A membrane-chromatography combined purification process for directional enrichment of blue pricklyash pepper amide

Abstract

The application discloses a kind of blue flower pepper pungent amide directional enrichment membrane-chromatography combined purification process, belong to natural product separation and purification technical field. Including: S1 with ethanol aqueous solution extracts blue flower pepper, and solid-liquid separation is extracted clear liquid;S2 the extracted clear liquid is carried out cross-flow filtration by modified polyvinylidene fluoride ultrafiltration membrane, and the membrane permeate is obtained;S3 the membrane permeate is carried out low-temperature gas stripping under inert gas bubbling condition;S4 supplement anhydrous ethanol and adjust ethanol concentration;S5 after adjusting, sample to macroporous adsorption resin column with the feed liquid, and obtain crude product concentrated solution;S6 the crude product concentrated solution is loaded to chromatographic column with modified microspheres as filler, and the fraction rich in hydroxyl-alpha-sichuan peppercorn essence and the fraction rich in hydroxyl-beta-sichuan peppercorn essence are collected respectively, and solvent is recovered, and the hydroxyl-alpha-sichuan peppercorn essence product with the purity ≥97% and the hydroxyl-beta-sichuan peppercorn essence product with the purity ≥91% (cis configuration retention rate ≥91%) can be obtained by the cooperation of the above six steps.

Description

Technical Field

[0001] This invention belongs to the field of natural product separation and purification technology, specifically a membrane-chromatographic purification process for the directional enrichment of Sichuan pepper numbing flavor amide. Background Technology

[0002] Sichuan peppercorns ( Zanthoxylum armatum Sichuan pepper (DC.) is an important spice resource in southwestern my country. Its signature numbing flavor components are mainly amides such as hydroxy-α-sanshool and hydroxy-β-sanshool. Recent studies have confirmed that these substances are not only the core of Sichuan pepper's flavor chemistry but also exhibit outstanding performance in TRP channel activation, analgesia, and anti-inflammation, demonstrating clear application value in functional foods and natural medicines. Achieving high-purity, large-scale preparation of these substances is a key prerequisite for industrial development.

[0003] Currently, the extraction and purification of Sichuan pepper numbing amide generally follows the conventional process route of "solvent extraction—macroporous adsorption resin coarse separation—silica gel column chromatography purification—recrystallization". This route is relatively mature in the industrial production of most natural products. However, when the target is set as a purity of hydroxy-α-sanshool higher than 95% and to achieve efficient separation of hydroxy-α-sanshool and hydroxy-β-sanshool, while requiring high-fidelity enrichment that retains the natural configuration and composition ratio to the greatest extent, the existing technology reveals the following technical problems: Firstly, conventional alcohol extraction uses ethanol or aqueous alcohol to extract the raw material, driving the dissolution of the target compound through concentration gradient. However, Sichuan pepper raw materials are rich in volatile oils, phospholipids, and amphiphilic glycolipids. During alcohol extraction, these byproducts spontaneously assemble with the amide to form nanoscale microemulsion micelles. Superficially, the extract appears clear and homogeneous, and conventional tests such as UV or refractive index show no abnormalities. In reality, a significant proportion of amide molecules are encapsulated within the micelle core or interfacial film, existing in a state of "pseudo-dissolution" rather than as free molecules. When the extract enters the subsequent adsorption process, the amides shielded by the micelles cannot effectively contact the resin, resulting in an adsorption capacity far lower than theoretically expected.

[0004] Secondly, the function of macroporous adsorption resin in coarse separation is to enrich amides and preliminarily remove polysaccharides and some pigments. Given that amides contain multiple pairs of cis-conjugated double bonds, nitrogen purging, light protection, and the addition of antioxidants are typically employed for protection during production. However, trace amounts of iron and copper ions remaining in the resin pores, along with localized dissolved oxygen, and unavoidable light and temperature fluctuations during operation, continue to exert their effects during the long-term processes of sample loading, elution, desorption, and concentration, inducing free radical-mediated configuration inversion and causing the cis double bonds to slowly isomerize to the trans configuration. The cis and trans isomers have the same molecular weight and identical functional groups, and their physicochemical properties are extremely similar, making effective separation difficult under conventional reversed-phase high-performance liquid chromatography conditions.

[0005] Thirdly, silica gel column chromatography performs the function of purification and separation, achieving the separation of amide homologues based on polarity differences. However, the feed solution treated with macroporous resin often still contains microemulsion micelles formed during the extraction stage. After entering the silica gel column with the feed solution, these micelles form numerous hydrogen bonds with silanol groups in the column head region, resulting in strong adsorption and forming a strongly adsorbed contamination layer at the column head. This contamination layer leads to abnormally high column pressure and uneven mobile phase permeation. At the same time, the micelles encapsulate and drag amide molecules, disrupting the adsorption-desorption balance between them and the stationary phase. Consequently, the target analyte cannot be effectively separated from impurities, and cross-contamination of fractions is severe, significantly reducing the purification and separation efficiency.

[0006] Fourth, after the amide purity is increased to approximately 90% by silica gel column chromatography, the conventional approach uses recrystallization as the final purification method, utilizing the difference in solubility between the main component and impurities in the solvent to gradually improve purity through repeated dissolution and precipitation. However, in the Sichuan pepper-flavored amide system, hydroxy-α-sanshool and hydroxy-β-sanshool have highly similar molecular structures, with differences limited to the double bond position or hydroxyl spatial orientation. During crystallization, the β-isomer does not exist as a mechanical inclusion or adsorption on crystal faces, but rather replaces the α-isomer in the crystal lattice through molecular-level miscibility, forming a thermodynamically stable substitutional solid solution. This phenomenon results in almost no selective partitioning of the target substance and impurities during crystallization; the ratio of the two in the mother liquor and crystals remains essentially unchanged, and the yield decreases exponentially with repeated recrystallization, while the product purity shows almost no improvement.

[0007] Based on this, we developed an integrated purification process that can effectively eliminate the loss of extract nanoemulsions, inhibit the cis-trans isomerism of active ingredients, avoid interference of colloidal impurities in the feed solution with chromatographic separation, and efficiently separate sanshool homologues without recrystallization. This process is of great practical significance for the large-scale preparation of high-purity, naturally intact Sichuan pepper flavor amide. Summary of the Invention

[0008] The purpose of this invention is to provide a membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper numbing amide, in order to solve the problems of amide leakage caused by microemulsion encapsulation of extract in existing conventional processes, quality deviation due to configurational isomerization in macroporous resin processes, column efficiency decline due to micelle contamination in silica gel column chromatography, and purification barriers of recrystallized homologue solid solutions, thereby achieving the preparation of high-purity Sichuan pepper numbing amide products with high configurational fidelity.

[0009] The objective of this invention is achieved through the following technical solution: A membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides includes the following steps: S1. Extraction and primary filtration: Using Sichuan pepper as raw material, the extract was obtained by ethanol-water extraction, followed by solid-liquid separation to obtain a clear extract; S2. The extract obtained in step S1 is filtered through a modified polyvinylidene fluoride ultrafiltration membrane via cross-flow filtration to obtain membrane permeate; the modified polyvinylidene fluoride ultrafiltration membrane is obtained by successively modifying a polyvinylidene fluoride membrane with hydrophilic polyvinyl alcohol and hydrophobic stearic acid. S3. The membrane permeate obtained in step S2 is subjected to air stripping under inert gas bubbling conditions to obtain the air-stripped membrane permeate. S4. Measure the ethanol concentration of the membrane permeate obtained after air stripping in step S3, and add anhydrous ethanol to restore its ethanol concentration to a level comparable to that of the membrane permeate obtained in step S2, thus obtaining the recovery solution. S5. Load the refluxing solution obtained in step S4 onto a macroporous adsorption resin column. After adsorption, elution and desorption, collect the eluent and concentrate it under reduced pressure to obtain crude concentrate. S6. The crude concentrate obtained in step S5 is loaded onto a chromatography column packed with modified microspheres. Gradient elution is performed using an ethanol-water system as the mobile phase. The fractions rich in hydroxy-α-sanshool and the fractions rich in hydroxy-β-sanshool are collected separately. The solvent is recovered to obtain the high-purity Sichuan pepper numbing flavor amide product. The modified microspheres are obtained by covalently linking carboxylated polymethyl methacrylate rigid microspheres with amino-derived quercetin via amide bonds.

[0010] Preferably, in step S1, the concentration of the ethanol-water solution is 60–85 vol%, the extraction temperature is 35–50°C, and the extraction time is 1.5–3.0 h.

[0011] Preferably, in step S2, the operating pressure of the cross-flow filtration is 0.15–0.35 MPa, the membrane surface flow rate is 1.5–3.5 m / s, and the feed temperature is 20–35°C. Cross-flow filtration maintains the concentration polarization factor ≤1.2 by adjusting the membrane surface flow rate and operating pressure. In this scheme, by limiting the cross-flow filtration parameters, the adverse effects of excessively high operating pressure or temperature on the molecular structure stability of amide can be avoided while ensuring membrane flux and demulsification efficiency. Maintaining the concentration polarization factor ≤1.2 by adjusting the membrane surface flow rate and operating pressure reduces the dragging and retention of amide molecules by the membrane gel layer, thereby improving the permeability of free amide.

[0012] Preferably, in step S2, the water contact angle of the modified polyvinylidene fluoride ultrafiltration membrane is 55° to 70°.

[0013] Preferably, in step S2, the modified polyvinylidene fluoride ultrafiltration membrane is prepared by the following method: A polyvinylidene fluoride (PVDF) membrane is immersed in an aqueous solution containing 0.5–3.0 wt% polyvinyl alcohol (PVA), and the pH is adjusted to 8.0–9.0 with the addition of an appropriate amount of sodium bicarbonate. After removal, it is cross-linked and dried at 80–100 °C to form a PVA hydrophilic layer. Then, it is immersed in an anhydrous ethanol solution containing 0.5–2.0 wt% stearic acid and reacted at 40–60 °C for 2–6 hours. After washing and drying, the membrane is obtained. In this method, the PVA hydrophilic layer enhances the hydration capacity of the membrane surface and weakens the affinity between the hydrophilic groups in the microemulsion micelles and the solution, thus reducing the stability of the micelle interface film. The C18 long-chain alkyl group of stearic acid imparts appropriate hydrophobic microregions to the membrane surface, which can insert into the hydrophobic core or interface film of the microemulsion micelles, disrupting the structural integrity of the micelles and releasing the benzoyl amide originally encapsulated inside the micelles into a free molecular state, thereby achieving the function of breaking down microemulsion micelles.

[0014] Preferably, in step S2, the method for determining the membrane treatment effect is as follows: take the membrane permeate, and measure its turbidity at 660 nm by ≥60% compared with the extract, and after separating the free and micellar encapsulated amides with an ultrafiltration centrifuge tube with a molecular weight cutoff of 10 kDa, measure the ratio of the content of free hydroxy-α-sanshool in the membrane permeate to the total content of hydroxy-α-sanshool in the extract by ≥0.75.

[0015] Preferably, in step S3, the temperature of the air stripping is 35–40°C. These parameters effectively remove free volatile oils while avoiding excessively high temperatures that could lead to thermal degradation or isomerization of the flavonoids.

[0016] Preferably, in step S5, the desorption uses a 60–85 vol% ethanol aqueous solution, with a concentration temperature ≤40℃, and the total residence time of the solution in the resin column does not exceed 3 hours. The macroporous adsorption resin is a non-polar styrene-divinylbenzene copolymer resin with a specific surface area ≥800 m² / g and an average pore size of 8–15 nm. In this scheme, the desorption uses a 60–85 vol% ethanol aqueous solution, which is consistent with the extraction solvent system, which helps simplify solvent recovery; the concentration temperature ≤40℃ can reduce the heating time of the heat-sensitive amide; the total residence time of the solution in the resin column does not exceed 3 hours, which can shorten the contact time between the amide and trace metal ions and dissolved oxygen in the resin channels, and help control the configurational isomerization of cis double bonds. The macroporous adsorption resin selected is a non-polar styrene-divinylbenzene copolymer resin with a specific surface area ≥800 m² / g and an average pore size of 8–15 nm. This pore structure is conducive to the adsorption and mass transfer of amide molecules, taking into account both adsorption capacity and desorption efficiency.

[0017] Preferably, in step S6, the volume fraction of ethanol in the gradient elution is gradually increased from 15% to 55%. This method allows hydroxy-α-sanshool and hydroxy-β-sanshool to elute sequentially based on the difference in the strength of their respective π-π interactions with quercetin, thereby achieving chromatographic separation of the two.

[0018] Preferably, in step S6, the modified microspheres have a particle size of 25–45 μm, an average pore size of 12–20 nm, and a quercetin loading of 0.8–2.0 mmol / g.

[0019] Preferably, in step S6, 0.05–0.2 wt% of tocopherol is added during the preparation of the modified microspheres.

[0020] Preferably, in step S6, 0.01 to 0.05 wt% of ascorbate palmitate is added to the mobile phase of the gradient elution.

[0021] In step S6, tocopherol is added during the preparation of the modified microspheres. The polymer backbone physically encapsulates the tocopherol, allowing it to be slowly released to the microsphere surface during chromatography, quenching any potentially generated reactive oxygen species in localized areas of the microsphere surface. Simultaneously, ascorbate palmitate is added to the mobile phase as a dissolved antioxidant, providing antioxidant protection to the bulk feed solution. Both act on the stationary and mobile phase sides respectively, forming a synergistic antioxidant system covering different regions within the chromatography column, helping to reduce the risk of cis-double bond isomerization during chromatography.

[0022] Preferably, in step S6, the pH of the mobile phase during the gradient elution process is maintained at 6.5 to 7.5.

[0023] Preferably, in step S6, the modified microspheres are prepared by the following method: Preferably, in step S6, the modified microspheres are prepared by the following method: Carboxylated polymethyl methacrylate rigid microspheres (particle size 25–45 μm, average pore size 12–20 nm) are dispersed in anhydrous ethanol, amino-derived quercetin and a condensing agent are added, and the mixture is reacted at room temperature to 60°C for 12–24 h. Afterwards, the microspheres are washed with ethanol and dried to obtain the modified microspheres. In this scheme, the carboxyl groups on the surface of the carboxylated polymethyl methacrylate rigid microspheres undergo an amidation coupling reaction with the primary amino groups on the amino-derived quercetin molecules under the activation of the condensing agent, forming stable amide bonds, thereby firmly anchoring quercetin to the surface of the microsphere matrix. This covalent bonding method avoids the dissolution and detachment of quercetin ligands during chromatography and elution, ensuring the long-term stability of the packing material's recognition and separation performance. The modified microspheres were adapted to an ethanol-water gradient elution system, and by relying on the differential π-π interaction between the conjugated aromatic ring of quercetin and the two sorbitan compounds, the efficient separation of hydroxy-α-sorbitan and hydroxy-β-sorbitan was achieved with excellent separation reproducibility.

[0024] Compared with the prior art, the beneficial effects of the present invention are: This invention addresses four types of technical problems in existing purification processes for Sichuan pepper numbing amide extracts: microemulsion encapsulation leading to hidden loss of extracts, configurational isomerization and quality deviation in macroporous resin processes, micelle contamination and column efficiency decline in silica gel column chromatography, and purification barriers for recrystallized homologues in solid solutions. The specific steps and overall effects are as follows: In step S1, the raw material of Sichuan pepper is extracted with an ethanol-water solution, which dissolves the numbing amide and associated volatile oils, phospholipids, amphiphilic glycolipids, etc. The extract is separated by solid-liquid separation to obtain the extract liquid, which provides the raw material liquid for subsequent membrane treatment.

[0025] In step S2, a polyvinylidene fluoride (PVDF) ultrafiltration membrane, modified sequentially with hydrophilic polyvinyl alcohol and hydrophobic stearic acid, is used for cross-flow filtration of the extract. The membrane surface, after two-step modification, forms an amphiphilic structure with alternating hydrophilic and hydrophobic microdomains. During filtration, this structure competitively adsorbs emulsified components at the interface of nanoscale microemulsion micelles formed spontaneously by volatile oils, phospholipids, amphiphilic glycolipids, and amides in the extract. This causes the micelle emulsion layer to desorb, destabilize, and rupture, releasing the amides and volatile oils originally encapsulated within the micelle core. The amides enter the permeate in free molecular form. This step alleviates the problem of low adsorption capacity caused by microemulsion shielding in conventional processes, which prevents effective contact between amides and subsequent resins. Simultaneously, during amide release, membrane demulsification also converts the volatile oils, previously stably dispersed within the micelles, into a free state. Direct entry into the downstream resin column could contaminate the resin.

[0026] In step S3, the membrane permeate is subjected to low-temperature gas stripping under inert gas bubbling conditions. Utilizing the window period after S2 membrane demulsification when the volatile oil is in a free state and has not yet contacted the resin, inert gas bubbling, with the carrying effect of the ethanol-water azeotropic system, efficiently carries out the free volatile oil released after demulsification and condenses it for recovery. This simultaneously recovers the volatile oil while protecting the adsorption performance of the downstream macroporous resin. A synergistic "release first, remove later" relationship is formed between S2 and S3: without S2 membrane demulsification, the volatile oil is encapsulated in micelles, resulting in low stripping efficiency; without S3 stripping, the free volatile oil released in S2 may be adsorbed by the non-polar macroporous resin and is difficult to desorb, affecting the resin's reusability. Simultaneously, the ethanol-water azeotropic process during the alcohol-water stripping process causes a decrease in the ethanol concentration of the feed solution, which needs to be adjusted in subsequent steps.

[0027] In step S4, the ethanol concentration of the membrane permeate after stripping is measured, and anhydrous ethanol is added to restore it to a level comparable to that of the membrane permeate in step S2. This step compensates for the change in solvent composition caused by ethanol co-evaporation during the S3 stripping process, ensuring that the ethanol concentration of the feed solution entering the S5 macroporous resin column remains consistent with that of the membrane permeate, maintaining the dissolution state of the amide and the resin adsorption conditions, and ensuring a smooth transition of the solvent environment between processes.

[0028] In step S5, the sample solution after refluxing is loaded onto a macroporous adsorption resin column. After adsorption, elution, desorption, and vacuum concentration, a crude concentrate is obtained. Since microemulsion micelles have been broken in S2, free volatile oils have been removed in S3, and alcohol concentration has been restored in S4, the sample solution entering the resin column is a true molecular solution without volatile oil interference. The resin can enrich amides under conditions of no micelle shielding and no competitive adsorption by volatile oils, resulting in improved adsorption capacity and separation efficiency compared to the conventional process of directly loading the extract.

[0029] In step S6, the crude concentrate is loaded onto a chromatography column packed with carboxylated polymethyl methacrylate rigid microspheres covalently linked to amino-derived quercetin, and gradient elution is performed using an ethanol-water mobile phase. Quercetin, a natural flavonoid, contains a benzopyranone structure and multiple phenolic hydroxyl groups that can generate π-π interactions and hydrogen bonds with the conjugated double bonds in hydroxy-α-sanshool and hydroxy-β-sanshool molecules. Due to the different double bond positions and conjugation degrees of the two sanshools, the π-π interaction strengths between them and quercetin differ, resulting in retention time differences within the column and achieving the separation of the two homologues. This step, using π-π interaction and hydrogen bond recognition chromatography instead of conventional recrystallization, bypasses the purification difficulties caused by the high similarity of the two sanshool molecular structures and the formation of thermodynamically stable solid solutions during crystallization, yielding high-purity hydroxy-α-sanshool and hydroxy-β-sanshool products.

[0030] The process of this invention, through the combination of the above six steps, finally obtains hydroxy-α-sanshool products with a purity ≥97% (cis configuration retention rate ≥92%) and hydroxy-β-sanshool products with a purity ≥91% (cis configuration retention rate ≥91%), providing technical support for the application of Sichuan pepper flavor amide in the fields of functional foods and natural medicines. Detailed Implementation

[0031] Example 1 S1 Extraction and Pre-filtration: Take 1 kg of dried Sichuan pepper fruit (HPLC analysis showed that its hydroxy-α-sanshool content was 1.12 wt%, hydroxy-β-sanshool content was 0.30 wt%, and total numbing amide content was 1.42 wt%), grind it to 20 mesh, add 10 L of 75 vol% ethanol aqueous solution, stir and extract at 45℃ for 2.0 h, filter through a 200 mesh filter cloth, and collect the filtrate; add another 8 L of 75 vol% ethanol aqueous solution to the filter residue, stir and extract at 45℃ for 1.5 h, filter through a 200 mesh filter cloth, combine the two filtrates, and obtain the clear extract.

[0032] S2 membrane treatment: Preparation of S21 modified polyvinylidene fluoride ultrafiltration membrane: A polyvinylidene fluoride membrane with a molecular weight cutoff of 100 kDa was immersed in an aqueous solution containing 1.5 wt% polyvinyl alcohol (PVA-1799 type, degree of alcoholysis ≥88%) (membrane area to solution volume ratio of 1 cm²:2 mL). Sodium bicarbonate was added to adjust the pH to 8.5. After soaking for 30 min, the membrane was removed and crosslinked and dried at 90℃ for 2 h to form a PVA hydrophilic layer. The membrane was then immersed in an anhydrous ethanol solution containing 1.0 wt% stearic acid (membrane area to solution volume ratio of 1 cm²:2 mL) and reacted at 50℃ for 4 h. After removal, the membrane surface was rinsed three times with anhydrous ethanol and dried at 40℃ to obtain the modified polyvinylidene fluoride ultrafiltration membrane with a water contact angle of 62° and an actual molecular weight cutoff of 80–120 kDa.

[0033] The modified polyvinylidene fluoride ultrafiltration membrane was analyzed by ATR-FTIR spectroscopy. The results showed that, compared with the unmodified PVDF base membrane, the modified membrane exhibited better performance at 3350 cm⁻¹. -1 A broad absorption peak of PVA hydroxyl groups appears nearby, at 1730 cm⁻¹. -1 A stretching vibration peak of the ester carbonyl C=O group appears at 2920 cm⁻¹. -1 and 2850 cm -1 The CH stretching vibration peak was significantly enhanced (attributable to the C18 long-chain alkyl group of stearic acid).

[0034] S22 Cross-flow filtration: The extract obtained in step S1 was subjected to cross-flow filtration through the modified membrane described above. The operating pressure was 0.25 MPa, the membrane surface flow rate was 2.5 m / s, and the feed temperature was 28℃. The concentration polarization factor (the ratio of solute concentration at the membrane surface to the concentration of the bulk feed solution) was maintained at ≤1.2 by adjusting the membrane surface flow rate and operating pressure. Approximately 15 L of permeate was collected. The permeate was analyzed, and compared with the extract, the turbidity at 660 nm decreased by 68%. After separation using an ultrafiltration centrifuge tube with a molecular weight cutoff of 10 kDa, the ratio of the free hydroxy-α-sanshool content in the permeate to the total hydroxy-α-sanshool content in the extract was measured to be 0.82.

[0035] S3 Low-Temperature Gas Stripping Oil Removal: The membrane permeate obtained in step S2 was transferred to an air-stripping device and air-stripped for 40 minutes at 38°C under nitrogen bubbling conditions (gas-liquid ratio 10:1, v / v). Volatile components were carried out with the nitrogen and recovered by condensation, yielding approximately 14.5 L of membrane permeate after air-stripping. The condensate was dehydrated to obtain approximately 32 g of Sichuan pepper volatile oil.

[0036] S4 Ethanol Concentration Adjustment: The ethanol concentration of the membrane permeate after air stripping was measured to be 68 vol%. Approximately 1.2 L of anhydrous ethanol was added to restore the ethanol concentration to 75 vol%, resulting in approximately 15.7 L of feed solution after refluxing.

[0037] S5 macroporous resin enrichment: The refluxing solution was loaded at a flow rate of 2 BV / h into a chromatography column packed with 1.5 L of macroporous adsorption resin (AB-8 type, styrene-divinylbenzene copolymer, specific surface area 850 m² / g, average pore size 12 nm). After loading, the column was washed with 2 BV of purified water, followed by desorption with 75 vol% ethanol aqueous solution at a flow rate of 1.5 BV / h, collecting approximately 4.5 L of eluent. The eluent was concentrated under reduced pressure at 38 °C to approximately 300 mL to obtain the crude concentrate. The total residence time of the solution in the resin column was 2.5 h.

[0038] S6 Chromatographic Purification: Preparation of S61 modified microspheres: (1) Preparation of amino-derived quercetin: 10 g of quercetin was dissolved in 500 mL of anhydrous ethanol under nitrogen protection and light-protected conditions. The solution was stirred and dissolved. Triethylamine was added to adjust the pH to 8.5. 20 mL of ethylenediamine was slowly added dropwise, and the mixture was refluxed at 75 °C for 6 h. The mixture was cooled to room temperature, and 2.2 g of sodium borohydride was added in three batches. The mixture was stirred and reacted for another 2 h. After the reaction was completed, the mixture was concentrated to dryness under reduced pressure at 40 °C and -0.08 MPa. The solution was dissolved by heating in a mixed solvent of ethyl acetate and n-hexane (1:3, v / v). The solution was slowly cooled to 0 °C and allowed to stand for crystallization for 12 h. The solution was filtered and dried under vacuum at 40 °C to obtain amino-derived quercetin. The product was detected by FT-IR at 3420 cm⁻¹. -1 A primary amine NH stretching vibration peak appears nearby, at 1630 cm⁻¹. -1 The characteristic peak of C=N (Schiff base intermediate) appears nearby and then disappears.

[0039] (2) Amidation coupling: 100 g of commercially available carboxylated polymethyl methacrylate rigid microspheres (particle size 30-40 μm, average pore size 15 nm, carboxyl content ≥0.5 mmol / g) were dispersed in 500 mL of anhydrous ethanol. 8 g of the above-mentioned amino-derived quercetin and a condensation system consisting of 9 g of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) and 5.5 g of N-hydroxysuccinimide (NHS) were added. The mixture was stirred and reacted at 40 °C under nitrogen protection for 18 h. After the reaction, the microspheres were thoroughly washed with anhydrous ethanol and dried under vacuum at 50 °C to obtain quercetin-bonded modified microspheres. The quercetin loading of the obtained modified microspheres was 1.2 mmol / g.

[0040] The modified microspheres were analyzed by Fourier transform infrared spectroscopy (FT-IR). Compared with unbonded quercetin-modified carboxylated PMMA microspheres, the modified microspheres showed better performance at 1655 cm⁻¹. -1 An amide bond C=O stretching vibration peak appears at 1608 cm⁻¹. -1 and 1510cm -1 The characteristic vibrational peak of the benzene ring skeleton of quercetin appears at 3400 cm⁻¹. -1 Enhanced absorption by nearby phenolic hydroxyl groups indicates that amino-derived quercetin has been covalently linked to the microsphere surface via amide bonds. Meanwhile, no characteristic peaks of residual condensing agents were detected in the infrared spectrum, indicating that they were effectively removed after washing.

[0041] S62 Chromatographic Separation: The modified microspheres were packed into a chromatography column (3 cm diameter, 30 cm height) and equilibrated to 2 BV with a 15 vol% ethanol-water solution. The crude concentrate obtained in step S5 was loaded onto the column, and gradient elution was performed using an ethanol-water system as the mobile phase. The ethanol volume fraction was gradually increased from 15% to 55% within 4 BV, the pH of the mobile phase was maintained at 7.0, and the elution flow rate was 1.0 BV / h. Fractions were collected, and the fractions rich in hydroxy-α-sanshool (eluted in the range of approximately 25%–35% ethanol concentration) and the fraction rich in hydroxy-β-sanshool (eluted in the range of approximately 40%–50% ethanol concentration) were combined. The two fractions were concentrated under reduced pressure at 38°C to recover the solvent, and then dried under vacuum to obtain 8.2 g of hydroxy-α-sanshool product and 2.3 g of hydroxy-β-sanshool product.

[0042] Product testing methods: (1) The cis configuration retention rate was determined by cyclodextrin-bonded stationary phase high-performance liquid chromatography (HPLC). Chromatographic conditions: cyclodextrin-bonded stationary phase column (250 mm × 4.6 mm, 5 μm); mobile phase: acetonitrile-water (35:65, v / v); detection wavelength: 254 nm; column temperature: 30℃; flow rate: 1.0 mL / min; injection volume: 10 μL. The obtained hydroxy-α-sanshool product and Sichuan pepper raw material were dissolved in chromatographic methanol, filtered through a 0.45 μm microporous membrane, and then injected for analysis. The cis and trans configurations of hydroxy-α-sanshool were separated on the cyclodextrin-bonded stationary phase due to differences in inclusion action. The retention time of the cis configuration was approximately 12.3 min, and the retention time of the trans configuration was approximately 14.1 min. The cis configuration percentage was calculated using the peak area normalization method, and the cis configuration retention rate was calculated using the following formula: Cis configuration retention rate (%) = (Cis configuration percentage in product / Cis configuration percentage in raw material) × 100%.

[0043] (2) Yield definition: Based on the initial total content of the corresponding amides in the green pepper raw material (hydroxy-α-sanshool 11.2g, hydroxy-β-sanshool 3.0g), the total yield of the whole process from raw material to final pure product is calculated.

[0044] Example 1 Product test results: The purity of hydroxy-α-sanshool was 97.6% (containing 1.8% hydroxy-β-sanshool), with a cis configuration retention rate of 93.5%; the purity of hydroxy-β-sanshool was 91.3% (containing 7.2% hydroxy-α-sanshool), with a cis configuration retention rate of 92.1%; the yield of hydroxy-α-sanshool was 8.2g × 97.6% / 11.2g = 71.5%; the yield of hydroxy-β-sanshool was 2.3g × 91.3% / 3.0g = 69.9%.

[0045] Example 2 Referring to Example 1, the difference is: In step S1, the concentration of the ethanol-water solution was 60 vol%, the extraction temperature was 50℃, and the extraction time was 1.5 h. In step S2, the PVA concentration and stearic acid concentration were 0.5 wt% during the preparation of the modified membrane, and the resulting membrane had a water contact angle of 68°. The cross-flow filtration operation pressure was 0.15 MPa, the membrane surface velocity was 1.5 m / s, and the feed temperature was 20℃. In step S5, the macroporous adsorption resin has a specific surface area of ​​800 m² / g and an average pore size of 8 nm; the desorbed ethanol concentration is 60 vol%, and the total residence time of the feed solution in the resin column is 1.8 h. In step S6, the modified microspheres had a particle size of 25–35 μm, an average pore size of 12 nm, and a quercetin loading of 0.8 mmol / g; the gradient elution mobile phase had a pH of 6.5.

[0046] Product testing results: The purity of hydroxy-α-sanshool was 97.1% (containing 2.2% hydroxy-β-sanshool), with a cis configuration retention rate of 94.2%; the purity of hydroxy-β-sanshool was 95.7% (containing 3.5% hydroxy-α-sanshool), with a cis configuration retention rate of 93.1%; the yield of hydroxy-α-sanshool was 70.2%, and the yield of hydroxy-β-sanshool was 71.8%.

[0047] Example 3 Referring to Example 1, the difference is: In step S1, the concentration of the ethanol-water solution was 85 vol%, the extraction temperature was 35℃, and the extraction time was 3.0 h. In step S2, the PVA concentration was 3.0 wt% and the stearic acid concentration was 2.0 wt% during the preparation of the modified membrane, and the resulting membrane had a water contact angle of 55°; the cross-flow filtration operation pressure was 0.35 MPa, the membrane surface flow velocity was 3.5 m / s, and the feed temperature was 35℃. In step S3, the air stripping temperature is 40℃; In step S5, the macroporous adsorption resin has a specific surface area of ​​900 m² / g and an average pore size of 15 nm; the desorbed ethanol concentration is 85 vol%, the concentration temperature is 40℃, and the total residence time of the feed solution in the resin column is 3.0 h. In step S6, the modified microspheres had a particle size of 35–45 μm, an average pore size of 20 nm, and a quercetin loading of 2.0 mmol / g; the gradient elution mobile phase had a pH of 7.5.

[0048] Product testing results: The purity of hydroxy-α-sanshool was 97.8% (containing 1.5% hydroxy-β-sanshool), with a cis configuration retention rate of 92.8%; the purity of hydroxy-β-sanshool was 96.9% (containing 2.3% hydroxy-α-sanshool), with a cis configuration retention rate of 91.8%; the yield of hydroxy-α-sanshool was 73.3%, and the yield of hydroxy-β-sanshool was 75.9%.

[0049] Example 4 Based on Example 1, in step S6, during the preparation of modified microspheres, 0.1 g of tocopherol (0.1 wt% based on the dry weight of 100 g of microspheres) was pre-dissolved in 10 mL of anhydrous ethanol and added together with amino-derived quercetin to the microsphere dispersion. After stirring and reacting at 40 °C for 18 h, the solvent was slowly evaporated under reduced pressure at 50 °C, allowing the tocopherol to fully penetrate into the polymer backbone of the microspheres during the solvent evaporation process.

[0050] 0.03 wt% ascorbate palmitate was added to the gradient elution mobile phase.

[0051] Product testing results: The purity of hydroxy-α-sanshool was 97.5% (containing 1.9% hydroxy-β-sanshool), with a cis configuration retention rate of 96.1%; the purity of hydroxy-β-sanshool was 96.2% (containing 3.0% hydroxy-α-sanshool), with a cis configuration retention rate of 94.8%; the yield of hydroxy-α-sanshool was 71.4%, and the yield of hydroxy-β-sanshool was 73.7%.

[0052] The residue and leaching of the modified ultrafiltration membranes and modified microsphere packings used in Examples 1-4 were verified as follows: 1. Verification of leachables from modified polyvinylidene fluoride ultrafiltration membranes: One piece each of the modified polyvinylidene fluoride ultrafiltration membranes (approximately 100 cm² in area) prepared in step S21 of Examples 1-4 was immersed in 200 mL of 75 vol% ethanol aqueous solution and shaken at 28°C for 24 h. The immersion solutions were then filtered through a 0.45 μm filter membrane and analyzed. The results showed that the polyvinyl alcohol (PVA) content in all four immersion solutions was below the detection limit (0.5 mg / L), the stearic acid content was below the detection limit (0.1 mg / L), and the total organic carbon (TOC) value was less than 1.0 mg / L with no significant difference from the blank control, indicating that no material was leached from the modified membranes under the operating conditions.

[0053] 2. Verification of modified microsphere filler residue and ligand detachment: Take 10 g of each of the modified microspheres prepared in step S61 of Examples 1-4, and perform the following verifications respectively: (1) Reagent residue verification: Each modified microsphere was placed in 200 mL of purified water and extracted by ultrasonication at 40℃ for 30 min. After filtration through a 0.45 μm filter membrane, borate was detected by ion chromatography, and EDC and NHS were detected by liquid chromatography-mass spectrometry (LC-MS / MS). The results showed that the borate content in all four extracts was below the detection limit (0.01 mg / L), and EDC and NHS were not detected (detection limit 0.005 mg / L), indicating that the condensing agent and reducing agent had been effectively removed and there was no risk of reagent residue.

[0054] (2) Ligand release verification: Modified microspheres prepared in each example were packed into chromatography columns and continuously washed with 75 vol% ethanol aqueous solution at a flow rate of 1.0 BV / h for 48 h (approximately 50 column volumes). The eluents were collected in fractions, and quercetin and amino-derived quercetin were detected by LC-MS / MS in multiple reaction monitoring mode (quantitatively quantified using the quercetin standard curve). The results showed that the contents of quercetin and amino-derived quercetin in the four eluents were all below the detection limit (0.001 μg / mL), indicating that the ligands were firmly anchored by covalent bonds and there was no release.

[0055] Based on the above verification results, the modified PVDF ultrafiltration membrane used in Examples 1 to 4 of this invention showed no material leaching, the reagents for preparing the modified microsphere packing were effectively removed, and the functional ligands were stably anchored without detachment. The final product of Sichuan pepper flavor amide met the relevant food safety requirements.

[0056] Comparative Example 1 Referring to Example 1, the difference is that the S2 membrane treatment step is omitted, and the extract obtained in S1 is directly introduced into S3 air stripping. The remaining steps are the same as in Example 1.

[0057] The results showed that the turbidity of the liquid at 660 nm decreased by 12% after air extraction, and the final purity of the hydroxy-α-sanshool product was 88.6% (containing 9.5% hydroxy-β-sanshool), with a yield of 33.6%; the purity of the hydroxy-β-sanshool product was 84.2% (containing 13.8% hydroxy-α-sanshool), with a yield of 32.9%.

[0058] Explanation: The lack of a membrane demulsification step means that the microemulsion micelles are not broken down, and a large amount of amide is shielded by the micelles and lost in subsequent resin processes. At the same time, the volatile oil is encapsulated by the micelles and cannot be removed by air stripping, which contaminates the resin and reduces separation efficiency.

[0059] Comparative Example 2 Referring to Example 1, the difference is that the S3 low-temperature gas stripping deoiling step is omitted, and the membrane permeate obtained in S2 is directly fed into the S5 macroporous resin for enrichment after the ethanol concentration is measured (no change). The remaining steps are the same as in Example 1.

[0060] The results showed that the purity of the hydroxy-α-sanshool product was 91.3% (containing 7.5% hydroxy-β-sanshool); the purity of the hydroxy-β-sanshool product was 87.6% (containing 10.8% hydroxy-α-sanshool).

[0061] Note: When the membrane breaks down and releases the amide, it also converts the volatile oil into a free state. If it is not removed by air stripping, the free volatile oil will be adsorbed and accumulated by the non-polar resin in the resin column, causing the resin adsorption capacity to decrease batch by batch. In addition, the volatile oil will co-elute with the amide during desorption, interfering with subsequent chromatographic separation.

[0062] Comparative Example 3 Referring to Example 1, the difference is that the S4 ethanol concentration adjustment step is omitted, and the membrane permeate after air stripping directly enters the S5 macroporous resin for enrichment. The remaining steps are the same as in Example 1.

[0063] The results showed that the purity of the hydroxy-α-sanshool product was 94.2% (containing 4.8% hydroxy-β-sanshool), and the purity of the hydroxy-β-sanshool product was 91.5% (containing 6.9% hydroxy-α-sanshool).

[0064] Note: During the gas stripping process of the alcohol-water system, the azeotropic volatilization of ethanol causes a decrease in the alcohol concentration of the feed solution. If it is not adjusted back, the insufficient ethanol concentration in the feed solution will cause the amide to change its dissolution state. During the resin loading and rinsing stages, some amide will precipitate or penetrate, resulting in the loss of the target substance.

[0065] Comparative Example 4 Referring to Example 1, the difference is that in step S6, conventional silica gel column chromatography (200-300 mesh silica gel, petroleum ether-ethyl acetate gradient elution) is used instead of quercetin-modified microsphere chromatography. The fraction rich in hydroxy-α-sanshool obtained from silica gel column chromatography is concentrated and then recrystallized with ethanol-water. The remaining steps are the same as in Example 1.

[0066] The results showed that the separation degree of silica gel column chromatography was lower than that of Example 1, and there was serious overlap between the hydroxy-α-sanshool and hydroxy-β-sanshool fractions. After recrystallization three times, the purity of the hydroxy-α-sanshool product only increased to 92.5% (containing 6.8% hydroxy-β-sanshool), and the yield decreased from 75% of the first crystallization to 28% of the third crystallization, with almost no further improvement in purity. The purity of the hydroxy-β-sanshool product was 89.3% (containing 9.5% hydroxy-α-sanshool).

[0067] Note: Conventional silica gel column chromatography has limited ability to separate isomers at double bond positions, and subsequent recrystallization results in extremely low separation efficiency due to the formation of a solid solution between the two sorbitols. Therefore, increasing the number of crystallization cycles cannot effectively improve purity.

[0068] Comparative Example 5 Referring to Example 1, the difference is that in step S2, an unmodified commercial PVDF ultrafiltration membrane (molecular weight cutoff 100 kDa, water contact angle 105°) is used, while the other steps are the same as in Example 1.

[0069] The results showed that the turbidity at 660 nm in the membrane permeate decreased by only 15%, and the proportion of free hydroxy-α-sanshool was 0.28%. The final purity of the hydroxy-α-sanshool product was 89.3% (containing 9.2% hydroxy-β-sanshool), with a yield of 36.5%; the purity of the hydroxy-β-sanshool product was 85.1% (containing 13.2% hydroxy-α-sanshool), with a yield of 35.6%.

[0070] Note: Conventional PVDF membranes without hydrophilic-hydrophobic modification have a strongly hydrophobic surface and do not have demulsification function. They cannot destroy the microemulsion micelle structure, and most of the amides are trapped or shielded by the micelles. As with conventional processes, there is a hidden loss problem caused by the "pseudo-dissolution" of the microemulsion.

[0071] Comparative Example 6 Referring to Example 1, the difference is that in step S6, carboxylated PMMA microspheres without quercetin bonding (same matrix as in Example 1, only without amino-derived quercetin bonding) are used as the chromatography packing material, while the rest of the steps are the same as in Example 1.

[0072] The results showed that hydroxy-α-sanshool and hydroxy-β-sanshool had almost the same retention time on the column, making effective separation impossible. The two fractions showed significant cross-contamination, with the purity of the hydroxy-α-sanshool product being only 78.2% (containing 19.5% hydroxy-β-sanshool) and the purity of the hydroxy-β-sanshool product being only 74.5% (containing 23.1% hydroxy-α-sanshool).

[0073] Note: PMMA microspheres without quercetin bonding only exhibit weak hydrophobic interactions and cannot provide π-π recognition or hydrogen bond recognition functions. They cannot distinguish between two salicornin homologues that differ only in the position of their double bonds, thus verifying the key role of quercetin as a recognition ligand.

[0074] Comparative Example 7 Referring to the raw materials and extraction conditions of Example 1, the difference is that all steps S2 to S6 are cancelled and a conventional full process is used instead: the extract obtained in S1 is directly loaded onto a macroporous adsorption resin column (AB-8 type, the same as in Example 1). After adsorption, elution, desorption and concentration, column chromatography is performed with 200-300 mesh silica gel (petroleum ether-ethyl acetate gradient elution). The fraction rich in hydroxy-α-sanshool is collected, concentrated and recrystallized three times with ethanol-water. All other conditions are the same as in Example 1.

[0075] Results: The purity of the hydroxy-α-sanshool product was 91.2% (containing 7.8% hydroxy-β-sanshool), with a yield of 27.2% and a cis configuration retention rate of 71.5%; the purity of the hydroxy-β-sanshool product was 87.3% (containing 11.2% hydroxy-α-sanshool), with a yield of 26.6% and a cis configuration retention rate of 69.8%.

[0076] Note: Compared with Example 1, the three indicators of the conventional whole process—total yield, product purity, and cis configuration retention rate—were significantly lower. The failure to break down the extract microemulsion led to amide loss during the resin process, the silica gel column suffered from micellar contamination resulting in low separation efficiency, recrystallization failed due to solid solution phenomena and the yield decreased sharply, and prolonged operation and repeated heating exacerbated configuration isomerization.

[0077] Comparative Example 8 Referring to Example 1, the difference lies in that the modified membrane used in step S2 was only hydrophilically modified with PVA and not hydrophobically modified with stearic acid. This membrane was prepared according to the method in S21 of Example 1, only completing the PVA coating crosslinking and then stopping, without further immersion in the stearic acid solution. The resulting membrane had a water contact angle of 38°.

[0078] Results: The turbidity at 660 nm in the permeate decreased by only 28%, and the proportion of free hydroxy-α-sanshool was 0.41%. The final purity of the hydroxy-α-sanshool product was 90.5% (containing 8.3% hydroxy-β-sanshool), with a yield of 41.5%; the purity of the hydroxy-β-sanshool product was 86.7% (containing 11.5% hydroxy-α-sanshool), with a yield of 40.5%.

[0079] Note: The surface of the membrane modified only by PVA hydrophilic modification lacks hydrophobic microregions and cannot form an amphiphilic structure with alternating hydrophilic and hydrophobic properties. Therefore, its ability to disrupt the microemulsion micelle interface membrane is limited, and its demulsification effect is far lower than that of the two-step modified membrane in Example 1.

[0080] Comparative Example 9 Referring to Example 1, the difference lies in that the modified membrane used in step S2 is only hydrophobically modified with stearic acid and not hydrophilically modified with PVA. The membrane preparation method is as follows: the PVDF base membrane is directly immersed in an anhydrous ethanol solution containing 1.0 wt% stearic acid, reacted at 50°C for 4 hours, then rinsed with anhydrous ethanol and dried at 40°C. The resulting membrane has a water contact angle of 98°.

[0081] Results: The turbidity at 660 nm in the permeate decreased by 18%, and the proportion of free hydroxy-α-sanshool was 0.32%. The final purity of the hydroxy-α-sanshool product was 89.8% (containing 8.9% hydroxy-β-sanshool), with a yield of 37.2%; the purity of the hydroxy-β-sanshool product was 85.3% (containing 12.9% hydroxy-α-sanshool), with a yield of 36.3%.

[0082] Note: PVDF base film is inherently highly hydrophobic. Direct treatment with stearic acid cannot effectively introduce hydrophilic microdomains, and the film surface remains predominantly hydrophobic, lacking the amphiphilic structure required for demulsification.

[0083] In summary, Examples 1-3, through the coordinated six-step process, ultimately yielded hydroxy-α-sanshool products with a purity ≥97% (cis configuration retention ≥92%) and hydroxy-β-sanshool products with a purity ≥91% (cis configuration retention ≥91%), providing technical support for the application of Sichuan pepper flavor amide in functional foods and natural medicines. Testing showed that the content of heavy metals such as lead and arsenic met the limits of GB 2762. Example 4, based on Example 1, introduced the antioxidant schemes of claims 8 and 9, further increasing the cis configuration retention rate from 93.5% to 96.1%, verifying the solid-liquid two-phase synergistic antioxidant effect of tocopherol on the filler side and ascorbate palmitate on the mobile phase side.

[0084] Comparative Example 1 eliminated the membrane treatment step, and Comparative Example 5 used an unmodified conventional PVDF membrane. Neither could break down the microemulsion micelles, resulting in significant amide loss, thus verifying the necessity of the modified membrane's demulsification function. Comparative Example 2 eliminated the gas stripping deoiling step, leading to the accumulation of free volatile oil in the resin column and a batch-by-batch decrease in adsorption capacity, verifying the necessity of the synergistic "release first, remove later" sequence between the membrane and gas stripping. Comparative Example 3 eliminated the ethanol concentration adjustment step, resulting in a decrease in the feed alcohol concentration and partial amide loss, verifying the role of this connecting step. Comparative Example 4 replaced quercetin-modified microsphere chromatography with conventional silica gel column chromatography followed by recrystallization; purification failed due to solid solution phenomena, verifying the crucial role of quercetin recognition chromatography in overcoming the solid solution barrier. Comparative Example 6 used blank microspheres without quercetin bonding; the two sorbitols could not be effectively separated, verifying the irreplaceable role of quercetin as a recognition ligand. Comparative Example 7, using conventional full-process operation, showed significantly lower product purity, yield, and configuration retention compared to the present invention, verifying the overall advantages of the six-step integrated process of the present invention. Comparative Examples 8 and 9 respectively verified that neither hydrophilic modification nor hydrophobic modification alone could achieve effective demulsification, confirming the necessity of a two-step composite modification of the PVA hydrophilic layer and the stearic acid hydrophobic layer for constructing an amphiphilic demulsifying surface.

[0085] In summary, this invention effectively solves problems such as microemulsion encapsulation loss, blocking negative coupling between processes, and inhibiting configurational isomerization by synergistically combining six steps: membrane demulsification, gas stripping deoiling, ethanol reversion, resin enrichment, and quercetin-modified microsphere chromatography purification. This achieves efficient, stable, and configurationally faithful directional enrichment of Sichuan pepper numbing flavor amide.

Claims

1. A membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides, characterized in that, Includes the following steps: S1. Extraction and primary filtration: Using Sichuan pepper as raw material, the extract was obtained by ethanol-water extraction, followed by solid-liquid separation to obtain a clear extract; S2. The extract obtained in step S1 is filtered through a modified polyvinylidene fluoride ultrafiltration membrane via cross-flow filtration to obtain membrane permeate; the modified polyvinylidene fluoride ultrafiltration membrane is obtained by successively modifying polyvinylidene fluoride membrane with hydrophilic polyvinyl alcohol and hydrophobic stearic acid; S3. The membrane permeate obtained in step S2 is subjected to gas stripping under inert gas bubbling conditions to obtain gas stripped membrane permeate. S4. Measure the ethanol concentration of the membrane permeate obtained after air stripping in step S3, and add anhydrous ethanol to restore its ethanol concentration to a level comparable to that of the membrane permeate obtained in step S2, thus obtaining the recovery solution. S5. Load the refluxing solution obtained in step S4 onto a macroporous adsorption resin column. After adsorption, elution and desorption, collect the eluent and concentrate it under reduced pressure to obtain crude concentrate. S6. The crude concentrate obtained in step S5 is loaded onto a chromatography column packed with modified microspheres. Gradient elution is performed using an ethanol-water system as the mobile phase. The fractions rich in hydroxy-α-sanshool and the fractions rich in hydroxy-β-sanshool are collected separately. The solvent is recovered to obtain the high-purity Sichuan pepper numbing flavor amide product. The modified microspheres are obtained by covalently linking carboxylated polymethyl methacrylate rigid microspheres with amino-derived quercetin via amide bonds.

2. The membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides according to claim 1, characterized in that, In step S1, the concentration of the ethanol-water solution is 60–85 vol%, the extraction temperature is 35–50°C, and the extraction time is 1.5–3.0 h.

3. The membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides according to claim 1, characterized in that, In step S2, the operating pressure of the cross-flow filtration is 0.15–0.35 MPa, the membrane surface velocity is 1.5–3.5 m / s, and the feed temperature is 20–35°C. The cross-flow filtration maintains the concentration polarization factor ≤1.2 by adjusting the membrane surface velocity and operating pressure.

4. The membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides according to claim 1, characterized in that, In step S2, the method for determining the membrane treatment effect is as follows: take the membrane permeate and measure its turbidity at 660 nm by ≥60% compared with the extract. After separating the free and micellar encapsulated amides with an ultrafiltration centrifuge tube with a molecular weight cutoff of 10 kDa, measure the ratio of the content of free hydroxy-α-sanshool in the membrane permeate to the total content of hydroxy-α-sanshool in the extract by ≥0.

75.

5. The membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper numbing flavor amide according to claim 1, characterized in that, In step S3, the temperature of the air stripping is 35-40°C.

6. The membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides according to claim 1, characterized in that, In step S5, the desorption is performed using a 60-85 vol% ethanol aqueous solution, with a concentration temperature ≤40℃, and the total residence time of the solution in the resin column does not exceed 3 hours; the macroporous adsorption resin is a non-polar styrene-divinylbenzene copolymer resin with a specific surface area ≥800 m² / g and an average pore size of 8-15 nm.

7. The membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides according to claim 1, characterized in that, In step S6, the modified microspheres have a particle size of 25–45 μm, an average pore size of 12–20 nm, and a quercetin loading of 0.8–2.0 mmol / g.

8. The membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides according to claim 1, characterized in that, In step S6, 0.05–0.2 wt% of tocopherol is added during the preparation of the modified microspheres.

9. The membrane-chromatographic purification process for the targeted enrichment of Sichuan pepper-flavored amides according to claim 8, characterized in that, In step S6, 0.01 to 0.05 wt% of ascorbate palmitate is added to the mobile phase of the gradient elution.