All reagents used in this study were of analytical grade. The products used were: L-cystine (98%), silver nitrate (AgNO₃, 99%), anhydrous sodium carbonate (Na₂CO₃, 99%), sodium chloride (NaCl, 98%), magnesium chloride (MgCl₂, 98%), calcium chloride (CaCl₂, 99%), sodium hydroxide (NaOH, 98%), and sodium borohydride (NaBH₄, 99%), supplied by Sigma Aldrich. Lemon powder, obtained by grinding lemon peel, was purchased commercially.

A grinder equipped with a robust JIASOUNG JS8000 C motor, with a power rating of 3000 W, a frequency of 50 Hz, and a maximum speed of 5 rpm, was used to obtain the lemon powder. A high-speed benchtop centrifuge from the company MRCLTD with a maximum capacity of 13,000 rpm, set at 4000 rpm, was also used for all centrifugation operations. UV-Vis spectral curves were recorded using an Ocean Optics FLAME spectrophotometer (USA). Fluorescence measurements of the CQDs were performed using a UV lamp with two wavelengths (254 and 365 nm). The demineralized water used in this work came from the water purification system of Sichuan Zhuoyue Water Treatment Equipment CO., Ltd. (Chengdu, China). This system produces demineralized water with a resistivity of 18.25 MΩ·cm.

The sensor was manufactured in three (3) main steps:

Step 1: Hydrothermal Synthesis of Cys-CQDs

A lemon powder was prepared by calcining lemon pulp at 200˚C for 1 hour. 250 mg of this powder and 150 mg of L-cysteine were dissolved in 50 mL of deionized water. The mixture was then transferred to a Teflon reactor, sealed in a stainless steel autoclave, and heated at 180˚C for 12 hours. After cooling, the solution was centrifuged at 10,000 rpm for 15 minutes and then filtered through a 0.22 μm membrane to remove aggregates and obtain a pure aqueous suspension of Cys-CQDs.

Step 2: Cys-CQD Reduction

10 mL of the Cys-CQD suspension is mixed with 50 mg of sodium borohydride (NaBH₄) under magnetic stirring (500 rpm) at room temperature for 8 hours. The excess NaBH₄ was removed by centrifugation, thereby significantly minimizing the presence of residual borohydride in the final suspension. This step reduces the surface functional groups and increases the passivation of the CQDs.

Step 3: Functionalization of AgNPs with reduced Cys-CQDs

1 mL of reduced Cys-CQDs was dispersed in 200 mL of distilled water under magnetic stirring (600 rpm) for 2 minutes. 0.02 g of silver nitrate (AgNO₃) was added under stirring for an additional 2 minutes. The mixture was then heated to 70˚C for 30 minutes, and 0.1 g of sodium hydroxide (NaOH) was added. The reaction was activated under stirring on a hot plate (70˚C, 600 rpm) for 8 hours to allow the formation of Cys-CQDs/AgNPs nanocomposites via silver thiol bonding [8].

To two milliliters of standard solution (or water sample), 50 μL of the sensor solution is added, followed by 1 minute of homogenization. After homogenization, the color change of the mixture is monitored over time, and then the absorption spectrum is recorded. For quantitative analysis, the absorbance was recorded after a fixed incubation time of 3 min, corresponding to the plateau region of the kinetic response curve. The calibration curve is subsequently established under the same conditions with different concentrations of the nitrate ion standard solution (0, 4, 8, 16, 40, 160, 200, 640, and 800 mg/L), between the difference in absorbance (A0) in the absence and maximum absorbance (Ai) in the presence of nitrate ions at a wavelength 400 nm, relative to the nitrate ion concentrations (A0 − Ai = f(C)).

The validation of an analytical method is a procedure that establishes, through experimental studies, that the method’s performance criteria meet the requirements of its intended use. It is worth recalling that the objective of an analytical assay method is to quantify each unknown quantity that the analyst will have to evaluate as accurately and precisely as possible. The objective of validation, on the other hand, is to provide the laboratory and regulatory authorities with sufficient assurance that each measurement performed routinely with a given method will be close to the truth. At this stage, it is important to define some frequently encountered terms.

2.4.1. Linearity

The linearity range represents the concentration range explored during method validation. This range must take into account all possible future results. Indeed, if, during routine use of the method, the unknown quantities that the laboratory will have to measure fall outside the linearity range, the results provided will not be valid considered reliable. The linearity of an analytical procedure is its ability, within a certain titration range, according to a predefined mathematical function, to obtain the analyte concentration in the sample. Linearity in this study was verified by plotting absorbance against concentration. The limit of detection and quantification was also determined using this same calibration curve.

2.4.2. Limit of Detection

The limit of detection (LOD) is the smallest concentration or quantity of a parameter or analyte that produces a detectable signal with a defined reliability and is statistically different from that produced by a blank under the same conditions [9]. It is calculated using Equation (1):

With s being the standard deviation of the deviation and S the slope of the calibration curve.

2.4.3. Limit of Quantification

The limit of quantification is the concentration equivalent to 10 times the standard deviation obtained when establishing the limit of detection of the method [10] as described by Equation (2).

It is also the smallest quantity of a substance to be examined in a sample that can be quantified under the described experimental conditions with an acceptable coefficient of variation of the response factor [11].

2.4.4. Accuracy

Accuracy, or recovery rate, or percentage recovery is the percentage difference between the measured concentration of a fortified sample and the measured concentration of the same unfortified sample, divided by the concentration of the added substance. This ratio takes into account any chemical transformation that has occurred. This rate is expressed by Equation (3).

where C_f: Measured concentration of the fortified sample; C: Measured concentration of the unfortified sample; C_a: Concentration of the added substance.

2.4.5. Precision

Precision expresses the closeness of agreement (degree of dispersion, coefficient of variation) between a series of measurements taken from multiple aliquots of a homogeneous sample under the prescribed conditions [11]. It is generally expressed by the coefficient of variation (Cv in %), which is obtained by dividing the standard deviation (SD) of the test results by the mean of the observed responses (2) [12]. Precision reflects only the distribution of random errors and has no relation to the true or specified value. Quantitative measures of precision are critically dependent on the stipulated conditions.

2.4.6. Sensitivity

The sensitivity of an analytical procedure can be defined as the ratio of the change in the analytical method’s response to the change in the amount of analyte. A procedure is considered sensitive if a small change in the concentration or amount of analyte results in a change in the response. In this study, sensitivity was verified by measuring the concentration of nitrate ions in the presence of other compounds that could interfere with its detection.

The different synthesized materials were analyzed by UV-Vis spectroscopy (Figure 1). First, the spectrum of Cys-CQDs (Figure 1(a)) shows a strong absorption picture at 242 nm. This peak is typical of π→π* electron transitions in aromatic carbon structures (sp² domains) that form during hydrothermal synthesis. A small shoulder is also observed around 320 nm, corresponding to n-π* transitions of heteroatoms such as carbonyl (C=O), amine, etc., groups on the surface. These two peaks are characteristic of carbon quantum dots. It should be noted that the presence of these heteroatoms contributes to the reactivity and passivation of CQDs. Furthermore, the spectrum of reduced Cys-CQDs (Figure 1(b)) shows changes at the shoulder at 320 nm. This

indicates that the surface groups have been modified, confirming that the reduction has worked well. This step was necessary to activate the reducing power of the CQDs. Finally, the spectrum of the Cys-CQDs/AgNPs nanocomposite (Figure 1(c)) is very revealing. It clearly shows the combination of the CQDs peak at 242 nm and a new, broad peak centered at 421 nm. This new peak is the signature of the silver nanoparticles (AgNPs): it is the surface plasmon resonance (SPR) phenomenon, which proves that the silver ions (Ag⁺) were indeed reduced to metallic silver (Ag⁰) by the Cys-CQDs, which in turn underwent oxidation. Thus, upon the addition of AgNO₃, the reduction of Ag⁺ ions is primarily attributed to the surface functional groups of CQDs-Cys (notably −OH, −COOH, and −NH₂), which can act as mild reducing and stabilizing agents. This good stability is essential for considering a reliable application as a sensor.

Figure 1. UV-Visible Spectrum of the synthesized sensors: (a) Cys-CQDs, (b) reduced Cys-CQDs, and (c) Cys-CQDs/AgNPs.

The study of photoluminescence properties revealed distinct behaviors for each nanomaterial (Figure 2(B)), providing crucial information on their structure and interactions. Cys-CQDs exhibit intense blue fluorescence under excitation at 365 nm. This strong emission results from the combination of two main mechanisms: the quantum confinement effect in nanometer-sized graphitic domains (sp₂), characteristic of CQDs [13], and the presence of heteroatoms (nitrogen, sulfur) from cysteine functionalization, which create surface electronic states amplifying light emission. For reduced Cys-CQDs, a significant increase in fluorescence intensity is observed. This improvement results from the reduction of carbonyl groups (C=O) to hydroxyl groups (−OH) by NaBH₄, thus decreasing the density of surface “non-radiative traps.” This passivation promotes radiative recombination, leading to a higher quantum yield, consistent with the mechanisms described by Shi and colleagues [14]. The most striking result concerns the Cys-CQDs/AgNPs nanocomposite, which exhibits near-total quenching of fluorescence. This phenomenon confirms the formation of an intimate nanocomposite, likely explained by fluorescence resonance energy transfer (FRET). In this By this mechanism, Cys-CQDs (donors) transfer their excitation energy to AgNPs (acceptors) in

Figure 2. Fluorescence of synthesized sensors a) Cys-CQDs, b) reduced Cys-CQDs, and c) Cys-CQDs/AgNPs.

a non-radiative manner [15]. The efficiency of this transfer, which is highly dependent on distance (1 - 10 nm), confirms the close proximity between AgNPs and CQDs [6]. This fluorescence quenching is essential for sensor development, as analyte recognition could modulate this signal, thus generating a detectable response.

A preliminary assessment of the nanomaterials’ detector potential was performed using a simple colorimetric test. As shown in Figure 3, the solutions of Cys-CQDs (a, b) and reduced Cys-CQDs (c, d) showed no visible color change after the addition of nitrate ions ( NO 3 − ) compared to the controls in distilled water. In contrast, a clear positive response was observed for the Cys-CQDs/AgNPs nanocomposite. Its color changed from a pale yellow (tube e, control) to a distinctive pink (tube f) in the presence of nitrate. This preliminary result indicates that the Cys-CQDs/AgNPs nanocomposite is the only one among the materials tested to exhibit colorimetric sensitivity to nitrate ions, validating its selection for further investigation. This response suggests that this sensor can be used for the colorimetric detection of nitrates [16][17]. Nitrate ions ( NO 3 − ) interact specifically with the sensor through the combined action of Cys-CQDs and AgNPs. The sensing response is likely associated with nitrate-induced modulation of the colloidal stability of AgNPs, possibly through ionic strength effects and electrostatic screening, leading to nanoparticle aggregation and consequent plasmonic changes. Similar behaviors have been reported in related AgNP-based sensing systems [18]-[20].

Figure 3. Solution of the synthesized sensors in the absence and presence of nitrate ions, respectively: (a), (b) Cys-CQDs, (c), (d) reduced Cys-CQDs, and (e), (f) Cys-CQDs/AgNPs.

The stability of Cys-CQDs/AgNPs was investigated by recording the absorption spectra of the freshly prepared Cys-CQDs/AgNP colloidal solution after 1; 14; 30 and 60 days of synthesis (Figure 4).

Figure 4. Stability of Cys-CQDs/AgNPs.

The absorbance of the colloidal Cys-CQDs/AgNPs solution, after 1, 14, 30, and 60 days of synthesis, exhibits the same absorption wavelength as that of the freshly prepared solution (400 nm). The results show that there is neither a significant decrease in the intensity of the absorbance peak nor a change in the position of the maximum wavelength, suggesting excellent stability of the prepared Cys-CQDs/AgNPs.

3.5.1. Calibration of the Detection Method

In order to determine the detection range of nitrate ions, different concentrations of the standard solution were evaluated (0; 4; 8; 16; 40; 160; 200; 640 and 800 mg/L) as shown in Figure 5.

The results indicate that the Cys-CQDs/AgNPs sensors exhibit a color change in the presence of nitrate ions. The color gradually changes from yellow to pink, intensifying over time and strongly dependent on the concentration. At high concentrations (160 - 800 mg/L), the pink coloration appears immediately. At low concentrations (4 - 40 mg/L), the complete change takes approximately 30 minutes. This encouraging visual observation motivated a more in-depth quantitative study. To this end, we established a calibration curve by measuring the change in absorbance (ΔA) as a function of concentration (Figure 5(B)).

Figure 5. Calibration curve: (a) 0 mg/L, (b) 4 mg/L, (c) 8 mg/L, (d) 16 mg/L, (e) 40 mg/L.

The results obtained show a linear curve with Equation (4):

The R² coefficient is 0.9995, extremely close to 1. This confirms that the relationship is not only linear but also highly reliable. Therefore, this equation can be used to determine the nitrate concentration of an unknown water sample simply by measuring the change in absorbance.

3.5.2. Limit of Detection and Quantification

The limits of detection and quantification were calculated using Equation (1) and Equation (2). The results obtained are LOD = 2.51 × 10⁻² mg/L and LOQ = 8.38 × 10⁻² mg/L. These results indicate that the sensor has an ultrasensitive detection capability, perfectly suited for environmental applications, as observed in other systems based on hybrid nanomaterials [21][22].

3.5.3. Sensor Reliability

The sensor’s reliability was evaluated through recovery rate experiments. Three (3) NaNO₃ solution samples at different concentrations were analyzed, with three (3) independent replicates for each. The results (Table 1) show recovery rates ranging from 97% to 104%, indicating excellent sensor reliability, in accordance with analytical criteria [23].

Table 1. Mean recovery rate and standard deviation (n = 3).

These results demonstrate that the Cys-CQDs/AgNPs sensor provides reliable measurements suitable for practical applications of monitoring nitrates in water.

3.5.4. Sensor Accuracy (Reliability)

The sensor accuracy was verified by inter-day and intra-day measurements of three (3) nitrate solution concentrations. Each concentration was measured three (3) times independently on the same day and on three (3) consecutive days. All values found (RSD) were less than 5% (Table 2). This high accuracy suggests that our chosen sensor can be used to detect nitrate ions.

Table 2. Relative recovery rate and standard deviation.

RSD: Relative Standard Deviation.

3.5.5. Interference Test on Nitrate Ion Detection by Our Sensor

Water contains many dissolved ions, the main ones being calcium (Ca²⁺), magnesium (Mg²⁺), sodium (Na⁺), potassium (K⁺), carbonates ( CO 3 − ), bicarbonates ( HCO 3 − ), sulfates ( SO 4 2 − ), chlorides (Cl⁻), and nitrates ( NO 3 − ). Therefore, compounds such as Na₂CO₃, NaCl, MgCl₂, and CaCl₂ were tested at concentrations 10 times higher than those of nitrate ions ( NO 3 − ) in the presence of our sensor to evaluate their interference with these different ions (Figure 6). The near 100% recovery rates confirm that variations in ionic strength do not affect the sensor’s response [24]-[27].

Figure 6. Recovery rates (A) of nitrate alone at 16 mg/L and of nitrate in the presence of 160 mg/L, including (B) NaCl, (C) MgCl₂, and (D) CaCl₂.

3.5.6. Nitrate Detection in the Real Matrix

To evaluate the effectiveness of the Cys-CQDs/AgNPs sensor under environmental conditions, water samples from wells, boreholes, lagoons, and taps were collected (2 mL each) and mixed with 50 μL of the Cys-CQDs/AgNPs nanocomposite. A distinct pink coloration was observed in all the tested matrix, visually confirming the presence of nitrate ions. Nitrate concentrations were then measured in each matrix (Table 3). The results show that the sensor effectively detects nitrates at both low and high concentrations, confirming its suitability for applications on complex environmental samples.

Table 3. Detection of nitrate in real matrix (n = 3).

These results confirm the effectiveness of nanomaterial-based sensors for the rapid and sensitive detection of nitrate ions in real water samples, consistent with previous work on environmental nitrate detection [28].

3.5.7. Comparison with Other Methods

To better illustrate the performance of the implemented technique, a comparison with results from the literature was conducted. The detection method, colorimetry using carbon quantum dots modified by silver nanoparticles (Cys-CQDs/AgNPs), showed a limit of detection (LOD) of 0.0251 mg/L, which is competitive with other methods. This colorimetric method is valued for its simplicity and efficiency, allowing direct visual detection without additional equipment [29]. Comparatively, other nitrate detection methods, such as high-performance liquid chromatography (HPLC), have higher limits of detection. For example, Uddin reported a limit of detection (LOD) of 2.26 mg/kg for nitrate determination in fruits and vegetables, which is significantly less sensitive than the Cys-CQDs/AgNPs method [30]. Furthermore, techniques such as flow injection analysis have also been used, with an LOD of 0.34 mg/kg, which is less effective than the aforementioned colorimetric method [31]. Other approaches, such as electrochemical sensors, have shown promising performance. For example, one study demonstrated that electrodes modified with palladium-gold nanoparticles detected nitrates with a sensitivity of 4.7 μA·mg⁻¹·L [32]. However, these methods are hampered by more complex experimental conditions and heavy equipment. Furthermore, the systems Lab-on-a-chip devices have been developed for in situ analyses, achieving LODs of 0.025 μM, but they are generally more difficult to implement in the field [33]. In summary, the Cys-CQDs/AgNPs nitrate ion detection method with an LOD of 0.0251 mg/L is competitive with other techniques such as HPLC, flow injection analysis, and electrochemical sensors. Its simplicity, speed, and efficiency make it an attractive option for nitrate detection in various settings.