both nanomaterials.

The morphology of g-C₃N₄ and QDs CdSe@g-C₃N₄ was analyzed by TEM. Figure 2A corresponds to bulk g-C₃N₄, in which overlapping sheets, characteristic of bulk graphitic nitride structures, were clearly observable. Figure 2B,C correspond to the heterostructure of QDs CdSe@g-C₃N₄, in which the QDs CdSe are uniformly distributed on sheets of g-C₃N₄. These findings are consistent with the growing of quantum dots on the surface of the g-C₃N₄ sheet during the in situ synthesis of the heterostructure.

Figure 3a shows absorption spectra of g-C₃N₄, QDs CdSe, and QDs CdSe@g-C₃N₄. The absorption spectrum of exfoliated g-C₃N₄ (magenta line) shows a broad band with two peaks centred at 218 nm and 313 nm, respectively. The first peak was attributed to a π → π* transition of the C-N double bond, and this peak did not appear in a solution of the bulk nitride. However, it is sharply defined when the nitride sheets are separated, e.g., by exfoliation. The second absorption peak corresponds to a transition of n → π* [45].

Absorption spectra of QDs CdSe and QDs CdSe@g-C₃N₄ (black and green lines) shown a band close to 435 nm, due to the electronic transition between valence and conduction bands of nanocrystals. However, in the heterostructure, the absorption band of g-C₃N₄ was not observed due to its low concentration compared to the concentration of the QDs CdSe.

On the other hand, when solutions were excited to 320 nm, fluorescence emission spectra were obtained (Figure 3b). The spectra of g-C₃N₄ (magenta line) presented a symmetric band centred at 414 nm. The fluorescence emission spectrum of QDs CdSe showed a band centred around 598 nm (black line). For QDs CdSe@g-C₃N₄, two bands centred at 414 nm and 598 nm were observed, corresponding to g-C₃N₄ and QDs CdSe (green line). However, a decrease of 94% in the intensity of fluorescence was observed for the g-C₃N₄ band, and an increase of 54% corresponding to QDs CdSe band was observed. These effects are due to RET, where the g-C₃N₄ acts as a donor and QDs CdSe act as acceptors of energy. This result is consistent with an overlap between the emission band of g-C₃N₄ and the absorption band of QDs CdSe (see Figure S2) [46]. The presence of g-C₃N₄ in the conjugated heterostructure facilitates the effective movement of charge carriers towards the QDs CdSe, thereby enhancing the emission of QDs CdSe (fluorescence or electrochemiluminescence). Time-resolved fluorescence decay spectra were gathered to investigate the charge carrier transfer dynamics (see Figure S2 in Supplementary Materials). Each sample showed two distinct radiative lifetimes (τ), whose relative percentages were calculated by fitting the decay spectra by means of a bi-exponential equation (see Supplementary Materials). Results are shown in

Table 1. Kinetic parameters of emission decay for g-C₃N₄, QDs CdSe, and QDs CdSe@g-C₃N₄.

Sample	Lifetime/ns	% Population	Lifetime/ns	% Population
	τ1	β1	τ2	β2
g-C3N4	0.41	12.62	2.84	87.38
QD CdSe	1.61	10.72	13.40	89.27
QD CdSe@g-C3N4	1.39	8.98	14.07	91.02

. The lifetimes of the conjugated heterostructure were similar to those of the unconjugated QDs CdSe, but a slight increase in τ₂ was observed for the heterostructure, suggesting that g-C₃N₄ may reduce non-radiative recombination.

When the synthesis of the heterostructure is carried out using twofold concentrations of the precursor of QDs CdSe, the fluorescence of the heterostructure at 598 nm decreases sharply. This effect is due to agglomerations of QDs CdSe on g-C₃N₄ sheets that produce non-radiative deactivations of the exited state of quantum dots.

Bandgap energies (Eg) of g-C₃N₄, QDs CdSe and QDs CdSe@g-C₃N₄ were determined by analyzing the absorption edge in the UV–visible spectra using the Tauc model [47] and are shown in

Table 2. Eg determined by UV–visible spectra and ΔE_p determined by cyclic voltammetry for g-C₃N₄, QDs CdSe, and QDs CdSe@g-C₃N₄.

Sample	aEg/eV	Ea,p/V vs. HNE	Ec,p/V vs. HNE	ΔEp/V
g-C3N4	2.55	0.966	−1.504	2.47
QD CdSe	2.23	0.594	−1.724	2.32
QD CdSe@g-C3N4	2.28	0.630	−1.704	2.34

^aEg determined by UV–visible spectroscopy.

. The detailed procedure for determining the Eg is described in the Supplementary Materials (Section S4). The Eg of g-C₃N₄ was higher than those obtained for QDs CdSe and QDs CdSe@g-C₃N₄. These results allow us to infer that the Eg of heterostructure is governed by QDs CdSe on g-C₃N₄ sheets.

3.2. Electrochemical Characterization of QD CdSe@g-C₃N₄

Electrochemical studies of g-C₃N₄, QDs CdSe and QDs CdSe@g-C₃N₄ were carried out to determine their reduction and oxidation potentials, aiming to identify possible mechanisms of electron and hole transfer between the two nanostructures. Thus, cyclic voltammograms were performed in 1000 µL of acetonitrile containing 0.1 M of tetrabutylammonium hexafluorophosphate as a supporting electrolyte + 50 µL of each sample, under an oxygen-free atmosphere, applying a scan rate (v) of 50 mV s⁻¹. As shown in Figure 4a, the cyclic voltammograms of g-C₃N₄ and QDs CdSe (represented by the magenta and black lines, respectively) exhibit an irreversible single peak in the anodic scan, corresponding to hole injection into the valence band (VB). When the cathodic scan was performed, g-C₃N₄ showed an irreversible single peak corresponding to electron injection into the conduction band (CB) (Figure 4b, magenta line). QDs CdSe showed a reduction peak during the forward potential scan due to reduction of Cd⁺² to Cd⁰ and an oxidation peak corresponding to oxidation of Cd⁰ when the potential scan is reversed (Figure 4b, black line). Anodic peak (E_a,p) and cathodic peak (E_c,p) potentials determined for QDs CdSe, g-C₃N₄ and QDs CdSe@g-C₃N₄ are listed in Table 2. As shown in Table 2, g-C₃N₄ is oxidized at a higher potential than QDs CdSe and reduced at a lower potential than QDs CdSe. The anodic voltammograms of QDs CdSe@g-C₃N₄ exhibit two distinct peaks, corresponding to QDs CdSe and g-C₃N₄, respectively, as these peaks appear close to the E_a,p of the individual nanostructures. When the cathodic scan was carried out, a single peak was observed in the direct scan at −1.51 V, with its corresponding anodic peak appearing when the scan direction was inverted due to Cd²⁺/Cd⁰.

The difference between E_a,p and E_c,p (ΔE_p) is the Eg determined electrochemically by Equation (1).

ΔE_p = E_a,p − E_c,p = Eg

(1)

The values determined are listed in Table 2. These values were very close to Eg determined by the Tauc method (see Table 2).

For the case of the QDs CdSe@g-C₃N₄, the E_a,p used to determine the ΔE_p corresponds to the first peak at 0.630 V vs. HNE. The difference in the Eg values calculated by the optical and electrochemical methods arises from the fact that the photophysical process occurs within the core of each nanomaterial, whereas the ΔE_p (potential difference) results from the injection of electrons or holes from the electrode into the nanomaterial, which is a surface-related process [39]. As shown in Figure 4a (green line), QDs CdSe on the heterostructure are oxidized at lower potential that g-C₃N₄, which allows us to infer, along with the spectral overlap, a possible RET mechanism. When a potential higher than 1 V is applied, QDs CdSe act as hole acceptors, thereby acting as light emitters.

3.3. Electrochemiluminescence Analysis of QDs CdSe@g-C₃N₄

To perform ECL measurements, 60 μL of the dispersion of either QDs CdSe, g-C₃N₄, or QDs CdSe@g-C₃N₄ was added to a 0.1 M Tris-HCl buffer solution, pH 7.5, + 100 μL of TPrA at a concentration of 5.25 M. TPrA was used as a co-reactant in the mechanism of light generation by electrochemiluminescence. ECL measurements were carried out as described in the Supplementary Materials (Section S2).

Figure 5 shows the ECL spectra of g-C₃N₄, QDs CdSe, and QDs CdSe@g-C₃N₄ when a potential of 1.1 V is applied. In comparison with the fluorescence spectrum (Figure 2B), the ECL spectrum of g-C₃N₄ shows a bathochromic shift close to 50 nm, which is common in electrochemical processes [48]. However, QDs CdSe exhibit a hypsochromic shift of approximately 55 nm in their ECL emission.

The ECL spectrum of QDs CdSe@g-C₃N₄ shows a predominant peak with a maximum at 578 nm corresponding to the emission of QDs CdSe in the heterostructure, as well as a small band close to 390 nm attributed to g-C₃N₄. The spectral shifts observed in the absorption band of QDs CdSe within the heterostructure, when compared to QDs CdSe solution, result from different synthesis conditions. Specifically, the first ones were grown directly on g-C₃N4 nanosheets, whereas the second ones were synthesized in colloidal solution. From the ECL spectrum of QDs CdSe@g-C₃N₄, it is possible to infer that there is a charge transfer from nitride to the QDs CdSe, where g-C₃N₄ acts as a hole donor to QDs CdSe, and QDs CdSe are the light-emitting species. This process is explained by the E_a,p of both nanostructures, as shown in Figure 6.

When the anodic potential exceeds 1.10 V, g-C₃N₄, QDs CdSe and TPrA are oxidized on the electrode. TPrA forms its radical cation (TPrA^•+), which is deprotonated to generate a highly reducing radical (TPrA^•). TPrA^• reduces the heterostructure, thereby increasing the population of electrons in the conduction band (CB). In addition, g-C₃N₄ is oxidized at a higher anodic potential than QDs CdSe, and a hole transfer occurs from g-C₃N₄ to the QDs CdSe, resulting in a higher population of holes in the valence band (VB) of the QDs CdSe, and favouring the recombination of electron–hole pairs. This process enhances the radiative recombination [49].

3.4. ECL Measurements of QDs CdSe@g-C₃N₄ Solution

Figure 7 shows ECL voltammograms of a QDs CdSe@g-C₃N₄ solution. An increase in ECL intensity was observed for higher potential of 0.520 V, reaching its maximum around 0.8 V (black line). This behaviour is in accordance with the electrochemical oxidation of both QDs CdSe and g-C₃N₄ within the heterostructure. When a low concentration of H₂O₂ is added into the cell (4 × 10⁻⁶ M), an increase in ECL intensity is observed (red line). This enhancement was attributed to the oxidative capacity of H₂O₂, where its reduction or decomposition generated highly reactive intermediates that participate in the redox cycle of the QDs CdSe@g-C₃N₄, promoting the formation of excited states that enhance the ECL signal. However, when higher concentrations of H₂O₂ were added, a decrease in ECL intensity was observed, as depicted in Figure 7 (green line). An overproduction of reactive oxygen species can be produced, which could lead to the oxidative degradation of the heterostructure, reducing its emissive properties.

On the basis of the data obtained, and from previous reports, at pH = 7.4, a probable ECL mechanism is proposed [50]:

$$\mathrm{TPrA}+{\mathrm{Tris}}^{+}\rightleftarrows {\mathrm{TPrAH}}^{+}+\mathrm{Tris}$$

(2)

$$\mathrm{TPrA}-{\mathrm{e}}^{-}\to {\mathrm{TPrA}}^{\bullet +}$$

(3)

$${\mathrm{TPrA}}^{\bullet +}+\mathrm{Tris}\rightleftarrows {\mathrm{TPrA}}^{\bullet }+{\mathrm{Tris}}^{+}$$

(4)

$$\mathrm{QDs} \mathrm{CdSe}@\mathrm{g-}{\mathrm{C}}_3{\mathrm{N}}_4-\mathrm{n}{\mathrm{e}}^{-}⟶{\mathrm{n}\mathrm{QDs} \mathrm{CdSe}@\mathrm{g-}{\mathrm{C}}_3{\mathrm{N}}_4}^{\bullet +}$$

(5)

$${\mathrm{QDs} \mathrm{CdSe}@\mathrm{g}{\mathrm{C}}_3{\mathrm{N}}_4}^{\bullet +}+\mathrm{TPrA}⟶{\mathrm{QD} \mathrm{CdSe}@\mathrm{g}{\mathrm{C}}_3{\mathrm{N}}_4+\mathrm{TPrA}}^{\bullet +}$$

(6)

$${\mathrm{QDs} \mathrm{CdSe}@\mathrm{g}{\mathrm{C}}_3{\mathrm{N}}_4}^{\bullet +}+{\mathrm{TPrA}}^{\bullet }⟶{\mathrm{QD} \mathrm{CdSe}@\mathrm{g}{\mathrm{C}}_3{\mathrm{N}}_4}^{*}+{\mathrm{Im}}^{+}$$

(7)

$${\mathrm{QDs} \mathrm{CdSe}@\mathrm{g}{\mathrm{C}}_3{\mathrm{N}}_4}^{*}⟶\mathrm{QD} \mathrm{CdSe}@\mathrm{g}{\mathrm{C}}_3{\mathrm{N}}_4+\mathrm{h}\nu$$

(8)

$$\mathrm{QDs} \mathrm{CdSe}@\mathrm{g}-{\mathrm{C}}_3{\mathrm{N}}_4+\mathrm{TPrA}+{\mathrm{H}}_2{\mathrm{O}}_2⟶\mathrm{QD} \mathrm{CdSe}@\mathrm{g}{\mathrm{C}}_3{\mathrm{N}}_{4\left(\mathrm{oxidized}\right)}+{\mathrm{TPrA}}_{\left(\mathrm{oxidized}\right)}+{\mathrm{H}}_2\mathrm{O}$$

(9)

At pH 7.4, Tris⁺ protonates TPrA via an acid–base reaction (Equation (2)) [51]. Then, when a potential of 0.60 V is applied, TPrA is oxidized to form the radical cation TPrA^+• (Equation (3)), which diffuses and is deprotonated by Tris (Equation (4)) [52]. When the potential reaches 1.10 V, CdSe QDs@g-C₃N₄ is oxidized through n-electron transfer, generating CdSe QDs@g-C₃N₄^+• (Equation (5)). Both g-C₃N₄ and QDs CdSe undergo an oxidation, initiating a resonance energy transfer (RET) process where n-holes from the valence band of g-C₃N₄ migrate to the valence band of CdSe QDs (Figure 6) [33]. Subsequently, CdSe QDs@g-C₃N₄^+• can react with TPrA or TPrA^• to produce TPrA^+• and a high-energy species (CdSe QDs@g-C₃N₄^*) along with an imidazole cation (Im⁺) (Equations (6) and (7), respectively). CdSe QDs@g-C₃N₄^* then returns to its ground state by emitting radiation (Equation (8)). However, in the presence of H₂O₂, both CdSe QDs@g-C₃N₄ and TPrA can be oxidized by H₂O₂ (Equation (9)), quenching the ECL process.

3.5. Determination of Optimal Ratio of QDs CdSe and g-C₃N₄

In order to determine the optimum ratio between QDs CdSe and g-C₃N₄ to maximize the RET effect, different syntheses were performed with varying concentrations of nitride while keeping the concentration of QDs CdSe constant. As shown Figure 8, an increase in fluorescence intensity at 570 nm was observed with increasing g-C₃N₄ concentration until reaching 0.1 mg mL⁻¹. However, a decrease in emission intensity was obtained for g-C₃N₄ concentrations higher than 0.1 mg mL⁻¹. Therefore, 0.1 mg mL⁻¹ of g-C₃N₄ was used for all measurements.

3.6. Analytical Performance of the Electrochemiluminiscence Sensor for H₂O₂

H₂O₂ was added at different concentrations to the solution composed of 60 µL of QDs CdSe@g-C₃N₄ + 100 µL of 5.25 M TPrA in 0.1 M Tris-HCl solution, pH = 7.4 (final volume of 2500 μL). The ECL measurements were performed by applying four cycles of potential pulses between −0.2 V and 1.2 V, where each pulse was applied for 5 s in a measurement known as the on–off response [53]. ECL emission was monitored at 570 nm using a fixed monochromator setting. Figure 9a shows the ECL transients where, as previously observed, the ECL intensity decreases with increasing H₂O₂ concentrations. In addition, high reproducibility of ECL responses was observed for all H₂O₂ concentrations. For each H₂O₂ concentration, the energy intensity values of ECL (${\mathrm{I}}_{\mathrm{ECL}}^{\mathrm{area}}$) were determined by calculating the area under each emission transient using the following equation:

$${\mathrm{I}}_{\mathrm{ECL}}^{\mathrm{area}}={\int }_{{\mathrm{t}}_{\mathrm{n}}}^{{\mathrm{t}}_{\mathrm{n}+5 \mathrm{s}}}\mathit{ECL} \mathit{int}.\times \mathrm{dt}$$

(10)

where t_n corresponds to the start time of each potential step, and ECL int. is the electrochemiluminescence value for each time. A calibration curve (Figure 9b) was constructed, represented as ${\mathrm{I}}_{\mathrm{ECL}}^{\mathrm{area}}/{\mathrm{I}}_{\mathrm{ECL}}^{{\mathrm{area}}^0}$ vs. Log ${\mathrm{c}}_{{\mathrm{H}}_2{\mathrm{O}}_2}^{*}$, where ${\mathrm{I}}_{\mathrm{ECL}}^{{\mathrm{area}}^0}$ corresponds to the energy intensity value in the absence of H₂O₂.

From the calibration curve, the analysis of the figures of merit was carried out [54]. The limit of detection (LOD), limit of quatification (LOQ), and sensitivity were 1.81 × 10⁻⁹ M, 5.37 × 10⁻⁹ M, and 1.52 × 10⁸ M⁻¹, respectively, with a linear range from 5.37 × 10⁻⁹ to 1 × 10⁻⁶ M. These results allow us to infer the excellent LOQ of the ECL sensor for detecting H₂O₂ in buffer Tris-HCl, pH = 7.4.

3.7. Quantification of H₂O₂ in Lyophilized Serum Samples

The detection of H₂O₂ was carried out, as a proof of concept, in reconstituted lyophilized serum samples because this is a complex matrix with known levels of specific analytes such as electrolytes, enzymes, metabolites, and proteins. Then, the effect of amount of serum added to the cell on ECL responses was determined. Therefore, different volumes of serum were added into the cell containing 60 µL of QDs CdSe@g-C₃N₄ in 2500 µL of Tris-HCl buffer, pH = 7.4, in the presence of the co-reactant. As shown in Figure 10a, an increase in ECL intensity was observed with increasing serum volume, reaching a maximum value of ECL intensity at 5 µL. For higher serum volumes, a significant decrease in ECL intensity was observed, indicating a clear matrix effect. Therefore, 5 µL of serum sample was chosen for the detection of H₂O₂. A calibration curve represented as ${\mathrm{I}}_{\mathrm{ECL}}^{\mathrm{area}}/{\mathrm{I}}_{\mathrm{ECL}}^{{\mathrm{area}}^0}$ vs. Log ${\mathrm{c}}_{{\mathrm{H}}_2{\mathrm{O}}_2}^{*}$ using serum samples as the matrix was performed for a H₂O₂ concentration range from 5.2 × 10⁻⁷ to 1 × 10⁻² M (see Figure 10b). A linear relationship between ${\mathrm{I}}_{\mathrm{ECL}}^{\mathrm{area}}/{\mathrm{I}}_{\mathrm{ECL}}^{{\mathrm{area}}^0}$ and Log ${\mathrm{c}}_{{\mathrm{H}}_2{\mathrm{O}}_2}^{*}$ was obtained in the concentration range from 5.2 × 10⁻⁷ to 5.2 × 10⁻⁶ M. Due to the matrix, the ECL intensity decays by approximately four times compared to the ECL intensity obtained in buffer solution. Due to this matrix effect, the ${\mathrm{I}}_{\mathrm{ECL}}^{\mathrm{area}}/{\mathrm{I}}_{\mathrm{ECL}}^{{\mathrm{area}}^0}$ ratio tends toward zero for H₂O₂ concentrations (in lyophilized serum) higher than 5 × 10⁻⁶ M. The lowest H₂O₂ concentration detected was 5.2 × 10⁻⁷ M.

The reproducibility of the electrochemiluminescent sensor in the linear concentration range was good. Since the calibration curve was created using the sample matrix, these results allow us to infer that the electrochemiluminescent sensor is a reliable tool for determining H₂O₂ in serum samples. Percentage coefficient variation values for H₂O₂ concentrations of 5.20 × 10⁻⁷ M, 1.30 × 10⁻⁶ M and 2.60 × 10⁻⁶ M were determined. The values obtained were 11%, 17% and 13%, respectively. These values are very good, even more so considering that they were obtained from serum samples without pre-treatment. The lyophilized serum contains glucose, urea, uric acid and cholesterol at concentrations higher than 2 × 10⁻⁴ M. These components are likely responsible for the observed matrix effects. Since ascorbic acid and dopamine are not present in the lyophilized serum, we conducted separate interference studies for both compounds at concentrations of 1 × 10⁻⁴ M. At this concentration (threefold higher than physiological serum levels), ascorbic acid produces a decrease in the ECL signal of 18%. However, the interfering effect diminishes substantially when the ascorbic acid concentration is reduced to the normal physiological range (close to 8%). Dopamine showed interference at 1 × 10⁻⁴ M by the reaction with H₂O₂, and the ECL intensity increased by 20%. However, this is irrelevant at the physiological concentration of dopamine because it is at sub-pM levels.

3.8. Comparison of the Performance of the ECL Sensor with Other Sensors for the Detection of H₂O₂

The performance of the ECL sensor was compared with other relevant sensors used for the detection of H₂O₂, as shown in

Table 3. Different methods proposed in the literature to quantify H₂O₂.

Materials/Probe	Technique	Sample	Linear Range /μM	LOD /μM	Ref.
PdNP/CNT fibre	Amperometry	PBS, pH 7.4	2–1300	2	[55]
Pd-NPs/BGFs/GCE	Amperometry	PBS, pH 7.4	4–13,500	1.5	[56]
DHLA-CuNCs	Fluorescence sensing methods	Aqueous solution	1–10	0.3	[57]
AuNPs-AuNCs		PBS, pH 8	1–100	0.8	[58]
CuNCs/ZIF-8		Acetate buffer, pH 7	0.5–30	0.5	[59]
Upconversion nanoparticles		PBS, pH 6.5	2.5–70	0.8	[60]
QDs CdTe		PBS, pH 7	10–125	0.3	[61]
CdSe@ZnS/AgNCs		Tris-HCl buffer	0.5–60	0.3	[62]
CuO/g-C3N4/GCE	DPV	PBS, pH 7	0.5–50	0.31	[17]
HRP/AuNPs/CESM	Amperometric	PBS, pH 7	10–2700	3	[63]
QDs CdSe@g-C3N4	ECL	Tris-HCl, pH 7.4	1 × 10−4–1	1.81 × 10−3	This sensor

DPV: differential pulse voltammetry.

. The ECL sensor demonstrated advantages such as high sensitivity, a wide linear range, and very low detection limits to detect H₂O₂ in Tris-HCl buffer. Additionally, its ability to operate in complex matrices, such as serum, without significant interference highlights its potential for practical applications. The ECL sensor’s reproducibility and reliability further support its suitability for accurate H₂O₂ quantification in biological samples.

4. Conclusions

In this work, we have developed an electrochemiluminescent sensor based on a novel probe composed of CdSe quantum dots and g-C₃N₄ nanosheets. The QDs CdSe@g-C₃N₄ heterostructure exhibits remarkable reproducibility and excellent sensitivity for the detection of H₂O₂ by electrochemiluminescence, demonstrating compatibility with both aqueous systems and complex biological matrices (e.g., serum). This performance demonstrates its potential for biochemical studies and human health applications. The high performance of the electrochemiluminescent probe is based on the effective energy transfer process between the two nanostructures. Two aspects are important to highlight. First, the electrochemiluminescence methodology developed is remarkably simple, requiring only the addition of the probe to the sample solution for analysis; second, the sensor allowed the detection of H₂O₂ in untreated reconstituted lyophilized serum samples, where, despite the matrix effect, sub-µM levels of H₂O₂ were successfully quantified. In this way, the electrochemiluminescent sensor has proven to be a reliable and simple analytical tool for detecting H₂O₂ in complex samples without the need for pre-treatment.