An Optimization of the Extraction of Phenolic Compounds from Grape Marc: A Comparison between Conventional and Ultrasound-Assisted Methods

Abstract

The green extraction of total phenolic compounds, flavonoids, anthocyanins, and tannins from grape marc was optimized using response surface methodology. The extracts were characterized and analyzed using LC-ESI-QTOF-MS/MS, and free radical scavenging capacity was evaluated. An efficient green extraction method is crucial for improving the recovery rates of these high-value phytochemicals and for sustainably reusing wine by-products. Our study optimized parameters for both conventional and ultrasound-assisted extraction methods, including solution pH, extraction temperature, liquid-to-solvent ratio, and ultrasonic amplitude. The optimized conditions for conventional extraction were identified as 60% ethanol with a pH of 2, a solvent-to-solid ratio of 50:1, extraction time of 16 h at a temperature of 49.2 °C. For ultrasound-assisted extraction, the optimized conditions were determined as 60% ethanol with a pH of 2, a solvent-to-solid ratio of 50:1, and an amplitude of 100% for 5.05 min at a temperature of 60 °C. We also demonstrated that lowering the temperature to 49.5 °C improves the energy efficiency of the extraction process with a minor reduction in recovery rates. Considering all factors, ultrasound-assisted extraction is more suitable for efficiently recovering bioactive compounds from grape marc.

1. Introduction

Grapes consistently occupy a leading position in global agricultural production [1]. According to a report by the Food and Agriculture Organization of the United Nations (FAO), the global production of grapes in 2021 was approximately 73.52 million tons [1], which contributed to the production of about 250.3 million hectoliters of wine [2]. The winemaking process generates a substantial amount of solid waste, known as grape marc (GM) or grape pomace, which remains after juice is extracted for white wines or grapes are pressed post-fermentation for red wines. GM accounts for nearly 25% of the total mass of pressed grapes and is mainly composed of stems, skins, crushed cells from the grape pulp, and seeds remaining after the pressing operations [3]. This byproduct is rich in alcohol, phenolics, tannins, pigments, unfermented sugars, and other valuable compounds and is thus a promising source of bioactive substances for nutraceuticals and functional foods [4]. GM possesses significant potential for value addition through ethanol fermentation into grape spirit and industrial ethanol [5], enhancing its economic worth and offering applications in cosmetics and pharmaceuticals [6]. Additionally, as a source of bioethanol, GM provides a sustainable alternative to fossil fuels [3]. Despite its application as an animal feed additive, its high phenolic content may impair digestibility by inhibiting key enzymes and microbial growth in the rumen [7]. GM also has potential as an organic soil amendment due to its nutrient richness [8], though direct application may adversely affect plant growth through phytotoxic effects from tannins and other phenolic compounds [7]. Value adding to GM for these applications requires addressing environmental impacts and the adoption of advanced techniques for the extraction of phenolic compounds in wineries and related sectors, which remains limited.

Compounds of value in GM vary structurally from those with simple aromatic rings to complex high molecular weight tannins. Phenolic compounds primarily recovered from GM include gallic acid, catechins, and epicatechins. Tannins, or condensed tannins, usually refer to phenolics (proanthocyanidins) found in grape skins and seeds, whereas anthocyanins (pigments in red grape skins) are considered monomeric flavonoid compounds [9]. Other compounds such as hydroxytyrosol, tyrosol, cyanidin glycosides, and various phenolic acids, including caffeic acid, protocatechuic, syringic, vanillic, o-coumaric, and p-coumaric acids, also have been identified [10]. Around 60% to 70% of the recoverable phenolic compounds are found within the seeds of grapes, making up about 5% to 8% of the seeds’ overall weight [11]. The extractable phenolic substances account for 10–11% of the dry weight of GM [6].

The recovery of phenolic compounds typically involves solvent extraction processes, where solvent concentration, time, temperature, and liquid-to-solid ratio are critical parameters for optimizing the maximum recovery of target compounds [12]. Ethanol or methanol solutions are commonly utilized as solvents; the choice of ethanol is justified both scientifically and practically, as it is considered a safe solvent for the food industry [13] and is recognized as an eco-friendly organic solvent capable of extracting a high proportion of phenolic compounds [14]. However, these processes may lead to the hydrolysis and oxidation of the compounds of interest due to the generally required lengthy extraction times [15]. In this context, novel methods such as ultrasound-assisted extraction (UAE) have been proposed to extract bioactive compounds from fruits and their by-products in shorter durations, with higher recovery efficiency. In UAE, ultrasonic waves penetrate the material at frequencies above 20 kHz, causing molecules in the medium to expand and contract continuously. This leads to the formation of cavities within the plant material, which can disrupt biological cell walls and promote the release of substances [16]. Ultrasound is considered a clean, green extraction technique for commercially important bioactive molecules [17].

The efficiency of industrial extraction can be limited by factors such as the thickness of grape skins, the climate in which the grapes were grown, the ripeness of the grapes, and different winemaking methods, all of which may influence the composition and content of phenolics [18]. To commercially exploit GM, it is imperative to develop effective extraction strategies. One advantage of Response Surface Methodology (RSM) is its ability to reduce the number of experiments while assessing the interactions between multiple parameters [19]. The objective of this study is to use RSM to optimize the liquid-solid extraction process for high-value phytochemicals. The resulting extracts are measured for total phenolics content (TPC), total flavonoid content (TFC), total anthocyanins content (TATC), total tannins content (TTC), and DPPH radical scavenging activity. This research supports the application of green extraction techniques to valorize winemaking by-products, aligning with the principles of the circular economy. A systematic investigation was conducted to assess how ultrasound conditions influence the efficiency of extracting valuable phenolic and antioxidant components from GM. This aligns with the goal of enhancing extraction yield and efficiency using green methodologies while preserving the integrity of bioactive compounds. Ultimately, our work aims to develop a commercially viable green extraction method and increase the conversion of GM into valuable products, thereby reducing the environmental impact of winemaking by-product disposal. Furthermore, this study characterizes the phenolic components of the extracts obtained and highlights the differences in composition between conventional and ultrasound-assisted extraction methods.

2. Materials and Methods

2.1. Materials and Reagents

The chemicals used for this study were of analytical grade. Folin and Ciocalteu’s phenol reagent, gallic acid, L-ascorbic acid, hexahydrate aluminum chloride, quercetin, trolox, DPPH, and polyvinylpolypyrrolidone were bought from the Sigma-Aldrich (Castle Hill, NSW, Australia). Sodium carbonate anhydrous was purchased from Chem-Supply Pty Ltd. (Adelaide, SA, Australia).

2.2. Grape Marc Preparation

Fresh GM (Tempranillo) was collected from Domlina Estate in Waurn Ponds on 2 May 2023 and then frozen at −80 °C for 20 h before freeze drying using Dynavac FD3 Freeze Dryer (Sydney, NSW, Australia). The GM was freeze-dried under vacuum (13.33 Pa absolute pressure) at −50 °C condenser temperature for 60 h [20]. Once dry, all GM was ground into a uniform powder using a commercial grinder (Laobenhang, model 400Y, Yongkang, Zhejiang, China) and stored at −20 °C for subsequent use.

2.3. Ethanol–Water Extraction

One gram of grape marc was mixed with a 60% ethanol (v/v) aqueous solution at a designed liquid-to-solid ratio (conditions based on RSM design). This mixture was then incubated on a shaker at a set temperature (conditions based on RSM design) for 16 h to facilitate the extraction process. Following incubation, the sample was centrifuged at 8000 rpm for 15 min at a temperature of 4 °C to separate solids from the liquid extract. The resulting clarified supernatant was then collected. Before analysis, this extract was further purified by passing it through a syringe filter with a pore size of 0.45 μm to ensure the removal of any remaining particulate matter.

2.4. Ultra-Sonic Assisted Extraction (UAE)

The process starts by mixing 1 g of grape marc with a 60% ethanol (v/v) solution, adhering to a specified liquid-to-solid ratio (conditions based on RSM design). The sample mix was then exposed to ultrasonic treatment at a predetermined temperature using a Sonicator (Q55 Sonicator, 20 kHz, 55W, Qsonica, CA, USA) with a standard 1/8″ (3.2 mm) diameter probe. After the ultrasonic treatment, the mixture was centrifuged at 8000 rpm for 15 min at a temperature of 4 °C, and the supernatant was carefully recovered. Subsequently, this supernatant was filtered through a syringe filter with a 0.45 μm pore size to ensure the removal of particulate matter, resulting in a clear and purified extract.

2.5. Analytical Methods

2.5.1. Total Phenolics Content

The Total Phenolics Content (TPC) of GM samples was quantified using an adapted Folin–Ciocalteu method [21]. In detail, 25 μL of GM extracts were mixed with an equal volume of Folin–Ciocalteu reagent in a well of a 96-well plate (Costar, Corning, NY, USA), which was prefilled with 200 μL of Milli-Q water. This mixture was incubated at a consistent temperature of 25 °C for 5 min. Subsequently, 25 μL of 10% (w/v) sodium carbonate solution was integrated into the mixture. The ensuing reaction was maintained for an hour at 25 °C under conditions devoid of light. The absorbance of the solution was then evaluated at a wavelength of 765 nm utilizing a spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). For calibration and quantification, a standard curve was constructed using gallic acid concentrations ranging from 0 to 200 μg/mL in ethanol. The TPC was reported as milligrams of gallic acid equivalent (GAE) per gram of dry sample, providing a precise measure of the phenolic content.

2.5.2. Total Flavonoids Content

The Total Flavonoids Content (TFC) in grape marc (GM) samples was measured using an adapted aluminum chloride method [20]. The assay involved adding 80 μL of the extract, 80 μL of 2% aluminum chloride (w/v), and 120 μL of 50 mg/mL sodium acetate into each well of a 96-well plate, conducted in triplicate. The mixture was incubated at 25 °C for 2.5 h in the dark, and the absorbance was recorded at 440 nm. TFC was quantified using a quercetin calibration curve (0–50 μg/mL) and expressed as mg of quercetin equivalents per gram of dry weight (mg QE/g).

2.5.3. Total Tannins Content

The Total Tannins Content (TTC) was assessed employing the polyvinylpolypyrrolidone (PVPP) binding method, referenced from a previous study [22]. The analysis commenced with the addition of 50 mg PVPP to a mixture containing 0.5 mL of GM extract and 0.5 mL of water. This mixture was vortexed, then left to settle at 4 °C for 15 min, followed by centrifugation at 8000 rpm for an identical time span at 4 °C. The clear supernatant extracted post-centrifugation was subjected to the Folin–Ciocalteu test described in 3.5.1. The TTC was deduced by subtracting the content of non-tannin phenolics (post-PVPP treatment) from the total phenolic content, with the results articulated as milligrams of gallic acid equivalents (GAE) per gram of the dry sample.

2.5.4. Total Anthocyanins Content

The Total Anthocyanins Content (TATC) was quantified using the pH differential method as described in [23]. In this procedure, 400 μL of the sample extract was added to a cuvette along with 2.8 mL of a pH 1.0 buffer (0.025 M potassium chloride). Similarly, another aliquot of 400 μL of sample extract was combined with 2.8 mL of a pH 4.5 buffer (0.4 M sodium acetate) in a different cuvette. Absorbance values were measured separately at 510 nm and 700 nm. The calculation of absorbance utilized the formula: Abs = (A510–A700 nm) at pH 1.0−(A510–A700 nm) at pH 4.5.

TAC calculation:

$$\mathrm{T}\mathrm{A}\mathrm{C}\left({\displaystyle \frac{\mathrm{m}\mathrm{g}}{\mathrm{g}}}\right)={\displaystyle \frac{\mathrm{A}\mathrm{b}\mathrm{s}}{\mathrm{e}\mathrm{L}}}\times \mathrm{M}\mathrm{W}\times \mathrm{D}\times {\displaystyle \frac{\mathrm{V}}{\mathrm{G}}}$$

Abs: absorbance, e: cyanidin 3-glucoside molar absorbance (26,900), L: cell path length (1 cm), MW is the molecular weight of anthocyanin (449.2 g/mol), D is a dilution factor, V is the final volume (mL), and G is the dry seed weight (g).

2.5.5. 2,2′-Diphenyl-1-Picrylhydrazyl (DPPH) Assay

The DPPH free radical scavenging capacity was evaluated using an adapted method [24]. Forty microliters of the extract and 40 μL of DPPH methanol solution (0.1 mM) were combined in a 96-well plate in triplicate. The mixture was then incubated in darkness at 25 °C for 30 min. Following incubation, absorbance was measured at 517 nm. A calibration curve was established using Trolox standards ranging from 0 to 200 μg/mL. The scavenging capacity results were expressed in terms of milligram Trolox equivalents per dry weight of the sample (mg TE/g), ensuring a clear and precise assessment of antioxidant activity.

2.5.6. Characterization of Phenolic Compounds Using LC-ESI-QTOF-MS/MS Analysis

Phenolic characterization was made by following the method of Yang et al. [20] with some modifications. Phenolic compounds were characterized using an integrated system comprising an Agilent 1200 series HPLC connected to an Agilent 6530 Accurate-Mass Quadrupole Time-of-Flight (Q-TOF) LC/MS via an electrospray ionization source (ESI), both from Agilent Technologies, Santa Clara, CA, USA. Chromatographic separation was achieved on a Synergi Hydro-Reverse Phase 80 Å, LC column (250 × 4.6 mm, 4 μm from Phenomenex, Torrance, CA, USA) maintained at 25 °C.

The liquid chromatography utilized a binary solvent system consisting of mobile phase A (water with 0.2% formic acid, v/v) and mobile phase B (50% acetonitrile, 49.8% water, and 0.2% formic acid, v/v/v). Before usage, the solvents were degassed by sonication at 23 °C for 10 min. A sample volume of 15 μL was injected into the HPLC system at a flow rate of 0.6 mL/min.

Gradient elution was programmed as follows: 5% B for 10.8 min; linear increase from 5% to 10% B over 8 min; then from 10% to 20% B over 4 min; followed by a ramp from 20% to 25% B over 7.2 min, from 25% to 30% B over 8 min, from 30% to 40% B over 4.2 min, from 40% to 45% B over 2.8 min, and finally from 45% to 100% B over 5 min. The system was then held at 100% B for 10 min before returning to 5% B over 2 min to re-equilibrate.

Peaks were identified employing both positive and negative ionization modes, with nitrogen gas serving as both a nebulizer and drying agent at a pressure of 45 psi and a flow rate of 0.5 mL/min. The capillary and nozzle voltages were set at 3.5 kV and 500 V, respectively. Mass spectral data were acquired across a range of 50–1300 amu, and fragmentation was facilitated using collision energies of 10, 15, and 30 eV. Data collection and analysis were executed using the Agilent LC-ESI-QTOF-MS² Mass Hunter Data Acquisition Software, Version B.03.01, from Agilent Technologies, Santa Clara, CA, USA.

2.6. Experimental Design and Statistical Analysis

The study investigated the effects of temperature, pH, and liquid-to-solid ratio on the recovery rates of TPC, TFC, TTC, TATC, and DPPH radical scavenging activity in GM. The experimental design employed a Box-Behnken factorial design for the second-order response surface, with each of the three factors varying across three levels and incorporating three central points for replication. Each complete response surface design consists of 17 measurement tests. For ethanol concentration and extraction duration, the study adhered to the optimized conditions suggested by Carla et al. [12], single-factor experiment results, and previous empirical methodologies from our laboratory [25], selecting a 60% ethanol concentration for a 16-h extraction period.

All results were pure results subtracted by blanking or control values and expressed as the mean of triple-independent analyses. RSM experiments were constructed and analyzed using Design Expert Software version 13.0. The adequacy of the developed polynomial model was evaluated by the coefficient of multiple determinations (R²). Furthermore, ANOVA was applied to ascertain the statistical significance of the model. The model and its variables were considered statistically significant if the p-value was less than 0.05.

3. Results and Discussion

3.1. Modeling of the Extraction Process from Grape Marc

The optimization of conventional extraction (CE) involved three experimental variables: temperature, pH, and liquid-to-solid ratio. For UAE, the amplitude, time, and temperature were optimized (

Table 1. Symbols and ranges of independent variables for the Box–Behnken experimental design with central replicate under both extractions.

Extraction Methods	Symbols	Independent Variables	Coded Levels
			−1	0	1
CE	A	Temperature (°C)	4	32	60
	B	solvent pH	2	4	6
	C	Liquid-to-solid ratio, L/S (mL/g DM)	10:1	30:1	50:1
UAE	A	Amplitude %	10	55	100
	B	Time (min)	2	5	8
	C	Temperature (°C)	4	32	60

). The results of the ANOVA, the goodness of fit, and the adequacy of the model are summarized in Supplementary Tables S1 and S2. The relationship between each experimental level and response is further depicted in surface plots for intuitive understanding (Figure 1). The diagnostic plots for each response surface model are included in the Supplementary Figures S1–S10.

The analysis utilized statistical parameters to bridge theoretical predictions with experimental results and the structure of the experiment. The regression coefficient (R²), indicative of the goodness of fit, exceeded 0.8500 in this study, suggesting reasonable model fit and forecasting ability. A p-value < 0.05 denotes the significant impact of the assessed factors on the dependent variables [26]. Based on the model indicators for the analyzed variables in the grape marc matrix, the consistency of the results suggests that the model has been effectively adjusted.

3.2. Effects of Extraction Variables on Total Phenolics Content and Total Flavonoids Content

For the conventional method, TPC was influenced by three variables, with quadratic effects being significant, indicating a non-linear relationship. In contrast, UAE exhibits a more pronounced interdependence between parameters, as illustrated in the surface plots in Figure 1. Under conventional conditions, the optimal TPC recovery was achieved at a temperature of 60 °C, pH of 2.0 ± 0.05, and a liquid-to-solid ratio of 50:1. During UAE, the maximum TPC recovery was achieved with an amplitude of 100, within a time range of 4–6 min, and a temperature range of 44–60 °C. The enhanced extraction efficiency under acidic pH in ethanol-rich solvents likely contributes to the improved extraction rates [27]. This finding is in line with previous reports that the optimal TPC was achieved at extraction temperatures between 60–80 °C for conventional methods [28].

However, high-intensity ultrasonication could lead to the generation of diverse free radicals that may interact with phenolics like quercetin, triggering unfavorable processes, including oxidation and polymerization [29]. Thus, UAE parameters have been refined to reduce processing time, aiming to decrease the generation of diverse free radicals.

In conventional extraction, optimal TFC recovery is observed at a temperature of 60 °C, a pH of 2.0 ± 0.05, and a 10:1 liquid-to-solid ratio, with the latter showing less significance on TFC compared to the other variables. For UAE methods, peak TFC yields are reached under settings similar to those for optimal TPC recovery: an amplitude of 100, a time span of 4–5 min, and a temperature of 60 °C (Figure 1c,d).

3.3. Effects of Extraction Variables on Total Tannins Content

The extraction of tannins depends on the influence of multiple interacting factors for both methodologies. The surface plots in Figure 1 illustrate that for conventional extraction, the optimum extraction efficiency is achieved when the temperature is maintained at 60 °C, pH at 2.0 ± 0.05, and the solid-to-solvent ratio at 1:50. Furthermore, the impact of pH on the outcome is significantly more pronounced. In contrast, for UAE, the highest extraction efficiency is observed when the extraction duration is approximately 5 min, temperature is controlled between 26 °C and 32 °C, and amplitude is set at 53%. However, the influence of amplitude on the TTC is relatively minor. As Das et al. noted [30], for solvents composed of ethanol and water mixtures, higher temperatures yield better results; in contrast, when employing an ultrasonic technique for extraction, lower processing temperatures can be maintained. From this perspective, it results in reduced energy consumption and lowers the risk of compound degradation.

3.4. Effects of Extraction Variables on Total Anthocyanins Content

For the conventional extraction approach of total anthocyanins, the corresponding surface plot (Figure 1g) displays minimal curvature, illustrating negligible quadratic effects from the independent variables. Optimal extraction conditions were identified as a 1:50 solid-to-liquid ratio, pH at 2.0 ± 0.05, and a direct correlation between higher temperatures and increased extraction efficiency. From response surface methodology, the TATC’s regression model boasts an R² of 0.9338, indicating that the model accounts for 93.38% of the total variability within the studied range. Conversely, in UAE, selected parameters showed no significant effects on outcomes, with a slight preference towards temperature but devoid of any discernible interaction effects. Optimal extraction was achieved at 100% amplitude over an 8-min duration, and the temperature was set at 60 °C.

The pH value of the solution is a crucial parameter in this study; employing HCl to acidify the solvent represents a strategy that allows for a more selective extraction of the compounds under investigation [7]. Türker et al. analyzed the impact of pH on the stability of anthocyanin pigments extracted from black carrots and reported that colored flavonoid compounds exhibit greater stability in acidic environments (pH = 2.0 ± 0.05) [31]. Consequently, pigment stability is compromised with increasing pH values. Furthermore, maintaining a lower pH value during storage can enhance the stability of the samples [32].

Research by Havlíková et al. focused on the stability of anthocyanins subjected to varying pH levels and temperatures [33]. Their findings indicated that pH markedly impacts the thermal stability of anthocyanins at temperatures between 50–60 °C. Yet, this influence wanes at higher temperatures, specifically those above 70 °C. The study also notes that the anthocyanins’ stability is not impacted by the pH if the oxygen concentration during the extraction is insignificantly low [33].

3.5. Effects of Extraction Variables on Radical Scavenging Capacity

The antioxidant capacity of grape marc extracts was evaluated using the DPPH radical scavenging assay, where traditional extraction methods displayed a quadratic effect, with the regression model reaching an R² of 0.9799. Optimal radical scavenging activity is observed when the solid-to-solvent ratio is maintained at 1:50, and temperature is controlled between 44–60 °C, with pH held steady at 2 (Figure 1h). However, the UAE method did not show a significant correlation. This optimum condition in traditional extraction closely parallels the best conditions determined for TPC and TFC, thus accounting for the pronounced free radical scavenging capability detected via DPPH.

The introduction of high-frequency waves leads to the disturbance of the solute-solvent mixture, resulting in the rupture of cell walls and enhanced solvent diffusion, thereby aiding in achieving higher extraction rates [34]. However, for such small systems, temperature control must also be considered. For instance, uneven contact of the sample with the probe or localized excessive heating could lead to instability in the final DPPH results.

3.6. Optimization of the Extraction Process

Through the application of RSM, optimal conditions for extracting TPC, TFC, TTC, TATC, and DPPH have been established. The regression models for the conventional extraction method demonstrated satisfactory statistical significance, whereas the UAE provided less guidance for the extraction of TATC and DPPH. In addition to the effects of localized excessive heating caused by ultrasonic probes as described earlier, a more plausible explanation lies in the unexpected substances released during UAE, which may contribute to the differences between the TATC and DPPH methods: TATC calculates anthocyanin content based on absorbance changes induced by pH variations; whereas the DPPH does not solely reflect the antioxidant capacity of phenolic compounds. Due to its lipophilic nature, the DPPH method lacks a certain degree of biological relevance. Considering the degree of fit for each response to the model, we prioritized responses with higher fit in the overall condition optimization sequence. The desirability function approach [19] provided predictions for the best set of conditions for all responses (

Table 2. Optimal extraction conditions.

Condition	Conventional Extraction	Ultrasonic-Assisted Extraction
Temperature (°C)	49.2	60.0
pH	2.0	2.0
Solid-to-Solvent Ratio	1:50	1:50
Amplitude (%)	N/A	100
Time (min)	16 h	5.05 min

), which were validated (

Table 3. Optimal predicted values and validation.

Response Variables	Predicted		Experimental
	Conventional Extraction	Ultrasonic-Assisted Extraction	Conventional Extraction *	Ultrasonic-Assisted Extraction *
Total phenolics content (mg GAE/g)	36.38	28.35	33.16 ± 0.49	28.29 ± 0.36
Total flavonoids content (mg QE/g)	2.08	4.09	2.51 ± 0.20	4.01 ± 0.04
Total tannins content (mg GAE/g)	24.78	20.13	25.75 ± 0.83	20.04 ± 0.31
Total anthocyanins content (mg/g)	0.77	1.04	0.92 ± 0.14	1.14 ± 0.05
DPPH (mg TE/g)	52.22	50.18	51.47 ± 0.87	54.14 ± 0.71

* Analytical results are means ± SD (n  =  3).

). Findings indicate that UAE surpasses conventional methods in terms of TFC, TATC, and DPPH (Figure 2). Notably, The TFC achieved via UAE was approximately 160% of that yielded by traditional methods, achieved in just 5.05 min. Typically, the antioxidant attributes of plant extracts are ascribed to their flavonoid concentrations, correlating higher contents with stronger antioxidant capacities and increased DPPH values [35], echoing the results obtained here.

The reported method significantly surpasses the universal conventional extraction techniques previously applied to grapes by our lab [36], and it also improves upon the extraction method optimized by Carla et al. [12], achieving higher yields of TPC, TFC, TTC, and TATC.

Considering the energy efficiency of the extraction process, apart from TFC, the impact of temperature on other responses is not significant. Even when the temperature is reduced to 49.5 °C, the extraction efficiency remains nearly consistent, with TFC still achieving 85% (3.4534 mg QE/g; compared to 60 °C). Excluding TFC, maintaining the temperature at 15 °C can still ensure that the effectiveness of other response factors remains at or above 85%. Notably, even at 15 °C, the predicted extraction efficiency of TFC can reach 2.219 mg/g, which is still higher than the optimized conventional extraction methods. To enhance the overall energy efficiency of the extraction, it is recommended to consider the target compounds and adopt 49.5 °C as a compromise. This allows for the extraction of 85% of TFC and is close to optimal conditions for TPC, TTC, TATC, and DPPH.

3.7. LC-ESI-QTOF-MS/MS Based Characterization of Phenolic Compounds

The measurement of antioxidant capabilities involves a range of reactions and methodologies aimed at determining the antioxidant capacity of plant-based materials. However, the intricate composition of phytochemicals precludes a unified approach that accurately reflects total antioxidant capacity. Hence, MS/MS characterization stands as a vital segment in phytochemical investigation, crucial for the quantification of total phenolics and their antioxidant efficacy.

Qualitative phenolic analysis of grape marc extracts was conducted utilizing LC-ESI-QTOF-MS/MS under negative and positive ionization settings with the Personal Compound Database and Library [37] (

Table 4. Characterization of phenolic compounds in extracted grape marc by LC-ESI-QTOF-MS/MS.

No.	Molecular Formula	Proposed Compounds	RT (min)	Ionization (ESI+/ESI−)	Theoretical Mass (m/z)	Observed Mass (m/z)	Mass Error (ppm)	MS/MS Production	Phenolic Compounds	Extraction Methods
	Phenolic acids
1	C26H26O12	1-Feruloyl-5-caffeoylquinic acid	3.73	[M − H]−	529.1340	529.1349	−1.7009	193.0506, 191.0561 179.0350, 135.0452	Hydroxycinnamic acids	CE
2	C39H58O4	Schottenol ferulate	43.20	[M − H]−	589.4291	589.4272	−3.2235	413.3789	Hydroxycinnamic acids	CE
3	C18H22O10	3-Sinapoylquinic acid	37.50	[M − H]−	397.1135	397.1120	−3.7773	191.0561, 173.0455	Hydroxycinnamic acids	CE
4	C9H8O4	Caffeic acid	45.80	[M − H]−	179.0360	179.0355	−1.1171	135.0519	Hydroxycinnamic acids	UAE
5	C16H20O9	Ferulic acid 4-O-glucoside	45.719	[M − H]−	355.1055	355.1049	−1.6896	193.0506, 177.0193, 147.0452, 135.0452	Hydroxycinnamic acids	UAE
6	C18H16O8	Rosmarinic acid	30.185	[M − H]−	359.0802	359.0804	0.5570	179.0355, 161.0225, 135.0519	Hydroxycinnamic acids	UAE
7	C14H6O8	Ellagic acid	48.44	[M − H]−	301.0006	300.9993	−4.3189	283.9956, 257.0079, 229.0126, 185.0229	Hydroxybenzoic acids	CE
	Flavonoids
1	C22H24O13	4′-O-Methyl-(-)-epigallocatechin 7-O-glucuronide	12.04	[M − H]−	495.1160	495.1161	0.2020	415.1035, 313.0565	Flavanols	CE
2	C21H22O13	(-)-Epigallocatechin 3′-O-glucuronide	25.104	[M − H]−	481.0999	481.0983	−3.3257	305.0667	Flavanols	CE, UAE
3	C43H32O20	Theaflavin 3,3′-O-digallate	25.324	[M − H]−	867.1405	867.1387	−2.0758	715.1305, 563.0826, 545.0725	Flavanols	CE, UAE
4	C45H38O18	Procyanidin trimer C1	45.787	[M − H]−	865.2015	865.2000	−1.7337	739.1668, 713.1512, 577.1351, 289.0844	Flavanols	CE, UAE
5	C16H16O6	3′-O-Methylcatechin	53.772	[M − H]−	303.089	303.0879	−2.6395	271.0612, 163.0401	Flavanols	UAE
6	C21H20O10	Apigenin 7-O- glucoside	30.811	[M − H]−	431.0979	431.0958	−4.87	431.0958, 268.0360	Flavones	UAE
7	C30H34O17	Naringin 6′-malonate	25.104	[M − H]−	665.1710	665.1723	1.9544	665.1723, 545.1148	Flavanones	UAE
8	C27H32O14	Narirutin	7.965	[M − H]−	579.1744	579.1759	2.5899	271.0591, 151.0038	Flavanones	CE, UAE
9	C28H34O15	Neohesperidin	18.903	[M − H]−	609.1880	609.1865	−2.4623	609.1865, 301.0710	Flavanones	UAE
10	C26H28O16	Quercetin 3-O-glucosyl-xyloside	19.820	[M − H]−	595.1278	595.1294	2.6885	265.0264, 138.0156, 115.9991	Flavonols	UAE
11	C33H40O21	Quercetin 3-O-glucosyl-rhamnosyl-glucoside	18.068	[M − H]−	771.1954	771.1946	−1.0374	753.1878, 301.0348	Flavonols	CE, UAE
12	C24H24O11	6″-O-Acetylglycitin	43.656	[M + H]+	489.1392	489.1413	4.2932	489.1391, 285.0758, 269.0808	Isoflavonoids	CE, UAE
13	C17H16O6	Violanone	49.102	[M + H]+	317.1020	317.1034	4.41	299.0914, 285.0758, 135.0441	Isoflavonoids	CE
14	C28H33O15	Peonidin 3-O-rutinoside	43.104	[M + H]+	610.1899	610.1877	−3.6054	301.0707	Anthocyanins	CE, UAE
15	C23H23O12	Cyanidin 3-O-(6″-acetyl-glucoside)	50.959	[M + H]+	491.1160	491.1172	2.4434	287.0550	Anthocyanins	CE, UAE
	Lignans
1	C20H20O7	7-Oxomatairesinol	18.192	[M + H]+	373.1282	373.1297	4.0200	357.0969, 343.1176, 327.1227	Lignans	UAE
2	C20H20O6	Conidendrin	42.546	[M − H]−	355.1196	355.1215	5.3503	337.0718, 311.0925, 309.1132, 295.0976	Lignans	CE
3	C20H26O6	Secoisolariciresinol	35.665	[M + H]+	363.1823	363.1802	−5.8145	327.1640, 163.0776, 137.0569, 133.0663	Lignans	CE
4	C28H36O8	Tigloylgomicin H	48.253	[M − H]−	499.2375	499.2367	−1.6024	499.2367, 401.1606, 385.1657	Lignans	CE
	Other phenolic compounds
1	C9H12O2	4-Ethylguaiacol	47.591	[M − H]−	151.0753	151.0753	0.0000	151.0753, 103.0189	Alkylmethoxyphenols	CE
2	C24H30O13	Demethyloleuropein	16.74	[M − H]−	525.162	525.1634	3.2371	363.1085, 319.1187, 249.0769	Tyrosols	UAE
3	C20H26O4	Carnosol	43.205	[M − H]−	329.1776	329.1779	0.9114	286.1847, 185.0586, 270.1605, 201.0885	Phenolic terpenes	CE
4	C27H22O12	Lithospermic acid	5.446	[M − H]−	537.1080	537.1089	1.6756	537.1033, 493.1135, 357.0610, 295.0606	Other phenolic compounds	UAE

). Through the CE approach, 19 types of compounds were recognized, including flavonoids, phenolic acids, and lignans. The UAE approach also resulted in the identification of 19 compounds, including flavonoids, phenolic acids, and lignans (Figure 3).

For UAE, only three types of phenolic acids were identified in GM, fewer compared to those detected through CE methods. This discrepancy is likely attributable to the duration of ultrasonic treatment. Iftikhar, et al. [38] reported that increasing UAE to durations such as 30 min can enhance the concentration of phenolic acids in the extract. In contrast, the optimized duration for ultrasound in this experiment was set at 5.05 min. For lignans, UAE also may not have allowed for their maximal release due to insufficient time or intensity to facilitate their optimal extraction.

3.7.1. Phenolic Acids

In our study, a total of seven phenolic acids were preliminarily identified (Table 4). Most phenolic acids show the loss of CO₂ (44 Da) and the hexosyl moiety (162 Da) [39]. Compound 1 was identified as 1-feruloyl-5-caffeoylquinic acid based on its m/z 529.1349 in negative ion mode, which has been previously reported to be isolated from grape marc [40]. Compound 2 was identified as schottenol ferulate at m/z 589.4272 in the ESI− mode. This compound has been noted to be secreted in large amounts by grapevine rootstocks, which exhibit a stronger increase in Fe availability at the rhizosphere, potentially playing a role in Fe(III) reduction [41]. Similarly, in negative ion mode, compound 4 was identified as caffeic acid at m/z 179.0355, which produced fragment ions at m/z 135.0519 via the removal of CO₂. Compound 7 was identified as ellagic acid at m/z 300.9993 and MS² fragments at m/z 257.0079 [M − H-CO₂] and 229.0126 [M − H-CO₂-CO]. These two compounds are important phenolic acids widely present in wine [42] and grapes [43]. Ferulic acid 4-O-glucoside (compound 5 with m/z 355.1049 in the ESI− mode) was identified by the presence of product ions at m/z 193, m/z 178, m/z 149, and m/z 134. These ions resulted from the loss of C₆H₁₀O₅ (162 Da), C₇H₁₃O₅ (177 Da), C₇H₁₀O₇ (206 Da), and C₈H₁₃O₇ (221 Da) from the parent ion, respectively. It has been previously reported to be isolated from grape pulp samples [44]. Compound 6 (rosmarinic acid) was preliminarily identified at m/z 359.0804 in negative ion mode [45]. Structurally, it is an ester of caffeic acid and 3,4-dihydroxyphenyllactic acid [46], also known as a caffeic acid dimer [47]. It produced the fragment ions at m/z 179.0355 (caffeic acid) and caffeic acid fragments at m/z 135.0519 via the removal of one unit of CO₂ moiety, respectively, from the precursor ion. It is widely present in plants as a natural antioxidant and is frequently studied for its role in stabilizing/enhancing the color of wine as an auxiliary pigment [46].

3.7.2. Flavonoids

Flavonoids commonly exist in foods in the form of O-glycosides or C-glycosides [48]. UAE released a total of 13 flavonoid compounds in this study, compared to 10 flavonoids from CE, with eight being common to both techniques. This advantage of UAE in extracting flavonoid compounds corresponds with the previously mentioned TFC and DPPH trends. Compounds 1 and 2 were preliminarily identified in negative ion mode at m/z 495.1161 and m/z 481.0983, respectively. Epigallocatechin derivatives are typically found in grape skins or stems [49,50]. Compound 3 (theaflavin 3,3′-O-digallate) was identified at m/z 867.1387 in ESI−, although it is not considered a typical compound found in grapes or wine. A possible explanation is that theaflavins are formed through the enzymatic oxidation and condensation of catechins with dihydroxy and trihydroxy B-rings [51]. This reaction involves oxidizing the B-ring to a quinone, followed by a Michael addition of catechin quinone to gallocatechin quinone, then carbonyl addition across the entire ring and subsequent decarboxylation [52].

Procyanidin trimer C1 (Compound 4) is an important compound found in grape seeds and was identified at m/z 865.2000 in negative ion mode [53]. This was confirmed by the major fragments at m/z 739.1668, m/z 713.1512, and m/z 577.1351, corresponding to a 126 Da loss due to the heterocyclic ring fission reaction, a 152 Da loss from the retro-Diels–Alder reaction and (epi)catechin dimer [M − H]⁻ ion produced by quinone methide reaction, respectively [54]. 3′-O-methylcatechin (Compound 5) was tentatively identified at precursor [M − H]⁻ m/z 303.0879 and is considered a circulating metabolite of grape seed phenolics [55]. Compound 10 (quercetin 3-O-glucosyl-xyloside with [M − H]⁻ m/z at 595.1294) was tentatively identified by its main product ions at m/z 265.0264 [M − H–glucoside–xyloside], 138.0156 [M − H–glucoside–xyloside–H₂O–C₆H₅O₂], and 115.9991 [M − H–glucoside–xyloside–C₈H₆O₃] [56]. And Compound 11 (quercetin 3-O-glucosyl-rhamnosyl-glucoside) was tentatively identified in negative ion mode at m/z 771.1946. These two compounds are present in white wine and are used to detect the oxidation state of the studied white wine [57]. 6″-O-acetylglycitin (compound 12) and cyanidin 3-O-(6″-acetyl-glucoside) (compound 15) were both identified at m/z 489.1413 and m/z 491.1172 in positive ionization mode. The presence of a fragment ion at m/z 287 in the mass spectrum of compound 15 suggests an acetyl glucoside moiety. Both compounds were reported in the detection of active substances in grape pomace [45]. The compound 13 violanone forms its ion at m/z 315.0893 and has been reported as an isoflavone compound identified from unripe grape samples [58]. Similarly, compound 6 (apigenin 7-O-glucoside) has also been detected in unripe grapes [58], and it was identified at m/z 431.0958 in negative ionization mode.

Compounds 7 (naringin 6′-malonate) and 8 (narirutin) were tentatively identified at m/z 665.1723 and m/z 579.1759, respectively, in negative ionization mode. Compound 9 (neohesperidin) was identified at m/z 609.1865. It is noteworthy that these phenolic compounds identified by LC-MS are commonly found in citrus fruits. When comparing the phenolic composition of citrus fruits and grapes, there is significant overlap [36,59]. Chen et al. [60] conducted chemometric analyses on 15 citrus and 12 grape varieties and found that the phenolic compounds in both fruits could be categorized into two major groups with considerable consistency. Additionally, the flavonoid pathway in grapes is subject to transcriptional regulation, which results in the production of a variety of compounds. The chemical structures of different flavonoids determine their properties, with anthocyanins commonly regarded as the end product of the flavonoid pathway [61]. This means that there are pathways and possibilities for the transformation of compounds, including narirutin, kaempferol, quercetin, and catechin, from the grape growth phase through to the wine fermentation process.

3.7.3. Lignans

Lignan compound 1 was preliminarily identified at m/z 373.1297 in positive ionization mode, and it has been reported to be present in grape pulp [44]. Conidendrin (compound 2) is a lignin present in wine and was identified at m/z 255.1215 in ESI− [62]. Compound 3 (secoisolariciresinol) forms ions at m/z 363.1802 and has been reported to be present in unripe grapes [58].

3.7.4. Other Phenolic Compounds

4-Ethylguaiacol (compound 1) was tentatively identified by its main product ions at m/z 151.0753 [M − H]⁻ and m/z 103.0189 [M–CH₅O₂]⁻ in negative ionization mode [63]. This compound is found in wine, and its potential sources include grapes, winemaking additives, and oak barrel aging [64]. Excessive concentrations may impart undesirable flavors to the wine [63]. Lithospermic acid (compound 4) was preliminarily identified at m/z 537.1089 in ESI− and is a compound of rosmarinic acid with caffeic acid [47].

4. Conclusions

The focus of this study is on the optimization of green extraction of phenolic compounds from GM using RSM. The Box–Behnken experimental design proved to be a valuable tool for assessing the effects of solvent pH, liquid-to-solid ratio, temperature, and their interactions, as well as for comparing traditional extraction methods with green UAE techniques, thereby optimizing the TPC, TFC, TTC, TATC, and DPPH in GM extracts. The optimal conditions for traditional extraction were found to be 60% ethanol (pH 2.0 ± 0.05) at a liquid-to-solid ratio of 50:1, extracting for 16 h at 49.2 °C, whereas for UAE, the optimal conditions were determined to be 60% ethanol (pH 2.0 ± 0.05) at a liquid-to-solid ratio of 50:1, with an amplitude of 100%, extracting for 5.05 min at 60 °C, but alternative lower temperature of 49.5 is still suitable. Under these conditions, the extraction efficiency of UAE is slightly higher than that of traditional methods, with a significant improvement in the extraction of flavonoids, anthocyanins, and DPPH. Additionally, the overall required time is greatly reduced. However, this observation is limited to the laboratory-scale system used in the current study. Further research is needed to determine its applicability in industrial production. This includes pilot-scale studies focused on maintaining the extraction efficiency observed at the laboratory scale while optimizing industrial conditions, conducting cost-benefit analyses after scale-up to evaluate the economic feasibility of scaling up and other relevant investigations.

Under optimized conditions, UAE and CE extraction resulted in different phenolic compound profiles. While both methods extracted similar numbers of phenolic acids, the compositions are different. UAE identified more flavonoid compounds than CE, with 8 of them being identical, particularly anthocyanins. Notably, the UAE significantly reduced lignin content. This indicates that UAE effectively extracts flavonoid compounds in a short time, with a similar chemical composition to CE. This discovery provides valuable insight into the use of UAE for targeted extraction of flavonoids. Given that the phenolic composition of the extracts varies, future research will target detailed qualitative and quantitative studies on the extracted compounds. The optimized UAE method proposed in this study offers an alternative approach to phenolic compound extraction from GM, with potential advantages in processing time and flavonoid recovery. Due to its simplicity, the recyclability of the solvents used, and the significantly shortened extraction time after optimization, UAE extraction shows great potential for green, large-scale industrial production. This not only promotes the further development of GM phenolic compound-based products, such as those in the food, pharmaceutical, and cosmetics industries but also mitigates the negative economic and environmental impacts associated with improper or inefficient GM processing.