Machine Learning and Experimental Design-Driven Fluorescence Detection of Metronidazole in Food Samples Using Mold-Derived Carbon Quantum Dots

Machine Learning and Experimental Design-Driven Fluorescence Detection of Metronidazole in Food Samples Using Mold-Derived Carbon Quantum Dots | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Machine Learning and Experimental Design-Driven Fluorescence Detection of Metronidazole in Food Samples Using Mold-Derived Carbon Quantum Dots Kimiya Khandestan, Azar Sabukbar, Bahareh Rahimian Zarif, Nahid Haghnazari, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7721529/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Carbon quantum dots (CQDs) were synthesized using mold as a green carbon source and characterized by a combination of techniques including fluorescence spectroscopy, Fourier-transform infrared spectroscopy (FTIR), Scanning Electron Microscopy (SEM), Energy-Dispersive X-ray Spectroscopy (EDX), elemental mapping, and line scan analysis. The CQDs showed strong fluorescence emission suitable for sensitive detection of the antibiotic metronidazole based on fluorescence quenching. After optimizing various parameters influencing sensor performance through experimental design, a linear detection ranges from 97.4 to 3215.3 µM with a detection limit of 50.5 µM was established. Machine learning algorithms were applied to enhance the accuracy and reliability of metronidazole quantification. This novel, eco-friendly sensor platform offers effective antibiotic monitoring for biomedical and environmental applications. Physical sciences/Chemistry Earth and environmental sciences/Environmental sciences Physical sciences/Nanoscience and technology Physical sciences/Optics and photonics Carbon quantum dots Mold synthesis Metronidazole detection Fluorescence sensor Experimental design Machine learning Antibiotic monitoring Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Introduction Ensuring the safety and quality of food is crucial for maintaining public health and promoting proper growth and development. Safe food not only requires adherence to hygiene standards during production and storage but also must be free from harmful contaminants, including antibiotic residues. In recent decades, the widespread use of antibiotics in agriculture and animal husbandry to prevent and treat infections has increased significantly. However, excessive and improper use of these drugs has led to contamination of food products with antibiotic residues, posing serious health risks to consumers[1,2]. While antibiotics play a vital role in treating bacterial infections, their overuse in livestock production has contributed to the emergence of antibiotic-resistant bacteria. This resistance not only makes infections harder and more expensive to treat but also threatens global health security. Antibiotic residues present in animal-derived foods such as meat, milk, and eggs can cause allergic reactions and other adverse health effects in humans. Therefore, monitoring and controlling antibiotic use in food production systems is essential to prevent these risks [3]. Beyond resistance issues, the overconsumption of antibiotics can disrupt the gut microbiome— the complex community of beneficial microorganisms in the human digestive system. The gut microbiota plays a key role in digestion, nutrient absorption, vitamin synthesis, and immune function. Excessive antibiotic exposure disturbs this delicate balance, potentially leading to gastrointestinal problems such as diarrhea, inflammation, and increased susceptibility to chronic diseases. Hence, maintaining food safety and regulating antibiotic use in animal farming are critical steps toward protecting human health and preserving the efficacy of antibiotics for future generations[4]. Metronidazole is a widely used antibiotic and antiprotozoal agent frequently employed in both human medicine and veterinary practices. Due to its effectiveness against anaerobic bacteria and certain parasites, it is also used in livestock to prevent and treat infections. However, the overuse and misuse of metronidazole in food-producing animals can lead to residual traces of the drug in meat, milk, and other animal-derived products. Consumption of these contaminated foods poses significant risks to human health, including allergic reactions, disruption of normal gut flora, and the potential development of antibiotic-resistant microorganisms[5,6]. Excessive intake of metronidazole residues through food can also cause toxic side effects such as nausea, vomiting, neurological disorders, and in some cases, carcinogenic risks have been debated. Furthermore, chronic exposure to antibiotic residues complicates the treatment of infections in humans by promoting the spread of resistant strains. These concerns underline the critical need for accurate and rapid detection methods to monitor metronidazole residues in food products and ensure compliance with safety regulations[7,8]. Traditional detection techniques, although effective, often require expensive instruments, lengthy sample preparation, and skilled personnel. Therefore, the development of novel, simple, and sensitive methods—such as fluorescence-based sensors combined with advanced data analysis like machine learning—has become highly important. These innovative approaches enable faster, cost-effective, and on-site monitoring of metronidazole residues, which not only protect consumer health but also help in controlling antibiotic usage in the food industry. Several analytical techniques have been employed for the detection of the antibiotic metronidazole, including high-performance liquid chromatography (HPLC), electrochemical methods, UV-Vis spectroscopy, and immunoassays. While these methods offer acceptable accuracy, they are often expensive, time-consuming, and require sophisticated instruments and trained personnel. Moreover, some of these methods suffer from low sensitivity at trace levels and face challenges in complex matrices such as biological or food samples due to interference, which can compromise the reliability of the results [9–12]. Fluorescence-based detection techniques are highly regarded for their superior sensitivity, precision, and rapid analysis. These methods are cost-effective, suitable for small-volume samples, and often do not require complex equipment. Additionally, they provide fast response times and allow for real-time (online) monitoring. When integrated with fluorescent nanosensors, these systems can detect even trace amounts of metronidazole with excellent accuracy and minimal background interference [13]. Recently, carbon quantum dots (CQDs) synthesized from bio-based and waste-derived materials have gained significant attention due to their eco-friendly nature, low cost, and outstanding optical properties [14]. In this study, mold was utilized as a novel carbon precursor to produce CQDs. This approach not only aligns with the principles of green chemistry but also enhances the fluorescence properties of the sensor for sensitive detection of metronidazole. The use of mold as a raw material provides added value to biological waste and opens new pathways for developing high-performance, sustainable, and biocompatible fluorescent sensors [15,16]. In this study, a novel fluorescence-based sensor was developed using carbon quantum dots (CQDs) synthesized from mold as a sustainable carbon source. The CQDs were thoroughly characterized using multiple analytical techniques to confirm their structural and optical properties. The sensor demonstrated high sensitivity for metronidazole detection via a fluorescence quenching mechanism. To improve analytical performance, experimental design was employed to optimize sensing parameters, and machine learning algorithms were integrated to enhance prediction accuracy. This eco-friendly sensing platform provides a promising approach for efficient monitoring of metronidazole in biomedical and environmental contexts. Experimental 2.1. Reagents and materials The chemicals employed in this research, including copper nitrate, silver nitrate, and Artemisia absinthium, were sourced from Merck. Furthermore, phosphate-buffered saline (PBS) at pH 7.4, KCl, NaCl, MnCl₂, MgCl₂, CaCl₂, as well as bovine serum albumin (BSA) and glucose, were also obtained from Merck. A range of antibiotics, including Ofloxacin, Amoxicillin, Azithromycin, Ceftriaxone, Ciprofloxacin, Clamoxin, Cotrimoxazole, Tetracycline, Metronidazole, Doxycycline, Cefixime, and Levofloxacin, were likewise procured from Merck. The pH of the solutions was adjusted to 6 by utilizing 0.1M phosphoric acid, sodium phosphate, hydrochloric acid, and sodium hydroxide. All solutions were prepared using double-distilled water. 2.2 Synthesis of Carbon Quantum Dots (CQDs) from Mold Carbon quantum dots (CQDs) were synthesized using mold developed on spoiled verjuice (ab-ghooreh), a sour grape juice, which naturally formed after three months of storage under ambient conditions. The mold biomass was collected, dried at 60 °C, and then carbonized at 200 °C for 2 hours in a muffle furnace to obtain a black carbonaceous powder. The powder was then dispersed in deionized water, sonicated for 30 minutes at 40 kHz, and filtered through a 0.22 µm membrane to isolate the CQDs with strong fluorescent properties. 2.3 Characterization Techniques The structural and morphological properties of the synthesized CQDs were characterized using several analytical techniques. Fluorescence emission spectra were recorded using a PerkinElmer LS 45 fluorescence spectrophotometer. UV-Vis absorption measurements were carried out with a Cary 100 UV-Vis spectrophotometer. Functional groups were identified using a Thermo Avatar Fourier-transform infrared spectrometer (FTIR). Surface morphology and elemental analysis were performed using a TESCAN MIRA III field emission scanning electron microscope (FE-SEM, Philips brand) operating at 15 kV, equipped with energy-dispersive X-ray spectroscopy (EDX), elemental mapping, and line scan analysis. 2.4 Experimental Design and Optimization The design of experiments (DOE) was employed to assess the impact of key parameters on the experimental conditions using the central composite design (CCD) method. This approach aimed to optimize three independent factors: pH, time, and temperature. The CCD methodology involves conducting 2n axial runs, 2n factorial runs, and nc center runs (with six replicates), where n represents the number of parameters being studied. Given that the response of the sensor is directly influenced by the levels of these parameters and the synergistic interactions among them, a total of 20 experiments were conducted based on a 15-run CCD design. The ranges considered for optimization were pH levels from 4.0 to 10.0, reaction times from 0.5 to 5.0 min, and temperatures ranging from 34 to 60 °C, as detailed in Table 1 . A quadratic model was developed to describe the mathematical relationship among the three independent parameters, represented by the following equation: Y = β0 + (β1 × A) + (β2 × B) + (β3 × C) + ( β11 × A2) + ( β22 × B2) + ( β33 × C2) + (β12 × AB) + (β13 × AC) + (β23 × BC) (1) In the provided equation, Y represents the predicted response, and β0 denotes the model constant. The coefficients β1, β2, β3, β11, β22, β33, β12, β13, and β23 within the statistical model illustrate the linear, quadratic, and interactive effects of factors A (pH), B (Temperature), and C (Time) on the response, respectively. The Central Composite Design (CCD) was conducted using Design-Expert software (Version 11.1.1.0 USA) to analyze the experimental data. An Analysis of Variance (ANOVA) was conducted to evaluate the statistical significance with a 95% confidence interval, utilizing a p-value threshold of 0.05. In addition, regression analysis and response surface plots were used to identify the optimal conditions for assessing the CQDs sensor. Table 1 2.5 Fluorescence Sensing Procedure To evaluate the fluorescence intensity in the presence of varying concentrations of Metronidazole, a specific protocol was followed. Initially, 30 µg of the synthesized CQDs was added to 3 mL of a solution adjusted to pH 7. The fluorescence excitation wavelength was set at 450 nm, and the corresponding fluorescence spectra were recorded, focusing on an emission wavelength of 538 nm. Measurements were subsequently taken both in the presence and absence of Metronidazole, with the fluorescence intensities denoted as F₀ (without Metronidazole) and F (with Metronidazole). The ratio of these fluorescence intensities, expressed as F₀/F, was calculated to establish a correlation between the fluorescence response and the concentration of Metronidazole present in the solution. For the assessment of the detection limit, a signal-to-noise ratio of 3 was applied, which is a standard criterion in analytical chemistry. This threshold helps determine the lowest concentrations of Metronidazole that can be reliably detected using the CQDs fluorescence system. 2.6 Machine Learning Analysis To improve the prediction accuracy and robustness of the sensing platform, machine learning algorithms were implemented using Python (version X.X) and the Scikit-learn library. Models such as support vector regression (SVR), random forest regression, and artificial neural networks (ANN) were trained on the experimental data. Model performance was evaluated using R² (coefficient of determination), root mean square error (RMSE), and mean absolute error (MAE) to select the most suitable model for reliable quantification of metronidazole. 2.7. Preparation of sample In this study, metronidazole levels were examined using serum samples obtained from the Kermanshah Blood Bank. The samples were provided under standard institutional protocols and did not contain any identifiable personal information. To remove protein interferences, 5 mL of serum was mixed with 5 mL of 10% (w/v) trichloroacetic acid and centrifuged at 4000 rpm for 30 minutes. The resulting supernatant was filtered through a 0.45-μm Millipore membrane to eliminate particulate matter. The treated serum samples were then diluted tenfold with a universal buffer solution adjusted to pH 7. Subsequently, 300 μL of the prepared solution was combined with 30 µg of the CQDs-based nanosensor and varying concentrations of metronidazole. Fluorescence spectra were recorded at an emission wavelength of 523 nm with excitation at 440 nm. Results and discussion 3.1 Structural and Morphological Characterization Scanning electron microscopy (SEM) analysis also provided insightful details, revealing uniformly distributed spherical nanoparticles with an average size ranging from 5 to 10 nm. As illustrated in Fig. 1a , these images demonstrate a high degree of dispersion with minimal aggregation, indicating the successful synthesis of well-defined quantum dots. The uniform morphology is essential for consistent performance in sensing applications. Additionally, Fig. 1b depicts elemental mapping that illustrates the spatial distribution of various elements on the surface of the nanoparticles. From the mapping, it is evident that carbon (C) and oxygen (O are present in notably high densities, confirming the organic nature of the quantum dots. Moreover, the Energy Dispersive X-ray Spectroscopy (EDX) spectrum shown in Fig. 1c quantifies the elemental composition of the CQDs. The analysis indicates that carbon, oxygen, silicon account for significant percentages at 35.5%, 24.6%, and 25.9% respectively, while trace amounts of, Mg, Ca, and Na among other elements are also detected within the structure, highlighting the successful incorporation of metal ions. To further elucidate the distribution of surface elements. This comprehensive mapping effectively demonstrates the diverse elemental composition within the CQDs structure, underscoring its potential for multifaceted applications in biosensing and nanotechnology. FTIR spectra ( Fig. 2 ) revealed characteristic peaks corresponding to various surface functional groups. A broad absorption around 3400 cm⁻¹ was attributed to O–H and N–H stretching vibrations, indicating the presence of hydroxyl and amine groups. Peaks at ~1700 cm⁻¹ and ~1600 cm⁻¹ were assigned to C=O (carboxyl) and C=C stretching, respectively. These functional groups play a critical role in enhancing the solubility and fluorescence properties of CQDs. The UV-Vis absorption spectra of the mold-derived CQDs showed a main absorption band near 280 nm, attributed to π–π* transitions of C=C aromatic domains, and a shoulder around 330 nm due to n–π* transitions of C=O bonds. Upon excitation at 440 nm, the CQDs exhibited a strong green fluorescence emission peak centered at 520 nm. This large Stokes shift (80 nm) indicates efficient energy transfer and low self-absorption, which are favorable for sensing applications. The intense emission confirms the high quantum yield and good surface passivation of the CQDs synthesized from mold ( Fig. 3). In this study, we focused on the analysis of fluorescence spectra specifically designed for fluorescence-based sensors. The importance of obtaining the optimal fluorescence spectrum is crucial, particularly in detection and measurement applications. Therefore, we investigated various excitation wavelengths ranging from 250 nm to 460 nm. Our findings revealed that the best fluorescence spectrum was observed at 460 nm, corresponding to an excitation wavelength of 450 nm. In contrast, the other excitation wavelengths did not yield satisfactory intensity or sharpness, failing to achieve the desired quality and resolution. These results not only enhance the design of fluorescence sensors but also pave the way for the development of new applications in the field ) Fig. 4 .). Fig. 1 Fig. 2 Fig . 3 Fig. 4 3.2. Statistical analysis To optimize the fluorescence response (F₀/F) for metronidazole detection, a total of 16 experimental runs were designed using the Central Composite Design (CCD) under the framework of Response Surface Methodology (RSM). The experimental conditions and corresponding responses are summarized in Table 2 . The F₀/F values ranged from 1.00 to 1.66, reflecting the influence of pH (A), temperature (B), and time (C) on the fluorescence quenching behavior. CCD was employed for its efficiency in modeling complex interactions with a reduced number of experiments. To accurately capture the non-linear effects and interactions among the variables, a quadratic regression model was fitted to the data. As shown in Table 3 , the quadratic model demonstrated a high degree of correlation with the experimental data, as indicated by an adjusted R² value of 0.9646 and a predicted R² of 0.8665, confirming its predictive reliability. The resulting quadratic equation, expressed in terms of coded variables, is as follows: Y = 3.9 + (0.1625 × A) + (0.7366 × B) - (0.0297 × C) - (0.2233 × A²) - (0.5777 × B²) - (0.0536 × C²) + (0.0612 × AB) + (0.0662 × AC) - (0.1113 × BC) In this model, positive coefficients indicate synergistic effects that enhance the fluorescence response, while negative coefficients reflect antagonistic or diminishing effects. The significance of the quadratic term over the linear and interaction models was statistically validated by the p-values reported in Table 3 , where the quadratic model yielded a p-value < 0.0001, confirming its superiority. Furthermore, the lack-of-fit p-value (0.0005) suggests that the model is statistically significant with an acceptable deviation from the actual experimental data. Table 2 Table 3 3.3. ANOVA Analysis of variance (ANOVA) was applied to the quadratic model described earlier to evaluate the statistical significance of individual and interactive factors influencing the fluorescence response (F₀/F) of metronidazole. The ANOVA results, summarized in Table 4 , reveal that the overall model is highly significant, as evidenced by an F-value of 58.59 and a p-value < 0.0001, confirming the robustness of the regression model in describing the experimental data. Among the individual terms, the linear effects of pH (A) and time (B) exhibited strong statistical significance, with p-values < 0.0001. However, temperature (C) was found to have no significant individual effect (p = 0.8040). Regarding interaction effects, only the AB interaction (p = 0.0021) showed statistical significance, whereas AC and BC were not significant contributors (p > 0.76). In terms of curvature, the squared terms A² and B² were highly significant (p = 0.0001 and < 0.0001, respectively), indicating nonlinear relationships with the response. Conversely, the squared term C² was not significant (p = 0.5329), suggesting that quadratic effects of temperature have minimal influence on fluorescence variation. Despite a slight lack of fit detected (p = 0.0005), the model still explains a substantial portion of variability, supported by a low residual mean square (0.0022) and a high coefficient of determination (R² = 0.9130). This confirms the model’s predictive capability and suitability for response surface optimization. The final fitted quadratic equation, derived in terms of actual values of the variables, is presented below: Y = –0.6954 + (0.4200 × A) + (0.2581 × B) + (0.0170 × C) – (0.0357 × A²) – (0.0071 × B²) – (0.00023 × C²) + (0.0027 × AB) + (0.0017 × AC) – (0.00084 × BC) (2) This equation illustrates how each factor contributes to the fluorescence intensity and allows for effective prediction and optimization of conditions for metronidazole analysis. Table 4 3.4. 3D response surface plots To investigate the interactive effects of the studied variables on the fluorescence response (F₀/F), three-dimensional response surface plots were constructed based on the quadratic regression model. The plots illustrate the combined influence of two factors while the third is held constant at its central level. Fig . 5a shows the interactive effect of pH (A) and Time (B) on the fluorescence intensity. The curved surface and elliptical contour lines suggest a strong interaction between pH and reaction time. The response increases with increasing pH and time, reaching an optimal region before declining, indicating a nonlinear relationship. This trend may be attributed to the enhanced binding or quenching environment at optimal pH and exposure time. Fig. 5b displays the influence of pH (A) and Temperature (C). A clear curvature is observed, but the contour lines are more parallel, indicating a relatively weaker interaction between these two variables. The response increases with pH and decreases slightly with temperature at higher pH levels. This implies that while pH plays a dominant role, elevated temperatures may reduce the stability of the sensing complex. Fig. 5c illustrates the combined effect of Time (B) and Temperature (C). The surface is slightly curved, but the contours are nearly straight and parallel, suggesting a minimal interaction between time and temperature. The fluorescence response increases with time but remains relatively stable over the temperature range studied. This indicates that reaction time has a more pronounced effect compared to temperature under these conditions. Overall, these plots confirm that pH and time are the most influential parameters affecting the fluorescence response, with temperature showing a moderate to minimal effect. The optimal sensing conditions are located within the curved regions of the surfaces, supporting the use of response surface methodology (RSM) in identifying the best operational parameters. Fig. 5 3.5. Accuracy of the Model The accuracy and reliability of the quadratic model developed for predicting amoxicillin removal efficiency were evaluated using several key statistical parameters. The coefficient of determination (R²) was found to be 0.9814, indicating that approximately 98.14% of the variability in the response could be explained by the model. The adjusted R² value of 0.9646 confirms a strong fit, even after accounting for the number of predictors in the model. The predicted R² of 0.8665 is in reasonable agreement with the adjusted R², with a difference of less than 0.2. This close alignment demonstrates the model’s robustness and its ability to generalize well to new data points. Additionally, the standard deviation (SD) of 0.0464 and a relatively low coefficient of variation (C.V.) of 3.23% reflect high precision and low dispersion of the experimental data around the mean. Furthermore, the adequate precision value, which measures the signal-to-noise ratio, was calculated to be 19.0136—substantially greater than the minimum acceptable value of 4. This suggests that the model has a strong signal and is capable of navigating the design space efficiently. Collectively, these statistical indicators validate the high quality and predictive power of the model, supporting its suitability for optimizing the operational parameters in the amoxicillin removal process ( Table 5 ). Table 5 3.6 Model Diagnostics and Assumptions Verification To evaluate the reliability and statistical adequacy of the developed model for predicting amoxicillin removal efficiency, several diagnostic plots were analyzed (Fig. 4a–d). These visual tools are essential for validating the underlying assumptions of regression modeling, including normality, constant variance, and independence of residuals. Fig . 6a (Residuals vs. Predicted Values) illustrates the distribution of residuals around the predicted responses. The lack of any systematic pattern and the relatively uniform spread of points above and below the zero line suggest homoscedasticity—indicating constant variance across the range of predicted values. This confirms that the model errors are not dependent on the magnitude of predictions, which is a critical assumption in regression analysis. Fig . 6b (Normal Probability Plot of Residuals) examines the normality of residuals. The residuals follow a near-linear trend along the reference line, indicating that they are approximately normally distributed. This supports the appropriateness of the regression model and suggests that no major deviations from normality exist. Fig . 6c (Predicted vs. Actual Values) demonstrates the relationship between the observed and model-predicted responses. Most points lie close to the diagonal line, confirming a strong agreement between actual and predicted values. While the model performs exceptionally well for low to mid-range response levels, it shows slightly less accuracy at the highest removal efficiencies. This could indicate the need for more training data in that range or the potential for minor model refinement. Fig . 6d (Residuals vs. Run Order) helps assess the independence and randomness of residuals over time. The random scatter of residuals around zero with no evident trend or cyclic pattern confirms the absence of time-related bias or systematic experimental errors. However, one or two points that slightly deviate may merit further investigation to ensure they are not influential outliers. In summary, these diagnostic plots collectively demonstrate that the regression model satisfies the key statistical assumptions. The model exhibits strong predictive capabilities, consistent residual behavior, and robustness across different runs, reinforcing its suitability for optimizing amoxicillin removal processes in future applications. Fig.6 3.7. Machine Learning Models for Predicting Metronidazole Detection To assess the performance of machine learning models in predicting metronidazole detection, three algorithms were investigated: Linear Regression, Random Forest, and Support Vector Regression (SVR). Their predictive capabilities were evaluated based on statistical metrics including the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE). The comparison of model performance is visualized in Fig.7(a), which presents a 3D scatter plot based on R², MSE, and MAE. As shown, the Random Forest model significantly outperforms the other two models, occupying a favorable position with high R² (0.9633) and low error values (RMSE = 0.0446, MAE = 0.0355). This indicates strong agreement between predicted and actual values with minimal error. On the other hand, the Linear Regression and SVR models demonstrated much weaker performance. The Linear Regression model showed a negative R² value (-0.1103), along with the highest RMSE (0.2451) and MAE (0.1988), implying that it fails to capture the true relationship between input variables and target values. Similarly, SVR also yielded a negative R² (-0.0163) with relatively high error metrics (RMSE = 0.2345, MAE = 0.1904), though slightly better than Linear Regression. Further insights can be seen in Fig.7(b) , which plots actual vs. predicted values for all three models. Here again, the predictions made by the Random Forest model lie very close to the 45-degree reference line, indicating high predictive accuracy. In contrast, the predictions from Linear Regression and SVR display noticeable deviations from the ideal line, highlighting their limited predictive power. the Random Forest model demonstrated the most reliable and accurate performance in predicting metronidazole detection. These findings highlight the advantage of non-linear ensemble methods over traditional linear or kernel-based models in capturing complex patterns within analytical chemistry datasets. Fig.7 3.8. Method Selectivity Selectivity is a key parameter in evaluating the performance of a fluorescence-based sensor, particularly when used in complex matrices. In this study, the selectivity of the CQDs sensor for metronidazole detection was systematically investigated against a panel of potential interfering substances. As illustrated in Fig.8 , various compounds, including common pharmaceuticals and biological molecules, were individually tested at a concentration of 1000 μM under the optimized sensing conditions. The fluorescence response was expressed as the F₀/F ratio, where F₀ is the fluorescence intensity in the absence of the analyte and F is the intensity in the presence of each test compound. The results clearly indicate that metronidazole produces a distinct and significantly higher fluorescence quenching effect compared to the other substances tested. While most interferents resulted in minimal or negligible changes in fluorescence, only metronidazole caused a marked decrease in signal intensity. This pronounced difference highlights the excellent selectivity of the CQDs sensor towards metronidazole. The sharp contrast in F₀/F values across the tested compounds underscores the sensor's ability to discriminate metronidazole even in the presence of structurally or chemically similar substances. Such selectivity is essential for reliable detection in real-world applications, and these results confirm that the developed CQDs sensor is a highly selective and robust platform for accurate metronidazole quantification. Fig. 8 3.9. Calibration and Detection of Metronidazole The CQDs nanocomposite exhibited a notable fluorescence enhancement at 523 nm upon excitation at 440 nm in the presence of metronidazole, forming the basis for its application as a novel fluorescence nanoprobe. To evaluate its analytical sensitivity, different concentrations of metronidazole were introduced under optimized experimental conditions. As depicted in Fig. 9A , a clear and direct linear increase in fluorescence intensity was observed with rising metronidazole concentrations within the range of 97.4 to 3215.3 µM. This linear relationship confirms the probe's capacity for consistent and quantifiable detection of metronidazole across this dynamic range. To further validate the performance of the sensor, a calibration curve was established by plotting the ratio of fluorescence intensity in the absence (F₀) and presence (F) of metronidazole (F₀/F) versus metronidazole concentration. As shown in Figure 9B, the resulting plot demonstrated excellent linearity with a correlation coefficient of R² = 0.9769, indicating high precision and reliability in response. The limit of detection (LOD) for metronidazole was calculated to be 50.5 µM, based on a signal-to-noise ratio of 3. This relatively low LOD, combined with the strong linear correlation, highlights the high sensitivity and robustness of the developed CQDs nanoprobe. Furthermore, the analytical performance of this sensor was benchmarked against previously reported quantum dot-based sensors for metronidazole detection. The comparison emphasizes the superior capabilities of the present approach in terms of sensitivity and dynamic range, confirming its promising potential for practical applications in metronidazole analysis. Fig. 9 3.10. Interference Study for Metronidazole Detection To ensure the specificity of the CQDs-based nanoprobe for metronidazole detection, a comprehensive interference study was conducted. Various potentially interfering substances were examined, including PBS (pH 7.4), KCl, NaCl, MnCl₂, MgCl₂, CaCl₂, bovine serum albumin (BSA), and glucose, which are commonly encountered in biological or environmental samples. Each of these compounds was tested at concentrations significantly higher than those of metronidazole to rigorously challenge the sensor’s performance under realistic conditions (Fig. 10). The findings reveal that the fluorescence response of the CQDs/Ag/Cu nanoprobe toward metronidazole remains stable and consistent even in the presence of these potential interferents. The negligible variation in the sensor’s output suggests a high level of selectivity and interference resistance, confirming that the presence of common ions, proteins, and sugars does not significantly affect metronidazole detection. This robustness against interference highlights the practical applicability of the CQDs nanoprobe, particularly in complex matrices such as biological fluids or environmental samples, where such compounds are prevalent. The ability to detect metronidazole accurately without significant interference eliminates the necessity for extensive sample pretreatment, further enhancing the probe's value for real-world analytical applications. Fig. 10 3.11. Application To evaluate the practical applicability of the developed CQDs-based nanoprobe for the detection of metronidazole in real biological samples, serum samples was chosen as the representative matrix. The serum samples underwent an initial preparation process to ensure they were suitable for analysis. This included isolating the serum and subsequently diluting it tenfold using a buffer solution adjusted to pH 7. To validate the method’s accuracy, a standard spiking approach was employed. Known concentrations of metronidazole were added to the diluted serum samples under optimized experimental conditions. This allowed for the assessment of the nanoprobe’s performance in complex biological environments. Following the spiking process, fluorescence intensity measurements were taken to determine the concentration of metronidazole present in the samples. The experimental results, presented in Table 6 , indicate excellent analytical recovery, with values of 100.12%, 100.7%, and 100.5%. Moreover, the method exhibited strong precision, as reflected in low standard deviation values of 0.301, 0.625, and 0.740. These findings confirm the reliability and reproducibility of the CQDs-based sensor for metronidazole quantification in serum. Additionally, a comparison with previously reported quantum dot-based detection systems, as shown in Table 7 , highlights the superior sensitivity of the proposed CQDs probe. Specifically, the lower limit of detection observed in this study underscores the enhanced sensing capabilities of the CQDs nanocomposite, likely due to its improved fluorescence properties and stability in biological matrices [17–19]. Table 7 Table 8 Conclusions In this study, a highly sensitive and selective carbon quantum dot (CQDs)-based fluorescent nanoprobe was successfully developed for the detection of metronidazole. Synthesized via a hydrothermal method with mold as the carbon source, this nanocomposite showcases remarkable properties upon thorough characterization using techniques such as scanning electron microscopy (SEM), Fourier-transform infrared spectroscopy (FTIR), and transmission electron microscopy (TEM). The sensor exhibited excellent photoluminescent behavior, with a significant fluorescence response upon interaction with metronidazole, enabling precise quantification. Among the machine learning models evaluated for predictive analysis, the Random Forest algorithm demonstrated superior performance, achieving high accuracy and minimal error in comparison to linear regression and SVR models. This highlights the potential of integrating data-driven approaches for optimizing sensor response and analytical prediction. The CQDs nanoprobe showed strong selectivity toward metronidazole, even in the presence of structurally similar or commonly encountered interfering compounds. Its robustness was further validated through interference studies, where the presence of excess ions, proteins, and sugars did not significantly affect detection performance. Analytical calibration revealed a strong linear response over a wide concentration range, with a limit of detection as low as 50.5 µM. The applicability of the method was confirmed through successful detection of metronidazole in spiked serum samples, with high recovery rates and low standard deviations, underscoring the method’s accuracy and precision in complex biological matrices. Overall, the developed CQDs-based sensor provides a rapid, reliable, and cost-effective platform for metronidazole detection. Its excellent analytical performance, combined with high selectivity and compatibility with real samples, makes it a promising candidate for use in pharmaceutical, clinical, and environmental monitoring applications. Declarations Acknowledgements The authors acknowledge the support and laboratory facilities provided by Islamic Azad University of Karaj and Sanandaj. Author contributions Kimiya Khandestana: Conducted the experimental work and collected fluorescence data. Azar Sabukbara: Assisted in experimental design and sample preparation. Bahareh Rahimian Zarif: Conceived and designed the study, supervised the research, performed data analysis, and critically revised the manuscript. Nahid Haghnazari: Contributed to data interpretation and manuscript editing. Nasser Harzandi: Provided technical support and contributed to material characterization. All authors reviewed and approved the final version of the manuscript. Bahareh Rahimian Zarif is the corresponding author and will serve as the main contact throughout the review process. Competing interests Kimiya Khandestana, Azar Sabukbara, Bahareh Rahimian Zarif, Nahid Haghnazari, and Nasser Harzandi declare no competing interests. Data availability All data generated or analyzed during this study are available from the corresponding author upon reasonable request. Ethics declarations All serum samples used in this study were obtained from the Kermanshah Blood Bank in accordance with institutional guidelines. The samples were anonymized and commercially provided for research purposes, without direct involvement of human subjects or collection of personal data. Therefore, ethical approval and informed consent were not required for this study. Consent to participate / Consent for publication Not applicable. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References Shahid, A. et al. Antibiotic residues in food chains; impact on the environment and human health: a review. Applied Ecology & Environmental Research 19, (2021). Kyuchukova, R. Antibiotic residues and human health hazard-review. Bulgarian Journal of Agricultural Science 26, (2020). Van Boeckel, T. P. et al. Global trends in antimicrobial use in food animals. Proceedings of the National Academy of Sciences 112, 5649–5654 (2015). Francino, M. Antibiotics and the human gut microbiome: dysbioses and accumulation of resistances. Frontiers in microbiology 6, 164577 (2016). Ljubojević Pelić, D. et al. Antibiotic Residues in Cultured Fish: Implications for Food Safety and Regulatory Concerns. Fishes 9, 484 (2024). Baynes, R. E. et al. Health concerns and management of select veterinary drug residues. Food and chemical toxicology 88, 112–122 (2016). Baynes, R. E. et al. Health concerns and management of select veterinary drug residues. Food and chemical toxicology 88, 112–122 (2016). Kyuchukova, R. Antibiotic residues and human health hazard-review. Bulgarian Journal of Agricultural Science 26, (2020). Karami, C. & Taher, M. A. Colorimetric Sensor of Cobalt Ions in Aqueous Solution Using Gold Nanoparticles Modified with Glycyrrhizic Acid. Plasmonics 13, (2018). Deymehkar, E., Taher, M. A., Karami, C. & Arman, A. Synthesis of SPR Nanosensor using Gold Nanoparticles and its Application to Copper (II) Determination. Silicon 10, (2018). Karami, C., Mehr, S. Y., Deymehkar, E. & Taher, M. A. Naked Eye Detection of Cr 3+ and Co 2+ Ions by Gold Nanoparticle Modified with Azomethine. Plasmonics 13, (2018). Karami, C., Taher, M. A. & Shahlaei, M. A simple method for determination of mercury (II) ions by PNBS-doped carbon dots as a fluorescent probe. Journal of Materials Science: Materials in Electronics 31, 5975–5983 (2020). Jain, P. K., Huang, X., El-Sayed, I. H. & El-Sayed, M. A. Noble metals on the nanoscale: optical and photothermal properties and some applications in imaging, sensing, biology, and medicine. Accounts of chemical research 41, 1578–1586 (2008). Kneipp, J., Kneipp, H., Wittig, B. & Kneipp, K. Novel optical nanosensors for probing and imaging live cells. Nanomedicine: Nanotechnology, Biology and Medicine 6, 214–226 (2010). Hou, J. et al. Rapid microwave-assisted synthesis of molecularly imprinted polymers on carbon quantum dots for fluorescent sensing of tetracycline in milk. Talanta 146, 34–40 (2016). Liu, G. et al. In-situ hydrothermal synthesis of molecularly imprinted polymers coated carbon dots for fluorescent detection of bisphenol A. Sensors and Actuators B: Chemical 228, 302–307 (2016). Alvarenga, L. M. et al. Preparation of a composite sensor based on a fluorescent and magnetic molecular imprint polymer for metronidazole extraction–detection. Journal of Molecular Liquids 390, 123027 (2023). Xia, Z. & Li, Q. Application of Metronidazole detection by antibiotic ampicillin sodium based-carbon quantum dots. International Journal of Environmental Analytical Chemistry 102, 4178–4190 (2022). Ren, G. et al. Efficient preparation of nitrogen-doped fluorescent carbon dots for highly sensitive detection of metronidazole and live cell imaging. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 234, 118251 (2020). Tables Table 1: Experimental parameters and levels in the 16 CCD for the optimization of pH, Temp, and Time Factor Name Level Low Level High Level Std. Dev. Coding A pH 7.00 4.00 10.00 0.0000 Actual B Time 2.75 0.5000 5.00 0.0000 Actual C Temperatures 47.00 34.00 60.00 0.0000 Actual Table 2: Experiment runs and responses for optimizing parameters evaluation Factor 1 Factor 2 Factor 3 Response 1 Run A: pH C: Time B: Temp F 0 /F 1 10 0.5 34 1.1 2 10 5 34 1.62 3 7 2.75 47 1.65 4 10 5 60 1.64 5 7 2.75 47 1.65 6 12 2.75 47 1.65 7 7 2.75 47 1.66 8 7 2.75 47 1.65 9 2 2.75 47 1.25 10 7 0.2 47 1 11 4 5 34 1.35 12 7 6.5 47 1.35 13 4 0.5 34 1.1 14 7 2.75 25 1.63 15 7 2.75 47 1.64 16 10 0.5 60 1.1 Table 3: Model summary statistic. Source Sequential p-value Lack of Fit p-value Adjusted R² Predicted R² Linear 0.0618 < 0.0001 0.2395 -0.0058 2FI 0.8772 < 0.0001 0.1102 -0.8665 Quadratic < 0.0001 0.0005 0.9646 0.8665 Suggested Cubic 0.0005 0.9983 Aliased Table 4 : ANOVA for response surface quadratic model for F 0 /F Source Sum of Squares df Mean Square F-value p-value Model 1.14 9 0.1263 58.59 < 0.0001 significant A-pH 0.1074 1 0.1074 49.81 < 0.0001 B-Time 0.5344 1 0.5344 247.83 < 0.0001 C-teTmperatures 0.0001 1 0.0001 0.0649 0.8040 AB 0.0365 1 0.0365 16.90 0.0021 AC 0.0000 1 0.0000 0.0000 1.0000 BC 0.0002 1 0.0002 0.0928 0.7670 A² 0.0753 1 0.0753 34.91 0.0001 B² 0.6127 1 0.6127 284.12 < 0.0001 C² 0.0009 1 0.0009 0.4172 0.5329 Residual 0.0216 10 0.0022 Lack of Fit 0.0210 5 0.0042 39.43 0.0005 significant Pure Error 0.0005 5 0.0001 Cor Total 1.16 19 Table 5: Standard deviation and R 2 of the response. Std. Dev. 0.0464 R² 0.9814 Mean 1.44 Adjusted R² 0.9646 C.V. % 3.23 Predicted R² 0.8665 Adeq Precision 19.0136 Table 6: Determination of metronidazole concentration in real samples. By fluorescence method (n = 5) sample Added (µM) found(µM) Recovery (%) RSD% 1 2 3 100 98.9 98.9 1.46 500 489.925 97.99 5.53 1000 1018 101.8 2.11 Table 7: Comparison of the performances of various sensors for detection of cefixim Probe Linear range LOD Antibiotics Real sample [Ref] sodium based-carbon quantum dots 20–100 μM 62.5 nM metronidazole water samples [17] molecular imprint polymer for metronidazole extraction–detection 5.0-60.0 μM 1.28 μM metronidazole real samples [18] nitrogen-doped fluorescent carbon dots 0.5–22 μM 0.22 μM metronidazole urine samples [19] CQDs of mod 97.4 -3215.3 µM µM 0.2 nM metronidazole serum samples This work Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7721529","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":531787299,"identity":"0b38424b-adfb-499d-ad72-837aa03027c4","order_by":0,"name":"Kimiya Khandestan","email":"","orcid":"","institution":"Islamic Azad University","correspondingAuthor":false,"prefix":"","firstName":"Kimiya","middleName":"","lastName":"Khandestan","suffix":""},{"id":531787301,"identity":"798c7c68-b68a-4534-9530-bce3a7ef8ebc","order_by":1,"name":"Azar Sabukbar","email":"","orcid":"","institution":"Islamic Azad University","correspondingAuthor":false,"prefix":"","firstName":"Azar","middleName":"","lastName":"Sabukbar","suffix":""},{"id":531787303,"identity":"47b5fb60-9141-41df-b770-53ee874184c7","order_by":2,"name":"Bahareh Rahimian Zarif","email":"data:image/png;base64,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","orcid":"","institution":"Islamic Azad University","correspondingAuthor":true,"prefix":"","firstName":"Bahareh","middleName":"Rahimian","lastName":"Zarif","suffix":""},{"id":531787304,"identity":"1626f310-e152-4a9a-ba74-85158fdee525","order_by":3,"name":"Nahid Haghnazari","email":"","orcid":"","institution":"Islamic Azad University","correspondingAuthor":false,"prefix":"","firstName":"Nahid","middleName":"","lastName":"Haghnazari","suffix":""},{"id":531787305,"identity":"608cb167-7125-47dc-929b-abca8858a41f","order_by":4,"name":"Nasser Harzandi","email":"","orcid":"","institution":"Islamic Azad University","correspondingAuthor":false,"prefix":"","firstName":"Nasser","middleName":"","lastName":"Harzandi","suffix":""}],"badges":[],"createdAt":"2025-09-26 11:53:34","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7721529/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7721529/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94359552,"identity":"a71c0a9d-8244-4e86-8a3e-d498bdd63342","added_by":"auto","created_at":"2025-10-27 13:02:09","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2834506,"visible":true,"origin":"","legend":"","description":"","filename":"figs.docx","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/329d747da44cfc3e6e8ed75b.docx"},{"id":94358644,"identity":"0e268384-1265-4277-9d0a-401893b7cd9e","added_by":"auto","created_at":"2025-10-27 13:01:26","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":49513,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/9c24797f486f9074ff1e4e74.docx"},{"id":94359417,"identity":"2efe1437-b715-480b-9bb4-21bade4aa3a5","added_by":"auto","created_at":"2025-10-27 13:02:02","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23747,"visible":true,"origin":"","legend":"","description":"","filename":"tabels.docx","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/400097a97d847f7e43f36391.docx"},{"id":94358811,"identity":"8af59a25-69cb-4269-af2a-33fc8da95bc1","added_by":"auto","created_at":"2025-10-27 13:01:33","extension":"json","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6540,"visible":true,"origin":"","legend":"","description":"","filename":"35c76b3a76bd4a2cb7afb464e3c6a93f.json","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/d03e73bee0f8a628fcc3b89d.json"},{"id":94359498,"identity":"f59b46a2-68ca-4368-be60-fbff46a40e78","added_by":"auto","created_at":"2025-10-27 13:02:06","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":619008,"visible":true,"origin":"","legend":"\u003cp\u003ea) SEM images, b) mapping imaging, c) EDS of CQDs\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/8a180e590fd94ea98384a97a.png"},{"id":94358865,"identity":"cd334ec2-91b0-42ff-b8a2-9fd160c811ae","added_by":"auto","created_at":"2025-10-27 13:01:39","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":5713,"visible":true,"origin":"","legend":"\u003cp\u003eFTIR spectra of CQDs\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/825e520723e04bd13712fba3.png"},{"id":94359495,"identity":"f9136a72-d30c-43a6-bb42-43dd455ea84a","added_by":"auto","created_at":"2025-10-27 13:02:05","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":5713,"visible":true,"origin":"","legend":"\u003cp\u003ea) Uv. Vis spectra of CQDs, b) Uv. Vis spectra of CQDs-mold\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/600a4eb778f1cc46f9e72c23.png"},{"id":94359280,"identity":"1a58fc83-1943-495f-9282-d4a026bb8863","added_by":"auto","created_at":"2025-10-27 13:01:55","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":5713,"visible":true,"origin":"","legend":"\u003cp\u003eFluorescence spectrum of CQDs of mold\u003cstrong\u003e \u003c/strong\u003ein excitation with different wavelengths from 250 nm to 460 nm\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/77434bf5ae43cf418158f40f.png"},{"id":94489251,"identity":"dc3ad940-4740-44a9-92cb-d093339f063a","added_by":"auto","created_at":"2025-10-27 17:03:59","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":5713,"visible":true,"origin":"","legend":"\u003cp\u003eThe 3D plot for interaction effects between pH and temperature at 2.75 min (a), pH and time at 47 ◦C (b) and temp and time at pH = 7 (c) for the response surface quadratic model.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/f9047157f270fa4b75b87099.png"},{"id":94359376,"identity":"9f86c103-1162-4b24-a1f8-e05468352ae5","added_by":"auto","created_at":"2025-10-27 13:01:59","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":5713,"visible":true,"origin":"","legend":"\u003cp\u003eDiagnostic plots for the Central Composite Design (CCD) methodology adequacy (a ) Residuals vs. Predicted plot (b) The Normal Plot of Residuals (c) residual versus fitted values and (d) Residuals vs. Run plot.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/20d069648b5b62aaf1efe4d1.png"},{"id":94359235,"identity":"4d603912-bcb6-49e2-821d-0553f9bea4c2","added_by":"auto","created_at":"2025-10-27 13:01:53","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":258726,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e(a)\u003c/strong\u003e 3D Scatter Plot of Model Performance. The plot visualizes the Mean Squared Error (MSE) and R-squared values for each model, with the z-axis representing the model indices, (b) Model Performance: Actual vs. Predicted values\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/2fc6e29c335b5de39c624dde.png"},{"id":94358536,"identity":"6fb615ad-da66-4923-bb7e-4e895afc66c3","added_by":"auto","created_at":"2025-10-27 13:01:17","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":37191,"visible":true,"origin":"","legend":"\u003cp\u003eFluorescence response of CQDs to various compounds\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/f966781714d5e19024eba66a.png"},{"id":94359252,"identity":"dc7bb8b2-7935-4ad8-9d7d-91daa9fca0d4","added_by":"auto","created_at":"2025-10-27 13:01:53","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":186524,"visible":true,"origin":"","legend":"\u003cp\u003ea) Change in the fluorescence intensity of the CQDs/Ag/Cu compound in the presence of different concentrations of cefixime from 97.4 to 3215.3 μM. b) Increasing the relative sensitivity of the detection system with different concentrations of metronidazole, 97.4 to 3215.3 µM\u003c/p\u003e","description":"","filename":"9.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/908bbdaac47e4fe036558441.png"},{"id":94359358,"identity":"c49bf085-6563-4b19-abf2-a11d7808850b","added_by":"auto","created_at":"2025-10-27 13:01:57","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":39550,"visible":true,"origin":"","legend":"\u003cp\u003eF\u003csub\u003eo\u003c/sub\u003e/F ratio of CQDs/Ag/Cu in the presence of various compounds of blue rods represent metronidazole alone, orang rods represent a mixture of metronidazole (295 µM) with other compounds and gray rods represent a mixture of metronidazole (1000 µM) with other compounds.\u003c/p\u003e","description":"","filename":"10.png","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/fce12f6a84e2d3cc21716552.png"},{"id":96638172,"identity":"de555155-de75-4987-9c5a-476d4a87491c","added_by":"auto","created_at":"2025-11-24 13:54:04","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2322932,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7721529/v1/adf45ba0-afb0-4397-81b8-a8b44b20efbc.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine Learning and Experimental Design-Driven Fluorescence Detection of Metronidazole in Food Samples Using Mold-Derived Carbon Quantum Dots","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEnsuring the safety and quality of food is crucial for maintaining public health and promoting proper growth and development. Safe food not only requires adherence to hygiene standards during production and storage but also must be free from harmful contaminants, including antibiotic residues. In recent decades, the widespread use of antibiotics in agriculture and animal husbandry to prevent and treat infections has increased significantly. However, excessive and improper use of these drugs has led to contamination of food products with antibiotic residues, posing serious health risks to consumers[1,2].\u003c/p\u003e\n\u003cp\u003eWhile antibiotics play a vital role in treating bacterial infections, their overuse in livestock production has contributed to the emergence of antibiotic-resistant bacteria. This resistance not only makes infections harder and more expensive to treat but also threatens global health security. Antibiotic residues present in animal-derived foods such as meat, milk, and eggs can cause allergic reactions and other adverse health effects in humans. Therefore, monitoring and controlling antibiotic use in food production systems is essential to prevent these risks\u003cspan dir=\"RTL\"\u003e\u0026nbsp;\u003c/span\u003e[3].\u003c/p\u003e\n\u003cp\u003eBeyond resistance issues, the overconsumption of antibiotics can disrupt the gut microbiome\u0026mdash; the complex community of beneficial microorganisms in the human digestive system. The gut microbiota plays a key role in digestion, nutrient absorption, vitamin synthesis, and immune function. Excessive antibiotic exposure disturbs this delicate balance, potentially leading to gastrointestinal problems such as diarrhea, inflammation, and increased susceptibility to chronic diseases. Hence, maintaining food safety and regulating antibiotic use in animal farming are critical steps toward protecting human health and preserving the efficacy of antibiotics for future generations[4].\u003c/p\u003e\n\u003cp\u003eMetronidazole is a widely used antibiotic and antiprotozoal agent frequently employed in both human medicine and veterinary practices. Due to its effectiveness against anaerobic bacteria and certain parasites, it is also used in livestock to prevent and treat infections. However, the overuse and misuse of metronidazole in food-producing animals can lead to residual traces of the drug in meat, milk, and other animal-derived products. Consumption of these contaminated foods poses significant risks to human health, including allergic reactions, disruption of normal gut flora, and the potential development of antibiotic-resistant microorganisms[5,6].\u003cspan dir=\"RTL\"\u003e\u0026nbsp;\u003c/span\u003eExcessive intake of metronidazole residues through food can also cause toxic side effects such as nausea, vomiting, neurological disorders, and in some cases, carcinogenic risks have been debated. Furthermore, chronic exposure to antibiotic residues complicates the treatment of infections in humans by promoting the spread of resistant strains. These concerns underline the critical need for accurate and rapid detection methods to monitor metronidazole residues in food products and ensure compliance with safety regulations[7,8].\u003c/p\u003e\n\u003cp\u003eTraditional detection techniques, although effective, often require expensive instruments, lengthy sample preparation, and skilled personnel. Therefore, the development of novel, simple, and sensitive methods\u0026mdash;such as fluorescence-based sensors combined with advanced data analysis like machine learning\u0026mdash;has become highly important. These innovative approaches enable faster, cost-effective, and on-site monitoring of metronidazole residues, which not only protect consumer health but also help in controlling antibiotic usage in the food industry.\u003cspan dir=\"RTL\"\u003e\u0026nbsp;\u003c/span\u003eSeveral analytical techniques have been employed for the detection of the antibiotic metronidazole, including high-performance liquid chromatography (HPLC), electrochemical methods, UV-Vis spectroscopy, and immunoassays. While these methods offer acceptable accuracy, they are often expensive, time-consuming, and require sophisticated instruments and trained personnel. Moreover, some of these methods suffer from low sensitivity at trace levels and face challenges in complex matrices such as biological or food samples due to interference, which can compromise the reliability of the results\u003cspan dir=\"RTL\"\u003e\u0026nbsp;\u003c/span\u003e[9\u0026ndash;12].\u003c/p\u003e\n\u003cp\u003eFluorescence-based detection techniques are highly regarded for their superior sensitivity, precision, and rapid analysis. These methods are cost-effective, suitable for small-volume samples, and often do not require complex equipment. Additionally, they provide fast response times and allow for real-time (online) monitoring. When integrated with fluorescent nanosensors, these systems can detect even trace amounts of metronidazole with excellent accuracy and minimal background interference\u003cspan dir=\"RTL\"\u003e\u0026nbsp;\u003c/span\u003e[13].\u003c/p\u003e\n\u003cp\u003eRecently, carbon quantum dots (CQDs) synthesized from bio-based and waste-derived materials have gained significant attention due to their eco-friendly nature, low cost, and outstanding optical properties\u003cspan dir=\"RTL\"\u003e\u0026nbsp;\u003c/span\u003e[14]. In this study, mold was utilized as a novel carbon precursor to produce CQDs. This approach not only aligns with the principles of green chemistry but also enhances the fluorescence properties of the sensor for sensitive detection of metronidazole. The use of mold as a raw material provides added value to biological waste and opens new pathways for developing high-performance, sustainable, and biocompatible fluorescent sensors\u003cspan dir=\"RTL\"\u003e\u0026nbsp;\u003c/span\u003e[15,16].\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;In this study, a novel fluorescence-based sensor was developed using carbon quantum dots (CQDs) synthesized from mold as a sustainable carbon source. The CQDs were thoroughly characterized using multiple analytical techniques to confirm their structural and optical properties. The sensor demonstrated high sensitivity for metronidazole detection via a fluorescence quenching mechanism. To improve analytical performance, experimental design was employed to optimize sensing parameters, and machine learning algorithms were integrated to enhance prediction accuracy. This eco-friendly sensing platform provides a promising approach for efficient monitoring of metronidazole in biomedical and environmental contexts.\u003c/p\u003e"},{"header":"Experimental ","content":"\u003cp\u003e2.1. Reagents and materials\u003c/p\u003e\n\u003cp\u003eThe chemicals employed in this research, including copper nitrate, silver nitrate, and\u0026nbsp;Artemisia absinthium, were sourced from Merck. Furthermore, phosphate-buffered saline (PBS) at pH 7.4, KCl, NaCl, MnCl₂, MgCl₂, CaCl₂, as well as bovine serum albumin (BSA) and glucose, were also obtained from Merck. A range of antibiotics, including Ofloxacin, Amoxicillin, Azithromycin, Ceftriaxone, Ciprofloxacin, Clamoxin, Cotrimoxazole, Tetracycline, Metronidazole, Doxycycline, Cefixime, and Levofloxacin, were likewise procured from Merck. The pH of the solutions was adjusted to 6 by utilizing 0.1M phosphoric acid, sodium phosphate, hydrochloric acid, and sodium hydroxide. All solutions were prepared using double-distilled water.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2 Synthesis of Carbon Quantum Dots (CQDs) from Mold\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCarbon quantum dots (CQDs) were synthesized using mold developed on spoiled verjuice (ab-ghooreh), a sour grape juice, which naturally formed after three months of storage under ambient conditions. The mold biomass was collected, dried at 60 °C, and then carbonized at 200 °C for 2 hours in a muffle furnace to obtain a black carbonaceous powder. The powder was then dispersed in deionized water, sonicated for 30 minutes at 40 kHz, and filtered through a 0.22 µm membrane to isolate the CQDs with strong fluorescent properties.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3 Characterization Techniques\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe structural and morphological properties of the synthesized CQDs were characterized using several analytical techniques. Fluorescence emission spectra were recorded using a PerkinElmer LS 45 fluorescence spectrophotometer. UV-Vis absorption measurements were carried out with a Cary 100 UV-Vis spectrophotometer. Functional groups were identified using a Thermo Avatar Fourier-transform infrared spectrometer (FTIR). Surface morphology and elemental analysis were performed using a TESCAN MIRA III field emission scanning electron microscope (FE-SEM, Philips brand) operating at 15 kV, equipped with energy-dispersive X-ray spectroscopy (EDX), elemental mapping, and line scan analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4 Experimental Design and Optimization\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe design of experiments (DOE) was employed to assess the impact of key parameters on the experimental conditions using the central composite design (CCD) method. This approach aimed to optimize three independent factors: pH, time, and temperature. The CCD methodology involves conducting 2n axial runs, 2n factorial runs, and nc center runs (with six replicates), where n represents the number of parameters being studied. Given that the response of the sensor is directly influenced by the levels of these parameters and the synergistic interactions among them, a total of 20 experiments were conducted based on a 15-run CCD design. The ranges considered for optimization were pH levels from 4.0 to 10.0, reaction times from 0.5 to 5.0 min, and temperatures ranging from 34 to 60 °C, as detailed in \u003cstrong\u003eTable 1\u003c/strong\u003e. A quadratic model was developed to describe the mathematical relationship among the three independent parameters, represented by the following equation:\u003c/p\u003e\n\u003cp\u003eY = β0 + (β1 × A) + (β2 × B) + (β3 × C) + ( β11 × A2) + ( β22 × B2) + ( β33 × C2) + (β12 × AB) + (β13 × AC) + (β23 × BC) \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;(1)\u003c/p\u003e\n\u003cp\u003eIn the provided equation, Y represents the predicted response, and β0 denotes the model constant. The coefficients β1, β2, β3, β11, β22, β33, β12, β13, and β23 within the statistical model illustrate the linear, quadratic, and interactive effects of factors A (pH), B (Temperature), and C (Time) on the response, respectively. The Central Composite Design (CCD) was conducted using Design-Expert software (Version 11.1.1.0 USA) to analyze the experimental data. An Analysis of Variance (ANOVA) was conducted to evaluate the statistical significance with a 95% confidence interval, utilizing a p-value threshold of 0.05. In addition, regression analysis and response surface plots were used to identify the optimal conditions for assessing the CQDs sensor.\u003c/p\u003e\n\u003cp\u003eTable 1\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.5 Fluorescence Sensing Procedure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the fluorescence intensity in the presence of varying concentrations of Metronidazole, a specific protocol was followed. Initially, 30 µg of the synthesized CQDs was added to 3 mL of a solution adjusted to pH 7. The fluorescence excitation wavelength was set at 450 nm, and the corresponding fluorescence spectra were recorded, focusing on an emission wavelength of 538 nm. Measurements were subsequently taken both in the presence and absence of Metronidazole, with the fluorescence intensities denoted as F₀ (without Metronidazole) and F (with Metronidazole). The ratio of these fluorescence intensities, expressed as F₀/F, was calculated to establish a correlation between the fluorescence response and the concentration of Metronidazole present in the solution. For the assessment of the detection limit, a signal-to-noise ratio of 3 was applied, which is a standard criterion in analytical chemistry. This threshold helps determine the lowest concentrations of Metronidazole that can be reliably detected using the CQDs fluorescence system.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.6 Machine Learning Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo improve the prediction accuracy and robustness of the sensing platform, machine learning algorithms were implemented using Python (version X.X) and the Scikit-learn library. Models such as support vector regression (SVR), random forest regression, and artificial neural networks (ANN) were trained on the experimental data. Model performance was evaluated using R² (coefficient of determination), root mean square error (RMSE), and mean absolute error (MAE) to select the most suitable model for reliable quantification of metronidazole.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.7. Preparation of sample\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, metronidazole levels were examined using serum samples obtained from the Kermanshah Blood Bank. The samples were provided under standard institutional protocols and did not contain any identifiable personal information. To remove protein interferences, 5 mL of serum was mixed with 5 mL of 10% (w/v) trichloroacetic acid and centrifuged at 4000 rpm for 30 minutes. The resulting supernatant was filtered through a 0.45-μm Millipore membrane to eliminate particulate matter. The treated serum samples were then diluted tenfold with a universal buffer solution adjusted to pH 7. Subsequently, 300 μL of the prepared solution was combined with 30 µg of the CQDs-based nanosensor and varying concentrations of metronidazole. Fluorescence spectra were recorded at an emission wavelength of 523 nm with excitation at 440 nm.\u003c/p\u003e"},{"header":"Results and discussion ","content":"\u003cp\u003e\u003cstrong\u003e3.1 Structural and Morphological Characterization\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eScanning electron microscopy (SEM) analysis also provided insightful details, revealing uniformly distributed spherical nanoparticles with an average size ranging from 5 to 10 nm. As illustrated in \u003cstrong\u003eFig. 1a\u003c/strong\u003e, these images demonstrate a high degree of dispersion with minimal aggregation, indicating the successful synthesis of well-defined quantum dots. The uniform morphology is essential for consistent performance in sensing applications. Additionally, \u003cstrong\u003eFig. 1b\u003c/strong\u003e depicts elemental mapping that illustrates the spatial distribution of various elements on the surface of the nanoparticles. From the mapping, it is evident that carbon (C) and oxygen (O are present in notably high densities, confirming the organic nature of the quantum dots. Moreover, the Energy Dispersive X-ray Spectroscopy (EDX) spectrum shown in \u003cstrong\u003eFig. 1c\u003c/strong\u003e quantifies the elemental composition of the CQDs. The analysis indicates that carbon, oxygen, silicon account for significant percentages at 35.5%, 24.6%, and 25.9% respectively, while trace amounts of, Mg, Ca, and Na among other elements are also detected within the structure, highlighting the successful incorporation of metal ions. To further elucidate the distribution of surface elements. This comprehensive mapping effectively demonstrates the diverse elemental composition within the CQDs structure, underscoring its potential for multifaceted applications in biosensing and nanotechnology. FTIR spectra (\u003cstrong\u003eFig. 2\u003c/strong\u003e) revealed characteristic peaks corresponding to various surface functional groups. A broad absorption around 3400 cm⁻¹ was attributed to O–H and N–H stretching vibrations, indicating the presence of hydroxyl and amine groups. Peaks at ~1700 cm⁻¹ and ~1600 cm⁻¹ were assigned to C=O (carboxyl) and C=C stretching, respectively. These functional groups play a critical role in enhancing the solubility and fluorescence properties of CQDs. The UV-Vis absorption spectra of the mold-derived CQDs showed a main absorption band near 280 nm, attributed to π–π* transitions of C=C aromatic domains, and a shoulder around 330 nm due to n–π* transitions of C=O bonds. Upon excitation at 440 nm, the CQDs exhibited a strong green fluorescence emission peak centered at 520 nm. This large Stokes shift (80 nm) indicates efficient energy transfer and low self-absorption, which are favorable for sensing applications. The intense emission confirms the high quantum yield and good surface passivation of the CQDs synthesized from mold (\u003cstrong\u003eFig. 3).\u003c/strong\u003e In this study, we focused on the analysis of fluorescence spectra specifically designed for fluorescence-based sensors. The importance of obtaining the optimal fluorescence spectrum is crucial, particularly in detection and measurement applications. Therefore, we investigated various excitation wavelengths ranging from 250 nm to 460 nm. Our findings revealed that the best fluorescence spectrum was observed at 460 nm, corresponding to an excitation wavelength of 450 nm. In contrast, the other excitation wavelengths did not yield satisfactory intensity or sharpness, failing to achieve the desired quality and resolution. These results not only enhance the design of fluorescence sensors but also pave the way for the development of new applications in the field\u003cstrong\u003e) \u003c/strong\u003e\u003cstrong\u003eFig. 4\u003c/strong\u003e.).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig. 1\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig. 2\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;3\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig. 4\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2. Statistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo optimize the fluorescence response (F₀/F) for metronidazole detection, a total of 16 experimental runs were designed using the Central Composite Design (CCD) under the framework of Response Surface Methodology (RSM). The experimental conditions and corresponding responses are summarized in \u003cstrong\u003eTable 2\u003c/strong\u003e. The F₀/F values ranged from 1.00 to 1.66, reflecting the influence of pH (A), temperature (B), and time (C) on the fluorescence quenching behavior. CCD was employed for its efficiency in modeling complex interactions with a reduced number of experiments. To accurately capture the non-linear effects and interactions among the variables, a quadratic regression model was fitted to the data. As shown in \u003cstrong\u003eTable 3\u003c/strong\u003e, the quadratic model demonstrated a high degree of correlation with the experimental data, as indicated by an adjusted R² value of 0.9646 and a predicted R² of 0.8665, confirming its predictive reliability. The resulting quadratic equation, expressed in terms of coded variables, is as follows:\u003c/p\u003e\n\u003cp\u003eY = 3.9 + (0.1625 × A) + (0.7366 × B) - (0.0297 × C) - (0.2233 × A²) - (0.5777 × B²) - (0.0536 × C²) + (0.0612 × AB) + (0.0662 × AC) - (0.1113 × BC)\u003c/p\u003e\n\u003cp\u003eIn this model, positive coefficients indicate synergistic effects that enhance the fluorescence response, while negative coefficients reflect antagonistic or diminishing effects. The significance of the quadratic term over the linear and interaction models was statistically validated by the p-values reported in \u003cstrong\u003eTable 3\u003c/strong\u003e, where the quadratic model yielded a p-value \u0026lt; 0.0001, confirming its superiority. Furthermore, the lack-of-fit p-value (0.0005) suggests that the model is statistically significant with an acceptable deviation from the actual experimental data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3. ANOVA\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAnalysis of variance (ANOVA) was applied to the quadratic model described earlier to evaluate the statistical significance of individual and interactive factors influencing the fluorescence response (F₀/F) of metronidazole. The ANOVA results, summarized in \u003cstrong\u003eTable 4\u003c/strong\u003e, reveal that the overall model is highly significant, as evidenced by an F-value of 58.59 and a p-value \u0026lt; 0.0001, confirming the robustness of the regression model in describing the experimental data. Among the individual terms, the linear effects of pH (A) and time (B) exhibited strong statistical significance, with p-values \u0026lt; 0.0001. However, temperature (C) was found to have no significant individual effect (p = 0.8040). Regarding interaction effects, only the AB interaction (p = 0.0021) showed statistical significance, whereas AC and BC were not significant contributors (p \u0026gt; 0.76). In terms of curvature, the squared terms A² and B² were highly significant (p = 0.0001 and \u0026lt; 0.0001, respectively), indicating nonlinear relationships with the response. Conversely, the squared term C² was not significant (p = 0.5329), suggesting that quadratic effects of temperature have minimal influence on fluorescence variation.\u003c/p\u003e\n\u003cp\u003eDespite a slight lack of fit detected (p = 0.0005), the model still explains a substantial portion of variability, supported by a low residual mean square (0.0022) and a high coefficient of determination (R² = 0.9130). This confirms the model’s predictive capability and suitability for response surface optimization. The final fitted quadratic equation, derived in terms of actual values of the variables, is presented below:\u003c/p\u003e\n\u003cp\u003eY = –0.6954 + (0.4200 × A) + (0.2581 × B) + (0.0170 × C) – (0.0357 × A²) – (0.0071 × B²) – (0.00023 × C²) + (0.0027 × AB) + (0.0017 × AC) – (0.00084 × BC) \u0026nbsp; \u0026nbsp; (2)\u003c/p\u003e\n\u003cp\u003eThis equation illustrates how each factor contributes to the fluorescence intensity and allows for effective prediction and optimization of conditions for metronidazole analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4. 3D response surface plots\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate the interactive effects of the studied variables on the fluorescence response (F₀/F), three-dimensional response surface plots were constructed based on the quadratic regression model. The plots illustrate the combined influence of two factors while the third is held constant at its central level. \u003cstrong\u003eFig\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;5a\u003c/strong\u003e shows the interactive effect of pH (A) and Time (B) on the fluorescence intensity. The curved surface and elliptical contour lines suggest a strong interaction between pH and reaction time. The response increases with increasing pH and time, reaching an optimal region before declining, indicating a nonlinear relationship. This trend may be attributed to the enhanced binding or quenching environment at optimal pH and exposure time. \u003cstrong\u003eFig. 5b\u003c/strong\u003e displays the influence of pH (A) and Temperature (C). A clear curvature is observed, but the contour lines are more parallel, indicating a relatively weaker interaction between these two variables. The response increases with pH and decreases slightly with temperature at higher pH levels. This implies that while pH plays a dominant role, elevated temperatures may reduce the stability of the sensing complex. \u003cstrong\u003eFig. 5c\u003c/strong\u003e illustrates the combined effect of Time (B) and Temperature (C). The surface is slightly curved, but the contours are nearly straight and parallel, suggesting a minimal interaction between time and temperature. The fluorescence response increases with time but remains relatively stable over the temperature range studied. This indicates that reaction time has a more pronounced effect compared to temperature under these conditions. Overall, these plots confirm that pH and time are the most influential parameters affecting the fluorescence response, with temperature showing a moderate to minimal effect. The optimal sensing conditions are located within the curved regions of the surfaces, supporting the use of response surface methodology (RSM) in identifying the best operational parameters.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig. 5\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5. Accuracy of the Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe accuracy and reliability of the quadratic model developed for predicting amoxicillin removal efficiency were evaluated using several key statistical parameters. The coefficient of determination (R²) was found to be 0.9814, indicating that approximately 98.14% of the variability in the response could be explained by the model. The adjusted R² value of 0.9646 confirms a strong fit, even after accounting for the number of predictors in the model. The predicted R² of 0.8665 is in reasonable agreement with the adjusted R², with a difference of less than 0.2. This close alignment demonstrates the model’s robustness and its ability to generalize well to new data points. Additionally, the standard deviation (SD) of 0.0464 and a relatively low coefficient of variation (C.V.) of 3.23% reflect high precision and low dispersion of the experimental data around the mean. Furthermore, the adequate precision value, which measures the signal-to-noise ratio, was calculated to be 19.0136—substantially greater than the minimum acceptable value of 4. This suggests that the model has a strong signal and is capable of navigating the design space efficiently. Collectively, these statistical indicators validate the high quality and predictive power of the model, supporting its suitability for optimizing the operational parameters in the amoxicillin removal process (\u003cstrong\u003eTable 5\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 5\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.6 Model Diagnostics and Assumptions Verification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the reliability and statistical adequacy of the developed model for predicting amoxicillin removal efficiency, several diagnostic plots were analyzed (Fig. 4a–d). These visual tools are essential for validating the underlying assumptions of regression modeling, including normality, constant variance, and independence of residuals. \u003cstrong\u003eFig\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003e6a\u003c/strong\u003e (Residuals vs. Predicted Values) illustrates the distribution of residuals around the predicted responses. The lack of any systematic pattern and the relatively uniform spread of points above and below the zero line suggest homoscedasticity—indicating constant variance across the range of predicted values. This confirms that the model errors are not dependent on the magnitude of predictions, which is a critical assumption in regression analysis. \u003cstrong\u003eFig\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003e6b\u003c/strong\u003e (Normal Probability Plot of Residuals) examines the normality of residuals. The residuals follow a near-linear trend along the reference line, indicating that they are approximately normally distributed. This supports the appropriateness of the regression model and suggests that no major deviations from normality exist. \u003cstrong\u003eFig\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003e6c\u003c/strong\u003e (Predicted vs. Actual Values) demonstrates the relationship between the observed and model-predicted responses. Most points lie close to the diagonal line, confirming a strong agreement between actual and predicted values. While the model performs exceptionally well for low to mid-range response levels, it shows slightly less accuracy at the highest removal efficiencies. This could indicate the need for more training data in that range or the potential for minor model refinement. \u003cstrong\u003eFig\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003e6d\u003c/strong\u003e (Residuals vs. Run Order) helps assess the independence and randomness of residuals over time. The random scatter of residuals around zero with no evident trend or cyclic pattern confirms the absence of time-related bias or systematic experimental errors. However, one or two points that slightly deviate may merit further investigation to ensure they are not influential outliers. In summary, these diagnostic plots collectively demonstrate that the regression model satisfies the key statistical assumptions. The model exhibits strong predictive capabilities, consistent residual behavior, and robustness across different runs, reinforcing its suitability for optimizing amoxicillin removal processes in future applications.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig.6\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.7. Machine Learning Models for Predicting Metronidazole Detection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo assess the performance of machine learning models in predicting metronidazole detection, three algorithms were investigated: Linear Regression, Random Forest, and Support Vector Regression (SVR). Their predictive capabilities were evaluated based on statistical metrics including the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE). The comparison of model performance is visualized in \u003cstrong\u003eFig.7(a),\u003c/strong\u003e which presents a 3D scatter plot based on R², MSE, and MAE. As shown, the Random Forest model significantly outperforms the other two models, occupying a favorable position with high R² (0.9633) and low error values (RMSE = 0.0446, MAE = 0.0355). This indicates strong agreement between predicted and actual values with minimal error. On the other hand, the Linear Regression and SVR models demonstrated much weaker performance. The Linear Regression model showed a negative R² value (-0.1103), along with the highest RMSE (0.2451) and MAE (0.1988), implying that it fails to capture the true relationship between input variables and target values. Similarly, SVR also yielded a negative R² (-0.0163) with relatively high error metrics (RMSE = 0.2345, MAE = 0.1904), though slightly better than Linear Regression. Further insights can be seen in \u003cstrong\u003eFig.7(b)\u003c/strong\u003e, which plots actual vs. predicted values for all three models. Here again, the predictions made by the Random Forest model lie very close to the 45-degree reference line, indicating high predictive accuracy. In contrast, the predictions from Linear Regression and SVR display noticeable deviations from the ideal line, highlighting their limited predictive power. the Random Forest model demonstrated the most reliable and accurate performance in predicting metronidazole detection. These findings highlight the advantage of non-linear ensemble methods over traditional linear or kernel-based models in capturing complex patterns within analytical chemistry datasets.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig.7\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.8. Method Selectivity\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSelectivity is a key parameter in evaluating the performance of a fluorescence-based sensor, particularly when used in complex matrices. In this study, the selectivity of the CQDs sensor for metronidazole detection was systematically investigated against a panel of potential interfering substances. As illustrated in \u003cstrong\u003eFig.8\u003c/strong\u003e, various compounds, including common pharmaceuticals and biological molecules, were individually tested at a concentration of 1000 μM under the optimized sensing conditions. The fluorescence response was expressed as the F₀/F ratio, where F₀ is the fluorescence intensity in the absence of the analyte and F is the intensity in the presence of each test compound. The results clearly indicate that metronidazole produces a distinct and significantly higher fluorescence quenching effect compared to the other substances tested. While most interferents resulted in minimal or negligible changes in fluorescence, only metronidazole caused a marked decrease in signal intensity. This pronounced difference highlights the excellent selectivity of the CQDs sensor towards metronidazole. The sharp contrast in F₀/F values across the tested compounds underscores the sensor's ability to discriminate metronidazole even in the presence of structurally or chemically similar substances. Such selectivity is essential for reliable detection in real-world applications, and these results confirm that the developed CQDs sensor is a highly selective and robust platform for accurate metronidazole quantification.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig. 8\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.9. Calibration and Detection of Metronidazole\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe CQDs nanocomposite exhibited a notable fluorescence enhancement at 523 nm upon excitation at 440 nm in the presence of metronidazole, forming the basis for its application as a novel fluorescence nanoprobe. To evaluate its analytical sensitivity, different concentrations of metronidazole were introduced under optimized experimental conditions. As depicted in \u003cstrong\u003eFig. 9A\u003c/strong\u003e, a clear and direct linear increase in fluorescence intensity was observed with rising metronidazole concentrations within the range of 97.4 to 3215.3 µM. This linear relationship confirms the probe's capacity for consistent and quantifiable detection of metronidazole across this dynamic range. To further validate the performance of the sensor, a calibration curve was established by plotting the ratio of fluorescence intensity in the absence (F₀) and presence (F) of metronidazole (F₀/F) versus metronidazole concentration. As shown in Figure 9B, the resulting plot demonstrated excellent linearity with a correlation coefficient of R² = 0.9769, indicating high precision and reliability in response. The limit of detection (LOD) for metronidazole was calculated to be 50.5 µM, based on a signal-to-noise ratio of 3. This relatively low LOD, combined with the strong linear correlation, highlights the high sensitivity and robustness of the developed CQDs nanoprobe. Furthermore, the analytical performance of this sensor was benchmarked against previously reported quantum dot-based sensors for metronidazole detection. The comparison emphasizes the superior capabilities of the present approach in terms of sensitivity and dynamic range, confirming its promising potential for practical applications in metronidazole analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig. 9\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.10. Interference Study for Metronidazole Detection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo ensure the specificity of the CQDs-based nanoprobe for metronidazole detection, a comprehensive interference study was conducted. Various potentially interfering substances were examined, including PBS (pH 7.4), KCl, NaCl, MnCl₂, MgCl₂, CaCl₂, bovine serum albumin (BSA), and glucose, which are commonly encountered in biological or environmental samples. Each of these compounds was tested at concentrations significantly higher than those of metronidazole to rigorously challenge the sensor’s performance under realistic conditions (Fig. 10). The findings reveal that the fluorescence response of the CQDs/Ag/Cu nanoprobe toward metronidazole remains stable and consistent even in the presence of these potential interferents. The negligible variation in the sensor’s output suggests a high level of selectivity and interference resistance, confirming that the presence of common ions, proteins, and sugars does not significantly affect metronidazole detection. This robustness against interference highlights the practical applicability of the CQDs nanoprobe, particularly in complex matrices such as biological fluids or environmental samples, where such compounds are prevalent. The ability to detect metronidazole accurately without significant interference eliminates the necessity for extensive sample pretreatment, further enhancing the probe's value for real-world analytical applications.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFig. 10\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.11. Application\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the practical applicability of the developed CQDs-based nanoprobe for the detection of metronidazole in real biological samples, serum samples was chosen as the representative matrix. The serum samples underwent an initial preparation process to ensure they were suitable for analysis. This included isolating the serum and subsequently diluting it tenfold using a buffer solution adjusted to pH 7. To validate the method’s accuracy, a standard spiking approach was employed. Known concentrations of metronidazole were added to the diluted serum samples under optimized experimental conditions. This allowed for the assessment of the nanoprobe’s performance in complex biological environments. Following the spiking process, fluorescence intensity measurements were taken to determine the concentration of metronidazole present in the samples. The experimental results, presented in \u003cstrong\u003eTable 6\u003c/strong\u003e, indicate excellent analytical recovery, with values of 100.12%, 100.7%, and 100.5%. Moreover, the method exhibited strong precision, as reflected in low standard deviation values of 0.301, 0.625, and 0.740. These findings confirm the reliability and reproducibility of the CQDs-based sensor for metronidazole quantification in serum. Additionally, a comparison with previously reported quantum dot-based detection systems, as shown in \u003cstrong\u003eTable 7\u003c/strong\u003e, highlights the superior sensitivity of the proposed CQDs probe. Specifically, the lower limit of detection observed in this study underscores the enhanced sensing capabilities of the CQDs nanocomposite, likely due to its improved fluorescence properties and stability in biological matrices\u0026nbsp;[17–19].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 7\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 8\u003c/strong\u003e\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eIn this study, a highly sensitive and selective carbon quantum dot (CQDs)-based fluorescent nanoprobe was successfully developed for the detection of metronidazole. Synthesized via a hydrothermal method with mold as the carbon source, this nanocomposite showcases remarkable properties upon thorough characterization using techniques such as scanning electron microscopy (SEM), Fourier-transform infrared spectroscopy (FTIR), and transmission electron microscopy (TEM). The sensor exhibited excellent photoluminescent behavior, with a significant fluorescence response upon interaction with metronidazole, enabling precise quantification. Among the machine learning models evaluated for predictive analysis, the Random Forest algorithm demonstrated superior performance, achieving high accuracy and minimal error in comparison to linear regression and SVR models. This highlights the potential of integrating data-driven approaches for optimizing sensor response and analytical prediction. The CQDs nanoprobe showed strong selectivity toward metronidazole, even in the presence of structurally similar or commonly encountered interfering compounds. Its robustness was further validated through interference studies, where the presence of excess ions, proteins, and sugars did not significantly affect detection performance. Analytical calibration revealed a strong linear response over a wide concentration range, with a limit of detection as low as 50.5 \u0026micro;M. The applicability of the method was confirmed through successful detection of metronidazole in spiked serum samples, with high recovery rates and low standard deviations, underscoring the method\u0026rsquo;s accuracy and precision in complex biological matrices. Overall, the developed CQDs-based sensor provides a rapid, reliable, and cost-effective platform for metronidazole detection. Its excellent analytical performance, combined with high selectivity and compatibility with real samples, makes it a promising candidate for use in pharmaceutical, clinical, and environmental monitoring applications.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e The authors acknowledge the support and laboratory facilities provided by Islamic Azad University of Karaj and Sanandaj.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e Kimiya Khandestana: Conducted the experimental work and collected fluorescence data. Azar Sabukbara: Assisted in experimental design and sample preparation. Bahareh Rahimian Zarif: Conceived and designed the study, supervised the research, performed data analysis, and critically revised the manuscript. Nahid Haghnazari: Contributed to data interpretation and manuscript editing. Nasser Harzandi: Provided technical support and contributed to material characterization. All authors reviewed and approved the final version of the manuscript. Bahareh Rahimian Zarif is the corresponding author and will serve as the main contact throughout the review process.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e Kimiya Khandestana, Azar Sabukbara, Bahareh Rahimian Zarif, Nahid Haghnazari, and Nasser Harzandi declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e All data generated or analyzed during this study are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics declarations\u003c/strong\u003e All serum samples used in this study were obtained from the Kermanshah Blood Bank in accordance with institutional guidelines. The samples were anonymized and commercially provided for research purposes, without direct involvement of human subjects or collection of personal data. Therefore, ethical approval and informed consent were not required for this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate / Consent for publication\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eShahid, A. \u003cem\u003eet al.\u003c/em\u003e Antibiotic residues in food chains; impact on the environment and human health: a review. \u003cem\u003eApplied Ecology \u0026amp; Environmental Research\u003c/em\u003e \u003cstrong\u003e19,\u003c/strong\u003e (2021).\u003c/li\u003e\n\u003cli\u003eKyuchukova, R. Antibiotic residues and human health hazard-review. \u003cem\u003eBulgarian Journal of Agricultural Science\u003c/em\u003e \u003cstrong\u003e26,\u003c/strong\u003e (2020).\u003c/li\u003e\n\u003cli\u003eVan Boeckel, T. P. \u003cem\u003eet al.\u003c/em\u003e Global trends in antimicrobial use in food animals. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e \u003cstrong\u003e112,\u003c/strong\u003e 5649\u0026ndash;5654 (2015).\u003c/li\u003e\n\u003cli\u003eFrancino, M. Antibiotics and the human gut microbiome: dysbioses and accumulation of resistances. \u003cem\u003eFrontiers in microbiology\u003c/em\u003e \u003cstrong\u003e6,\u003c/strong\u003e 164577 (2016).\u003c/li\u003e\n\u003cli\u003eLjubojević Pelić, D. \u003cem\u003eet al.\u003c/em\u003e Antibiotic Residues in Cultured Fish: Implications for Food Safety and Regulatory Concerns. \u003cem\u003eFishes\u003c/em\u003e \u003cstrong\u003e9,\u003c/strong\u003e 484 (2024).\u003c/li\u003e\n\u003cli\u003eBaynes, R. E. \u003cem\u003eet al.\u003c/em\u003e Health concerns and management of select veterinary drug residues. \u003cem\u003eFood and chemical toxicology\u003c/em\u003e \u003cstrong\u003e88,\u003c/strong\u003e 112\u0026ndash;122 (2016).\u003c/li\u003e\n\u003cli\u003eBaynes, R. E. \u003cem\u003eet al.\u003c/em\u003e Health concerns and management of select veterinary drug residues. \u003cem\u003eFood and chemical toxicology\u003c/em\u003e \u003cstrong\u003e88,\u003c/strong\u003e 112\u0026ndash;122 (2016).\u003c/li\u003e\n\u003cli\u003eKyuchukova, R. Antibiotic residues and human health hazard-review. \u003cem\u003eBulgarian Journal of Agricultural Science\u003c/em\u003e \u003cstrong\u003e26,\u003c/strong\u003e (2020).\u003c/li\u003e\n\u003cli\u003eKarami, C. \u0026amp; Taher, M. A. Colorimetric Sensor of Cobalt Ions in Aqueous Solution Using Gold Nanoparticles Modified with Glycyrrhizic Acid. \u003cem\u003ePlasmonics\u003c/em\u003e \u003cstrong\u003e13,\u003c/strong\u003e (2018).\u003c/li\u003e\n\u003cli\u003eDeymehkar, E., Taher, M. A., Karami, C. \u0026amp; Arman, A. Synthesis of SPR Nanosensor using Gold Nanoparticles and its Application to Copper (II) Determination. \u003cem\u003eSilicon\u003c/em\u003e \u003cstrong\u003e10,\u003c/strong\u003e (2018).\u003c/li\u003e\n\u003cli\u003eKarami, C., Mehr, S. Y., Deymehkar, E. \u0026amp; Taher, M. A. Naked Eye Detection of Cr\u003csup\u003e3+\u003c/sup\u003eand Co\u003csup\u003e2+\u003c/sup\u003eIons by Gold Nanoparticle Modified with Azomethine. \u003cem\u003ePlasmonics\u003c/em\u003e \u003cstrong\u003e13,\u003c/strong\u003e (2018).\u003c/li\u003e\n\u003cli\u003eKarami, C., Taher, M. A. \u0026amp; Shahlaei, M. A simple method for determination of mercury (II) ions by PNBS-doped carbon dots as a fluorescent probe. \u003cem\u003eJournal of Materials Science: Materials in Electronics\u003c/em\u003e \u003cstrong\u003e31,\u003c/strong\u003e 5975\u0026ndash;5983 (2020).\u003c/li\u003e\n\u003cli\u003eJain, P. K., Huang, X., El-Sayed, I. H. \u0026amp; El-Sayed, M. A. Noble metals on the nanoscale: optical and photothermal properties and some applications in imaging, sensing, biology, and medicine. \u003cem\u003eAccounts of chemical research\u003c/em\u003e \u003cstrong\u003e41,\u003c/strong\u003e 1578\u0026ndash;1586 (2008).\u003c/li\u003e\n\u003cli\u003eKneipp, J., Kneipp, H., Wittig, B. \u0026amp; Kneipp, K. Novel optical nanosensors for probing and imaging live cells. \u003cem\u003eNanomedicine: Nanotechnology, Biology and Medicine\u003c/em\u003e \u003cstrong\u003e6,\u003c/strong\u003e 214\u0026ndash;226 (2010).\u003c/li\u003e\n\u003cli\u003eHou, J. \u003cem\u003eet al.\u003c/em\u003e Rapid microwave-assisted synthesis of molecularly imprinted polymers on carbon quantum dots for fluorescent sensing of tetracycline in milk. \u003cem\u003eTalanta\u003c/em\u003e \u003cstrong\u003e146,\u003c/strong\u003e 34\u0026ndash;40 (2016).\u003c/li\u003e\n\u003cli\u003eLiu, G. \u003cem\u003eet al.\u003c/em\u003e In-situ hydrothermal synthesis of molecularly imprinted polymers coated carbon dots for fluorescent detection of bisphenol A. \u003cem\u003eSensors and Actuators B: Chemical\u003c/em\u003e \u003cstrong\u003e228,\u003c/strong\u003e 302\u0026ndash;307 (2016).\u003c/li\u003e\n\u003cli\u003eAlvarenga, L. M. \u003cem\u003eet al.\u003c/em\u003e Preparation of a composite sensor based on a fluorescent and magnetic molecular imprint polymer for metronidazole extraction\u0026ndash;detection. \u003cem\u003eJournal of Molecular Liquids\u003c/em\u003e \u003cstrong\u003e390,\u003c/strong\u003e 123027 (2023).\u003c/li\u003e\n\u003cli\u003eXia, Z. \u0026amp; Li, Q. Application of Metronidazole detection by antibiotic ampicillin sodium based-carbon quantum dots. \u003cem\u003eInternational Journal of Environmental Analytical Chemistry\u003c/em\u003e \u003cstrong\u003e102,\u003c/strong\u003e 4178\u0026ndash;4190 (2022).\u003c/li\u003e\n\u003cli\u003eRen, G. \u003cem\u003eet al.\u003c/em\u003e Efficient preparation of nitrogen-doped fluorescent carbon dots for highly sensitive detection of metronidazole and live cell imaging. \u003cem\u003eSpectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy\u003c/em\u003e \u003cstrong\u003e234,\u003c/strong\u003e 118251 (2020).\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eTable 1: \u003c/span\u003e\u003c/strong\u003e\u003cspan dir=\"LTR\"\u003eExperimental parameters and levels in the 16 CCD for the optimization of pH,\u0026nbsp;\u003c/span\u003eTemp, and Time\u003c/p\u003e\n\u003cdiv align=\"left\" dir=\"ltr\"\u003e\n \u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\" width=\"532\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eFactor\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eName\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eLevel\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eLow Level\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eHigh Level\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eStd. Dev.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eCoding\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003epH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e4.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e10.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eActual\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eTime\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.5000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e5.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eActual\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eTemperatures\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e34.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e60.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eActual\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp dir=\"LTR\"\u003e\u003cstrong\u003eTable 2:\u003c/strong\u003e Experiment runs and responses for optimizing parameters evaluation\u003c/p\u003e\n\u003cdiv align=\"\" dir=\"ltr\"\u003e\n \u003ctable border=\"1\" cellspacing=\"3\" cellpadding=\"0\" width=\"529\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eFactor 1\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eFactor 2\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eFactor 3\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eResponse 1\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eRun\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eA: pH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eC: Time\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eB: Temp\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eF\u003csub\u003e0\u003c/sub\u003e/F\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.62\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.65\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.64\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.65\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.65\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.66\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.65\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.35\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e6.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.35\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.63\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.64\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp dir=\"LTR\"\u003e\u003cstrong\u003eTable 3:\u003c/strong\u003e Model summary statistic.\u0026nbsp;\u003c/p\u003e\n\u003cdiv align=\"left\" dir=\"ltr\"\u003e\n \u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\" width=\"591\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eSource\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eSequential p-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eLack of Fit p-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eAdjusted R\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003ePredicted R\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eLinear\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0618\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u0026lt; 0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.2395\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e-0.0058\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e2FI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.8772\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u0026lt; 0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.1102\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e-0.8665\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eQuadratic\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u0026lt; 0.0001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e0.0005\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e0.9646\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e0.8665\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eSuggested\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eCubic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.9983\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eAliased\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eTable 4\u003c/span\u003e\u003c/strong\u003e\u003cspan dir=\"LTR\"\u003e: ANOVA for response surface quadratic model for F\u003csub\u003e0\u003c/sub\u003e/F\u003c/span\u003e\u0026nbsp;\u003c/p\u003e\n\u003cdiv align=\"left\" dir=\"ltr\"\u003e\n \u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\" width=\"631\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eSource\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eSum of Squares\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003edf\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eMean Square\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eF-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003ep-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eModel\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.1263\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e58.59\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u0026lt; 0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003esignificant\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eA-pH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.1074\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.1074\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e49.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u0026lt; 0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eB-Time\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.5344\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.5344\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e247.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u0026lt; 0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eC-teTmperatures\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0649\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.8040\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eAB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0365\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0365\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e16.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eAC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.0000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eBC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0928\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.7670\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eA\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0753\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0753\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e34.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eB\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.6127\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.6127\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e284.12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u0026lt; 0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eC\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0009\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0009\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.4172\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.5329\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eResidual\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0216\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003eLack of Fit\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0210\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0042\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e39.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003esignificant\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003ePure Error\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eCor Total\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eTable 5:\u003c/span\u003e\u003c/strong\u003e\u003cspan dir=\"LTR\"\u003e\u0026nbsp;Standard deviation and R\u003csup\u003e2\u003c/sup\u003e of the response.\u003c/span\u003e\u003cspan dir=\"LTR\"\u003e\u0026nbsp;\u003c/span\u003e\u003c/p\u003e\n\u003cdiv align=\"left\" dir=\"ltr\"\u003e\n \u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\" width=\"563\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eStd. Dev.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.0464\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eR\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.9814\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e1.44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eAdjusted R\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.9646\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eC.V. %\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e3.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003ePredicted R\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e0.8665\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eAdeq Precision\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp dir=\"LTR\"\u003e19.0136\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp dir=\"LTR\"\u003e\u003cstrong\u003eTable 6:\u003c/strong\u003e Determination of metronidazole concentration in real samples. By fluorescence method (n = 5)\u003c/p\u003e\n\u003cdiv align=\"\" dir=\"ltr\"\u003e\n \u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003esample\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eAdded (\u0026micro;M)\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 218px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003efound(\u0026micro;M)\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 112px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eRecovery (%)\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 87px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eRSD%\u003c/span\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 61px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e1\u003c/span\u003e\u003c/p\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e2\u003c/span\u003e\u003c/p\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e3\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e100\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 218px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e98.9\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 112px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e98.9\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e1.46\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e500\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 218px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e489.925\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 112px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e97.99\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e5.53\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e1000\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 218px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e1018\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 112px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e101.8\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cspan dir=\"LTR\"\u003e2.11\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp dir=\"LTR\"\u003e\u003cstrong\u003e\u003cspan dir=\"LTR\"\u003eTable 7:\u003c/span\u003e\u003c/strong\u003e\u003cspan dir=\"LTR\"\u003e\u0026nbsp;Comparison of the performances of various sensors for detection of cefixim\u003c/span\u003e\u003c/p\u003e\n\u003cdiv align=\"left\" dir=\"ltr\"\u003e\n \u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eProbe\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 145px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eLinear range\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eLOD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eAntibiotics\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003eReal sample\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp dir=\"LTR\"\u003e\u003cstrong\u003e[Ref]\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp dir=\"LTR\"\u003esodium based-carbon quantum dots\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 145px;\"\u003e\n \u003cp dir=\"LTR\"\u003e20\u0026ndash;100\u0026nbsp;\u0026mu;M\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp dir=\"LTR\"\u003e62.5\u0026nbsp;nM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003emetronidazole\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003ewater samples\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp dir=\"LTR\"\u003e[17]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp dir=\"LTR\"\u003emolecular imprint polymer for metronidazole extraction\u0026ndash;detection\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 145px;\"\u003e\n \u003cp dir=\"LTR\"\u003e5.0-60.0 \u0026mu;M\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp dir=\"LTR\"\u003e1.28\u0026nbsp;\u0026mu;M\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003emetronidazole\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003ereal samples\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp dir=\"LTR\"\u003e[18]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp dir=\"LTR\"\u003enitrogen-doped fluorescent carbon dots\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 145px;\"\u003e\n \u003cp dir=\"LTR\"\u003e0.5\u0026ndash;22 \u0026mu;M\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp dir=\"LTR\"\u003e0.22 \u0026mu;M\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003emetronidazole\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003eurine samples\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp dir=\"LTR\"\u003e[19]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 146px;\"\u003e\n \u003cp dir=\"LTR\"\u003eCQDs of mod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 145px;\"\u003e\n \u003cp dir=\"LTR\"\u003e97.4 -3215.3 \u0026micro;M\u0026nbsp;\u0026micro;M\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp dir=\"LTR\"\u003e0.2 nM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003emetronidazole\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp dir=\"LTR\"\u003eserum samples\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp dir=\"LTR\"\u003eThis\u003c/p\u003e\n \u003cp dir=\"LTR\"\u003ework\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Carbon quantum dots, Mold synthesis, Metronidazole detection, Fluorescence sensor, Experimental design, Machine learning, Antibiotic monitoring","lastPublishedDoi":"10.21203/rs.3.rs-7721529/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7721529/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Carbon quantum dots (CQDs) were synthesized using mold as a green carbon source and characterized by a combination of techniques including fluorescence spectroscopy, Fourier-transform infrared spectroscopy (FTIR), Scanning Electron Microscopy (SEM), Energy-Dispersive X-ray Spectroscopy (EDX), elemental mapping, and line scan analysis. The CQDs showed strong fluorescence emission suitable for sensitive detection of the antibiotic metronidazole based on fluorescence quenching. After optimizing various parameters influencing sensor performance through experimental design, a linear detection ranges from 97.4 to 3215.3 µM with a detection limit of 50.5 µM was established. Machine learning algorithms were applied to enhance the accuracy and reliability of metronidazole quantification. This novel, eco-friendly sensor platform offers effective antibiotic monitoring for biomedical and environmental applications.","manuscriptTitle":"Machine Learning and Experimental Design-Driven Fluorescence Detection of Metronidazole in Food Samples Using Mold-Derived Carbon Quantum Dots","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-24 16:20:20","doi":"10.21203/rs.3.rs-7721529/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"34491518-1baa-46dd-94dc-a8183d07ac9d","owner":[],"postedDate":"October 24th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":56534834,"name":"Physical sciences/Chemistry"},{"id":56534835,"name":"Earth and environmental sciences/Environmental sciences"},{"id":56534836,"name":"Physical sciences/Nanoscience and technology"},{"id":56534837,"name":"Physical sciences/Optics and photonics"}],"tags":[],"updatedAt":"2025-11-24T13:53:52+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-24 16:20:20","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7721529","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7721529","identity":"rs-7721529","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}