Liquorice dreg extract and extraction method

Abstract

The invention relates to the technical field of natural product extraction, in particular to a licorice herb residue extract and an extraction method.The extraction method comprises the steps that firstly, choline chloride and glycerinum are mixed, heated and stirred, then pure water is added into a DES solvent, and a DES solution is obtained; then adding the licorice herb residue powder into a DES solution, and carrying out ultrasonic extraction at room temperature to obtain a primary extracting solution; centrifuging and precipitating the primary extracting solution, and freeze-drying the supernate to obtain the liquorice herb residue extract. Compared with a traditional ethanol reflux extraction method, the method disclosed by the invention has the advantages that the extraction of the active substances in the licorice herb residues is remarkably improved, and the clearance rates of the licorice herb residue extract prepared by the method to DPPH, superoxide anions and superoxide anions are 93.02%, 80.95% and 73.22% respectively, and are improved by 17.08%, 15.73% and 21.08% respectively compared with the ethanol reflux extraction method; the biological activity and the antioxidant activity are excellent.

Description

Technical Field

[0001] The invention relates to the technical field of natural product extraction, and in particular to a liquorice residue extract and an extraction method. Background Art

[0002] Licorice, derived from the roots of the Glycyrrhiza uralensis plant, is considered a versatile herbal medicine with a rich history spanning centuries. Modern pharmacological studies have highlighted the pharmacological activities of licorice components, such as flavonoids, saponins, and polysaccharides, which include anti-inflammatory, antibacterial, antioxidant, and antitumor effects. In particular, saponins have more significant effects on cancer, immune dysfunction, and endocrine dysfunction. In recent years, the licorice industry has expanded rapidly and research interest has surged. With the rapid growth in demand for licorice extracts, the amount of waste residues after its extraction process has also been increasing. The wastewater and solvents generated during the secondary extraction of waste residues are also polluting the environment. Therefore, there is an urgent need to find a convenient and efficient extraction method.

[0003] The traditional licorice extraction method mainly utilizes ethanol reflux, which is simple and low-cost, but has the problems of low extraction rate, long extraction time and high solvent consumption. The extraction rate of total licorice by traditional ethanol extraction is only 0.54 mg / g (Caroline Metal., 2023), among which the extraction rates of total flavonoids, saponins and polysaccharides of licorice are approximately 11.32 mg / g, 2.27 mg / g and 117.30 mg / g, respectively (Li, 2022; Dong et al., 2023, Liu, 2023). The problem of low extraction rate is a serious waste of resources, which limits the industrial production of licorice extract. Therefore, finding an environmentally friendly, reusable extraction solvent with high extraction rate is the focus of current research.

[0004] Deep eutectic solvents (DES) are composed of two or three components in a specific ratio. Compared with pure substances, DES is more conducive to promoting the efficient extraction and separation of target compounds. At the same time, DES has the advantages of high stability, low melting point, easy synthesis, low cost, and environmental friendliness, and has been widely used in green chemistry. In recent years, DES has been increasingly used as a solvent for extracting bioactive components of natural medicines. Researchers believe that it can replace traditional solvents to extract natural compounds from plants. The study of ultrasonic-assisted DES extraction of perilla flavonoids found that under the same conditions, it can increase the extraction rate of flavonoids by 75%. Tang et al. extracted saponins from camellia seed residues using l-proline, glycerol and sucrose ternary DES, and the extraction rate reached 23.22%, which was 505% higher than that of traditional ethanol extraction. More studies have found that the extraction rate of polysaccharides in natural products can be increased by 150% using DES as a solvent (Yahaya et al., 2024). However, few studies have comprehensively considered the simultaneous extraction of multiple active ingredients in natural medicines, especially the simultaneous extraction of multiple active ingredients from licorice residues while maintaining a high extraction rate and good biological activity. Summary of the invention

[0005] Based on this, the present invention provides a liquorice residue extract and an extraction method to solve the problem in the prior art that multiple effective components of liquorice residue cannot be extracted simultaneously while maintaining a high extraction rate and good biological activity.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for extracting liquorice residues comprises the following steps:

[0008] S01. Mixing choline chloride and a hydrogen bond donor in a molar ratio of 1:3, heating and stirring, until a stable, uniform, transparent liquid is formed, which is a DES solvent; adding pure water to the DES solvent to obtain a DES solution; wherein the water content of the DES solution is 20% to 60%;

[0009] S02. The licorice residue powder is added to the DES solution, and ultrasonic extraction is performed at room temperature for 10 min to 50 min, with an ultrasonic power of 100 W to 500 W to obtain a primary extract; wherein the solid-liquid ratio of the licorice residue powder to the DES solution is 1: (10 to 50);

[0010] S03. Centrifuge the primary extract, precipitate, and freeze-dry the supernatant to obtain a liquorice residue extract.

[0011] Preferably, in the above-mentioned method for extracting liquorice residue, in step S01, the molar ratio of choline chloride to the hydrogen bond donor is 1:3.

[0012] Preferably, in the above-mentioned method for extracting licorice residue, in step S01, the hydrogen bond donor is selected from any one of 1,2-propylene glycol, 1,4-butanediol, ethylene glycol, glycerol, lactic acid, citric acid, malic acid, glucose, fructose, and sucrose.

[0013] Preferably, in the above-mentioned method for extracting liquorice residue, in step S01, the hydrogen bond donor is glycerol.

[0014] Preferably, in the above-mentioned method for extracting liquorice residue, in step S01, the water content of the DES solution is 50%.

[0015] Preferably, in the above-mentioned method for extracting liquorice residue, in step S01, the ultrasonic power is 300W.

[0016] Preferably, in the above-mentioned method for extracting liquorice residue, in step S02, the solid-liquid ratio of the liquorice residue powder to the DES solution is 1:40.

[0017] Preferably, in the above-mentioned method for extracting liquorice residue, in step S02, the extraction time is 30 minutes.

[0018] A liquorice residue extract is prepared by any of the extraction methods described above.

[0019] A use of the liquorice residue extract as an antioxidant.

[0020] The technical solution of the present invention has at least the following advantages:

[0021] The invention provides a method for extracting liquorice residues. First, choline chloride and a hydrogen bond donor are mixed in a molar ratio of 1:3, heated and stirred, and then pure water is added to a DES solvent to obtain a DES solution with a water content of 20% to 60%; then liquorice residue powder is added to the DES solution, the solid-liquid ratio of the liquorice residue powder to the DES solution is 1:(10 to 50), ultrasonic extraction is performed at room temperature for 10 minutes to 50 minutes, and the ultrasonic power is 100W to 500W to obtain a primary extract; and then the primary extract is centrifuged and precipitated, and then the supernatant is freeze-dried to obtain a liquorice residue extract. The present application establishes an optimal deep eutectic solvent (DES) system for extracting licorice residue. Through ultrasound-assisted extraction, multiple effective ingredients in licorice residue can be extracted simultaneously with a high extraction rate. Experiments show that the extraction rates of flavonoids, saponins and polysaccharides are 8.840±0.412 mg / g, 41.680±0.776 mg / g and 441.997±11.630 mg / g, respectively, and the traditional ethanol extraction rates are 0.735±0.092, 3.291±0.323, and 3.469±0.462, respectively. Compared with the traditional ethanol reflux extraction method, the total extract extraction rate of licorice residue in the DES system is increased by 1077.517%, the total flavonoid extraction rate is increased by 176.596%, the saponin extraction rate is increased by 389.892%, and the polysaccharide extraction rate is increased by 1367.317%.

[0022] The liquorice residue extract prepared by the above method disclosed in the present invention has excellent biological activity. Experimental data show that the liquorice residue extract (DES extract) extracted by the scheme has a scavenging rate of 93.02% for DPPH, which is significantly higher than 79.45% of the liquorice residue extract extracted by ethanol extraction (hereinafter referred to as ethanol extract), and the improvement rate is 17.08%; the scavenging rate of the DES extract for superoxide anion is 80.95%, which is significantly higher than 69.95% of the ethanol extract, and the improvement rate is 15.73%; the scavenging rate of the DES extract for superoxide anion is 83.04%, which is also higher than 70.98% of the ethanol extract, and the improvement rate is 16.99%; the scavenging rate of the DES extract for ABTS free radicals is 73.22%, which is significantly higher than 60.47% of the ethanol extract ABTS, and the improvement rate is 21.08%. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 Effects of different DES and ethanol on the extraction rate of effective components of licorice residue, including flavonoid extraction rate (a), saponin extraction rate (b), polysaccharide extraction rate (c); (compared with 60% ethanol, *P<0.05, **P<0.01, ***P<0.001; compared with DES-4, #P<0.01, ###P<0.001); Effects of different extraction conditions on the effective components of licorice residue in single factor experiment, including time (d), power (e), water content (f), liquid-to-solid ratio (g); (compared with the initial group, *P<0.05, **P<0.01, ***P<0.001).

[0025] Figure 2 Response surface plots and contour plots illustrate the interaction of various factors, including time and power (a), time and water content (b), time and liquid-solid ratio (c), water content and power (d), liquid-solid ratio and power (e), liquid-solid ratio and water content (f); schematic diagram of RSM and ANN extraction results, including RSM extraction results (g), ANN partial extraction results (h).

[0026] Figure 3 The antioxidant capacity of the active ingredients of licorice residue extracted with different solvents, as well as the recovery rate and reuse extraction rate of DES, including DPPH (a), superoxide anion (b), hydroxyl radical (c), ABTS (d); the first extraction rate and reuse extraction rate of DES (e) (compared with 60% ethanol, *P < 0.05, **P < 0.01, ***P < 0.001)

[0027] Figure 4 SEM images and molecular dynamics simulations of licorice residue powder, including untreated licorice residue powder (a-a'), licorice residue powder extracted with 60% ethanol (b-b'), licorice residue powder extracted with DES (c-c'), snapshot of glycyrrhizin dissolution in 60% ethanol solution (d), snapshot of glycyrrhizin dissolution in DES solution (d'), snapshot of glycyrrhizic acid dissolution in 60% ethanol solution (e), snapshot of glycyrrhizic acid dissolution in DES solution (e'), snapshot of glucose dissolution in 60% ethanol solution (f), snapshot of glucose dissolution in DES solution (f').

[0028] Figure 5 Schematic diagram of molecular dynamics analysis results, including the density of glycyrrhizin (a), the number of hydrogen bonds of glycyrrhizin (b), the average lifetime of hydrogen bonds of glycyrrhizin (c), the density of glycyrrhizic acid (d), the number of hydrogen bonds of glycyrrhizic acid (e), the average lifetime of hydrogen bonds of glycyrrhizic acid (f), the density of glucose (g), the number of hydrogen bonds of glucose (h), and the number of hydrogen bonds of glucose (i).

[0029] Figure 6It is a quantum chemical analysis, in which the molecular structure of ethanol extract glycyrrhizin (a); the molecular structure of DES extract glycyrrhizin (a'); the molecular structure of ethanol extract glycyrrhizic acid (b); the molecular structure of DES extract glycyrrhizic acid (b'); the molecular structure of ethanol extract glucose (c); the molecular structure of DES extract glucose (c'); the electrostatic distribution of ethanol extract glycyrrhizin (d); the electrostatic distribution of DES extract glycyrrhizic acid (d'); the electrostatic distribution of ethanol extract glycyrrhizic acid (e); the electrostatic distribution of DES extracted glycyrrhizic acid (e'); the electrostatic distribution of ethanol extract glucose (f); the electrostatic distribution of DES extracted glucose (f'); the RDG isomer of ethanol extracted glycyrrhizin (g); the RDG isosurface of DES extracted glycyrrhizic acid (g'); the RDG isosurface of ethanol extracted glycyrrhizic acid (h); the RDG isosurface of DES extracted glycyrrhizic acid (h'); the RDG isosurface of ethanol extracted glucose (i);

[0030] RDG isosurface of glucose extracted by DES method (i'); Weak interaction analysis of licorice extract by ethanol (j);

[0031] Weak interaction analysis of glycyrrhizin extracted by DES (j'); weak interaction analysis of glycyrrhizic acid extracted by ethanol (k); weak interaction analysis of glycyrrhizic acid extracted by DES (k'); weak interaction analysis of glucose extracted by ethanol (l); weak interaction analysis of glucose extracted by DES (l').

[0032] Figure 7 Schematic diagram of the artificial neural network structure. DETAILED DESCRIPTION

[0033] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present invention can be combined with each other. The technical solution of the present invention will be further described below in conjunction with the drawings of the embodiments of the present invention, and the present invention is not limited to the following specific implementation methods.

[0034] It should be understood that the same or similar reference numerals in the drawings of the embodiments correspond to the same or similar components. In the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "front", "rear", "left", "right", "top", "bottom" and the like indicating an orientation or positional relationship, it is based on the orientation or positional relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as a limitation on this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0035] In a specific embodiment, a method for extracting liquorice residue comprises the following steps:

[0036] S01. Mixing choline chloride and a hydrogen bond donor in a molar ratio of 1:3, heating and stirring, until a stable, uniform, transparent liquid is formed, which is a DES solvent; adding pure water to the DES solvent to obtain a DES solution; wherein the water content of the DES solution is 50%;

[0037] Deep eutectic solvents (DES) are composed of two or three components in a specific ratio. Compared with pure substances, DES is more conducive to promoting the efficient extraction and separation of target compounds. At the same time, DES has the advantages of high stability, low melting point, easy synthesis, low cost, and environmental friendliness, and has been widely used in green chemistry. DES is synthesized by reacting hydrogen bond acceptors and hydrogen bond donors. In this application, choline chloride is used as a hydrogen bond acceptor, and three types of hydrogen bond donors are selected, including alcohols, acids, and sugars, namely 1,2-propylene glycol, 1,4-butanediol, ethylene glycol, glycerol, lactic acid, citric acid, malic acid, glucose, fructose, and sucrose. Preferably, the hydrogen bond donor is glycerol. The change in the molar ratio of hydrogen bond acceptors to hydrogen bond donors will affect the viscosity of the DES system, thereby changing its extraction rate. Previous studies have shown that when the molar ratio of hydrogen bond acceptors to hydrogen bond donors is too small, the reaction cannot proceed, and excessive hydrogen bond donors may aggravate the internal interference of the system, reduce the interaction ability of the target chemical substance with chloride anions, and lead to a decrease in the extraction rate. The molar ratio of hydrogen bond acceptors to hydrogen bond donors in the commonly used DES system is mostly between 1:1 and 1:5. Furthermore, the present solution selects a molar ratio of choline chloride to hydrogen bond donors of 1:3. Excessively high or low water content in the DES solution affects the extraction rate. The water content of the DES solution in the present application is 20% to 60%. Through experiments, the maximum extraction rate is achieved when the water content is 50%, so the water content of the DES solution is preferably 50%.

[0038] S02. Add the licorice residue powder to the DES solution, perform ultrasonic extraction at room temperature for 10 to 50 minutes, and the ultrasonic power is 100W to 500W to obtain a primary extract; wherein the solid-liquid ratio of the licorice residue powder to the DES solution is 1: (10 to 50); if the ultrasonic power is too small, the extraction rate is low, and if the ultrasonic power is too large, the molecular structure of some polysaccharides and flavonoids is damaged. Therefore, the present application scheme preferably uses an ultrasonic power of 250W to 350W, and further, the ultrasonic power is 300W. In the process of adding the licorice residue powder to the DES solution for extraction, the increase in the volume of the DES solution promotes the dissolution of polysaccharides, thereby improving the total extraction rate. However, an excessively high liquid-to-solid ratio will cause too much impurity to dissolve, thereby affecting the extraction rate. Therefore, the solid-liquid ratio of the licorice residue powder to the DES solution is 1: (35 to 45), and preferably, the solid-liquid ratio of the licorice powder to the DES solution is 1: 40; further, the extraction time is 30min.

[0039] S03. Centrifuge the primary extract, precipitate, and freeze-dry the supernatant to obtain a liquorice residue extract.

[0040] The optimization of the extraction process used TOPSIS method and entropy weight method to prioritize the three main chemical components of licorice residue, and clarified the influence of each component on the total extraction rate. The results showed that the extraction weights of flavonoids, saponins and polysaccharides were 0.188215, 0.438488 and 0.373297, respectively. The response surface methodology (RSM) and artificial neural network (ANN) were used to verify each other, and the optimal extraction process parameters for licorice residue extraction were finally determined as follows: extraction time 30min, ultrasonic power 300W, water content in DES solution 50%, solid-liquid ratio of licorice residue powder and DES solution 40, and molar ratio of choline chloride and hydrogen bond donor 1:3. This process can extract multiple effective components of licorice residue at the same time, and maintain a high extraction rate and good biological activity.

[0041] For example: choline chloride and a hydrogen bond donor are mixed in a molar ratio of 1:3, heated and stirred until a stable, uniform, transparent liquid is formed, which is a DES solvent; pure water is added to the DES solvent to obtain a DES solution with a water content of 50%; licorice residue powder is added to the DES solution, with the solid-liquid ratio of the powder to the DES solution being 40, and ultrasonic extraction is performed at room temperature for 30 minutes with an ultrasonic power of 300 W to obtain a primary extract; the primary extract is centrifuged and precipitated, and the supernatant is freeze-dried to obtain a licorice residue extract.

[0042] In another specific embodiment of the present application, a liquorice residue extract is prepared by any of the extraction methods described above.

[0043] In another specific embodiment of the present application, an extract of the above-mentioned liquorice residue is used as an antioxidant. Experimental data show that the scavenging rate of the DES extract of the present scheme for DPPH is 93.02%, which is significantly higher than the 79.45% of the ethanol extract extracted by ethanol extraction method, and the improvement rate is 17.08%; the scavenging rate of the DES extract for superoxide anions is 80.95%, which is significantly higher than the 69.95% of the ethanol extract, and the improvement rate is 15.73%; the scavenging rate of the DES extract for hydroxyl radicals is 83.04%, which is also higher than the 70.98% of the ethanol extract, and the improvement rate is 16.99%; the scavenging rate of the DES extract for ABTS free radicals is 73.22%, which is significantly higher than the 60.47% of the ethanol extract ABTS, and the improvement rate is 21.08%. It shows that the liquorice residue extract prepared by the present application scheme can be used as a natural antioxidant in the medical field.

[0044] The technical solution and technical effects of the present invention are further illustrated below through specific experimental examples.

[0045] 1. Materials and Methods

[0046] (1) Chemicals and reagents:

[0047] The licorice used in this study was sourced from Tianfeng Licorice Planting Professional Cooperative in Yanchi County, Wuzhong City, Ningxia. Professor Fu Xueyan of the Department of Traditional Chinese Medicine of Ningxia Medical University identified the specimen as Ural Licorice (glycyrhiza uralensis Fisch) according to the Pharmacopoeia of the People's Republic of China.

[0048] Choline chloride (ChCl) (>99%), 1,2-propylene glycol (>99%), 1,4-butanediol (>99%), ethylene glycol (>99%), glycerol (>99%), lactic acid (>99%), citric acid (>99%), malic acid (>99%), glucose (>99%), fructose (>99%), sucrose (>99%), and AB-8 macroporous adsorption resin were purchased from Beijing Sun Biotechnology Co., Ltd. Anhydrous ethanol, sodium nitrite, aluminum nitrate, sodium hydroxide, phloroglucinol, sodium carbonate, vanillin, acetic acid, perchloric acid, phenol, concentrated sulfuric acid, and chloroform were produced by Tianjin Damao Chemical Reagent Factory.

[0049] The standard compounds rutin, gallic acid, anhydrous glucose, and oleanic acid were purchased from Shanghai Yuanye Biotechnology Co., Ltd.

[0050] (2) Active ingredient content detection method:

[0051] a. Flavonoid content detection method: Accurately weigh 2.5mg of rutin standard reference material into a conical flask, dissolve it in 50% ethanol aqueous solution, and transfer it to a 25ml volumetric flask. Dilute with solvent to the mark, shake well to obtain a 0.1mg / mL rutin standard solution. Next, accurately measure 0.5mL, 1.0mL, 2.0mL, 3.0mL, 4.0mL and 5.0mL of standard solution into different 10ml volumetric flasks. Dilute the solution to the mark to obtain solutions of different concentrations. Transfer each solution to a 10ml volumetric flask, add 0.3mL5% sodium nitrite solution, then shake gently and incubate for 6 minutes. Then add 0.3mL 10% aluminum nitrate solution, shake again, and let stand for 6 minutes. Finally, add 4ml 4% sodium hydroxide solution, dilute to the mark with 50% ethanol, shake well, and let stand for 12 minutes. Prepare a blank control solution with reagents. The absorbance was measured at 507 nm using a UV-visible spectrophotometer, with rutin concentration (C) as the abscissa and absorbance (a) as the ordinate, to obtain a standard curve: Y=5.9529X+0.0189 (r=0.9999), with good linearity.

[0052] Then, transfer 0.5 mL of the liquorice residue extract solution to a 10 mL volumetric flask and dilute to the mark with 50% ethanol aqueous solution. Accurately measure 2 mL of the mixed solution and determine the total flavonoid absorbance of the liquorice residue extract solution using the above method. Calculate the extraction rate of total flavonoids in the extract according to the following formula:

[0053] Extraction rate (%) = C × V × D / M × 100%

[0054] Where: C: mass concentration of sample solution after volume is fixed (mg / mL); V: total volume of solution to be tested (mL); D: dilution factor; M: mass of licorice residue (g).

[0055] b. Detection method of saponin content: 10.5 mg of oleanic acid standard was accurately weighed into a 50 mL container, dissolved in methanol, and diluted to volume to obtain a concentration of 0.21 mg mL -1 5 ml of 0.21 mg·mL -1 Transfer the oleanic acid solution to a 10 ml volumetric flask and dilute to the mark with methanol to obtain a concentration of 0.105 mg mL -1 Oleanolic acid standard solution. 0.105 mg mL -10, 0.2, 0.4, 0.6, 0.8, 1.0mL of oleanic acid standard solution was accurately transferred into a 10mL volumetric flask, evaporated to dryness in an 80°C water bath, and then 0.2mL of 5% vanillin-acetic acid solution and 0.8mL of perchloric acid were added in sequence. Heat in a 70°C water bath for 15 minutes, take out, cool in ice water for 5 minutes, add 5ml of acetic acid, shake well, and let stand for 5 minutes. Measure absorbance at a wavelength of 548nm, with concentration as the horizontal coordinate and absorbance as the vertical coordinate, and obtain the regression equation: Y=39.687X+0.0048 (r=0.9995), with a good linear relationship.

[0056] Take 2.0 mL of the liquorice residue extract, transfer it to a separate 10 mL container, evaporate it to dryness in a water bath at 80°C, and measure the total saponin absorbance of the liquorice residue extract using the above method. Calculate the extraction rate of total saponins in the extract according to the following formula.

[0057] Extraction rate (%) = C × V × D / M × 100%

[0058] Where: C: mass concentration of sample solution after volume is fixed (mg / mL); V: total volume of solution to be tested (mL); D: dilution factor; M: mass of licorice residue (g).

[0059] c. Polysaccharide content detection method: accurately weigh 10 mg of anhydrous d-glucose standard, put it into a 100 mL volumetric flask, and dilute it to volume with distilled water to obtain a concentration of 0.1 mg mL -1 Glucose standard solution. Accurately pipette 0, 0.2, 0.4, 0.6, 0.8 and 1.0 mL of the above solution into separate 10 ml volumetric flasks. Add 1 ml of 5% phenol to each flask, mix well, and then quickly add 5 ml of concentrated sulfuric acid. Mix well and allow the reaction to proceed in an 80°C water bath for 15 minutes. Remove the flask, cool it in cold water, and measure the absorbance at a wavelength of 487 nm. Draw a standard curve with absorbance as the ordinate and concentration as the abscissa. The regression equation obtained is Y=5.9529X+0.0189 (r=0.9998), and the linear relationship is good.

[0060] Take 20 μL of the liquorice residue extract and add 380 μL of distilled water. Determine the absorbance of the total polysaccharide in the liquorice residue extract solution using the above method. Calculate the extraction rate of the total polysaccharide in the extract according to the following formula.

[0061] Extraction rate (%) = C × V × D / M × 100%

[0062] Where: C: mass concentration of sample solution after volume is fixed (mg / mL); V: total volume of solution to be tested (mL); D: dilution factor; M: mass of licorice residue (g).

[0063] 1.1. Preparation of DESs

[0064] According to the combination in Table 1, the DESs were synthesized by the stirring-heating method. Specifically, the hydrogen bond acceptor and the hydrogen bond donor were heated and stirred at 80°C simultaneously until a stable, uniform, transparent liquid was formed. Thereafter, pure water was introduced at a volume ratio of 30% to generate an extract, which was then retained for further application.

[0065] Table 1 Synthesis and molar ratio of DESs

[0066]

[0067]

[0068] 1.2. Ultrasound-assisted DESs extraction method

[0069] The same batch of liquorice root residue was crushed and sieved with a 60-mesh sieve, and the obtained powder was sealed for later use. Ultrasonic extraction was performed using an ultrasonic cell disruptor (Scientz-IID, power: 600 W; Ningbo Xinzhi Biotechnology Co., Ltd., China).

[0070] The extraction process is as follows: 3g of licorice residue powder is taken and mixed with 30mL of pre-prepared DESA solution at a solid-liquid ratio of 1:10. The pulse sequence interval is set to 2s ON and 2s OFF, and the probe tip is immersed in the solution for 1cm. The ultrasonic power is set to 100W, and the extraction is carried out at room temperature for 10 minutes. The other parameters of the ultrasonic device are kept constant, and the random amplitude change is set to φ6. After extraction, the solution is transferred to a 50mL centrifuge tube and centrifuged at 4500×g for 20 minutes using a TD5 multi-tube balanced centrifuge. The supernatant is collected, sealed, and stored at 4°C.

[0071] 1.3. Entropy weight TOPSIS method to determine the entropy weight of each active ingredient

[0072] The entropy-weighted TOPSIS mathematical model mainly includes establishing the initial decision matrix, normalizing the decision matrix, calculating the entropy weight of the evaluation index, constructing a weighted decision matrix, determining the optimal vector and the worst vector of the weighted decision matrix, calculating the distance between each evaluation index and the positive and negative ideal solutions, and finally obtaining the evaluation results based on the Euclidean closeness of each evaluation index to the optimal solution.

[0073] 1.3.1. Establishment of the original data matrix

[0074] Excel was used to establish the original data matrix of different varieties, where Y1 represented the flavonoid extraction rate, Y2 represented the saponin extraction rate, and Y3 represented the polysaccharide extraction rate to construct the original data matrix (1).

[0075] Y=(A ij )m×n Formula (1)

[0076] Where A is the evaluation value of the i-th experiment under the j-th indicator, m is the number of experiments (m=29), and n is the evaluation indicator (n=4).

[0077] 1.3.2. Establishment of standardized data matrix

[0078] In the evaluation criteria of this study, the four indicators Y1-flavonoid extraction rate, Y2-saponin extraction rate, and Y3-polysaccharide extraction rate are all positive indicators. Referring to the standardized formula (2), the original data matrix (1) is converted into a standardized data matrix (4).

[0079] X ij =(A ij -A ij min ) / (A ij max -A ij min ) Formula (2)

[0080] X ij =(A ij max -A ij ) / (A ij max -A ij min ) Formula (3)

[0081] B=(X ij ) m×n Formula (4)

[0082] 1.3.3. Establishment of the Standardized Data Probability Matrix

[0083] Referring to transformation formula (5), the standardized data matrix (4) of each indicator is transformed into a standardized data probability matrix (6):

[0084]

[0085] P=(P ij ) m×n Formula (6)

[0086] Where P ij represents the probability of the i-th item appearing under the j-th indicator, and 0≤P ij ≤1.

[0087] 1.3.4. Calculate the information entropy (Hj), coefficient of variation (gj), and entropy weight coefficient (Wj) of each indicator

[0088] The information entropy (Hj) of each indicator is calculated according to different types of standardized data probability matrices (6) and formula (7), and the coefficient of variation (gj) and entropy weight coefficient (Wj) are calculated according to formulas (8) and (9).

[0089]

[0090] g j =1-H j Formula (8)

[0091]

[0092] Where: 0≤W j ≤1

[0093] 1.3.5. Establishment of weighting matrix

[0094] Each item in the standardized data matrix (3) of different experiments is multiplied by the corresponding entropy weight coefficient to obtain the weighted matrix (10) of each indicator.

[0095] Z=(W j ×X ij ) m×n Formula (10)

[0096] 1.3.6.Positive deviation squared Negative squared deviation Euclidean distance of the optimal solution Euclidean distance of the worst solution Relative proximity (C i )

[0097] According to the weighted matrix (10) of different experiments for each indicator and formulas (11), (12), (13), (14), (15), the square of the positive deviation of the optimal solution is calculated Negative squared deviation Euclidean distance to the optimal solution Euclidean distance of the worst solution and relative proximity (C i ).

[0098]

[0099]

[0100]

[0101]

[0102]

[0103] Finally, the entropy weight TOPSIS method was used to evaluate the four indicators under different experimental conditions using Excel.

[0104] 1.4. Customized design of artificial neural network

[0105] This model is implemented by defining a class that inherits from "torch.nn.Module". The specific formula of the model can be decomposed into several parts:

[0106] 1.4.1. Network layer

[0107] The model consists of three fully connected layers: the first fully connected layer (fc1) has 4 input features (A: time (min), B: power (W), C: water content (%), D: liquid-to-solid ratio) and 128 output features. The second fully connected layer (fc2) has 128 input features and 64 output features. The third fully connected layer (fc3) has 64 input features and 3 output features, corresponding to 4 target variables: Y1-flavonoid extraction rate (mg / g), Y2-saponin extraction rate (mg / g), Y3-polysaccharide extraction rate (mg / g).

[0108] Activation Function

[0109] After each layer is fully connected except the last layer, the ReLU (Rectified Linear Unit) activation function is applied. The ReLU function is defined as f(x) = max(0, x).

[0110] 1.4.3. Forward propagation of the model

[0111] During the forward propagation, the input data goes through the following steps: First, the input data goes through the first fully connected layer, followed by the ReLU activation function. Next, the output data is passed to the second fully connected layer, followed by the ReLU activation function. Finally, the data passes through the third fully connected layer without an activation function, effectively making it a linear transformation. In formula notation, the forward propagation of the model is represented as:

[0112] Output=fc3(ReLU(fc2(ReLU(fc1))input)))))

[0113] Where each "fc" represents a linear transformation of a fully connected layer, that is, fc(x)=Wx+b (W is the weight matrix, b is the bias term), and ReLU(x) is the activation function.

[0114] 1.4.4. Loss Function

[0115] The loss function of the model is Mean Squared Error, which calculates the average squared difference between the predicted value and the true value. The formula is:

[0116]

[0117] Where y i is the true value, is the predicted value, and n is the sample size.

[0118] 1.4.5. Optimizer

[0119] The model uses the Adam optimizer, a first-order gradient-based optimization algorithm, to update the network weights during training to minimize the loss function. The Adam optimizer combines the advantages of momentum and RMSprop, allowing automatic adjustment of the learning rate.

[0120] In order to obtain the optimal experimental conditions, the maximum time value was set to 60 minutes with an interval of 5 minutes; the maximum power was set to 600w with an interval of 50w; the maximum water content value was set to 60% with an interval of 5%; and the maximum liquid-to-solid ratio was set to 60 with an interval of 5. These conditions were then input into the neural network to predict the flavonoid extraction rate (Y1), saponin extraction rate (Y2), and polysaccharide extraction rate (Y3) under different conditions. The comprehensive score was then calculated based on the predicted extraction rate.

[0121] 1.5. Recovery and purification of DES

[0122] The extracted DES was centrifuged to remove the precipitate, and ethanol 8 times the volume of DES was added to precipitate at 4°C for 24 hours, and then the precipitate was centrifuged to remove. Subsequently, water 10 times the volume of DES and macroporous adsorption resin 1 / 4 the volume of DES were added to the solution, the mixture was stirred, adsorbed at room temperature for 4 hours, and then incubated at 4°C for 24 hours, and the precipitate was removed by centrifugation to obtain the recovered DES.

[0123] 1.6. Extraction mechanism

[0124] 1.6.1SEM inspection

[0125] The unextracted liquorice residue powder and the liquorice residue powder extracted with ethanol and DES solvents were ground and dried in a freeze dryer for 24 hours. The microscopic morphology of the dried raw materials and the extracts obtained by different extraction methods was analyzed using SEM at 500 and 1000 times magnification (S4800, HITACHI Japan).

[0126] Molecular dynamics simulation

[0127] Classical molecular dynamics (MD) simulations were performed using Gromacs 2020.6, and the number of molecules in each box was determined by the molar ratios of the individual systems, which were housed in a 60 × 60 × 60 periodic box. The ET-Li system consists of 41,300 ethanol molecules, 88,900 water molecules and 6 glycyrrhizin molecules; the DES-Li system consists of 4,800 choline chloride molecules, 14,500 glycerol molecules, 111,111 water molecules and 6 glycyrrhizin molecules; the ET-GA system contains 41,300 ethanol molecules, 88,900 water molecules and 38 glycyrrhizic acid molecules; the DES-GA system includes 4,800 choline chloride molecules, 14,500 glycerol molecules, 111,111 water molecules and 38 glycyrrhizic acid molecules; the ET-GL system contains 41,300 ethanol molecules, 88,900 water molecules, 84 glucose molecules; the DES-GL system includes 4,800 choline chloride molecules, 14,500 glycerol molecules, 111,111 water molecules and 84 glucose molecules. During the MD simulation, all bonds involving hydrogen atoms were constrained using the LINCS algorithm, electrostatic interactions were calculated using the PME method, and the Fourier spacing and cutoff for non-bonded interactions were set to 0.16 and 1 nm, respectively. The simulation temperature was controlled at 298.15 K using the V-rescale temperature coupling method, and the pressure was set to 1 bar using the Parrinello-Rahman method. The entire simulation process included energy minimization using the conjugate gradient method, 10 ns NPT equilibration, and 50 ns NPT production runs.

[0128] 1.6.3. Quantum chemistry research

[0129] The molecular configurations of DES components, licorice, glycyrrhizic acid and glucose were optimized by self-consistent field molecular orbital method based on DFT. The single point energy, frequency and geometric configuration of M06-2X / 6-311G**-grade DES were optimized by Gaussian16 software, and the dynamic stable structure with the lowest energy was obtained. The complex was optimized by Kohn-Sham wave function, which described the interaction between molecular structure and compound in DES.

[0130] The electrostatic potential (ESP) and reduced density gradient (RDG) were calculated by Multiwfn and VMD software, respectively, to analyze the interactions between DES molecules and between solvent and compound. RDG is a basic dimensionless quantity that describes the deviation from uniform electron distribution. The relationship between electron density (ρ) and RDG was plotted according to the equation, where ρ(r) is the critical point of weak interaction and represents the bonding strength. In actual analysis, the RDG value reflects the location of the interaction; but the ρ and RDG values ​​in the non-covalent interaction region are too low to identify the type of interaction. Analyzing the symbolic λ2 eigenvalue of the electron density Hessian (second-order derivative) matrix helps to distinguish different types of non-covalent interactions. Using the product of ρ and λ2 as a variant to reflect the RDG isosurface in different colors helps to visualize the location, strength and type of weak non-covalent interactions. The RDG value is calculated according to the formula:

[0131]

[0132] In the formula, ρ is the quantum mechanical electron density, r is the coordinate vector of a field, and π is the circumference ratio constant. 2 ) 1 / 3 ) is a constant, and the 4 / 3 power of the density ensures that ρ is a dimensionless quantity.

[0133] Statistical analysis

[0134] IBM SPSS 24.0 statistical software (IBM, USA) was used to calculate standard deviations, and GraphPad Prism 8 and Origin 2018 software were used to create graphs.

[0135] 2. Results and Discussion

[0136] 2.1.DES selection

[0137] In this study, choline chloride was used as a hydrogen bond acceptor and three types of hydrogen bond donors were selected, including alcohol, acid and sugar. Ten DES systems (hydrogen bond acceptor to hydrogen bond donor molar ratio of 1:3) were established as candidate systems for extraction solvents. Using 60% ethanol as a control, ultrasonic extraction was carried out under the conditions of ultrasonic power 100W, DES water content 30%, liquid-solid ratio 1:10, and extraction time 10min. Taking the extraction rate of total flavonoids, saponins and polysaccharides of licorice as an indicator, the DES system with the highest extraction rate was selected.

[0138] like Figure 1-(a), (b), (c) show that the extraction rates of total flavonoids, saponins, and polysaccharides from licorice residues in different solvents were 0.735-3.968 mg / g, 3.291-21.592 mg / g, and 2.458-40.076 mg / g, respectively. Among the 10 solvents, the type of DES had a significant effect on the extraction rate of licorice residues. Among them, DES-1 to DES-4 used alcohol hydrogen bond donors to significantly improve the extraction rate of effective components of licorice residues, and the extraction rates of flavonoids, saponins, and polysaccharides were significantly higher than those of the 60% ethanol control. DES-5 to DES-7, which were acidic hydrogen bond donors, had the least effect on the extraction rate, which was equivalent to the 60% ethanol control. The extraction rates of flavonoids and saponins of DES-8 to DES-10 were higher than those of 60% ethanol, but the increase was not as high as that of alcohol donors. Licorice polysaccharides are sensitive to alcohols and acidic hydrogen bond donors, but not to sugar hydrogen bond donors.

[0139] Most DES have high viscosity at room temperature, which has a dual effect on the extraction efficiency. High-viscosity DES tends to have more hydrogen-bonded network structures, which can improve the extraction rate; however, high viscosity also reduces the mobility of free molecules inside the DES, resulting in low extraction efficiency. If the viscosity is too low, the hydrogen-bonded network in the DES system will be reduced, which will also lead to a decrease in extraction efficiency. The viscosity of DES is affected by the molar ratio of hydrogen-bond acceptors and hydrogen-bond donors and the water content. The molar ratio of hydrogen-bond acceptors to hydrogen-bond donors in the DES system used in current research is mostly between 1:1 and 1:5. Changes in the molar ratio of hydrogen-bond acceptors to hydrogen-bond donors will affect the viscosity of the DES system, thereby changing its extraction rate. Previous studies have shown that when the molar ratio of hydrogen-bond acceptors to hydrogen-bond donors is too small, the reaction cannot proceed, and excessive hydrogen-bond donors may aggravate the internal interference of the system, reduce the interaction ability of the target chemical with chloride anions, and lead to a decrease in extraction rate. Therefore, this study selected a suitable ratio of DES system (choline chloride: glycerol = 1:3) to maximize the extraction rate ratio. According to the above results, DES-4 choline chloride: glycerol = 1:3 had the greatest impact on the extraction rate and was used as the extraction solvent in subsequent experiments.

[0140] Optimization of extraction conditions

[0141] 2.2.1. Single-factor experimental design

[0142] Factors such as solvent type, solid-liquid ratio, solvent water content, ultrasonic power, and extraction time have a significant effect on the yield of various active ingredients in licorice residue. Therefore, this single-factor extraction experiment designed four single-factor variables: ultrasonic power 100-500W, DES water content 20%-60%, liquid-solid ratio 1:10-1:50, and extraction time 10-50 minutes. When estimating the effect of one variable on yield, the other variables remain unchanged. The specific experimental design is shown in Table 2.

[0143] Table 2 Single factor experimental design

[0144]

[0145] 2.2.2. Single factor experimental results

[0146] 2.2.2.1. Effect of extraction time on the extraction rate of active ingredients from licorice residue

[0147] Figure 1 -(d) shows the trends observed during the experiment. It is noteworthy that the extraction rate of flavonoids increased significantly within 40 minutes, while saponins reached their peak extraction rate at 20 minutes. The polysaccharide extraction rate did not change significantly with the extraction time, showing a trend of first increasing and then decreasing. There are many reasons for this trend, such as the increase in extraction temperature and long-term exposure to air, which will cause some components to degrade and affect the change in extraction rate.

[0148] 2.2.2.2. Effect of ultrasonic power on the extraction rate of active ingredients from licorice residue

[0149] like Figure 1 -(e), the observation results show that the extraction rate of glycyrrhizin gradually increases with the increase of ultrasonic power. This is due to the amplification of sound intensity and the simultaneous increase of negative pressure in the liquid medium, which makes the cavitation bubble expand to the maximum, resulting in an increase in cavitation events per unit time under ultrasonic exposure, thereby improving the extraction efficiency of natural medicinal plant components. However, flavonoids and polysaccharides show a biphasic trend of first strengthening and then weakening, which may be due to the excessive ultrasonic power causing the molecular structure of some polysaccharides and flavonoids to be damaged, resulting in their degradation into simpler components, followed by a decrease in extraction rate.

[0150] 2.2.2.3. Effect of moisture content on the extraction rate of active ingredients from licorice residue

[0151] like Figure 1 -(f), all three active ingredients exhibited a biphasic response to the increase in water content, reaching maximum extraction yield at a water content of 50%. Statistical analysis showed significant differences in the extraction yields of flavonoids, saponins, and polysaccharides between 10% and 50% water content. The observed trends may be attributed to the effect of water content on the viscosity of the deep eutectic solvent (DES), where insufficient water content leads to increased viscosity, hindering the uniform distribution of licorice residue within the DES and hindering the effective contact of ultrasound-induced shear forces with the optimal cell wall, thereby limiting the extraction efficiency. Conversely, excessive water content leads to a decrease in DES concentration, thereby reducing the concentration of hydrogen bond donors and acceptors, thereby reducing the extraction yield.

[0152] 2.2.2.4. Effect of solid-liquid ratio on the extraction rate of active ingredients from licorice residue

[0153] like Figure 1 -(g) shows that under different liquid-to-solid ratios, the fluctuations of licorice flavonoids and saponins were relatively mild, while polysaccharides showed a slight biphasic trend. The increase in solvent volume promoted the dissolution of polysaccharides, thereby improving the total extraction rate, and the maximum extraction efficiency was obtained at a ratio of 1:40. However, too high a liquid-to-solid ratio would cause excessive dissolution of impurities, thus affecting the extraction rate.

[0154] Response surface optimization

[0155] 2.2.2.1 Response surface design

[0156] Response surface methodology (RSM) is a commonly used experimental design method in natural active products. It requires relatively few design points and experiments to generate a comprehensive data set for analysis. Compared with other designs, this design is more advantageous both experimentally and economically, has the advantages of high efficiency and low cost, and is widely used in the optimization of natural product extraction processes. Therefore, this study also used the RSM method to optimize the extraction of active ingredients from licorice residue. The Box-Behnken combination design was used to optimize the best comprehensive score determined under different factor conditions. Time, power, water content, and liquid-to-solid ratio were selected as the main influencing factors, and the comprehensive score was used as the response value. The process was optimized using the RSM method. The Box-Behnken experimental design is shown in Table 3. A Box-Behnken combination design with a 4-factor 3-level response surface experiment was used, and 29 experiments were carried out according to the process parameters in Table 3. The experimental design is shown in Table 4. According to the experimental design, the extraction rates and scores of flavonoids (Y1), saponins (Y2), and polysaccharides (Y3) under different experimental conditions were obtained, and the numerical predictions and scores obtained are shown in Table 4. Figure 2 -(g) as shown.

[0157] Table 3 Factors and levels in RSM experimental design

[0158]

[0159] Table 4 Experimental design of RSM comprehensive score

[0160]

[0161]

[0162] 2.2.2.2 Establishment and analysis of comprehensive scoring model

[0163] Taking time (A), power (B), water content (C), liquid-to-solid ratio (D) as representative variables and comprehensive score as response value, multiple regression fitting was performed. This analysis resulted in a quadratic polynomial regression equation:

[0164] Y 综合评价=89.92+5.61*A-2.63*B+5.11*C-3.37*D+11.68*AB-5.47*C+11.59*AD+5.95*BC+2.66*BD-0.3*CD-22.49*A 2 -23.25*B 2 -22.24*C 2 -21.35*D 2

[0165] Table 5 shows that the model is significantly different (P<0.001), and the coefficient of determination R2 is 0.9768, indicating that the model fits the experimental data well. There is no significant lack of fit (P>0.05), emphasizing the accuracy of the model. The adjusted coefficient of determination RAdj2 is 0.9536, confirming the effectiveness of the model in theoretical analysis and predictive ability.

[0166] The f value is a key indicator for evaluating the impact of each variable on the response value. The higher the f value, the greater the contribution of the model component to the response. When the significance test probability (P<0.05), it means that the variable has a significant impact on the response value. The results of testing the impact of each factor on the comprehensive score show that the first-order terms of time and water content have extremely significant effects on the comprehensive score (P<0.01), and the effects of power and liquid-to-solid ratio on the comprehensive score are significant (P<0.05). Among the influencing factors, the degree of influence on the comprehensive score is A>C>D>B, time>water content>liquid-to-solid ratio>power. The second-order term A 2 , B 2 , C 2 , D 2 The effect on the comprehensive score was extremely significant (P<0.01), indicating its nonlinear effect. The interaction terms AB and AD had extremely significant effects on the comprehensive score (P<0.01), and AC and BC had significant effects on the comprehensive score (P<0.05).

[0167] Table 5 Regression analysis results are comprehensive score model and regression coefficients

[0168]

[0169] Note: * P<0.05, ** P<0.01, ns means P>0.05, no statistical significance.

[0170] Design-Expert 10.0.3 software was used to generate response surface plots to visualize the experimental results. In the 3D graphs and contour plots, the color gradient from blue to red indicates the change from low to high response values, and a steeper slope indicates a more obvious impact on the experimental results. These three-dimensional response surface plots intuitively describe the effect of the interaction between factors on the comprehensive score, making it easier to determine the optimal process parameters and clarify the interaction between parameters. The response surface analysis results are shown in Figure 2. Figure 2 shown.

[0171] AB interaction diagram ( Figure 2 -a) Depicts the interaction surface between Factor A and Factor B. An increase in time initially increases the composite score, which then decreases. In contrast, at lower intervals, an increase in power leads to a gradual increase in the score, which then decreases significantly. However, with longer duration, an increase in power initially improves the composite score, which then decreases slightly. This highlights the significant interaction between time and power, with higher composite scores observed at power levels of 250-350 W and intervals of 25-35 min.

[0172] AC interaction diagram ( Figure 2 -b) Depicts the interaction surface between factors A and C. With increasing water content, the composite score increases and then decreases. At lower water content levels, time increments initially increase the score, which eventually levels off. In contrast, elevated water content coupled with extended durations results in an initial surge in scores followed by a decrease. This indicates a significant interaction between time and water content, with higher composite scores observed when water content was between 45%-55% and the time span was 25-35 minutes.

[0173] AD interaction diagram ( Figure 2 -c) The interaction surface between factors A and D is depicted. The time variation also shows a similar trend, with an initial score that rises first and then decreases. At shorter time intervals, higher liquid-to-cooking ratios caused the scores to level off and then decrease. In contrast, increasing time and liquid-to-cooking ratios initially caused the scores to rise, followed by a slight decrease. This highlights a significant interaction between time and liquid-to-cooking ratio, with higher composite scores observed in the liquid-to-cooking ratio range of 35-45 and time intervals of 25-35 minutes.

[0174] BC interaction diagram ( Figure 2-d) shows the interaction surface between Factor B and Factor C. Here, an increase in water content first increases the composite score and then decreases. Similarly, an increase in power level also shows an initial increase in the score followed by a decrease. In addition, the difference in composite score with increasing power also differs by water content, indicating a significant interaction between water content and power, with power levels between 250-350W and water contents between 45%-55% resulting in higher composite scores.

[0175] BD interaction diagram ( Figure 2 -e) shows the interaction surface between factor B and factor D. It can be seen that with increasing liquid-to-solid ratio and power, the comprehensive score first increases and then decreases. A higher comprehensive score is observed between the power level of 250-350W and the liquid-to-solid ratio of 35-45.

[0176] CD interaction diagram ( Figure 2 -f) depicts the interaction surface between factors C and D. With the increase of water content and liquid-to-solid ratio, the comprehensive score first increases and then decreases. The higher the comprehensive score, the higher the score is observed when the liquid-to-solid ratio is between 35-45 and the water content is between 45%-55%.

[0177] 2.2.3. Optimization of artificial neural network extraction process

[0178] The network architecture used in this experiment is Figure 7 As shown in the figure, it includes input layer, output layer and hidden layer. Each node in the input layer corresponds to a different feature, while the nodes in the output layer represent the target data, which can be multiple. The hidden layer is located between the input layer and the output layer, and the complexity of the neural network is determined by the number of layers and nodes in each layer.

[0179] After experiments, the features of the input layer include time, power, moisture content and liquid-to-solid ratio, while the output layer data include Y1-flavonoid extraction rate, Y2-saponin extraction rate and Y3-polysaccharide extraction rate. The hidden layer includes two layers. In order to facilitate the learning of neural networks and perform nonlinear complex tasks, the rectified linear unit (ReLU) activation function is integrated in the middle of each layer. This function invalidates negative values ​​while retaining positive values. In order to evaluate the predictive performance of the model, the mean square error (MSE) is used as the loss function to calculate the average square difference between the predicted value and the actual value.

[0180] To determine the optimal experimental conditions, parameters such as extraction time, power, water content, and liquid-to-solid ratio were input into the neural network to predict the extraction rates of flavonoids (Y1), saponins (Y2), and polysaccharides (Y3) under different conditions. The composite scores were calculated based on the predicted extraction rates, and the numerical predictions and scores were obtained as shown in Figure 2. Figure 2 -(h) shown.

[0181] 2.2.4. Result evaluation based on entropy weight-TOPSIS method

[0182] The quality evaluation results of different response surface experiments are shown in Table S6. The closer the relative closeness (Ci) is to 1, the better the experimental results are. Each item of the standardized data matrix of the 29 response surface experimental data is multiplied by the corresponding entropy weight coefficient to establish the weighting matrix of each indicator. Referring to the weighting matrix of each indicator in each experimental group, the positive deviation square (η i + ) 2 and negative squared deviation (η i - ) 2 Then, determine the Euclidean distance (D i + ), the Euclidean distance of the worst solution (D i - ) and the relative proximity of different experimental groups (C i ). The results of the Euclidean distance of the optimal solution, the Euclidean distance of the worst solution, and the relative proximity are detailed in Table 6. According to the calculated relative proximity, the experimental group with the most favorable results is determined as the center point experimental group.

[0183] Table 6 Quality evaluation results of different response surface experiments

[0184]

[0185]

[0186] The entropy TOPSIS method is a comprehensive evaluation method that integrates the entropy method and the TOPSIS method. It is an objective value assignment method that can reduce the subjective bias of value assignment. In this study, the quality information entropy (Hj) of different experimental groups reflects the degree of dispersion between indicators. A higher entropy means a lower degree of dispersion between indicators, resulting in a lower decision weight assigned to them. After normalizing the indicators of each experimental group, the information entropy (Hj) of each indicator is calculated using formula (7). Subsequently, based on the information entropy (Hj), the coefficient of variation (gj) and the entropy weight coefficient (Wj) are derived using formulas (8) and (9). The calculation results of the information entropy (Hj), coefficient of variation (gj) and entropy weight coefficient (Wj) of each indicator in each experimental group are shown in Table 7. Checking the calculation results in Table 7, it can be found that the indicator weights of each experimental group are as follows: saponin extraction rate > polysaccharide extraction rate > flavonoid extraction rate.

[0187] Table 7 Calculation results of information entropy, difference coefficient and entropy weight coefficient of four indicators in different experiments

[0188]

[0189]

[0190] The comprehensive score of flavonoid extraction rate, saponin extraction rate and polysaccharide extraction rate was used as the reference index, and the following formula was used:

[0191] Y=100*W1*(Y1-Y 1min ) / (Y 1max -Y 1min )+100*w2*(Y2-Y 2min ) / (Y 2max -Y 2min )

[0192] +100*W3*(Y3-Y 3min ) / (Y 3max -Y 3min )

[0193] In the formula, Y1-Y3 represent the extraction rate of flavonoids, saponin and polysaccharide, respectively. 1max -Y 3max Represents the maximum value of the four indicators.

[0194] 2.3. Optimal process and verification experimental results

[0195] In RSM, we used Design-Expert 10.0.3 experimental optimization software, and the optimization goal was to maximize the score. Through prediction, we obtained a predicted comprehensive score of 90.564, and the predicted values ​​of the four factors were time 30.919min, power 297.660W, water content 51.003%, and liquid-solid ratio 39.437. In the ANN prediction, the highest score can reach 92.534 points. At this time, the extraction conditions are 30min, 300W power, 50% water content, and 40 liquid-solid ratio, which is basically consistent with the RSM prediction results. But what is more noteworthy is that ANN can also predict the extraction rate of each component. In the prediction results, the extraction rate of flavonoids can reach 8.018mg / kg, saponins 42.958mg / kg, and polysaccharides 455.534mg / kg. To verify the accuracy of the two models, repeated experiments were carried out under the above three conditions.

[0196] like Figure 3-(e) shows that the experimental results show that the extraction rate of total flavonoids of Y1 is 8.840±0.412mg / g, the extraction rate of saponins of Y2 is 41.680±0.776mg / g, the extraction rate of polysaccharides of Y3 is 441.997±11.630mg / g, and the comprehensive average score is 92.863±1.213. The deviation of the result from the ANN model 92.534 is 0.36%, and the deviation from the RSM model 90.564 is 2.54%. The deviation between the predicted value and the actual value of the two models is within 5.00%, which confirms that the process parameters optimized by the two models have strong reliability.

[0197] After comparison, it was noted that ANNs provided a wider range of predictions for extraction conditions compared to response surface methodology. They provided comprehensive predictions and analyses based on the existing conditions, predicting not only the extraction conditions with the highest comprehensive scores, but also the extraction yield and maximum value of each component under each condition. Therefore, ANNs are able to select the most appropriate extraction conditions for each component and select suitable extraction conditions for different extraction needs and preferences. Compared with ANNs, RSMs can only make optimization predictions for a single indicator, but cannot predict the specific value of a single component, so they are limited in extraction applications. However, compared with ANNs, RSMs are much simpler in design and operation. Therefore, both methods are effective prediction methods with important reference value and application prospects in natural product and industrial research. On the other hand, ANNs have a wider range of applications and more accurate predictions than RSMs.

[0198] 2.3.1. Antioxidant activity verification

[0199] Studies have shown that DES can efficiently extract natural products. Therefore, due to the increase in chemical components, the biological activity of DES extracts is much greater than that of normal extracts. Therefore, this study compared the antioxidant activity of traditional licorice ethanol extracts with that of DES extracts to explore the effect of DES on the biological activity of licorice ethanol extracts. 2- The antioxidant capacity was evaluated by its scavenging ability of ·), hydroxyl radical (·OH) and ABTS free radical.

[0200] 2.3.1.1 DPPH free radical scavenging activity

[0201] DPPH free radical scavenging activity was determined using a DPPH detection kit (Shanghai ELISA Biotechnology Co., Ltd., China). Briefly, 20 μL of the extract was mixed with 380 μL of methanol-DPPH working solution according to the instructions. Incubated at room temperature in the dark for 20 min, and the absorbance was measured at 515 nm in a 96-well plate.

[0202] The calculation formula of DPPH free radical scavenging ability is as follows:

[0203] Clearance rate (%) = (A0 – A i ) / A0×100%

[0204] Where A0 is the absorbance of DPPH, A i is the absorbance of DPPH+ sample.

[0205] 2.3.1.2 Superoxide anion (O 2- ·) Scavenging activity

[0206] The superoxide anion scavenging activity was determined using a superoxide anion content detection kit (Shanghai ELISA Biotechnology Co., Ltd., China). Simply follow the instructions to mix 200 μL of the extract with 160 μL of hydroxylamine hydrochloride. Incubate the mixture in a 37 °C water bath for 20 min. Then add 120 μL of p-aminobenzenesulfonic acid and 120 μL of α-naphthylamine, mix, and incubate in a 37 °C water bath for 20 min. Finally, add 200 μL of chloroform, mix, and centrifuge at 8000 × g for 5 min at 25 °C. Carefully transfer 200 μL of the upper aqueous phase to a 96-well plate and measure the absorbance at 530 nm. The free radical scavenging activity was calculated in the same way as for DPPH.

[0207] 2.3.1.3 Activity in scavenging hydroxyl radicals (·OH)

[0208] The hydroxyl radical scavenging activity was determined using a hydroxyl radical detection kit (Shanghai ELISA Biotechnology Co., Ltd., China). Simply mix 60 μL of the extract with 60 μL of acetic acid and 120 μL of 2-deoxyribose according to the instructions. Incubate the mixture at 35 °C in the dark for 1 h. Then add 240 μL of tert-butyl alcohol and mix, and incubate in a 95 °C water bath for 10 min. After cooling to 25 °C on ice, centrifuge at 8000 × g for 3 min, take 200 μL of the supernatant and transfer to a 96-well plate, and measure the absorbance at 532 nm.

[0209] The free radical scavenging activity was calculated as follows:

[0210] Clearance rate (%) = (A s –A n ) / (A b –A n )×100%

[0211] is the absorbance of each mixed solution; A b Indicates the absorbance of the sample, replaced by distilled water; A n Represents the absorbance of distilled water.

[0212] 2.3.1.4 Free radical scavenging activity

[0213] ABTS free radical scavenging activity was determined using a total antioxidant capacity assay kit (Shanghai ELISA Biotechnology Co., Ltd., China). Briefly, 10 μL of the extract was mixed with 190 μL of ABTS and MnO2 mixed working solution according to the instructions. This was repeated four times in parallel, and 200 μL was taken within 10 min and measured at 734 nm in a 96-well plate. Subsequently, the free radical scavenging activity was calculated using the same method as for DPPH.

[0214] Depend on Figure 3 -(a) It can be seen that under the same conditions, the scavenging rate of DES extract on DPPH was 93.02%, which was significantly higher than 79.45% of ethanol extract. Similarly, the antioxidant activity evaluation based on superoxide anion scavenging rate, such as Figure 3 -(b) shows that the scavenging rate of DES extract was 80.95%, which was statistically significantly higher than that of 60% ethanol extract (69.95%). Figure 3 -(c) shows that the hydroxyl radical scavenging rate of the DES extract was 83.04%, which was slightly higher than that of the 60% ethanol extract (70.98%). Figure 3 -(d), the scavenging rate of ABTS free radicals by ultrasound-assisted DES extract was 73.22%, which was significantly higher than that of ethanol extract (60.47%). This indicates that ultrasound-assisted DES is a feasible method for extracting effective components from licorice residue without affecting the biological activity of the extract.

[0215] 2.4. Results of DES recovery, purification and reuse

[0216] In recycling, we first remove the extracted liquid by centrifugation, add ethanol, and separate the purified polysaccharides to precipitate them for subsequent use. The mixture is placed at low temperature, which will reduce the solubility of DES and ensure that the precipitate is completely precipitated. Then, water and macroporous adsorption resin are added to the solution, and the mixture is stirred at room temperature and adsorbed for 4 hours. The macroporous adsorption resin can adsorb the flavonoids and saponins in the DES solution, and then it is allowed to stand at low temperature and remove excess impurities by centrifugation. The recovered DES is obtained by removing water by rotary evaporation. From Figure 3 -(e) The results show that under the same conditions, the extraction yields of the first and second extractions with DES solvent were significantly higher than those with ethanol, and the difference was statistically significant. When polysaccharides were extracted from the recovered DES, the extraction yield of the second extraction was slightly lower than that of the first extraction. However, the extraction efficiency of DES for glycyrrhizic flavonoids and saponins was basically unaffected, which highlights the potential of DES as a new environmentally friendly solvent that can be reused through a series of purification methods. Therefore, DES is a reusable and environmentally sustainable product.

[0217] 2.5. Exploration of extraction mechanism

[0218] Microstructure analysis

[0219] Studies have shown that ultrasound can improve the extraction rate of natural products through cavitation, mechanical and thermal effects during the extraction process. The extraction process and mechanism were studied in depth. DES, we conducted SEM analysis to examine the microstructure of unextracted licorice residue, ethanol extraction and DES solvent extraction powder, such as Figure 4 -(a-c') shown. Figure 4 -(a) and (a') show that the surface of the unextracted licorice residue powder appears relatively smooth and flat, lacking obvious fracture structures. After extraction with 60% ethanol, Figure 4 -(b) and (b'), the surface of the sample powder exhibits obvious roughness and obvious cracks, indicating that the cells and cell walls are destroyed under the action of ultrasound, thereby exposing the active ingredients to the solvent. Figure 4 -(c) and (c') show the consequences of DES extraction, where the sample showed greater fragmentation, completely broken into smaller fragments, and its surface was covered with cracks. This observation indicates that compared with the 60% ethanol solution, the DES solvent has a stronger ability to penetrate and destroy the licorice residue powder, which can completely destroy the licorice tissue cells and cell walls, thereby releasing more effective ingredients. Therefore, the extraction efficiency of DES is better than that of ethanol.

[0220] Molecular dynamics analysis

[0221] MD is a computer simulation method that studies the microscopic and macroscopic properties of a system by simulating the motion trajectories of atoms and molecules in the system, and can be used to infer the interaction between chemical components and solvents. In order to reveal the microscopic mechanism of DES in effectively extracting target compounds, we used glycyrrhizin (Li), glycyrrhizic acid (GA) and glucose (GL) as representative active ingredients, and DES and 60% ethanol as solvents to perform molecular dynamics simulations.

[0222] Figure 4 -(a–f’) show that the distribution of Li, GA, and GL molecules in DES is more uniform and the molecular density in DES is higher than that in ethanol, which indicates that the dissolution rate of Li, GA, and GL in DES is higher. Figure 5 -(a, d, g) shows the stable density of the three components in DES and ethanol. The density curve is approximately a straight line, which verifies the reliability of the model. In addition, hydrogen bonding is one of the strongest non-covalent binding interactions. The higher the number of hydrogen bonds, the better the binding effect. Figure 5-(b,e,h) show that compared with ethanol, Li, GA, GL and DES solvents formed more hydrogen bonds during the extraction process, and these hydrogen bonds were observed to last longer in DES. This indicates that DES binds more tightly to the active ingredients and has a higher extraction efficiency. Since DES contains a large number of hydroxyl and amino groups, it is easy to form hydrogen bonds with the phenolic and alcoholic hydroxyl groups of flavonoids and polysaccharides, and it is also easy to form hydrogen bonds with the glycosides in saponins. It can increase the solubility in DESs, and the three components are quickly overflowed at the same time with the assistance of ultrasound-enzyme, and the extraction rate of the three components is increased. In addition, Figure 5 -(c,f,i) show that the average lifetime of hydrogen bonds formed during Li, GA, and GL extractions with the DES solvent is longer than with ethanol, indicating more stable bonding. The DES system exhibits more hydrogen bonds and greater variability throughout the simulation compared to the ethanol system, which is relatively stable with fewer hydrogen bonds. This difference suggests that the DES system may have a more dynamic or flexible environment in which more hydrogen bonds can be formed and broken, or that it contains more hydrogen bond donors and acceptors. The ethanol system has a more stable or rigid structure with less fluctuation in the number of hydrogen bonds, suggesting that its environment is less flexible or has fewer sites capable of forming hydrogen bonds. These observations may be related to the chemical composition and molecular interactions unique to each system, which affect their properties and behaviors in different ways.

[0223] 2.5.3. Quantum chemical analysis of three active components extracted from liquorice residue by DES

[0224] Quantum chemistry is the basic principle of quantum mechanics. It is an intensity theory method that accurately studies the intramolecular and intermolecular interactions at the microscopic level at the molecular electronic structure level. In this study, by comparing the interaction characteristics of the two systems, we found that choline chloride has significant advantages as a DES for extracting the active ingredients from licorice residue.

[0225] like Figure 6 -(a-c') structural formula, gray represents carbon atoms, red represents oxygen, white represents hydrogen, blue represents nitrogen atoms, and green represents chlorine atoms. From the molecular structure, it can be seen that the molecular structure of the DES system is relatively complex and has more interaction sites. This shows that DES can form more stable complexes through more interactions with target molecules. In the DES system, the hydrogen bond donor sites, ion-π interaction sites, etc. of DES can strongly interact with the target molecules, and these multi-point and multi-type interactions further enhance the stability of the system. However, due to its simple structure and limited interaction sites provided by ethanol, the molecular compounds formed are relatively loose. Figure 6-(d-f') The potential distribution diagram shows the electrostatic potential distribution on the molecular surface. The red area represents the area with higher electron density (negative potential), and the blue area represents the area with lower electron density (positive potential). It can be seen from the figure that the electrostatic potential distribution on the surface of each molecule extracted by ethanol shows obvious polar regions, but these regions are relatively dispersed. As a polar molecule, ethanol interacts with compounds mainly through hydrogen bonds and van der Waals forces. However, due to the weak polarity of ethanol molecules, this relatively weak interaction leads to limited extraction effect. By comparison, it can be found that the electrostatic potential distribution of the compounds extracted by DES on the molecular surface is more uniform and concentrated, which indicates that DES can interact more effectively with the active ingredient molecules of licorice residue. Due to its high polarity and the characteristics of ions in DES, it can be achieved by interacting with compound molecules through stronger electrostatic interactions with ions;

[0226] The extraction process is more thorough and efficient. Figure 6 -(g-l') are the RDG contour maps of ethanol-effective component and DES-effective component systems, respectively. By comparison, it can be found that the RDG contour surface density of the DES system is larger and covers a wider range, which indicates that there are more weak interactions in the DES system.

[0227] These weak interactions include hydrogen bonds, van der Waals forces, and π-π interactions, leading to the formation of more stable molecular complexes. In the weak interaction analysis diagram, the symbol (lambda2) is on the right side of the x-axis, rho, and RDG is on the y-axis, and the color is determined according to the strength of the interaction between the items (color range 0.035-0.020 au). For the DES system, there are more red and green areas in the figure, indicating that the attractive interactions are stronger and the types of interactions are more significant. This shows that not only the number of interactions between DES and the compounds is larger, but also the intensity of the interactions is greater and the types are richer. In the ethanol system, the distribution of weak interactions is relatively sparse and the intensity is low, and the binding between molecules is weak. Therefore, due to the stronger interactions, more complex interaction types, and denser distribution of RDG and other surfaces in the DES system, the effective components of licorice residue are more effectively extracted than the ethanol system. Through multi-point and multi-type interactions, DES can bind more firmly to compound molecules, thereby improving the extraction efficiency. Applying this result to the extraction of natural products provides theoretical support for practical applications and also demonstrates its potential advantages.

[0228] In this study, we established a DES system to extract the active ingredients from licorice residue, including flavonoids, saponins, and polysaccharides. The DES system significantly improved the extraction rate of the total extract and each component of licorice residue. In addition, the extract obtained using the DES system showed strong antioxidant activity. The extraction efficiency of licorice residue by DES may be due to the destruction of the surface structure, which promotes the release of active ingredients, the formation of additional hydrogen bonds to promote the precipitation of active ingredients, or the promotion of the dissolution of active ingredients through a more uniform electrostatic distribution. This study provides a clean, efficient, and environmentally friendly extraction system for the green industrial extraction of licorice, which significantly reduces environmental pollution and energy consumption. In addition, this method can provide a reference for the extraction of other natural medicines.

[0229] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0230] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A method for extracting liquorice residue, characterized in that: The following steps are involved: S01. Mixing choline chloride and a hydrogen bond donor in a molar ratio of 1:3, heating and stirring, until a stable, uniform, transparent liquid is formed, which is a DES solvent; adding pure water to the DES solvent to obtain a DES solution; wherein the water content of the DES solution is 20% to 60%; S02. The licorice residue powder is added to the DES solution, and ultrasonic extraction is performed at room temperature for 10 min to 50 min, with an ultrasonic power of 100 W to 500 W to obtain a primary extract; wherein the solid-liquid ratio of the licorice residue powder to the DES solution is 1: (10 to 50); S03. Centrifuge the primary extract, precipitate, and freeze-dry the supernatant to obtain a liquorice residue extract.

2. The method for extracting liquorice residue according to claim 1, characterized in that: In step S01, the molar ratio of choline chloride to the hydrogen bond donor is 1:

3.

3. The method for extracting liquorice residue according to claim 1, characterized in that: In step S01, the hydrogen bond donor is selected from any one of 1,2-propylene glycol, 1,4-butanediol, ethylene glycol, glycerol, lactic acid, citric acid, malic acid, glucose, fructose, and sucrose.

4. The method for extracting liquorice residue according to claim 3, characterized in that: In step S01, the hydrogen bond donor is glycerol.

5. The method for extracting liquorice residue according to claim 1, characterized in that: In step S01, the water content of the DES solution is 50%.

6. The method for extracting liquorice residue according to claim 1, characterized in that: In step S01, the ultrasonic power is 300 W.

7. The method for extracting liquorice residue according to claim 1, characterized in that: In step S02, the solid-liquid ratio of the licorice residue powder to the DES solution is 1:

40.

8. The method for extracting liquorice residue according to claim 1, characterized in that: In step S02, the extraction time is 30 min.

9. A liquorice residue extract, characterized in that Prepared by the extraction method according to any one of claims 1 to 8.

10. Use of the liquorice residue extract according to claim 9 as an antioxidant.

Images