A comprehensive insight into optimized ultrasound-assisted extraction of bioactive compounds and antioxidant activities from dandelion (Taraxacum officinale) plant extract using response surface methodology

Abstract

Dandelion (Taraxacum officinale) is an edible medicinal herb having an extended history for its traditional usage owing to the health promoting benefits associated with this plant. Nevertheless, traditional extraction methods limit the recovery of bioactive compounds from different parts of dandelion and insufficient research is available on process optimization. Hence, current research developed an effective ultrasound-assisted extraction method for maximum recovery of total phenolics contents (TPC), total flavonoids contents (TFC), antioxidant activities (DPPH, ABTS, FRAP assays), and bioactive compounds (HPLC) from dandelion plant using response surface methodology in combination with Box-Behnken design (BBD). The optimal extraction conditions determined by RSM were as follows: sonication time, 30 min; ultrasound amplitude, 70%, and ultrasound temperature, 40 °C. At these conditions, the recovery of total phenolics and total flavonoids reached 40.77 mg GAE/g and 22.68 mg RE/g, respectively. Moreover, the antioxidant activities determined based on DPPH-scavenging, ABTS⁺-scavenging and FRAP were reported as 88.55%, 445.39 µM TE/mg, and 30.64 mg TE/g, respectively. Additionally, a total of 11 major bioactive compounds were quantified using HPLC, including 6 phenolic acids and 5 flavonoids. Under optimized conditions major bioactive compounds identified and quantified were chlorogenic acid, quercetin, apigenin, luteolin-7-O-glycoside, luteolin, p-coumaric acid, caffeic acid, ferulic acid, cichoric acid, isoetin, and caftaric acid. Conclusively, results of present study give a comprehensive insight into optimized ultrasound-assisted extraction method for recovery of maximum antioxidants and bioactive compounds from dandelion using a combination of BBD and RSM. Furthermore, this study may provide a reference in utilizing optimized extraction process for dandelion bioactive compounds in food and pharmaceutical industry.

1. Introduction

Green chemistry focuses on promoting efficient, eco-friendly, and sustainable extraction methods mainly aimed at reducing resource consumption and environmental burden. Current thematic and industrial research are primarily focusing on under-utilized plants and plant parts as potential resource of bioactive compounds in medicinal applications. Dandelion (Taraxacum officinale) is a flowering plant with a wide range of valuable bioactive compounds, including phenolic acids (such as chlorogenic and chicoric acids), triterpenes, flavonoids, and inulin. All of these compounds are known for their antioxidative, anti-inflammatory, hepatoprotective, anticancer, and prebiotic potentials [1], [2]. Traditionally used extraction methods such as maceration, decoction, percolation, and Soxhlet extraction, have been employed to extract the phytochemicals from a diverse array of plants and plant parts. But these conventional extraction methods have significant associated disadvantages like higher energy consumption, large volumes of organic solvents, extended sonication time, and the potential thermal degradation of labile compounds. All of these disadvantages contradict the basic principles of green and sustainable and eco-friendly processing. Therefore, there is a dire need for advanced green extraction technologies that can enhance the efficiency of extraction by utilizing the resources in an economical way with minimal environmental impact [3].

The main aim of current research work was to provide a comprehensive insight into optimized ultrasound-assisted extraction to enhance the recovery of bioactive compounds and antioxidant properties from dandelion (Taraxacum officinale) plant extract using response surface methodology. Hence, the primary objective of the current research was focused on understanding the effects and interactions among different operating conditions of ultrasonication instead of comparison among different extraction techniques. Nonetheless, optimized ultrasonication with reference to other latest extraction procedures is significant to illuminate its practical relevance. Among various modern extraction techniques, microwave-assisted extraction (MAE) technique is mainly classified as an effective green extraction technique responsible for enhancing extraction yields through quick volumetric heating that aids in penetration of extraction solvent and disruption of plant cellular matrices [4], [5]. MAE is employed by researchers owing to its decreased solvent usage and sonication times along with increased recovery of bioactive compounds from plant cellular structures. Nevertheless, uneven distribution of energy and localized elevated temperatures results in detrimental effects on heat-labile phytomolecules if processing conditions are not properly controlled [6], [7]. Similarly, supercritical carbon dioxide (SC-CO₂) is another modern extraction technique well-recognized among scientific community owing to its adjustable selectivity, solvent-free nature and suitability towards volatile and non-polar constituents [8], [9]. In spite of all these merits, this modern extraction technique requires elevated operating-pressure conditions, costly operational expenses, and high capital investment hinders its overall application in effective recovery of bioactive compounds from plant matrices [10], [11].

As compared to MAE and SC-CO₂, ultrasound-assisted extraction provides a reasonable balanced extraction efficiency, easy operational handling, feasible operational costs, and overall simplicity. During ultrasonication, the acoustic cavitation results in disruption of plant cellular structure due to generated shock waves that has significant effect on mass transfer, facilitating efficient release of plant bioactive compounds at relatively low temperatures [12], [13]. These advantages of employing ultrasound assisted extraction (UAE) makes it a suitable technique for extraction of heat-sensitive bioactive and antioxidants compounds [14], [15], [16], [17]. The optimized ultrasound-assisted extraction conditions established in this study yielded maximum recovery of bioactive compounds and showed significant antioxidant properties, highlighting the importance of UAE as a suitable, attractive, and sustainable alternative to other modern extraction techniques (MAE & SC-CO₂) for extraction of compounds from T. officinale.

Recently, ultrasound-assisted extraction (UAE) has emerged as one of the main eco-friendly, green extraction techniques that perfectly align with the principles of green extraction [18]. UAE primarily works on the principles of acoustic cavitation that produces high-frequency ultrasonic waves in the liquid medium generating successive cycles of compression and rarefaction. The principle of cavitation leads to the formation, growth, and forceful breakdown of microscopic bubbles similar to the conditions of high temperatures, pressures, and intense shear forces within the medium. These intense conditions involved in the disruption of plant cell walls, which ultimately facilitating solvent penetration into the cellular matrix, and thus resulting in mass transfer of intracellular compounds into the solvent leading to improve yields [19]. Several interconnected parameters and variables like ultrasonic power/amplitude, frequency, extraction temperature, time, and the composition of the solvent system are linked to improve the effectiveness of UAE. It is crucial to optimize these extraction parameters for economize the recovery of bioactive compounds with the focus on preserving their integrity [20], [21].

UAE is extensively used for extraction of bioactive compounds from various medicinal plants. Even though phytochemical composition and biological characterization of Taraxacum officinale have been studied extensively, however, previous studies mainly focus on individual dandelion plant part like roots, leaves, or flower and either these studies employed traditional extraction methods or single parameter ultrasound-assisted extraction. Furthermore, dandelion-based ultrasonication studies have used single-factor approaches or usually employed variables like solvent concentration, solid-liquid ratio, sonication time, and ultrasound power. Hence, the effects of multi-variable ultrasonic parameters on diverse-responses remain insufficiently explored. Therefore, the novelty of current study is extraction from whole dandelion plant rather than individual plant part, multi-variable RSM design (sonication temperature (25–55 °C), amplitude (50–90%), & sonication time (15–45 min)) instead of solvent concentration, solid-liquid ratio, sonication time, and ultrasound power and simultaneous multiple-responses rather than single response (yield, quantification of one compound, or single antioxidant assay). Scientifically, current independent variables were selected in this study based on their direct effect on effectiveness of ultrasonication. As amplitude directly governs the cavitation phenomena that directly influences the disruption of cell wall and effectiveness of mass transfer. Whereas, sonication time and temperature regulate the diffusion equilibrium and solute solubility, respectively. On the other hand, practically, amplitude, sonication temperature and time are relatively considered adjustable in commercial sonication systems, while keeping solid-liquid ratio and solvent system fixed to enhance the overall reproducibility of the process. Further, the selection of independent variables in this study is also supported by various earlier studies on UAE optimization [22], [23], [24]. To date, no previous studies have studied whole dandelion plant, multi-variable combination of ultrasonication factors (sonication temperature, amplitude, & time), comprehensive antioxidant assays (TPC, TFC, DPPH-scavenging, ABTS+-scavenging and FRAP activities), and quantification of bioactive compounds as a comprehensive optimization framework. Most of the literature mainly focused on a narrow range of compounds and lack a systematic optimization approach. Therefore, this study mainly aims to optimize UAE conditions for the extraction of vital bioactive compounds present in dandelion. The effect of sonication conditions such as amplitude (%), time (min), and temperature (°C) on the extraction yield, antioxidant activity, and recovery of bioactive compounds is investigated and optimized using Response Surface Methodology (RSM) with a Box-Behnken Design (BBD). This research is carried out to establish UAE as a robust, green, and efficient platform for valorizing dandelion biomass, contributing to the development of sustainable nutraceutical and pharmaceutical ingredients.

2. Material and methods

2.1. Procurement of Raw Material

The fresh dandelion (Taraxacum officinale) plant was collected from botanical garden in Lahore, Pakistan. Procured dandelion was shifted to the laboratory, and then washed cleaned to remove dust and dirt particles adhering to the surface. The cleaned sample was then dried under shad at room temperature (25–30 °C) for seven to ten days. During drying period, the plant material was continuously overturned to prevent growth of microorganisms. The dried dandelion was converted in to less than 85-micron size fine powder through a hammer mill. This fine powder was packed in zipper bags and stored at 4 °C in a refrigerator. The further analysis was carried out within a 48 hours’ time after the grinding process.

2.2. Extraction Optimization

In this experiment, the bioactive compounds of dandelion were extracted by sonicator (Model VCX 750, Sonics & Materials, Inc., Newtown, CT, USA) using ethanol as solvent. The operating conditions included 20 kHz of ultrasound frequency with nominal power output of 750 W (max). For this experimentation, a titanium alloy probe having ≈13 mm diameter was used. Digital weighing machine was used to accurately measure the weight of dandelion powder (100 ± 0.1 g). Response surface methodology (RSM) with a Box-Behnken Design (BBD) was used to observe the effects of three independent variables; sonication time (15–45 min), ultrasonic amplitude (50–90%), and temperature (25–55 °C) on extraction efficiency. It is worth mentioning here that amplitude is instrument-dependent and may limit cross-platform comparability. Even though the calorimetric calibrations were not performed in this study, hence to ensure internal reproducibility, all the experimentations were conducted using consistent settings of ultrasonicator. Levels of the independent variables were mainly aimed at exploring broader experimental window in generating novel predictive insights instead of constraining the study to earlier reported ranges. Desirability function approach in Design-Expert software (Stat-Ease Inc., Minneapolis, USA) was used to conduct multi-response numerical optimization for the determination of optimal UAE conditions. Response surface methodology (RSM) with a Box-Behnken Design (BBD) was used to observe the effects of three independent variables at three levels for the development of quadric regression models against each response variable. Factors used in Box–Behnken Design are presented in Table 1. A maximize goal was assigned to all response variables for simultaneous optimization, whereas the goals for three independent variables (sonication time, sonication temperature, and amplitude) were in experimental ranges. For the identification of optimal conditions, composite desirability value predicted by RSM were maximized. The extractions were performed using an UAE system with 70% ethanol as the solvent. Temperature during extraction was regulated by a digital thermometer installed in ultrasonic apparatus within a narrow range of ±1.5 °C using thermostatically water bath. During ultrasonication, including high amplitude and longer duration runs, the temperature overshoot was regulated and maintained by thermostatically controlled water bath. After completion of each run, filtration of extract was carried out through a 40- mesh size screen. Afterwards, the centrifugation process was carried out at approximately 4200 × g with centrifuge machine (320R, Hettich, Germany) for 10 minutes to separate the dandelion extract from the solution [25].

Table 1.

Factors used in Box–Behnken Design.

Sonication variables	Coded levels
	−1	0	+1
Sonication time (min)	15	30	45
Sonication amplitude (%)	50	70	90
Sonication temperature (°C)	25	40	55

2.3. Total phenolic content (TPC) of dandelion extracts

TPC in dandelion extract was determined using the microplate-adapted Folin-Ciocalteu assay with gallic acid as a standard, where 20 µL of the diluted dandelion extract was mixed with 100 µL of Folin-Ciocalteu reagent (diluted 1:10) and 75 µL of sodium carbonate solution (75 g/L), incubated in the dark at room temperature for 2 hours. The absorbance was measured at 740 nm using a UV–visible spectrophotometer. Gallic acid is used as standard and the phenolic content calculated by comparing with the gallic acid standard curve (10–200 mg/L) and is expressed as mg of Gallic Acid Equivalents (GAE) per gram of dry dandelion extract (mg GAE/g) [26], [27].

2.4 Total. flavonoid content (TFC) of dandelion extracts

The aluminum chloride colorimetric method was used to estimate the TFC as described by Xiang et al. [28] with minor modifications. Briefly, 0.5 mL of dandelion extract was mixed with 1.5 mL of 95% (v/v) ethanol, 0.1 mL of 10% (w/v) aluminum chloride, 0.1 mL of 1 M potassium acetate, and 2.8 mL of distilled water. The mixture was then stored for half an hour at room temperature. UV–visible spectrophotometer at a wavelength 415 nm was used to measure the absorbance of the mixture. Calibration curve (20–100 µg mL⁻¹) by using Rutin as standard was prepared and the obtained results were expressed as milligrams of Rutin equivalents per gram of dry extract (mg RE g⁻¹) using the equation:

$$\mathit{TFC}=\frac{\mathit{CxV}}{m}$$

where C is the concentration obtained from the calibration curve (mg mL⁻¹), V is the volume of extract (mL), and m is the dry weight of the sample (g).

2.5 Antioxidant. potential of dandelion extract

2.5.1. 2,2-Diphenyl-1-picrylhydrazyl (DPPH) assay

The method of Ray et al. [29] with slight modifications was employed to assess the DPPH radical scavenging activity as an estimation of antioxidant potentials of ultrasound assisted dandelion extracts. 1000 µL of 0.1 mM DPPH solution is prepared in methanol for each run of ultrasonically assisted extract. The prepared solution was then rested for 30 minutes at room temperature. UV–visible spectrophotometer was used to observe the absorbance of sample solution and blank. Following formula was used to estimate the percent inhibition of DPPH assay:

$$\mathrm{I}\mathrm{n}\mathrm{h}\mathrm{i}\mathrm{b}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\left(\%\right)=\frac{\mathrm{a}\mathrm{b}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{b}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{k}-\mathrm{a}\mathrm{b}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{b}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{e}\mathrm{x}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{t}}{\mathrm{a}\mathrm{b}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{b}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{k}}\times 100$$

2.5.2. 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) radical scavenging assay

ABTS radical scavenging activities of dandelion extracts were estimated by using the method of Cano et al. [30] with minor modifications. The extract solution (25 μL) at different concentrations was added into 0.2 mL ABTS⁺ solution. Multi-detection microplate reader (SpectraMax M5, Molecular Device) was used to record the absorbance at 734 nm, after 6 min reaction. Standard curve was used to quantify the results by using Trolox as standard. Results were expressed as μmol TE per mg extract.

2.5.3. Ferric reducing antioxidant power (FRAP) assay

Methods of Benzie and Strain et al. [31] with slight modifications was used to estimate the FRAP assay to observe the ferric reducing antioxidant power of ultrasound assisted-dandelion extracts. In this assay, the FRAP reagent comprised of 300 mM acetate buffer [sodium acetate trihydrate (3.1 g) & acetic acid (1.6 mL)], 2,4,6-Tripyridyl-S-triazine (TPTZ) solution (10 mM) prepared in hydrochloric acid (40 mM), and Iron (III) chloride hexahydrate (20 mM) with a 10:1:1 (v/v) volumetric ratio was used. A volume of 400 μL sample solution was mixed with a volume of 3 mL FRAP reagent and this mixture was then placed in a water bath at a constant temperature of 37 °C for 30 minutes. UV–Visible spectrophotometer was used to observe the absorbance of solutions at 593 nm wavelength. Trolox reagent was used as standard and the final result was expressed as mg Trolox equivalents/g.

2.6. Quantification of Bioactive Compounds via HPLC

High-Performance Liquid Chromatography (HPLC) was used to separate and quantify the major bioactive compounds in the optimized dandelion extract, including polyphenols (chlorogenic acid, caffeic acid, p-coumaric acid, ferulic acid, caftaric acid, cichoric acid) and flavonoids (luteolin, apigenin, quercetin, luteolin-7-O-glycoside, isoetin). Content of bioactive compounds in dandelion extracts was assessed by Waters HPLC system that is equipped with PDA (photodiode array) detector. Separation of compounds was achieved using Waters Symmetry C18 column (4.6 × 250 mm, 5 µm). 0.1% formic acid in water (A) and methanol (B) is used as mobile phase. Gradient elution is performed, in which mobile phase B (10%) run for 0–5 min, followed by B (10–20%), B (20–30%), B (30–40%), B (40–45%), B (45–60%), B (60–70%), B (70–95%) for 5–15 min, 15–30 min, 30–35 min, 35–40 min, 40–45 min, and 45–50 min, respectively. The injection volume used in this analysis was 10 µL. Detection of phenolic acids and flavonoids were observed at 280 nm and 350 nm, respectively. Calibration curves were prepared using stock solutions in methanol, diluted to obtain concentrations ranging from 0.5 to 100 µg/mL.

2.7. Statistical analysis

The quadratic equation was used to describe the behavior of the Box–Behnken design. Three sonication runs were carried out for each treatment. The level of significance for each variable was evaluated through data analysis performed using the Design- Expert software package [32]. A 5% level of significance was employed to determine the statistical significance among various treatments.

3. Results and discussions

3.1. Extraction yield

The effects of extraction conditions used in this experiment i.e., sonication time (min), sonication amplitude (%), and sonication temperature (°C) on extraction yield (%) was investigated using a three level and three factor Box-Behnken design. The maximum extraction yield (12.25±0.82%) was observed at mid-range conditions i.e., sonication time 30 minutes, amplitude 70%, and temperature 40 °C, suggesting that these mid-range conditions are optimal for extract yield (%) from dandelion (Table 2). Reproducibility and authenticity of the model was also verified by multiple center-point runs (Runs 7, 12, 13, 15, 17), which consistently produced extraction yields above 11.8%. On the other hand, at 15 minutes, 70% amplitude, and 25 °C, the lowest yield (10.06±0.56%) was obtained, indicating that the inadequate time and low temperature have an adverse effect on extraction efficiency of yield. Sonication amplitude exhibited a pivotal role among all extraction parameters, as 70% amplitude outperformed both lower (50%) and higher (90%) ranges, suggesting optimal efficiency at moderate energy input. Comparable efficiency has also been observed at central point temperature and time as compared to the extreme ranges. The optimal yield obtained with moderate conditions suggested clear evidence that higher or lower levels of any extraction variable may negatively affect the extraction. The optimal yield at 70% as compared to 50% and 90% amplitude suggests inadequate energy production to disrupt cell walls effectively, while excessive energy may lead to degradation of plant matrix. The superior performance at 40 °C rather than lower (25 °C) or higher (55 °C) temperatures suggest that moderate heat enhances solubility and diffusion without degrading thermolabile compounds. The 30-minute extraction was optimal, as shorter times likely left compounds unextracted, while longer durations may have led to solvent saturation or compound degradation. The slight variations that have been observed in center-point runs (e.g., 11.84%–12.25%) were due to experimental noise, but the consistency still validates the model’s reliability and authenticity. Previous literature supports that the extraction yield increases with the gradual increase in sonication time and temperature, but after a specific interval it showed a nonlinear decline. The results obtained in terms of sonication time and temperature are consistent with the results of earlier studies [33], [34], [35].

Table 2.

Percentage yield of dandelion extract as carried out through Box-Behnkin Design.

Run	A: Time	B: Amp	C: Temp	Yield (%)
1	45	90	40	11.31 ± 0.52
2	30	90	25	11.53 ± 0.61
3	45	70	55	11.41 ± 0.47
4	45	70	25	11.57 ± 0.73
5	15	70	25	10.06 ± 0.56
6	15	90	40	11.35 ± 0.71
7	30	70	40	11.99 ± 0.74
8	15	50	40	10.11 ± 0.43
9	30	50	55	11.44 ± 0.82
10	45	50	40	10.08 ± 0.76
11	30	50	25	10.31 ± 0.77
12	30	70	40	12.25 ± 0.82
13	30	70	40	11.95 ± 0.67
14	15	70	55	11.71 ± 0.63
15	30	70	40	11.93 ± 0.74
16	30	90	55	11.37 ± 0.70
17	30	70	40	11.84 ± 0.59

Values are expressed as mean ± SD (n  =  3).

3.1.1. Fitting the Experimental Model

The results in analysis of variance (ANOVA) table illustrate that the regression model for dandelion extract is statistically significant (p = 0.0024). Among the individual factors, amplitude (B) and temperature (C) show significant linear effects (p = 0.0026 and p = 0.0174, respectively), while on the other hand, time (A) shows no significant behavior (p = 0.1942) (Table 3). However, the quadratic effects for time (A², p = 0.0025) and amplitude (B², p = 0.0020) showing significant behavior indicating that these factors have nonlinear relationships with extraction yield, suggesting optimal extraction occurs at intermediate levels rather than at the extremes of the tested ranges. While considering the mutual interactions, only significant behavior was shown between sonication time and temperature (AC, p = 0.0146), indicating that the mutual effects must also be considered for process optimization. On the other hand, other mutual interactions (AB & BC) were observed to be non-significant (p>0.05). Hence, the observed curvature in respective response surface plots were mainly driven by significant quadratic effects. In RSM, even in absence of significant interactions, curvature may arise from non-linear individual factor effects. The lack of fit test (p  =  0.0526) confirms the model’s adequacy, as it does not show significant variation (Table 3). Mutual interaction effects have also been elaborated in Fig. 1. These results suggested that the highest yields are achieved at moderate amplitude (70%) and temperature (40 °C), with time playing a secondary role. The results highlight the importance of balancing these parameters, particularly avoiding excessive amplitude or temperature, which could diminish extraction efficiency. It was clearly understood from the results that there was a significant effect of model on extract yield (Table 3). The linear and quadratic effects showed highly significant behavior as compared to interaction effects. The behavior of different model effects noticed for extraction yield from dandelion was linear > quadratic > interaction. It was also very much clear from Table 3 that linear coefficients and quadratic coefficient were mostly significant.

Table 3.

Analysis of Variance (ANOVA) for extract yield (%) from dandelion.

Source	Sum of Squares	df	Mean Square	F-value	p-value
Model	7.69	9	0.8547	10.85	0.0024	significant
A-Time	0.1625	1	0.1625	2.06	0.1942
B-Amp	1.64	1	1.64	20.78	0.0026	significant
C-Temp	0.7565	1	0.7565	9.60	0.0174	significant
AB	0.0000	1	0.0000	0.0003	0.9863
AC	0.8190	1	0.8190	10.39	0.0146	significant
BC	0.4160	1	0.4160	5.28	0.0552
A2	1.66	1	1.66	21.02	0.0025	significant
B2	1.79	1	1.79	22.73	0.0020	significant
C2	0.1323	1	0.1323	1.68	0.2362
Residual	0.5517	7	0.0788
Lack of Fit	0.4564	3	0.1521	6.39	0.0526	not significant
Pure Error	0.0953	4	0.0238
Cor Total	8.24	16

Fig. 1.

Effects of mutual interactions of different sonication extraction conditions on yield (%) of dandelion.

3.2. Analysis of antioxidant activity and phenolic content parameters of dandelion extract using RSM

The effects of various extraction conditions like sonication time (min), amplitude (%), and temperature (°C) on TPC, TFC, DPPH radical scavenging (%), ABTS (µM TE/mg), and FRAP (mg TE/g) was investigated using a three level and three factor Box-Behnken design. The maximum antioxidative potential was achieved at 30 minutes sonication time, 70% sonication amplitude level, and 40 °C extraction temperature at mid-range conditions (Runs 7, 12, 13, 15, 17), with TPC: ∼40 mg GAE/g; TFC: ∼21–22.68 mg RE/g; DPPH Scavenging: ∼86–88.55%; ABTS: ∼420–445 µM TE/mg; FRAP: ∼28–30.64 mg TE/g (Table 4). This suggests that moderate extraction conditions maximize the release of phenolic and flavonoid compounds, which are directly linked to antioxidant capacity.

Table 4.

Antioxidant potentials in dandelion extract using RSM as carried out through Box-Behnkin Design.

Run	A: Time (°C)	B: Amp (%)	C: Temp (min)	TPC (mg GAE/g)	TFC (mg RE/g)	DPPH (%)	ABTS (µM TE/mg)	FRAP (mg TE/g)
1	45	90	40	36.51 ± 2.87	11.76 ± 0.84	54.43 ± 2.83	262.20 ± 9.49	20.93 ± 1.61
2	30	90	25	38.31 ± 3.11	16.68 ± 1.03	67.42 ± 3.78	324.30 ± 9.20	26.78 ± 2.09
3	45	70	55	36.88 ± 2.64	12.32 ± 0.91	64.82 ± 3.65	293.25 ± 8.06	25.98 ± 1.92
4	45	70	25	38.44 ± 2.99	14.56 ± 1.11	69.84 ± 3.21	324.32 ± 9.75	27.05 ± 2.07
5	15	70	25	32.77 ± 1.93	8.06 ± 0.76	51.44 ± 2.44	235.60 ± 7.63	18.45 ± 1.86
6	15	90	40	36.66 ± 2.76	12.16 ± 1.09	59.77 ± 3.09	292.56 ± 7.97	21.55 ± 2.01
7	30	70	40	39.91 ± 3.09	21.00 ± 1.15	86.48 ± 4.78	420.90 ± 9.94	29.62 ± 2.42
8	15	50	40	32.97 ± 2.07	9.64 ± 0.92	52.12 ± 2.82	255.99 ± 6.62	20.48 ± 1.53
9	30	50	55	36.98 ± 2.47	14.28 ± 1.12	64.93 ± 3.14	308.19 ± 7.05	26.75 ± 2.11
10	45	50	40	32.85 ± 1.95	7.84 ± 0.78	51.39 ± 2.84	248.40 ± 5.83	18.02 ± 1.12
11	30	50	25	33.76 ± 2.31	10.92 ± 0.94	54.15 ± 2.95	263.58 ± 6.42	20.59 ± 1.48
12	30	70	40	40.77 ± 3.43	22.68 ± 1.17	88.55 ± 4.99	445.39 ± 9.92	30.64 ± 2.64
13	30	70	40	39.76 ± 3.01	22.12 ± 1.21	86.17 ± 4.92	411.95 ± 9.64	28.99 ± 2.17
14	15	70	55	38.93 ± 3.36	16.21 ± 1.06	76.47 ± 3.99	384.33 ± 7.11	27.82 ± 2.16
15	30	70	40	39.69 ± 3.51	21.28 ± 1.22	86.23 ± 4.69	403.65 ± 8.89	28.41 ± 2.25
16	30	90	55	36.73 ± 2.92	13.16 ± 1.07	61.55 ± 3.76	297.39 ± 6.73	22.82 ± 1.67
17	30	70	40	39.38 ± 3.76	21.08 ± 1.19	83.49 ± 4.48	398.13 ± 7.09	28.02 ± 2.33

Values are expressed as mean ± SD (n  =  3).

On the other hand, the lowest values were observed at 15 min, 50% amplitude, and 40 °C (Run 8) and 45 min, 50% amplitude, 40 °C (Run 10), indicating, insufficient sonication time (15 min) and low amplitude (50%) result in poor compound release. Longer extraction (45 min) with low amplitude (50%) also reduces efficiency, possibly due to compound degradation or solvent saturation. Sonication time of 30 minutes consistently outperformed 15 minutes and 45 minutes, suggesting that shorter sonication time (15 minutes) extraction of polyphenols/flavonoids is incomplete. On the other hand, longer sonication time (45 minutes), resulted in possible thermal degradation or oxidation of antioxidants. The other factor of extraction i.e., sonication amplitude level outperformed at 70% amplitude by yielding the highest TPC, TFC, and antioxidant activity, while 50% and 90% were less effective. At 50% amplitude there might be an insufficient cell disruption resulting in lower extraction efficiency. While, 90% amplitude causes excessive cavitation, leading to compound degradation (Table 4). Similarly, extraction temperature also showed a similar trend. At 40 °C optimal results were obtained, while 25 °C and 55 °C extraction temperature showed mixed effects. Extraction temperature of 25 °C showed lower solubility and diffusion rate resulting in reduced extraction. On the other hand, at higher extraction temperature (55 °C), there is possible thermal degradation of heat-sensitive antioxidants.

The bioactivity of dandelion extract in terms of TPC, TFC, and the antioxidant activity assays (DPPH, ABTS, and FRAP) is affected by interaction of different extraction parameters like sonication time (min), amplitude level (%), and extraction temperature (°C) as revealed by the Box-Behnken design. Sonication time (min) and extraction temperature (°C) have more promising effect, as has been observed in Run 12 (30 min, 70% amp, 40 °C), that resulted in higher values of TPC (40.77±3.43 mg GAE/g) and TFC (22.68±1.17 mg RE/g) (Table 4). This was verified by multiple center-point runs (Runs 7, 12, 13, 15, 17), suggesting reproducibility and authenticity of the model. On the other hand, longer sonication time (45 min) had adverse effect likely due to thermal degradation of heat-sensitive bioactive compounds. Similarly, with reduction in extraction temperature and time (AC) the recovery of phenolic and flavonoid compounds drastically reduced due to insufficient energy and exposure time for recovery of these phytochemicals as evident in run 5 and 8 (Table 4). While studying the interaction between amplitude and temperature (BC), where higher level of amplitude (90%) together with elevated temperature (55 °C) resulted in lower values of flavonoid content and antioxidant activity (Run 2 and Run 16), due to the fact that excessive mechanical energy and heat resulted in degradation of sensitive compounds. However, time and amplitude (AB) showing non-significant effects mainly due to independent behavior of their interactions.

These findings highlight the importance of balancing the extraction condition as moderate amplitude (70%) ensures efficient cell disruption without degradation by careful pairing of time and temperature. The study of mutual interaction highlights the importance of RSM in identifying synergistic and antagonistic effects which are definitely missed in single-factor experiments. The results of this study showed a strong correlation among TPC, TFC, and the antioxidant activity assays (DPPH, ABTS, and FRAP), indicating that phenolics and flavonoids are the primary contributors to the antioxidant capacity of extract. Higher values of TPC and TFC linked with enhanced radical scavenging (DPPH), electron transfer (ABTS), and reducing power (FRAP), as demonstrated by Run 12 (30 min, 70% Amp, 40 °C), which yielded the highest TPC (40.77±3.43 mg GAE/g), TFC (22.68±1.17 mg RE/g), DPPH (88.55±4.99 %), ABTS (445.39±9.92 µM TE/mg), and FRAP (30.64±2.64 mg TE/g) as clearly seen in Table 4. These results suggested that phenolic acids and flavonoids synergistically enhance the antioxidant activity, with flavonoids playing a predominant role in radical scavenging (DPPH) due to their hydrogen-donating ability, while in contrast phenolic compounds dominate in electron transfer mechanisms (ABTS/FRAP). Notably, extraction runs with lower phenolic and flavonoid contents (e.g., Run 10: 45 min, 50% Amp, 40 °C) showed markedly reduced antioxidant activity across all assays, which further validating this association. All the results of mutual interaction for TPC, TFC, and the antioxidant activity assays (DPPH, ABTS, and FRAP) have also been represented graphically in 3-D surface plots in Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6. These results indicated that that optimization of extraction variables have direct influence on overall antioxidant potentials, making them reliable indicators for bioactive quality in dandelion extracts (Fig. 7, Fig. 8, Fig. 9, Fig. 10, Fig. 11, Fig. 12, Fig. 13, Fig. 14, Fig. 15, Fig. 16, Fig. 17).

Fig. 2.

Effects of mutual interactions of different sonication extraction conditions on TPC (mgGAE/g) of dandelion.

Fig. 3.

Effects of mutual interactions of different sonication extraction conditions on TFC (mgRE/g) of dandelion.

Fig. 4.

Effects of mutual interactions of different sonication extraction conditions on DPPH (%) of dandelion.

Fig. 5.

Effects of mutual interactions of different sonication extraction conditions on ABTS (µM TE/mg of extract) of dandelion.

Fig. 6.

Effects of mutual interactions of different sonication extraction conditions on FRAP (mg TE/g) of dandelion.

Fig. 7.

Effects of mutual interactions of different sonication extraction conditions on chlorogenic acid (mg/g) of dandelion.

Fig. 8.

Effects of mutual interactions of different sonication extraction conditions on quercetin (mg/g) of dandelion.

Fig. 9.

Effects of mutual interactions of different sonication extraction conditions on apigenin (mg/g) of dandelion.

Fig. 10.

Effects of mutual interactions of different sonication extraction conditions on luteolin-7-O-glycoside (mg/g) of dandelion.

Fig. 11.

Effects of mutual interactions of different sonication extraction conditions on luteolin (mg/g) of dandelion.

Fig. 12.

Effects of mutual interactions of different sonication extraction conditions on p-coumaric acid (mg/g) of dandelion.

Fig. 13.

Effects of mutual interactions of different sonication extraction conditions on caffeic acid (mg/g) of dandelion.

Fig. 14.

Effects of mutual interactions of different sonication extraction conditions on ferulic acid (mg/g) of dandelion.

Fig. 15.

Effects of mutual interactions of different sonication extraction conditions on cichoric acid (mg/g) of dandelion.

Fig. 16.

Effects of mutual interactions of different sonication extraction conditions on isoetin (mg/g) of dandelion.

Fig. 17.

Effects of mutual interactions of different sonication extraction conditions on caftaric acid (mg/g) of dandelion.

Previously published literature supported current finding on the optimization of antioxidative potentials from plant extracts. A number of studies have reported that moderate extraction conditions resulted in optimal recovery of phenolics and flavonoids, which are correlated with antioxidant activity [35], [36], [37], [38]. For instance, research on ultrasound-assisted extraction (UAE) of polyphenols from other medicinal plants was identified 70–80% amplitude as optimal, as higher amplitudes can induce oxidative degradation of sensitive compounds [39]. The observed time-temperature interaction in this study mirrors findings in extractions of grape seed and olive leaf extracts, where shorter times at higher temperatures preserved flavonoids better than prolonged heating [39], [40], [41]. Moreover, the strong relationship of TPC and TFC with antioxidant activity assays are well established in literature, as phenolics and flavonoids are primary contributors to DPPH, ABTS, and FRAP responses [42]. However, the current research provides novel insights which are specific to dandelion by demonstrating that extraction conditions as 30 minutes, 70% amplitude and 40 °C create a critical balance between extraction efficiency and compound stability. The results of current research thus bridge gaps in existing knowledge by quantifying parameter interactions unique to dandelion while reinforcing broader principles of phytochemical extraction.

3.2.1. Fitting the Experimental Model

The results of ANOVA for TPC, TFC, and the antioxidant activity assays (DPPH, ABTS, and FRAP) demonstrate that the quadratic models are statistically significant (p < 0.05) which indicating the validity and authenticity of by describing the relationship between extraction parameters and antioxidant assays. These results are in accordance with the previous findings on plant extracts, such as those by Luo et al. [43] on green tea polyphenols and Irakli et al. [44] on rosemary antioxidants, in which the quadratic models constantly outperformed linear models in predicting antioxidant assays. As highly significant behavior has been shown by TFC and DPPH (p < 0.0001), indicating these responses are much more sensitive to the extraction parameters, which correlates with previous published literature showing flavonoids are primary contributors to radical scavenging activity (DPPH) in plant extracts [45].

Sonication amplitude (%) (B) seems to be the most significant linear factor across all responses (p = 0.0021 for TPC, p = 0.0137 for TFC), with 70% amplitude emerging as optimal. These results validate the work of Chen et al. [46] on ultrasound-assisted extraction of polyphenols, where moderate sonication amplitude level (70–80%) outperformed the polyphenol recovery with minimum degradation as observed at higher amplitude levels. In Table 5 the quadratic terms (A² and B²) show highly significant behavior (p < 0.001 for most responses), representing nonlinear associations that are optimal at intermediate values. This phenomenon has widely been observed in extraction studies of phytochemicals, including Brahmi et al. [47] research on apple and grape seed extracts, in which a deviation from intermediate range resulted in reduction the recovery of antioxidant assays.

Table 5.

Analysis of variance (ANOVA) for antioxidant potentials in dandelion extract.

Source	df	TPC p-value	TFC p-value	DPPH p-value	ABTS p-value	FRAP p-value
Model	9	0.0007	<0.0001	<0.0001	0.0003	0.0040
A-Time	1	0.2133	0.9072	0.9379	0.4979	0.4812
B-Amp	1	0.0021	0.0137	0.0445	0.1186	0.2476
C-Temp	1	0.0381	0.1339	0.0211	0.0475	0.0715
AB	1	0.9867	0.5777	0.4642	0.5859	0.6151
AC	1	0.0029	0.0034	0.0015	0.0028	0.0204
BC	1	0.0275	0.0240	0.0267	0.1160	0.0232
A2	1	0.0007	<0.0001	<0.0001	0.0002	0.0018
B2	1	0.0003	<0.0001	<0.0001	<0.0001	0.0008
C2	1	0.1294	0.0023	0.0029	0.0071	0.8560
Residual	7
Lack of Fit	3	0.0754	0.0802	0.0756	0.3738	0.0700
Pure Error	4
Cor Total	16

Notably, the significant results have been observed in mutual interactions of time and temperature (p < 0.05) for all responses except for FRAP. It may be concluded that higher temperatures can compensate for shorter sonication times, as seen in Run 14 (15 min, 55 °C) achieving comparable recovery to longer sonication time. This aligns with thermodynamic principles in extraction kinetics and matches observations from Melastoma malabathricum leaf extraction studies [48]. However, the non-significant results from mutual interaction time and amplitude (p>0.05) suggests these parameters act independently. On the other hand, mutual interaction between A and C was noticed to be significant (p < 0.05) across all the analyzed parameters. Likewise, B and C interaction was also significant (p < 0.05) for all the examined parameters except for ABTS, suggesting their pivotal role in modulation of cavitation intensity and effectiveness of mass transfer. Even though AB interaction was non-significant, the observed curvature is mainly driven by quadratic terms. In RSM methodology, curvatures may arise from non-linear individual factor effects rather than statistically significant interactions.

The lack of fit tests showing non-significant results (p>0.05) confirms model adequacy (Table 5), that is consistent with proper RSM design as described by Mushtaq et al. [49]. The strong correlations among TPC, TFC and antioxidant assays (particularly DPPH and ABTS) emphasize the principles that phenolics and flavonoids are primary antioxidants in plant matrix [45]. The optimal range of extraction process (30 min, 70% amplitude, 40 °C) balances compound release with stability which is consider as an improvement over earlier studies that used more extreme conditions.

3.3. Analysis of bioactive compounds in dandelion extract using RSM

The effects of various extraction variables demonstrate maximum recovery of phenolic compounds like chlorogenic acid (29.14±1.64 mg/g), p-coumaric acid (21.34±1.39 mg/g), caffeic acid (62.19±3.79 mg/g), ferulic acid (8.09±0.87 mg/g), cichoric acid (14.89±1.12 mg/g), and for flavonoids like quercetin (11.82±0.87 mg/g), apigenin (0.21±0.02 mg/g), luteolin-7-O-glycoside (9.98±0.69 mg/g), luteolin (92.37±3.99 mg/g), and isoetin (0.66±0.07 mg/g) in experimental run of 12 as observed through BBD (Table 6). The presence of diverse array of bioactive compounds is supported by previous studies identifying dandelion as a rich source of these bioactive compounds [50], though the concentration in this study is normally high than that of found in literature. This high concentration is more likely due to optimized ultrasound-assisted extraction (UAE) parameters. Highest recovery of bioactive compounds has been observed at optimal conditions (30 min, 70% amplitude, 40 °C) in runs 7, 12–13, 15, and 17, with luteolin recovery exceeding 90 mg/g which is almost two-fold more than as reported earlier [51]. The higher recovery rates are due to the fact that moderate UAE parameters maximize phenolic extraction while minimizing degradation as concluded by Chen et al. [52].

Table 6.

Bioactive components in dandelion extract using RSM as carried out through Box-Behnkin Design.

Run	A: Time (°C)	B: Amp (%)	C: Temp (min)	Chlorogenic acid (mg/g)	p-coumaric acid (mg/g)	Caffeic acid (mg/g)	Ferulic acid (mg/g)	Cichoric acid (mg/g)	Caftaric acid (mg/g)	Quercetin (mg/g)	Apigenin (mg/g)	luteolin-7-O-glycoside (mg/g)	Luteolin (mg/g)	Isoetin (mg/g)
1	45	90	40	18.36 ± 0.97	15.97 ± 1.06	38.65 ± 3.25	4.97 ± 0.49	7.72 ± 0.93	1.31 ± 0.15	7.37 ± 0.45	0.13 ± 0.01	6.19 ± 0.39	64.96 ± 2.89	0.36 ± 0.03
2	30	90	25	22.29 ± 1.22	18.31 ± 1.29	47.60 ± 3.49	5.71 ± 0.57	8.82 ± 0.96	1.55 ± 0.19	8.95 ± 0.73	0.16 ± 0.01	7.01 ± 0.45	75.52 ± 3.41	0.42 ± 0.04
3	45	70	55	21.10 ± 1.17	16.34 ± 1.18	45.82 ± 3.38	5.34 ± 0.53	6.27 ± 0.78	1.45 ± 0.20	8.33 ± 0.71	0.14 ± 0.01	6.25 ± 0.41	72.64 ± 3.18	0.33 ± 0.03
4	45	70	25	24.22 ± 1.48	17.62 ± 1.27	49.29 ± 3.45	5.04 ± 0.49	8.16 ± 0.84	1.56 ± 0.31	9.51 ± 0.87	0.19 ± 0.02	7.06 ± 0.47	77.34 ± 3.25	0.38 ± 0.03
5	15	70	25	14.78 ± 0.95	13.07 ± 1.21	36.61 ± 2.99	4.27 ± 0.43	7.15 ± 0.87	1.20 ± 0.14	7.32 ± 0.50	0.12 ± 0.01	4.81 ± 0.36	60.15 ± 2.79	0.33 ± 0.03
6	15	90	40	18.48 ± 1.02	16.18 ± 1.15	42.34 ± 2.96	5.06 ± 0.51	8.17 ± 0.89	1.39 ± 0.15	8.41 ± 0.67	0.14 ± 0.01	6.42 ± 0.43	68.88 ± 3.11	0.38 ± 0.03
7	30	70	40	29.29 ± 1.73	20.41 ± 1.40	61.74 ± 3.76	8.12 ± 0.85	14.68 ± 1.09	1.94 ± 0.25	11.59 ± 0.93	0.20 ± 0.02	9.96 ± 0.72	90.28 ± 4.02	0.65 ± 0.05
8	15	50	40	15.88 ± 1.04	13.68 ± 1.12	37.06 ± 2.73	4.13 ± 0.42	6.25 ± 0.69	1.20 ± 0.13	7.15 ± 0.53	0.10 ± 0.01	5.07 ± 0.29	60.78 ± 2.78	0.30 ± 0.03
9	30	50	55	23.97 ± 1.39	17.01 ± 1.29	45.91 ± 3.26	5.93 ± 0.49	6.66 ± 0.73	1.47 ± 0.15	8.85 ± 0.67	0.16 ± 0.01	6.72 ± 0.38	72.79 ± 3.09	0.36 ± 0.03
10	45	50	40	15.38 ± 1.03	13.01 ± 1.01	36.58 ± 2.91	4.63 ± 0.41	8.75 ± 0.88	1.22 ± 0.14	6.74 ± 0.48	0.09 ± 0.01	4.79 ± 0.33	60.17 ± 2.52	0.38 ± 0.04
11	30	50	25	16.28 ± 1.11	14.42 ± 1.02	38.48 ± 2.95	4.71 ± 0.38	7.73 ± 0.79	1.27 ± 0.15	7.58 ± 0.53	0.11 ± 0.01	5.05 ± 0.38	62.79 ± 2.86	0.35 ± 0.03
12	30	70	40	29.14 ± 1.64	21.34 ± 1.39	62.19 ± 3.79	8.09 ± 0.87	14.89 ± 1.12	1.98 ± 0.28	11.82 ± 0.87	0.21 ± 0.02	9.98 ± 0.69	92.37 ± 3.99	0.66 ± 0.07
13	30	70	40	28.51 ± 1.59	20.72 ± 1.41	61.55 ± 3.87	7.98 ± 0.57	14.51 ± 1.17	1.93 ± 0.23	11.81 ± 0.91	0.20 ± 0.02	9.89 ± 0.71	89.95 ± 3.96	0.64 ± 0.06
14	15	70	55	27.24 ± 1.45	18.39 ± 1.09	54.58 ± 2.99	5.34 ± 0.44	6.16 ± 0.71	1.66 ± 0.19	11.10 ± 0.86	0.20 ± 0.02	7.81 ± 0.55	83.14 ± 3.23	0.33 ± 0.03
15	30	70	40	28.54 ± 1.62	20.41 ± 1.17	61.61 ± 3.86	8.06 ± 0.79	14.56 ± 1.08	1.93 ± 0.23	11.31 ± 0.89	0.21 ± 0.02	9.64 ± 0.73	89.94 ± 4.01	0.65 ± 0.07
16	30	90	55	21.16 ± 1.07	16.53 ± 0.99	43.55 ± 2.45	5.88 ± 0.46	6.01 ± 0.68	1.41 ± 0.17	8.11 ± 0.56	0.14 ± 0.01	6.39 ± 0.44	70.21 ± 3.39	0.34 ± 0.03
17	30	70	40	28.31 ± 1.71	20.21 ± 1.25	58.69 ± 3.73	8.01 ± 0.82	14.61 ± 1.15	1.88 ± 0.18	11.34 ± 0.92	0.22 ± 0.02	9.61 ± 0.92	87.76 ± 3.76	0.65 ± 0.07

Values are expressed as mean ± SD (n  =  3).

The effects of mutual interaction time and temperature (AC) as observed in run 14 (15 min, 70% amplitude, 55 °C), showed that shorter sonication time at elevated temperature achieved 85–90% of maximum yields for most bioactive compounds. On the other hand, the mutual interactions of amplitude and temperature (BC) showed compound-specific patterns, while the recovery of luteolin is non-significantly affected at 70% amplitude regardless of temperature, while caffeic acid recovery increased at higher temperatures at fixed amplitude. These nuanced responses underscore the importance of multi-parameter optimization for comprehensive phytochemical recovery.

These diversified results have several practical insights for industrial applications. Firstly, the stagnant recovery rates beyond 70% amplitude suggest diminishing returns at higher energy inputs, supporting economic optimization. Secondly, the optimal recovery rates at mid-range temperature (40 °C) challenges conventional high-temperature extraction paradigms, possibly enabling it as energy efficient process. Thirdly, the compound-specific recovery allows for tailored processing, for example, a slightly higher temperature (45–50 °C) could be used when targeting chlorogenic acid specifically. These findings collectively demonstrate that RSM-optimized UAE can achieve significantly higher recovery of bioactive compounds from dandelion than previously reported methods, while providing new mechanistic insights into the recovery of specific compound.

3.3.1. Fitting the experimental model

The ANOVA table for bioactive compounds show highly significant behavior for quadratic models (p < 0.0001) representing excellent fit with the experimental data. This highly significant behavior of quadratic model aligns with previous RSM studies, such as studied by Dai and Mumper [53], who also found quadratic model essential for precisely predicting recovery of bioactive compounds. The high significance of luteolin and its glycosides (p < 0.0001) support the previous findings from Liu et al. [54], who observed the nonlinear extraction kinetics of flavonoid compounds. It is also noteworthy that the models for dandelion-specific markers i.e. cichoric acid and isoetin, showed high significance, although the limited literature is available about the extraction of these compounds [51].

The analysis of variance identified sonication amplitude (B) as highly significant linear factor for most bioactive compounds, particularly for chlorogenic acid (p = 0.0056) and apigenin (p = 0.0061). These results are being supported by another study conducted by Chen et al. [55], who concluded that sonication amplitude critically affects the disruption of cell wall in plant materials thus improving the recovery of bioactive compounds. On the other hand, extraction temperature (°C) also showed high significance for recovery of certain acids such as chlorogenic (p = 0.0002) and caffeic acid (p = 0.0206), revealing significant role of sonication temperature in increasing efficiency of extraction, validating the results of Wang et al. [37] about thermal stability variations among phenolic compounds. Surprisingly, sonication time (A) showed nominal linear effects (p>0.05 for most compounds), contradicting the conventional extraction studies but supported with recent studies on UAE depicting less significance of time when optimal amplitude is applied.

The quadratic models (A², B², C²) showed highly significant responses for all bioactive compounds (p < 0.0001 in most cases) confirming the existence of optimal extraction windows, as reported previously for other phytochemicals [37]. The highly significant quadratic effects for luteolin-7-O-glycoside (A²: p < 0.0001; B²: p < 0.0001) suggest this compound has a narrow optimal extraction range, elucidating why conventional methods often under-extract it. Statistical results for quadratic models confirm presence of non-linear behavior and supports existence of optimized conditions for extraction rather than a linear enhancement. The mutual interaction effects of time and temperature (AC) showed extraordinary significance for compounds like chlorogenic acid (p < 0.0001) and quercetin (p = 0.0004), supporting the thermal compensation effect observed in our experimental data but not previously quantified for dandelion phenolics [56]. Results of AC interaction highlights that the effect of sonication time on optimal recovery of bioactive compounds is strongly dependent on sonication temperature, revealing a synergism among thermal softening of cellular structure of plant and mass transfer induced by cavitation.

The mutual interaction of amplitude and temperature (BC) proved significance for most compounds (p < 0.05), especially ferulic (p = 0.0003) and cichoric acid (p = 0.0058). This reinforced the significant role of sonication temperature (C) in the modulation of cavitation effectiveness and overall diffusibility of bioactive compounds. The results from current research are in line with previous studies by showing energy input and thermal effects interact complexly in recovery of bioactive compounds. However, the interaction effects of time and amplitude (AB) showed more compound-specific relationship as it showed high significance for isoetin (p < 0.0001) and cichoric acid (p = 0.0003) but non-significant for other compounds. The observed curvature is driven by quadratic terms rather than a statistically significant AB interaction. These observations suggested that some compounds require specific coordination of sonication time and amplitude, mainly due to their specific cellular localization in dandelion. The non-significant results of lack of fit (p>0.05) for all bioactive compounds confirm model adequacy and validity as shown in Table 7. Current research suggested much higher recovery rates almost 20–50% higher than found in literature with conventional extraction methods as reviewed by González-Castejón et al. [51], indicating superiority of UAE for dandelion bioactive compounds. The high significance of quadratic model for all bioactive compounds suggesting that even compounds previously assumed to have linear extraction profiles (like p-coumaric acid) actually exhibit complex nonlinear behavior under optimized UAE conditions.

Table 7.

Analysis of variance (ANOVA) for bioactive components in dandelion extract.

Source	df	Chlorogenic acid p-value	p-coumaric acid p-value	Caffeic acid p-value	Ferulic acid p-value	Cichoric acid p-value	Caftaric acid p-value	Quercetin p-value	Apigenin p-value	luteolin-7-O-glycoside p-value	Luteolin p-value	Isoetin p-value
Model	9	<0.0001	0.0002	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001
A-Time	1	0.2676	0.4756	0.9679	0.0010	0.0015	0.5843	0.1162	0.7350	0.8375	0.7844	0.0004
B-Amp	1	0.0056	0.0044	0.0511	<0.0001	0.0723	0.0154	0.0614	0.0061	0.0013	0.0190	0.0004
C-Temp	1	0.0002	0.0586	0.0206	<0.0001	<0.0001	0.0348	0.0318	0.0724	0.0064	0.0192	0.0003
AB	1	0.8161	0.7709	0.4742	0.0067	0.0003	0.3975	0.4573	1.0000	0.9357	0.5573	<0.0001
AC	1	<0.0001	0.0034	0.0015	0.0016	0.0827	0.0013	0.0004	0.0003	0.0004	0.0013	0.0053
BC	1	0.0008	0.0238	0.0304	0.0003	0.0058	0.0182	0.0337	0.0102	0.0065	0.0247	0.0002
A2	1	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	0.0001	<0.0001	<0.0001	<0.0001
B2	1	<0.0001	0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001
C2	1	0.0056	0.0141	0.0027	<0.0001	<0.0001	0.0006	0.0053	0.1083	<0.0001	0.0062	<0.0001
Residual	7
Lack of Fit	3	0.0513	0.0659	0.1061	0.1608	0.1098	0.0956	0.0799	0.2531	0.0713	0.0780	0.7022
Pure Error	4
Cor Total	16

Earlier studies published regarding dandelion mainly focuses on extraction from individual dandelion plant parts and employed either traditional extraction [57], [58], [59] or ultrasound-assisted extraction within limited experimental parameters [60], [61], [62], [63], [64], [65], [66], [67]. On the other hand, current study employs whole dandelion plant and investigates the effect of multi-ultrasonication variables (sonication temperature, amplitude, and time) using BBD on multiple responses (TPC, TFC, DPPH-scavenging, ABTS+-scavenging, FRAP activities, and HPLC quantification of 11 major compounds). The multi-ultrasonication variables (sonication temperature; 25–55 °C, amplitude; 50–90%, and sonication time; 15–45 min) were selected to cover mild to intense ultrasound regimes, ensuring a better understanding of ultrasonication-based extraction behavior. Results validated that each factor had significant effect on recovery of antioxidant and bioactive compounds. Extraction using whole dandelion plant and multi-variable optimization framework reveals a unique methodological approach as compared to previous ultrasonication-based investigations on dandelion that mainly investigates single plant part or parameters. Therefore, current study provides an insight on enhanced extraction performance and comprehensive roadmap for sustained utilization of dandelion plant.

3.4. Numerical optimization using desirability function

For the identification of optimized UAE parameters in this study, desirability function approach was carried out for the determination of multi-response numerical optimization. Purposely, all analyzed parameters such as yield, TPC, TFC, DPPH, ABTS, FRAP, and individual bioactive compounds were optimized simultaneously. The model predicted the optimized operating conditions for sonication time, amplitude, and sonication temperature as 26.65 min, 71.74%, and 42.58 °C, respectively. The overall desirability value of 0.941 was corresponded under these conditions. Fig. 18 shows the desirability ramp plot, illustrating optimization of investigated responses within the experimental domain. The overall desirability value of 0.941 shows an excellent agreement between optimization objectives and responses (predicted). This confirmed the efficiency of RSM for identification of optimal extraction parameters.

Fig. 18.

Desirability ramp plot revealing numerical (multi-response) optimization of UAE parameter.

4. Conclusions

This study comprehensively examined an optimized and efficient ultrasound-assisted extraction method for the recovery of yield and bioactive compounds present in dandelion plant using response surface methodology by a three factor, three level process using Box-Behnken design. The study optimized the extraction parameters like time, amplitude and temperature. It was observed that the extraction yield, total phenolic contents, total flavonoid contents, antioxidant potentials (DPPH, ABTS, and FRAP) and recovery of bioactive compounds was significantly affected by extraction parameters. This study identified extraction conditions of 30 min, 70% amplitude, and 40 °C as the ideal conditions for maximizing yield, phenolic content, antioxidant activity, and recovery of bioactive compounds. It was also observed from the study that the recovery of bioactive compounds has significantly enhanced through optimization and the compounds like luteolin’s recovery exceeding 90 mg/g which is almost two-fold more than as reported earlier studies by conventional methods. The results obtained from this research offer a promising foundation for developing more sustainable ecofriendly extraction methods for the food, pharmaceutical, and nutraceutical industries. The UAE extraction method used in this study is in line with the principles of green processing and supports efficient resource utilization while minimizing environmental impact. Even though amplitude (%) was employed as one of the independent factors in present study, future studies must be designed to integrate ultrasonic acoustic intensity (W/cm²) as operational parameter to ensure enhanced reproducibility and scale-up assessment.

CRediT authorship contribution statement

Samee-Ullah: . Faryal Shaukat: Investigation, Formal analysis, Data curation, Conceptualization. Anees Ahmed Khalil: Resources, Methodology, Conceptualization. Muhammad Nadeem Akhtar: Writing – original draft, Visualization, Validation. Muhammad Bilal: Writing – review & editing, Writing – original draft, Supervision, Software, Resources, Methodology. Yemin Guo: Methodology, Formal analysis. Jingcheng Huang: Methodology, Investigation, Data curation. Rana Muhammad Aadil: Writing – original draft, Supervision, Resources. Gholamreza Abdi: Writing – review & editing, Writing – original draft, Supervision, Project administration, Conceptualization. Xia Sun: Writing – review & editing, Writing – original draft, Supervision.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.