Salt-saturated supramolecular solvent microextraction for spectrophotometric determination of curcumin in food samples

Abstract

A salt-saturated supramolecular solvent-based liquid–liquid microextraction method coupled with UV–Vis spectrophotometry was developed for the sensitive determination of trace levels of curcumin in food samples. Supramolecular solvents formed from decanol and tetrahydrofuran were employed as the extraction phase, while a mixed salt system (NaCl, Mg(NO₃)₂·6H₂O, and KNO₃) was used to enhance extraction efficiency through a salting-out effect. Key experimental parameters influencing extraction performance, including pH, extraction solvent volumes, mixing time, and centrifugation conditions, were optimized using Box–Behnken design combined with response surface methodology. Under optimal conditions, the proposed method exhibited a linear response over the concentration range of 1.0 µg/L − 5.0 mg/L with a limit of detection of 0.23 µg/L. An enrichment factor of 131 was achieved within a total extraction time of 6 min. The method demonstrated good precision, with relative standard deviations below 3.7%. Compared with supramolecular solvent microextraction without salt saturation, the proposed approach provided an average signal enhancement of 32.7%. The method was successfully applied to the determination of curcumin in tomato, apple and fish samples, yielding recoveries between 98.0% and 99.7%. Owing to its simplicity, rapidity, low solvent consumption, and satisfactory analytical performance, the developed method represents a green and efficient alternative for routine spectrophotometric analysis of curcumin in food matrices.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-46796-0.

Introduction

Various studies have recently focused on the investigation of natural compounds for use as natural medicines, owing to the increasing interest in alternative and plant-based therapies^1–4. The use of natural polyphenols in the treatment of different diseases, particularly cancer, is of great importance, and these compounds are considered promising candidates for the development of anticancer drugs. Polyphenols, including curcumin, possess at least one aromatic ring with one or more functional groups such as hydroxyl. They constitute a large class of secondary metabolites present in plants, beverages, and plant-based foods such as vegetables, spices, nuts, and fruits. Among natural polyphenols, curcumin has shown remarkable ability to inhibit the proliferation of cancer cells^5–7. Curcumin is a yellow phenolic compound obtained from the rhizomes of Curcuma longa (turmeric), which is extensively cultivated in Asia. It is commonly used as a food preservative, textile dye, cooking spice, and cosmetic ingredient. Curcumin is also employed as a coloring additive in the pharmaceutical, culinary, and confectionery industries, as well as in the dyeing of cotton, wool, and silk. Traditionally, Curcuma longa rhizomes have been used for the treatment of spleen disorders, bronchitis, and the common cold, and as blood purifiers, antiseptics, and expectorants. Curcumin has also been applied in the treatment of liver diseases, jaundice, cancer, and other medical conditions^8–11.

Although curcumin is not naturally present in tomato, apple, or fish, these matrices were selected to evaluate the applicability of the proposed method in diverse and complex food systems. Curcumin is widely used as a food additive, colorant, and dietary supplement, and thus may be introduced into food products during processing or fortification¹². Furthermore, polyphenolic compounds may enter aquatic environments and accumulate in fish tissues under certain conditions, as reported in previous studies¹³. Therefore, these samples provide suitable model matrices for assessing the performance and reliability of the developed method.

Since curcumin is often present at very low concentrations in complicated real matrices, an efficient preconcentration and extraction step is required. Conventional sample preparation techniques, such as liquid–liquid extraction and solid-phase extraction, are still widely used. Although these methods provide good analyte enrichment, large sample capacity, and high reproducibility, they suffer from several disadvantages, including being laborious, time-consuming, and requiring large amounts of hazardous organic solvents. To overcome these drawbacks, microextraction techniques have been introduced as alternative approaches. Different preconcentration based on microextraction approaches such as liquid–liquid microextraction (LLME)¹⁴, dispersive liquid–liquid microextraction¹⁵, and solid-phase microextraction¹⁶ have previously been employed for microextraction of polyphenols including curcumin. LLME is performed by injecting a suitable extractant (organic solvent) into an aqueous sample. After centrifugation, the organic phase containing the target analyte is separated and analyzed. LLME offers several advantages, including minimal organic solvent consumption, high enrichment capability, rapid operation, high efficiency, and lower cost compared to LLE and SPE. The solvents used in LLME include organic solvents, ionic liquids, deep eutectic solvents, supramolecular solvents^17–19, and fatty acids and fatty alcohols^20,21. The use of environmentally friendly green solvents, particularly supramolecular solvents (SUPRAS), is highly desirable for developing sustainable analytical methods^15,16,19.

SUPRAS are water-immiscible, non-volatile, non-flammable, and cost-effective green solvents formed through self-assembly of amphiphilic molecules in a continuous phase, resulting in organized nanostructures. These supramolecular assemblies exhibit excellent extraction capability owing to hydrogen bonding, ionic interactions, and hydrophobic effects. Long-chain alcohols such as decanol are commonly used as amphiphilic components in SUPRAS systems. In such systems, extraction occurs through partitioning of the analyte between the alcohol and an aqueous phase containing a polar organic modifier (commonly THF, although other suitable water-miscible solvents may also be used depending on the system). The alcohol spontaneously forms reverse micellar aggregates that organize into restricted-access supramolecular structures capable of efficiently extracting and enriching analytes^22,23. SUPRAS-LLME is a miniaturized version of this approach in which the volume of the extracting solvent is reduced to only a few microliters. Here, we introduce a modified form of SUPRAS-LLME, termed supramolecular solvent-based salt-saturated LLME (SUPRAS-SS-LLME), in which SUPRAS is employed as the extractant while the sample matrix is first saturated with a mixture of inorganic salts. Through the salting-out mechanism, the solubility of the analyte in the aqueous phase is reduced, enhancing its transfer into the extraction phase. The application of mixed salts further improves extraction efficiency by increasing ionic strength and strengthening analyte–solvent interactions^24,25. In the present work, a SUPRAS-SS-LLME procedure was developed and optimized by Box–Behnken design and the one-variable-at-a-time approaches for the efficient enrichment of curcumin prior to UV–Vis spectrophotometric determination in tomato, apple and fish samples. The synergistic combination of supramolecular solvent extraction and salt saturation was utilized to enhance the saturation effect and improve analyte partitioning, while chemometric optimization was employed to achieve rapid and robust extraction performance. The novelty of this work lies in the development of a SUPRAS-SS-LLME, in which a mixture of salts is used to achieve full saturation of the sample solution prior to extraction. This strategy provides a stronger salting-out effect compared to conventional single-salt systems and enhances the extraction efficiency of curcumin.

Experimental

Instruments and equipment

UV–Vis absorbance measurements were performed using a Unico model 2100 S spectrophotometer (Shanghai, China). A full spectral scan (325–500 nm) was performed to confirm the selectivity of the method, and the results indicate that the absorption at 425 nm is characteristic of curcumin with no significant spectral overlap, therefore a wavelength of 425 nm was employed for spectrophotometric determination. A Metrohm pH meter (model 630, Herisau, Switzerland) was used to measure the pH of the solutions. Fourier transform infrared (FTIR) spectra of samples were obtained by a Spectrum 100 FTIR spectrometer (PerkinElmer, Seer Green, Bucks, UK) and were applied for qualitative spectra interoperations as well as for structure elucidation. A centrifuge (Pole Ideal Tajhiz centrifuge, model PIT320R, Tehran, Iran) was applied for phase separation.

Reagents and chemicals

Curcumin (purity 99.3%), decanol, and tetrahydrofuran (THF) were obtained from Sigma-Aldrich (St. Louis, MO, USA). All other reagents used were of analytical grade and supplied by Merck (Darmstadt, Germany). A 500 mg/L stock solution of curcumin was prepared by dissolving 50 mg of the compound in 0.5 mL methanol and diluting to 100 mL with distilled water. Working standard solutions were freshly prepared daily by appropriate dilution of the stock solution with distilled water. For pH adjustment, 1.0 M NaOH or 1.0 M HCl solutions were used.

SUPRAS-SS-LLME procedure

The SUPRAS-SS-LLME procedure is illustrated in Fig. 1. First, the pH of the sample or standard solution was adjusted to 7.0 using 1.0 M NaOH or 1.0 M HCl, and solution was then saturated by a mixture of 1.12 g NaCl, 1.74 g Mg(NO₃)₂·6H₂O, and 3.32 g KNO₃ salts. After that, a 10.0 mL aliquot of this solution was taken, and a mixture of 100 µL of THF and 200 µL of decanol (1:2 decanol: THF, acting as the SUPRAS phase) were rapidly injected into the solution using a 100 µL glass syringe to form a cloudy dispersion. The solution was manually mixed for 1 min and subsequently centrifuged at 5000 rpm for 4 min. Finally, 50 µL of the separated SUPRAS phase was carefully collected from the upper layer using a microsyringe and transferred to the spectrophotometer for curcumin analysis.

Fig. 1.

Schematic illustration of the SUPRAS-SS-LLME procedure for the extraction of curcumin. The figure shows the formation of the supramolecular solvent (SUPRAS), addition of the salt mixture to achieve saturation, dispersion of the extraction solvent, phase separation, and collection of the enriched phase prior to UV–Vis measurement at 425 nm.

Recovery studies were carried out by spiking known amounts of curcumin into each sample matrix, and the recovery percentage was calculated according to Eq. 1:

1

Where C_found, is the curcumin concentration in the spiked sample, C_real is the initial concentration in the unspiked sample, and C_added is the concentration of added standard solution.

Pretreatment of fish samples

To evaluate the applicability of the SUPRAS-SS-LLME method, it was applied to the determination of curcumin in the muscle tissue of Parastromateus niger fish. The fish samples were obtained from a local commercial fish market (Haft Tir station, Chabahar Bay, a coast located in the southern east of Iran, bordering the Indian Ocean) as food-grade materials intended for human consumption. Extraction of curcumin from fish muscle was carried out based on EPA Method 3550 C with slight modifications²⁶. A dried muscle sample (1.0 g) was immersed in 5.0 mL of THF and sonicated for 20 min. The mixture was then centrifuged at 5000 rpm for 4 min, and the supernatant was transferred into a 10 mL conical tube and stored for subsequent analysis. After ultrasonic treatment and heating for 20 min, the volume of THF was reduced to 100 µL. Then, 200 µL of decanol and 10 mL of water were added, and the extraction was performed according to the procedure described in Sect. “SUPRAS-SS-LLME procedure”.

Pretreatment of tomato and apple samples

Extraction of curcumin from tomato and apple samples was carried out according to the method reported by Kanberoglu et al., with minor modifications²⁷. Tomato and apple samples were dried at 70 °C and powdered. 1.0 g of each sample was transferred to a stoppered tube. Then, 5 mL of THF was added, and the mixture was sonicated for 60 min. The solution was filtered and stored at 4 °C until analysis. After sonication and heating, the THF volume was reduced to 100 µL, and the SUPRAS-SS-LLME procedure was subsequently applied.

Results and discussion

Characterization of SUPRAS

Fourier transform infrared (FTIR) spectroscopy was employed to investigate the interactions involved in the formation of the supramolecular solvent (SUPRAS). The spectra of tetrahydrofuran (THF), 1-decanol, and the resulting SUPRAS are presented in Figure SI1. In the FTIR spectrum of 1-decanol, a broad band observed at approximately 3320 cm^− 1 is attributed to O–H stretching vibrations. The bands in the range of 2840–2930 cm^− 1 correspond to the C–H stretching vibrations of –CH₂ and –CH₃ groups. Additionally, the band at around 1465 cm^− 1 is assigned to C–H bending vibrations, while the peak near 1060 cm^− 1 is related to C–O stretching. The FTIR spectrum of THF shows characteristic bands in the range of 2860–2980 cm^− 1 due to symmetric and asymmetric C–H stretching vibrations, along with a prominent C–O stretching band at approximately 1090 cm^− 1.

The FTIR spectrum of the SUPRAS exhibits the characteristic peaks of both 1-decanol and THF, confirming the presence of both components in the final system. Notably, peak broadening, particularly in the O–H stretching region, is observed in the SUPRAS spectrum, indicating the formation of intermolecular interactions such as hydrogen bonding. These spectral changes support the successful formation of a supramolecular structure.

Optimization of SUPRAS-SS-LLME

To determine the optimum extraction conditions for the SUPRAS-SS-LLME method, five variables affecting the analytical signal were investigated using a combination of Box–Behnken design and the one-variable-at-a-time approach. The parameters studied included pH, volume of decanol, volume of THF, mixing (extraction) time, and centrifugation time and speed. All experiments were performed in triplicate.

Extraction (mixing) time

Extraction time is an important parameter in LLME because it determines the period required for the analyte to interact with the supramolecular solvent (the decanol-THF mixture). This interaction time directly influences extraction recovery (ER%), absorbance response, equilibration, selectivity, and reproducibility. Therefore, monitoring mixing time is essential to ensure efficient and reproducible analyte extraction. Manual mixing was examined in the range of 0–2 min, and the results indicated that ER% decreased beyond 1 min (Figure SI2), likely because of emulsion instability. Thus, 1 min was selected as the optimum extraction (mixing) time and was applied in subsequent experiments.

Centrifugation time and speed

The results indicated that extraction efficiency increased progressively with increasing centrifugation speed and time, reaching quantitative recovery at 5000 rpm (Figure SI3) and 4 min (Figure SI4), beyond which no further improvement was observed. Therefore, 4 min and 5000 rpm were selected as the optimum centrifugation conditions. The enhanced extraction efficiency (ER%) can be attributed to improved phase separation under these conditions, confirming the importance of centrifugation parameters in achieving effective separation in the SUPRAS-SS-LLME process. Under the optimized conditions, the total extraction time was approximately 6 min, highlighting the rapidity of the developed method.

Experimental design

In preliminary experiments, an OFAT strategy was used to identify suitable parameter ranges for subsequent chemometric evaluation. A Box–Behnken design (BBD) was then employed to investigate the interactions among key extraction parameters and to optimize the extraction efficiency of curcumin. The significant factors selected for optimization were pH (X₁/A), decanol volume (X₂/B), and THF volume (X₃/C). The low, middle, and high levels of each parameter were coded as − 1, 0, and + 1, respectively (Table 1).

Table 1.

Box–Behnken experimental design matrix and corresponding experimental and predicted absorbance values for curcumin extraction.

No.	pH	Volume of decanol	Volume of THF	Response	Predicted	Error
1	6	150	100	0.59	0.59	0.42
2	6	200	50	0.63	0.63	0.40
3	6	200	150	0.64	0.64	0.39
4	6	250	100	0.58	0.59	−1.29
5	7	150	50	0.67	0.68	−0.75
6	7	150	150	0.68	0.69	−0.74
7	7	200	100	0.78	0.77	1.54
8	7	200	100	0.77	0.77	0.26
9	7	200	100	0.77	0.77	0.26
10	7	200	100	0.75	0.77	−2.40
11	7	200	100	0.77	0.77	0.26
12	7	250	50	0.69	0.69	0.72
13	7	250	150	0.70	0.70	0.71
14	8	150	100	0.61	0.60	1.23
15	8	200	50	0.65	0.65	−0.38
16	8	200	150	0.66	0.66	−0.38
17	8	250	100	0.62	0.62	−0.40
In range	6.5	200	150	0.71	0.72	−1.65
In range	8	200	100	0.67	0.68	−0.97
In range	7	150	75	0.68	0.69	−1.73
In range	6.5	200	100	0.75	0.74	1.9
In range	7.3	150	150	0.67	0.68	−1.17
Out range	5.8	200	150	0.60	0.59	1.79
Out range	7	120	120	0.60	0.60	0.81
Out range	8.3	200	150	0.6	0.59	0.92
Out range	7	270	120	0.63	0.65	−2.94
Out range	8	250	40	0.58	0.59	−1.58

The relationship between the analytical response and the independent variables was modeled using a second-order polynomial equation expressed as Eq. 2:

2

Where Y is the predicted response, β₀ is the intercept, β_i, and β_ij represent the linear, quadratic, and interaction coefficients, respectively, X_i denotes the coded variables, and ε is the random error^25,28.

Quadratic regression fitting of the BBD results yielded the following model for absorbance (Eq. 3):

3

Optimum extraction conditions were determined by solving the partial derivative equations ∂Y/∂A = 0, ∂Y/∂B = 0, ∂Y/∂C = 0 and ∂Y/∂D = 0. The calculated optimum values were pH (A) = 7.06, volume of decanol (B) = 202.07 µL and volume of THF (C) = 106.65 µL.

The pH of the sample or standard solution is a critical parameter in the SUPRAS-SS-LLME process as it dictates the ionization state and, consequently, the hydrophobicity of curcumin. For a target analyte to be efficiently partitioned from the polar aqueous sample into the apolar SUPRAS phase (where decanol acts as the primary extraction medium), it must predominantly exist in its neutral form. Curcumin is a polyphenolic compound with acidic pKa values (typically pKa₁ ≈ 7.8 and pKa₂ ≈ 8.5). As shown in the optimization results (Fig. 2), at pH values above 8.0, curcumin undergoes deprotonation of its phenolic and enolic hydroxyl groups, converting into its ionized, negatively charged form. These anionic species exhibit increased water solubility and a significantly reduced affinity for the organic supramolecular phase, leading to a sharp decrease in extraction recovery and absorbance. Conversely, while curcumin remains in its molecular (neutral) state at acidic pH levels, a slight decrease in extraction efficiency was observed compared to neutral conditions. This may be attributed to shifts in the keto-enol tautomeric equilibrium; at lower pH, the keto form may be more prevalent, which can slightly alter its partitioning behavior compared to the enol form favored at neutral pH. Furthermore, at high pH values, partial deprotonation of the decanol hydroxyl groups may occur, potentially disrupting the self-assembled supramolecular structure and reducing the effective volume of the extraction phase. Based on these detailed chemical considerations, pH 7.0 was selected as the optimal value to maximize both analyte partitioning and the stability of the SUPRAS system.

Fig. 2.

Two-dimensional response surface contour plots showing the interaction effects of pH (A), decanol volume (B), and THF volume (C) on the absorbance of curcumin. The optimal conditions were found to be pH = 7.06, decanol volume = 202.07 µL, and THF volume = 106.65 µL.

Regarding the solvent volumes, the ratio of decanol (extractant) to THF (disperser) is vital for achieving high enrichment. Enhancing the extraction solvent volume beyond the calculated optimum results in a decrease in the analytical signal due to the dilution of the extracted curcumin within a larger organic phase. Conversely, when the solvent volume is below the optimum, the formed SUPRAS phase is insufficient to effectively extract the analyte from the 10 mL aqueous matrix, resulting in poor mass transfer and lower recovery.

Analysis of variance (ANOVA) confirmed that the quadratic model was statistically significant at the 95% confidence level (Table 2). There is only a 0.01% chance that a “Model F-Value” could occur due to noise. Values of “Prob > F” less than 0.0500 indicate model terms are significant. Values greater than 0.1000 indicate the model terms are not significant. The model F-value of 75.72 indicated strong model adequacy, while “Prob > F” values below 0.05 demonstrated the significance of key terms (A, A², B², C²). The lack-of-fit determined the failure of the model to represent values in the regression. The non-significant lack-of-fit value (F = 0.69) indicated good agreement between the model and experimental data. Non-significant of lack of fit is good. The coefficient of determination (R² = 0.9898, that only 1.02% of the total variations were not determined by the model), adjusted R² (0.9768), and predicted R² (0.9338) further supported the accuracy and reliability of the model. The R² value and the lack-of-fit test were utilized during protocol validation to assess the linearity of the calibration curve (details are available in the Supplementary Data). The F-value and p-value for the lack-of-fit in the ANOVA were 75.72 and 0.69, respectively, indicating that the model is valid. Specifically, there is a 60.19% chance that a lack-of-fit F-value this large could occur due to noise. Furthermore, high precision and reliability were demonstrated by a low coefficient of variation (CV = 1.5%). Two-dimensional contour plots illustrating the effects of variable interactions on curcumin absorbance (which correlates to extraction recovery) are shown in Fig. 2.

Table 2.

ANOVA results for the quadratic model describing curcumin extraction based on the BBD.

Source	Sum of squares	Degree of Freedom	Mean square	F Value	p-value Prob > F
Model	0.071	9	0.0079	75.72	< 0.0001
A	0.0012	1	0.0012	11.99	0.01
B	0.0002	1	0.0002	1.92	0.21
C	0.0002	1	0.0002	1.92	0.21
AB	0.0001	1	0.0001	0.96	0.36
AC	0	1	0	0	1.00
BC	0	1	0	0	1.00
A2	0.045	1	0.045	436.69	< 0.0001
B2	0.017	1	0.0172	165.38	< 0.0001
C2	0.00152	1	0.0015	14.57	0.007
Residual	0.00073	7	0.00010
Lack of Fit	0.00025	3	8.33 × 10− 5	0.69	0.60
Pure Error	0.00048	4	0.00012
Cor Total	0.072	16

Validation of the SUPRAS-SS-LLME method

To evaluate the performance of the proposed SUPRAS-SS-LLME method, its linearity, precision, accuracy, repeatability, enrichment factor (EF), limit of detection (LOD), and limit of quantitation (LOQ) were examined under optimized conditions. Calibration curves were constructed using absorbance values obtained from ten increasing spiked concentration levels ranging from 1 to 5000 µg/L. A good linear response was achieved for curcumin, with a correlation coefficient (R²) of 0.9904 and the calibration equation: A = 0.3539 C + 0.1244; where A is absorbance and C is curcumin concentration (mg/L).

Intra-assay and inter-assay precision were expressed as the RSD of five replicate analyses of control samples at three spiked levels. Measurements were performed on the same day and on three different days, giving RSD values below 3.1% and 4.4%, respectively. The intermediate precision (batch-to-batch) was 4.3% (RSD) at a curcumin concentration of 1.0 mg/L. Under the optimized conditions, the EF, calculated as the ratio of curcumin concentration in the organic (separated) phase (C_org) to its initial concentration in the sample (C₀)²³, was 131-fold. It should be noted that this value represents the experimental enrichment achieved and may differ from the theoretical volume ratio due to extraction efficiency and matrix effects. The enhanced signal demonstrated the effective purification and enrichment capability of the SUPRAS-SS-LLME procedure. The LOD and LOQ for curcumin were calculated as three and ten times the standard deviation of seven consecutive blank measurements divided by slope of the calibration curve²², respectively, and were determined to be 0.23 µg/L and 0.77 µg/L.

Comparison of the performance of extraction with and without salt saturation

The performance of SUPRAS-SS-LLME, supramolecular solvent-based liquid–liquid microextraction (SUPRAS-LLME), and conventional salting-out (addition of NaCl) was compared for the extraction of curcumin at concentrations of 0.1, 0.5, 1.0, 2.5, and 5.0 mg/L under optimized conditions. As shown in Fig. 3, saturation of the solution with a mixture of salts significantly enhanced the analytical signal, indicating improved extraction recovery. On average, a 32.7% increase in signal intensity was observed under salt-saturated conditions. This enhancement can be attributed to the salting-out effect, where high ionic strength reduces the solubility of curcumin in the aqueous phase and promotes its transfer into the supramolecular extraction solvent. Compared to salt saturation, the addition of 1.1 g NaCl alone resulted in a lower analytical response, indicating less efficient extraction (Fig. 3). The superior performance of mixed salts is likely due to the combined effect of different ions, which increases the ionic strength and enhances molecular interactions, thereby further decreasing the solubility of curcumin in the aqueous phase. Moreover, the use of mixed salts (e.g., 1.1 g NaCl and 3.3 g KNO₃) improved the extraction efficiency compared to the use of a single salt. However, complete salt saturation provided the highest analytical response among all tested conditions. The increase in analytical signal (32.7%) obtained under salt-saturated conditions was statistically significant (p < 0.05, Student’s t-test), confirming the effectiveness of the proposed approach.

Fig. 3.

Comparison of extraction performance for curcumin using supramolecular solvent-based microextraction under different conditions: without salt, with salt saturation, with a single salt (NaCl), and with a mixture of two salts (NaCl and KNO₃), under optimized conditions. The analytical signal corresponds to absorbance measured at 425 nm. Salt saturation significantly enhanced the extraction efficiency, resulting in a 32.7% increase in signal intensity.

Comparison of SUPRAS-SS-LLME with supramolecular solvent-based reversed-phase liquid–liquid microextraction

The performance of the proposed SUPRAS-SS-LLME method was also compared with supramolecular solvent-based reversed-phase liquid–liquid microextraction (SUPRAS-RPLLME). In SUPRAS-RPLLME, curcumin reacts with 4-vinylpyridine in an organic medium and is subsequently extracted into water as the extraction solvent. Briefly, 10.0 mL of standard n-hexane solution containing curcumin was placed in a glass tube, followed by the addition of 200 µL THF (as disperser), 50 µL of 0.7 M NaOH (as extraction solvent), and 50 µL of 1.4% w/v potassium peroxodisulfate. The pH was adjusted to 9.5, the solution was shaken for 10 min, and 50 µL of the separated aqueous phase was withdrawn and analyzed spectrophotometrically. As illustrated in Fig. 4, the proposed SUPRAS-SS-LLME method produced significantly higher analytical signals than SUPRAS-RPLLME, confirming its superior extraction capability and analytical performance.

Fig. 4.

Comparison of analytical signals obtained using SUPRAS-SS-LLME and SUPRAS-RPLLME for the extraction of curcumin under optimized conditions. The analytical signal corresponds to absorbance measured at 425 nm. The SUPRAS-SS-LLME method produced significantly higher signals than SUPRAS-RP-LLME, confirming its superior extraction efficiency and analytical performance.

Selectivity of the SUPRAS-SS-LLME method in the presence of potential interferences

Quercetin, methyl orange, brilliant green, acid red, malachite green, resveratrol and ions (Na⁺ and Cl^- as NaCl, Mg²⁺, K⁺, and NO₃^- as Mg(NO₃)₂·6H₂O and KNO₃), are commonly encountered alongside curcumin in many real samples and may potentially interfere with its determination. To assess selectivity, 10 mL aliquots containing 1.0 mg/L of curcumin were spiked with equal concentrations of each potential interferent (except interfering ions) and subjected to the suggested extraction procedure. The amount of applied ions were saturated in standard mixtures No noticeable interference was observed in the determination of curcumin, demonstrating that the developed method possesses good selectivity for curcumin extraction. The absence of interference from common co-existing compounds was further confirmed through real sample analysis (tomato, apple, and fish), suggesting that their matrix do not significantly affect the determination at 425 nm.

Comparison of the SUPRAS-SS-LLME method with other similar methods

The analytical performance of the developed SUPRAS-SS-LLME method was compared with previously reported methods for the determination of curcumin, as summarized in Table 3. The SUPRAS-SS-LLME approach offers several advantages, including a wide linear range, good precision with comparable RSD values, shorter analysis time, and reduced consumption of sample, eluent, lower LOD, and organic solvent. These features highlight its compatibility with the principles of green analytical chemistry and demonstrate its superiority over many conventional extraction techniques. Extraction with DLLME and deep eutectic solvent-microwave-assisted showed a wider linear range in compared with the developed method, practically because of using a more advanced analytical instrument.

Table 3.

Comparison of the analytical performance of the developed SUPRAS-SS-LLME method with previously reported methods for curcumin determination.

Extraction method	Detection system	Linear range (mg/L)	LOD (mg/L)	RSD (%)	Extraction time (min)	Ref.
DLLME	HPLC-MS/MS	0.740–18.500	0.0183	4.9	10	15
DES-MAE	HPLC	15.625–1000.0	0.320	2.42	20	29
LLUAE	HPLC-DAD-MS	5.260–10.220	0.590	1.28	10	30
SUPRAS-SS-LLME	Spectrophotometry	0.001–5.0.001.0	0.00023	3.7	6	This work

Abbreviations: DLLME, dispersive liquid–liquid microextraction; HPLC-MS/MS, ultrahigh-performance liquid chromatography-tandem mass spectrometry; DES-MAE, deep eutectic solvent-microwave-assisted extraction; HPLC, high performance liquid chromatography; LLUAE, liquid–liquid ultrasound-assisted extraction; HPLC-DAD-MS, high-performance liquid chromatography coupled with diode-array detection and electrospray ionization tandem mass spectrometry.

Applications

The applicability of the proposed method was evaluated by determining curcumin in real tomato, apple, and fish samples.

Three concentration levels were spiked into each matrix, and the corresponding recoveries obtained using the proposed SUPRAS-SS-LLME method are summarized in Table 4. Excellent recoveries of 99.1–99.6% with relative standard deviations (RSDs) of 1.1–3.7% were obtained for tomato samples, while recoveries of 98.0–99.7% with RSDs of 1.7–3.3% were achieved for fish samples. In addition, apple samples showed good recoveries in the range of 99.2–99.7% with RSDs between 1.3% and 3.3%. These results confirm the accuracy, reliability, and applicability of the developed SUPRAS-SS-LLME method for curcumin determination in real food matrices, also indicate negligible matrix effects. Figure 5 shows the UV–Vis absorption spectra of curcumin extracted from a tomato sample spiked at 2.0 mg/L, compared with the corresponding blank.

Table 4.

Determination of curcumin in tomato, apple and fish samples using the SUPRAS-SS-LLME method.

Sample	Analyte added (mg/L)	Analyte found (mg/L)	Recovery (%)	RSD % (n = 3)
Tomato	0	0.0401	-	1.1
	0.1	0.1392	99.1	3.1
	2.5	2.530	99.6	3.2
	4.5	4.5182	99.5	3.7
Apple	0	0.0484	-	1.3
	0.1	0.1476	99.2	2.6
	2.5	2.542	99.7	2.7
	4.5	4.5312	99.6	3.3
Fish	0	0.0467	-	1.7
	0.1	0.1447	98.0	3.3
	2.5	2.525	99.1	2.7
	4.5	4.5354	99.7	3.0

Fig. 5.

UV–Vis absorption spectra of curcumin after SUPRAS-SS-LLME extraction. Spectral range: 325–500 nm. The maximum absorbance observed at 425 nm was used for quantification. The spectra confirm method selectivity in the presence of potential co-extracted compounds.

Greenness considerations of the SUPRAS-SS-LLME method

The environmental and safety profile of the developed SUPRAS-SS-LLME method was evaluated according to the principles of Green Analytical Chemistry (GAC). A central concern in GAC is the toxicity and volume of organic solvents. While THF is classified as a hazardous solvent, the proposed protocol utilizes a strictly micro-volume of only 100 µL of it per extraction. By operating at this scale, the method significantly minimizes hazardous waste generation and reduces operator exposure compared to conventional liquid-liquid extraction techniques that typically require milliliters or liters of solvent. Furthermore, the protocol completely avoids the use of chlorinated solvents or highly persistent organic pollutants. To provide a quantitative evaluation, the method was assessed using the Analytical GREEnness (AGREE) metric that showed an estimated overall score of 0.78 for the developed method. The AGREE assessment highlights several green strengths of the method: waste minimization: the total organic waste produced is limited to approximately 300 µL (100 µL THF and 200 µL decanol), representing a very high score in waste reduction; energy efficiency: the entire extraction and separation process is completed within a total time of approximately less than 10 min, which minimizes energy consumption; methodological simplicity: the supramolecular solvent phase is formed through spontaneous self-assembly, eliminating the need for additional chemical reagents or high-energy synthesis steps; minimal sample handling: the method requires no chemical derivatization, heating, or complex clean-up procedures; instrumental footprint: use of spectrophotometer is inherently “greener” than HPLC as it consumes significantly less power and requires no mobile phase solvents during the analysis phase. These features confirm that the SUPRAS-SS-LLME approach aligns with the core principles of GAC and serves as a sustainable alternative for the routine determination of curcumin in food matrices.

Conclusion

A salt-saturated supramolecular solvent-based liquid–liquid microextraction method coupled with UV–Vis spectrophotometry was successfully developed for the determination of trace levels of curcumin in food samples. The combined use of supramolecular solvents and salt saturation significantly enhanced extraction efficiency through an intensified salting-out effect, enabling rapid and effective analyte enrichment within a total extraction time of only 6 min. The proposed method provided a high enrichment factor (131), low detection limit (0.23 µg/L), and good precision, demonstrating its suitability for trace-level analysis. Chemometric optimization using Box–Behnken design ensured reliable identification of optimal extraction conditions and confirmed the robustness of the analytical model. The method showed satisfactory selectivity and was successfully applied to tomato and fish samples, yielding recoveries between 98.0% and 99.7%, thereby confirming its accuracy in complex food matrices. Owing to its simplicity, low organic solvent consumption, minimal sample handling, and favorable analytical performance, the proposed SUPRAS-SS-LLME approach represents a green, cost-effective, and efficient alternative to conventional extraction techniques for the routine spectrophotometric determination of curcumin. The strategy may also be extended to the analysis of other hydrophobic bioactive compounds in food and related matrices.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

ZB: Investigation, Data curation. SHH: Conceptualization, Investigation, Writing – original draft, Writing – review and editing. AJK: Statistical analysis. MK: Conceptualization, Writing – original draft, Writing – review and editing. All authors read and approved the final manuscript.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.