Application of chemometrics for modeling and optimization of ultrasound-assisted dispersive liquid–liquid microextraction for the simultaneous determination of dyes

The study addresses the challenge of dye contamination in water, a significant barrier to achieving clean water and sanitation goals. With large-scale dye production and substantial discharge into wastewater, dyes such as malachite green (MG) and rhodamine B (RB) persist at trace levels, reduce light penetration, increase turbidity, and pose health hazards (e.g., allergenicity, mutagenicity, carcinogenicity). Reliable, accurate, and economical analytical methods are needed to detect and help remove such dyes from aquatic environments. While multiple analytical techniques exist, spectrophotometric methods are attractive for their simplicity and cost-effectiveness. Traditional liquid–liquid extraction (LLE) is limited by solvent consumption and time; therefore, microextraction approaches like DLLME are preferred. The research question is whether ultrasound-assisted DLLME, optimized via response surface methodology (RSM) and central composite design (CCD), can provide fast, efficient, and robust simultaneous extraction and determination of RB and MG in water samples.

The paper contextualizes dye pollution and its impacts, noting limited biodegradability and health risks of RB and MG. It reviews common analytical techniques (capillary electrophoresis, TLC, HPLC, electrochemistry/voltammetry, spectrophotometry) and highlights the advantages of spectrophotometry. Conventional LLE is critiqued for high solvent use and time, prompting interest in DLLME, which offers high sensitivity/efficiency with minimal solvent. The role of ultrasound is explained via cavitation, enhancing mass transfer and reducing extraction time and energy. The paper motivates multivariate optimization using RSM/CCD over one-factor-at-a-time approaches to capture interactions and efficiently identify optimal conditions. Comparative data with other extraction techniques (MSA-DLLME, CPE, DES-DLLME, MSPE, EME, classic DLLME) show UA-DLLME can achieve comparable or superior sensitivity and broader linear ranges with simpler, faster workflows.

Design: A UA-DLLME method coupled to UV/Vis spectrophotometry was developed and optimized using response surface methodology (RSM) with a central composite design (CCD) to simultaneously extract RB and MG from water. Reagents and instruments: Analytical-grade solvents and salts (Merck) including candidate extraction solvents (e.g., chloroform, carbon tetrachloride, dichloromethane, dichloroethane, chlorobenzene) and disperser solvents (ethanol, methanol, acetone, acetonitrile). Standard RB and MG stock solutions (100 µg mL⁻¹) in distilled water; working standards prepared by dilution. UV/Vis spectrophotometer for detection. Centrifugation used for phase separation. Solvent selection: Screening identified chloroform as the extraction solvent (highest recoveries attributed to solubility/polarity compatibility) and ethanol as the disperser solvent (good solubility in both phases). These were taken forward for optimization. Factors and ranges for optimization (CCD/RSM): - pH: 3–7 - Ultrasonication time: 1–5 min (time between injection of extraction mixture and centrifugation) - Extraction solvent (chloroform) volume: 50–250 µL - Disperser solvent (ethanol) volume: 200–1000 µL - Ionic strength (NaCl): 0–10% w/v - Centrifuge rate: evaluated 1000–4500 rpm (set at optimal 3500 rpm for further experiments) Procedure (optimized UA-DLLME): - A 10 mL aqueous sample (typically spiked at 500 ng mL⁻¹ for optimization) adjusted to target pH. - A mixture of ethanol (disperser) and chloroform (extractant) is rapidly injected into the sample to form a cloudy solution (fine extractant droplets dispersed by the disperser solvent). - Ultrasound is applied to enhance mass transfer via cavitation; after the selected ultrasonication time, the suspension is centrifuged (3500 rpm) to sediment the extractant phase at the tube bottom. - The sedimented chloroform phase is collected by microsyringe and analyzed by UV/Vis spectrophotometry. Modeling and statistics: - Quadratic polynomial models were fitted to extraction recovery responses for RB (Y_RB) and MG (Y_MG) as functions of coded variables A–E (where factors include pH, ultrasonication time, extraction solvent volume, disperser solvent volume, and ionic strength). Non-significant terms were removed. - Model adequacy assessed by regression (R² and adjusted R² > 0.99 for both dyes), lack-of-fit tests (not significant; p = 0.2919 for RB, 0.7371 for MG), residual diagnostics (normality, homoscedasticity, independence), and predicted vs. observed plots (R² > 99.8%). - Optimization with desirability functions in Design-Expert targeted maximal recovery for both analytes concurrently. Centrifugation optimization: Extraction efficiency increased with speed up to 3500 rpm and plateaued, indicating complete phase separation; 3500 rpm selected. Optimal conditions (from CCD/RSM): pH 5; ultrasonication time 4 min; extraction solvent (chloroform) 120 µL; disperser solvent (ethanol) 760 µL; no added salt (ionic strength 0%). Method validation: - Linearity: RB 7.5–1500 ng mL⁻¹ (r = 0.9979), MG 12–1000 ng mL⁻¹ (r = 0.9983). - LOD/LOQ: RB 1.45/4.83 ng mL⁻¹; MG 2.73/9.10 ng mL⁻¹. - Precision: RSD from replicate (n=5) measurements at 500 ng mL⁻¹; low RSD values reported. - Preconcentration factor (PF): 83.33 (both dyes); Enrichment factor (EF): RB 86.63, MG 97.54. Real sample analysis: - Matrices: deionized water, tap water, Sistan Lake water, and wastewater were spiked at 500 ng mL⁻¹ and processed under optimal conditions. - Recoveries ranged approximately 95.53–99.60% with low RSDs, demonstrating applicability to environmental waters. Comparative assessment: Performance compared favorably with literature methods (MSA-DLLME, CPE, DES-DLLME, MSPE, EME, traditional DLLME) in terms of LODs, linear ranges, simplicity, cost, and analysis time.

- Optimal UA-DLLME conditions: pH 5; ultrasonication 4 min; chloroform 120 µL; ethanol 760 µL; no salt; centrifugation 3500 rpm. - Linearity (LDR): RB 7.5–1500 ng mL⁻¹; MG 12–1000 ng mL⁻¹. - Sensitivity: LOD/LOQ for RB 1.45/4.83 ng mL⁻¹; for MG 2.73/9.10 ng mL⁻¹. - Precision: low RSDs in spiked water matrices (generally <~3.5%). - Efficiency: Extracted recoveries in real samples 95.53–99.60%. - Enrichment/Preconcentration: PF 83.33; EF RB 86.63, MG 97.54. - Model performance: Quadratic models with R² and Adj-R² > 0.99; predicted vs. measured R² > 99.8%; lack-of-fit not significant (RB p=0.2919; MG p=0.7371); overall desirability 1.00 with predicted recoveries ~99.07% (RB) and 99.74% (MG) vs. observed ~97.68% (RB) and 98.51% (MG). - Operational findings: Chloroform and ethanol were superior extractant/disperser; salt addition decreased recovery; centrifugation above 3500 rpm did not further improve extraction. - Comparative context: The method is simple, fast, economical, and exhibits LODs and linear ranges comparable or superior to several reported extraction/detection techniques.

The study demonstrates that UA-DLLME, optimized via CCD/RSM, effectively and simultaneously extracts RB and MG from water with high recoveries and low detection limits, addressing the need for rapid, reliable dye monitoring in environmental waters. Ultrasonication enhances mass transfer through cavitation, enabling equilibrium within minutes and reducing energy/time consumption. RSM captured interaction effects among critical factors (pH, solvent volumes, ultrasonication time, ionic strength) and efficiently identified optimal conditions with excellent model fit and predictive capability. The method's high enrichment and precision, minimal solvent volumes, and short extraction time make it suitable for routine monitoring and potentially for broader quality-control applications. Comparisons with other techniques indicate competitive analytical performance with simpler instrumentation (UV/Vis), supporting its relevance for resource-limited settings and for advancing water quality monitoring aligned with SDG goals.

A UA-DLLME method coupled with UV/Vis detection was developed and chemometrically optimized for simultaneous determination of RB and MG in water. Using CCD/RSM, optimal conditions (pH 5, 4 min ultrasonication, 120 µL chloroform, 760 µL ethanol, no salt, 3500 rpm centrifugation) delivered high recoveries (~96–100%), low LODs (RB 1.45 ng mL⁻¹; MG 2.73 ng mL⁻¹), broad linear ranges, and strong model predictability (R² > 0.99). Real water samples (tap, lake, wastewater) were successfully analyzed, confirming applicability. The method offers advantages of simplicity, speed, low cost, and high efficiency, making it a robust option for environmental dye monitoring. Potential future directions include extending the approach to other dye classes and contaminants, evaluating performance in more complex matrices (e.g., industrial effluents with diverse interferents), integrating greener extractants, and coupling with portable detection for field-deployable analysis.