JP7567335B2 - Fermentation liquid filtration method - Google Patents

Description

The present invention relates to a method for filtering fermentation liquid using a membrane module in the fermentation industry, food industry, and other fields.

In particular, in the treatment of fermented liquids in the food industry, diatomaceous earth has traditionally been used to remove yeast, solids, colloids, etc. from beer and wine after fermentation. However, there are issues with the safety of diatomaceous earth itself, the fact that used diatomaceous earth cannot be incinerated, and the high disposal costs associated with using large amounts of it. Therefore, in recent years, attention has been focused on the treatment of fermented liquids using separation membranes such as ultrafiltration membranes and microfiltration membranes, which are excellent for miniaturizing equipment.

When treating fermented liquids such as beer and wine with a separation membrane module, a layer consisting of the removed microorganisms, their crushed material, colloids, etc. forms on the membrane surface, clogging the membrane and causing an increase in filtration pressure and a decrease in filtration speed over time.

As a membrane structure that is less likely to clog the membrane surface and can exhibit filtering properties, a membrane capable of so-called depth filtration has been developed in which the surface pore size on one side of the separation membrane is larger than the substance to be removed, and the other surface or membrane thickness has a minimum pore size layer, capturing impurities inside the membrane. Patent documents 1 and 2 propose membranes with a gradient structure in which the pore size gradually increases from the outer surface of the hollow fiber membrane to the inner surface. The membranes disclosed in these documents either do not have sufficient water permeability, or have high water permeability but insufficient membrane breaking strength, and thus do not adequately suppress clogging.

In addition, relatively recently, hollow fiber membranes for beer filtration that contain hydrophilic polymers and have an asymmetric structure have been disclosed, as described in Patent Documents 3 and 4. The filtration performance of the membranes disclosed in these documents is relatively high, but they do not have sufficient washability with general-purpose chemicals such as aqueous sodium hydroxide solutions, and are not economical because expensive chemicals are required to achieve high washability.

Furthermore, Patent Document 5 proposes a hollow fiber membrane that has excellent chemical resistance, excellent mechanical breaking strength, and high pure water permeability, by heat treating a separation membrane made of polyvinylidene fluoride resin at a constant temperature near the membrane melting point. However, the manufacturing method disclosed therein causes excessive shrinkage of the solid parts, causing partial blockage of fine gaps between the solid parts, resulting in a membrane with poor branching of the flow paths. The above-mentioned membrane has the characteristic that the membrane area of the permeable flow paths, i.e., the effective membrane area, is greatly reduced by localized clogging, so when filtering liquids that are prone to clogging, such as fermentation liquid, the effective membrane area decreases at an accelerated rate as the filtration proceeds.

WO2002/058828 WO2010/035793 WO2016/182015 WO2017/155034 JP 2008-105016 A

Fermentation liquid is filtered through a separation membrane for various purposes, such as separating fermentation products and concentrating microorganisms. However, the fermentation liquid contains various organic substances, such as proteins, polysaccharides, and polyphenols, which cause fouling of the separation membrane. Fouled separation membranes can be reused by recovering their permeability through chemical cleaning, but if the separation membrane is prone to fouling, the frequency of chemical cleaning increases, leading to higher cleaning costs. Furthermore, repeated filtration of the fermentation liquid and chemical cleaning gradually accumulates fouling substances that cannot be completely washed away in the membrane, making it difficult to recover permeability and lengthening the cleaning time.

The present invention has been made in consideration of the above, and aims to provide a method for filtering fermentation liquid using a membrane module that allows for repeated filtration with stability over a long period of time.

In order to solve the above-mentioned problems, the present invention provides the following techniques [1] to [9].
[1] A method for filtering a fermentation liquid by carrying out the following steps a to d in this order: (a) a step of cross-flow filtering the fermentation liquid using a membrane module having a separation membrane;
(b) rinsing the separation membrane with water;
(c) permeating the chemical solution having a pH of 10 to 14 from the secondary side to the primary side of the separation membrane;
(d) rinsing the separation membrane with water;
The cross-flow membrane permeate in the step (a) has a total sugar concentration of 1,000 mg/L or more and 100,000 mg/L or less, a protein concentration of 50 mg/L or more and 10,000 mg/L or less, and a polyphenol concentration of 50 mg/L or more and 10,000 mg/L or less,
The step c
(c-1) A method for filtering a fermentation liquid, comprising a step of passing a chemical solution A having a pH of 10 or more and 14 or less and an oxidant concentration of 0.01% or more from the secondary side to the primary side of a membrane module.
[2] The step c is
(c-2) The method for filtering a fermentation liquid according to [1], further comprising a step of permeating a chemical solution B having an oxidant concentration of less than 0.01% from the secondary side to the primary side of the separation membrane prior to the step c-1.
[3] After the step d,
(e) passing a chemical solution C having a pH of 0 to 4 from the secondary side to the primary side of the separation membrane; and (f) rinsing the membrane module with water after the step e. The method for filtering a fermentation liquid according to any one of [1] to [2].
[4] The method for filtering a fermentation broth according to any one of [1] to [3], wherein in the steps c and e, the flux (m ³ /m ² /d) of the chemical solution passing from the secondary side to the primary side is 1.5 times or more and 10 times or less than the crossflow flux in the step a.
[5] The chemical solution C is an aqueous solution containing at least one compound selected from the group consisting of hydrochloric acid, nitric acid, sulfuric acid, phosphoric acid, formic acid, acetic acid, propionic acid, butyric acid, citric acid, oxalic acid, ascorbic acid and lactic acid. The fermentation liquor filtration method according to any one of [3] to [4].
[6] A method for filtering a fermentation liquid using the membrane module according to any one of [1] to [5], wherein the chemical solutions A, B and C contain at least one cleaning auxiliary selected from the group consisting of a surfactant, an enzyme and a chelating agent.
[7] The method for filtering a fermentation liquid according to any one of [1] to [6] above, wherein the polysaccharide concentration Ct in the raw water is in the range of 1.3 to 15 times the concentration C0 at the start of filtration, and the process proceeds from step a to step b.
[8] The method for filtering a fermentation liquid according to any one of [1] to [7] above, wherein the separation membrane is a hollow fiber membrane.
[9] The separation membrane contains a hydrophobic polymer resin,
The pore volume V2 of the pores having a pore diameter of less than 100 nm measured by a gas adsorption method is 0.02 cm3/g or more and less than 0.1 cm3/g,
The pore volume V1 of pores with diameters of 100 nm or more as measured by mercury intrusion porosimetry is 0.1 cm3/g or more,
The method for filtering a fermentation broth according to the above-mentioned [8], wherein the ratio V1/V2 of the pore volume V2 to the pore volume V1 is 3 or more and 60 or less.
V1: Pore volume with a pore diameter of 100 nm or more measured by mercury intrusion porosimetry V2: Pore volume with a pore diameter of less than 100 nm measured by gas adsorption [10] The method for filtering a fermentation broth according to [9] above, wherein the breaking elongation of the separation membrane is 38% or more.
[11] The separation membrane according to the above [9] or [10], wherein the shrinkage rate of the hollow fiber membrane length when the membrane is heat-treated in a water vapor atmosphere at 125° C. for 20 hours is 0.5% or more.
[12] The separation membrane contains a hydrophilic polymer,
The hydrophilic polymer contains a copolymer composed of N types of monomer units 1, 2, ... i ... N, and the hydration energy density of the hydrophilic polymer calculated based on the following formula (1) is 40 to 70 cal mol ^-1 Å ^-3 . The method for filtering a fermentation broth according to any one of the above [1] to [11].

[In formula (1), the hydration energy of any monomer unit i is the absolute value of the energy of the monomer unit i in water minus the energy of the monomer unit i in a vacuum, N is the total number of monomer species constituting the copolymer, i is an integer of 1 to N, and N is an integer of 2 or more.]
[13] The method for filtering a fermentation liquor according to any one of the above [1] to [12], wherein the hydrophobic polymer resin is a fluorine-based resin or a sulfone-based resin.
[14] The method for filtering a fermentation liquid according to any one of [1] to [13] above, wherein the filtration in the step a is an external pressure filtration in which filtration is performed from the outer periphery to the inner periphery of the hollow fiber membrane, and in the steps c and e, the flux of the permeating chemical solution per unit area of the inner surface of the hollow fiber membrane is 7.0 ^{m 3} /m ² /d or more and 10 m ³ /m ² /d or less.

The filtration method of the fermentation liquid using the membrane module of the present invention is a filtration method that repeats four steps: (a) cross-flow filtration of the fermentation liquid, (b) water rinsing, (c) alkaline chemical washing, and (d) water rinsing. Here, the cross-flow membrane permeate composition in step a has a total sugar concentration of 1000 mg/L to 100,000 mg/L, a protein concentration of 50 mg/L to 10,000 mg/L, and a polyphenol concentration of 50 mg/L to 10,000 mg/L. In step c, a chemical washing is performed by passing a chemical solution with a pH of 10 to 14 and containing an oxidizing agent from the secondary side to the primary side, thereby achieving a high cleaning effect in a short cleaning time and enabling stable filtration over a long period of time.

FIG. 1 is a schematic diagram of a filtration device according to an embodiment of the present invention. FIG. 2 is a flow chart showing an example of the method for filtering a fermentation broth of the present invention. FIG. 3 is a flow chart showing an example of the method for filtering a fermentation broth of the present invention. FIG. 4 is a perspective view of a hollow fiber membrane used in the present invention. FIG. 5 is a perspective view showing an outline of the spherical structure contained in the hollow fiber membrane.

The filtration method using a membrane module according to an embodiment of the present invention will be described in detail below with reference to the drawings. In this invention, "upper" and "lower" are based on the state shown in the drawings and are for convenience only, with the side into which the raw liquid flows being referred to as the "lower" direction and the side from which the filtrate flows out being referred to as the "upper" direction. Normally, when the membrane module is in use, the up-down direction coincides with the up-down direction in the drawings.

The configuration of a fermentation liquid filtration device according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic diagram of a filtration device according to an embodiment of the present invention, and FIG. 2 is a flow chart showing an example of a fermentation liquid filtration method.

First, we will use the flowchart in Figure 2 to provide an overview of the entire process.

The method for filtering fermentation liquid using the membrane module of the present invention includes the following steps a to d, and by repeating these steps a to d, it becomes possible to perform repeated filtration stably over a long period of time.

(a) a step of cross-flow filtering the fermentation liquid using a membrane module having a separation membrane; (b) a step of rinsing the separation membrane with water after step a; (c) a step of injecting a chemical solution having a pH of 10 or more and 14 or less from the secondary side to the primary side of the membrane of the membrane module after step b; and (d) a step of rinsing the separation membrane with water after step c. Furthermore, as shown in the flowchart of FIG. 3, by repeating steps a to f, which are steps a to d plus steps e and f, repeated filtration can be performed stably for a longer period of time.
(e) After the step d, a chemical solution having a pH of 0 or more and 4 or less is injected from the secondary side of the membrane of the membrane module to the primary side, and (f) after the step e, a step of rinsing the separation membrane with water. In the first step, step a, the fermentation liquid is cross-flow filtered by the membrane module. In this step a, the fouling state of the separation membrane can be judged from the transmembrane pressure difference obtained by subtracting the pressure on the filtrate side from the pressure on the raw liquid side of the separation membrane. When the transmembrane pressure difference value rises to a predetermined value, the process proceeds to step b, and when it is equal to or less than the predetermined value, the cross-flow filtration of step a is continued. In step b, the membrane module is washed by water rinsing. After the step b is completed, the process proceeds to step c, where the membrane module is washed with a chemical solution. After the step c is completed, the water rinsing of step d is performed. Here, if the washing recovery is insufficient, after the step d is completed, the process proceeds to step e, where the membrane module is washed with an additional chemical solution. After the step e is completed, the water rinsing of step f is performed, and the process returns to step a again. The fermentation liquid is filtered by repeating these steps a to d, or a to f.

＜Undiluted solution＞
In this embodiment, a fermentation liquid is used as the raw liquid. Examples of the fermentation liquid include beer, wine, sake, vinegar, soy sauce, amino acid fermentation liquid, organic acid fermentation liquid, and fermentation liquid of biopharmaceuticals. The fermentation liquid contains a fermentation product produced by a microorganism, but the type of the microorganism and the type of the fermentation product are not limited to the specific examples.

In the cross-flow filtration of step a, the filtrate obtained by permeating the fermentation liquid through the membrane module preferably has a total sugar concentration of 1,000 mg/L or more and 100,000 mg/L or less, a protein concentration of 50 mg/L or more and 10,000 mg/L or less, and a polyphenol concentration of 50 mg/L or more and 10,000 mg/L or less.

The total sugar concentration, protein concentration, and polyphenol concentration in the fermentation liquid (i.e., the raw liquid) before passing through the membrane module may be the same as or different from the permeated filtrate, depending on the pore size and structure of the separation membrane contained in the membrane module. When a part of the component is blocked by the separation membrane, the concentration of the component in the filtrate decreases compared to the concentration of the component in the raw liquid. Here, the total sugar concentration, protein concentration, and polyphenol concentration in the raw liquid or the filtrate can be calculated by the following methods. That is, the total sugar concentration is calculated by the phenol-sulfuric acid method or the monosaccharide quantification method by hydrolysis, the protein concentration is calculated by the BCA method, the Bradford method, the Lowry method, and the Pierce 660 nm Protein Assay method, and the polyphenol concentration is calculated by the Folin-Ciocalteu method, the Folin-Denis method, and the ferric tartrate absorptiometry method, but is not limited to these. (Step a)
<Cross-flow filtration>
Fermentation liquid contains various organic matter such as microorganisms, proteins, polysaccharides, and lipids, and fouling is likely to progress in dead-end filtration, which filters all of the liquid supplied to the membrane module. Therefore, cross-flow filtration, in which the raw fermentation liquid is flowed parallel to the membrane surface to perform filtration while cleaning the membrane surface, is suitable (step (a-1) of cross-flow filtering the fermentation liquid through the membrane module).

As shown in FIG. 1, when performing cross-flow filtration, the fermentation liquid (raw liquid) is supplied from the raw liquid tank 1 to the membrane module 2 by the raw liquid pump 3, and a portion of the raw liquid is returned to the raw liquid tank 1. The filtrate filtered by the separation membrane of the membrane module 2 is sent to the filtrate tank 6. Note that in order to improve the filtration operability of the fermentation liquid, a process such as centrifugation may be performed before the membrane filtration.

As mentioned above, in cross-flow filtration, the raw liquid is circulated between the raw liquid tank 1 and the membrane module 2 because the raw liquid flows parallel to the membrane surface. In this case, it is preferable to adjust the circulation flow rate of the raw liquid to a constant membrane surface linear velocity. The membrane surface linear velocity refers to the circulation flow rate of the raw liquid divided by the cross-sectional area through which the circulation flow in the membrane module flows. The membrane surface linear velocity may be set appropriately according to the properties of the raw liquid, but it is preferable to set it to 0.1 m/s or more and 3.0 m/s or less, and more preferably 0.3 m/s or more and 2.5 m/s or less. If the membrane surface linear velocity is less than 0.1 m/s, sufficient cleaning effect may not be obtained. In addition, if the membrane surface linear velocity exceeds 3.0 m/s, it is not preferable because the power cost becomes high. The flow rate of the raw liquid can be controlled by the raw liquid pump 3 and the circulation control valve 4.

The filtration flux during cross-flow filtration may be appropriately set according to the properties of the raw liquid, but is preferably 0.1 ^{m 3} /m ² /d or more and 2.5 ^{m 3} /m ² /d or less, and more preferably 0.3 ^{m 3} /m ² /d or more and 2.0 m ³ /m ² /d or less. If the filtration flux is less than 0.1 m ³ /m ² /d, the number of required membrane modules increases, resulting in high costs. On the other hand, if the filtration flux exceeds 2.5 m ³ /m ² /d, fouling of the separation membrane may progress rapidly, which is not preferable. The flow rate of the filtrate can be controlled by the raw liquid pump 3, the circulation control valve 4, and the filtrate control valve 5.

In filtration using a separation membrane, the fouling state of the separation membrane can be judged from the transmembrane pressure, which is the pressure on the raw liquid side of the separation membrane minus the pressure on the filtrate side. In the case of dead-end filtration, the transmembrane pressure can be calculated from the pressure gauge 21 on the raw liquid inlet side of the membrane module 2 and the pressure gauge 23 on the filtrate outlet side of the membrane module 2. At the same filtration flux, the transmembrane pressure increases as fouling of the separation membrane progresses. However, in the case of cross-flow filtration, the pressure loss is large when the raw liquid passes through the raw liquid side flow path of the membrane module 2 and flows back to the raw liquid tank 1, and the above-mentioned calculation method also includes the pressure loss of the raw liquid side flow path, making it difficult to calculate the transmembrane pressure appropriately. Therefore, the transmembrane pressure can be calculated from the difference between the average value of the pressure gauge 21 on the raw liquid inlet side of the membrane module 2 and the pressure gauge 22 on the outlet side and the value of the pressure gauge 23 on the filtrate outlet side of the membrane module 2.

In the cross-flow filtration of step a, when the fouling of the separation membrane progresses, the transmembrane pressure difference increases, so it is preferable to move to step b when the membrane is blocked to a certain extent. The timing of moving to step b is preferably when the concentration _Ct of polysaccharides in the raw water is in the range of 1.3 to 15 times the concentration _C0 at the start of filtration. Since polysaccharides tend to accumulate on the membrane surface when the concentration _Ct /concentration _C0 is 1.3 or more, a large cleaning effect can be obtained by washing in the opposite direction to filtration. On the other hand, if the concentration _Ct /concentration _C0 is 15 or less, the fouling of the separation membrane does not progress excessively, and there is a high possibility that the water permeability of the separation membrane will be sufficiently restored by performing chemical washing.

The polysaccharide may be selected appropriately depending on the type of fermentation liquid and the fermentation raw materials, but examples include β-glucan, amylose, amylopectin, pectin, dextran, cellulose, chitin, chitosan, glucomannan, etc.

The concentration of polysaccharides in raw water can be measured by extracting the polysaccharide fraction from the raw water by ethanol precipitation or other methods, then subjecting the polysaccharide fraction to hydrolysis using enzymes or alkali, etc., and analyzing the resulting constituent monosaccharides using high performance liquid chromatography (HPLC), but this is not a limited method.

<Backwashing>
The process a may include a step (a-1) of performing cross-flow filtration, and a step (a-2) of periodically stopping the cross-flow filtration and backwashing the membrane module. When the permeability is restored by backwashing, the filtration time can be extended, and the frequency of chemical washing can be reduced, thereby reducing the operating cost. Steps (a-1) and (a-2) can be repeated alternately any number of times.

Backwashing can be done with filtrate or other liquids such as water, but if you do not want other liquids such as water to get mixed into the raw solution, you can backwash with filtrate, or you can switch by operating a valve and backwash so that the backwash liquid does not get mixed into the raw solution.

The backwash flux during backwashing may be appropriately set according to the properties of the raw liquid and the fouling state of the separation membrane, but is preferably 1.0 ^{m 3} /m ² /d or more and 10.0 ^{m 3} /m ² /d or less, and more preferably 1.5 m ³ /m ² /d or more and 5.0 m ³ /m ² /d or less. If the backwash flux is less than 1.0 m ³ /m ² /d, the cleaning effect is reduced, which is not preferable. Also, if the backwash flux exceeds 10.0 ^{m 3} /m ² /d, the power cost increases and a large amount of liquid is required for backwashing, which is not preferable.

If the separation membrane is a hollow fiber membrane, the filtration in step a is preferably external pressure filtration in which filtration is performed from the outer periphery side of the hollow fiber membrane to the inner periphery side, and the backwashing in step a-2 is preferably performed by flowing a liquid from the inner periphery side of the hollow fiber membrane to the outer periphery side. In this case, the backwashing flux is preferably 4.0 m ³ /m ² /d or more and 10.0 ^{m 3} /m ² /d or less, and more preferably 7.0 m ³ /m ² /d or more and 10.0 ^{m 3} /m ² /d or less, calculated per inner area of the hollow fiber membrane. In step a, if a liquid is flowed from the outside to the inside of the hollow fiber membrane at a high flux, the hollow fiber membrane will be compressed and crushed, but by flowing a liquid from the inside to the outside of the hollow fiber membrane, washing at a high flux can be achieved, and a high washing effect can be obtained.

<Concentrated Liquid Recovery>
If cross-flow filtration is continued in step a, the amount of liquid in the stock solution tank will eventually decrease, making it difficult to circulate the stock solution between the stock solution tank and the membrane module. Here, in order to recover the stock solution remaining in the stock solution tank, piping, and membrane module, dead-end filtration or diafiltration can be performed. In the case of dead-end filtration, for example, water is charged into the stock solution tank 1, the stock solution is supplied to the membrane module 2 by the stock solution pump 3, and dead-end filtration is performed, and the filtration is stopped before the water reaches the membrane module. Dead-end filtration can also be performed by supplying gas to the stock solution tank 1 or piping. In diafiltration, water is charged into the stock solution tank 1 to perform cross-flow filtration, and when the amount of liquid in the stock solution tank 1 decreases, water is charged again and cross-flow filtration is repeated, thereby improving the recovery rate of the stock solution.

(Step b)
<Water rinse>
In step b, the membrane module 2 is rinsed with water, with the purpose of reducing contaminants such as organic matter before the chemical cleaning in step c. In the chemical cleaning, cleaning is performed with a chemical solution having a pH of 10 or more and 14 or less and containing an oxidizing agent, but if a large amount of organic matter remains, the oxidizing agent is consumed and the separation membrane may not be sufficiently cleaned, so the separation membrane is rinsed with water before the chemical cleaning.

As a method of water rinsing, for example, water can be charged into the cleaning liquid tank 10 and cross-flow filtration of the membrane module 2 can be performed to clean the entire inside of the membrane module 2, i.e., both the primary and secondary sides. At this time, water can be pumped from the backwash liquid tank 7 to perform backwashing with water at the same time, thereby further enhancing the cleaning effect. The membrane surface linear velocity, filtration flux, and backwash flux can be performed under the same conditions as in step a.

(Step c)
<Alkaline chemical cleaning>
In step c, the membrane module 2 is washed with a chemical solution to remove fouling components adhering to the separation membrane. Examples of chemical solutions for washing the separation membrane include acids, alkalis, oxidizing agents, surfactants, and chelating agents. It has been found that washing the separation membrane after filtration of the fermentation liquid is effective when using a combination of an oxidizing agent at a pH of 10 to 14 and injecting a chemical solution from the secondary side of the membrane to the primary side.

The pH of the chemical solution is preferably 10 to 14, and more preferably 11 to 13. The pH is adjusted with sodium hydroxide, potassium hydroxide, or the like. The fouling substances in the fermentation liquid are mainly organic matter, and high pH conditions enhance cleaning power, but conditions above pH 14 may cause deterioration of the membrane, membrane module components, piping, etc. On the other hand, a pH below 10 may result in insufficient cleaning effect.

Examples of oxidizing agents include hypochlorites such as sodium hypochlorite, calcium hypochlorite, and potassium hypochlorite, hydrogen persulfates such as sodium hydrogen persulfate and potassium hydrogen persulfate, persulfates such as sodium persulfate and potassium persulfate, percarbonates such as sodium percarbonate and potassium percarbonate, permanganates, ozone, and peroxides such as hydrogen peroxide. In order to obtain a sufficient cleaning effect, however, hypochlorites such as sodium hypochlorite, calcium hypochlorite, and potassium hypochlorite are preferred. These oxidizing agents can be used alone or in combination of two or more.

The oxidizing agent is dissolved in an alkaline aqueous solution having a pH of 10 or more and 14 or less and used as a cleaning solution. The oxidizing agent may be used in high concentration from the beginning, or in the order of low concentration and high concentration. Specifically, there is a method of performing step (c-1) alone in which a chemical solution A having a pH of 10 or more and 14 or less containing 0.01% or more of an oxidizing agent is injected from the secondary side to the primary side of the membrane, or a method of performing step (c-2) in sequence in which a chemical solution B having a pH of 10 or more and 14 or less containing less than 0.01% of an oxidizing agent is injected from the secondary side to the primary side of the membrane before c-1. In the present invention, when a chemical solution having a pH of 10 or more and 14 or less containing an oxidizing agent is injected from the secondary side to the primary side of the membrane of the membrane module, a synergistic effect is achieved to obtain a high cleaning effect in a short time. Furthermore, since the initial alkaline chemical solution injected from the secondary side to the primary side of the membrane of the membrane module in each of steps c-1 and c-2 contains a large amount of dirt components, it is possible to prevent reattachment to the membrane and obtain a higher cleaning effect by discarding it without reusing it. The recommended disposal time is about 5 to 10 minutes.

The concentration of the oxidizing agent used in the chemical cleaning is preferably 0.01% to 1%, more preferably 0.1% to 0.5%, when the chemical cleaning is performed in one step. On the other hand, when the chemical cleaning is performed in two or more steps, the concentration of the oxidizing agent in at least the first cleaning step is preferably less than 0.01%, and in at least one cleaning step from the second step onwards, the oxidizing agent concentration is preferably 0.01% to 1%, more preferably 0.1% to 0.5%. In step (c-1) where the oxidizing agent is used at a high concentration, a sufficient cleaning effect can be obtained by setting the oxidizing agent concentration to 0.01% or more. In addition, by setting the oxidizing agent concentration to 1% or less, a balance can be achieved between cost and cleaning effect. In addition, when chemical cleaning is performed in two or more steps, by performing cleaning with an oxidizing agent concentration of less than 0.01% in at least one step, the foulant components attached to the membrane can be removed without being denatured by the oxidizing agent, and then by performing further cleaning with an oxidizing agent concentration of 0.01% or more, the cleaning effect of the oxidizing agent can also be obtained. Whether chemical cleaning is performed in one step or two steps can be appropriately changed depending on the properties of the raw solution, the fouling components of the separation membrane, and the fouling state.

The temperature of the chemical solution is preferably 20°C or higher and 70°C or lower, and more preferably 40°C or higher and 60°C or lower. If the temperature of the chemical solution is below 20°C, it may not provide sufficient cleaning effect. On the other hand, if the temperature exceeds 70°C, it is not preferable because it accelerates the decomposition of the oxidizing agent.

The injection flux of the chemical solution from the secondary side to the primary side may be set appropriately depending on the properties of the raw liquid and the fouling state of the separation membrane, but is preferably 1.5 to 10 times the filtration flux during cross-flow filtration in step a, and more preferably 2.0 to 4.0 times. If the injection flux of the chemical solution from the secondary side to the primary side is less than 1.5 times, the cleaning effect will be reduced, which is not preferable. Furthermore, if the injection flux of the chemical solution from the secondary side to the primary side is more than 10 times, the power cost will increase and a large amount of chemical solution will be required, which is not preferable.

If the separation membrane is a hollow fiber membrane, the filtration in step a is preferably an external pressure filtration in which filtration is performed from the outer periphery side to the inner periphery side of the hollow fiber membrane, and the water rinse in step b and the chemical washing in step c are preferably performed by flowing water or a chemical solution from the inner periphery side to the outer periphery side of the hollow fiber membrane. In this case, the backwash flux, calculated per inner area of the hollow fiber membrane, is preferably 4.0 ^{m 3} /m ² /d or more and 10.0 ^{m 3} /m ² /d or less, and more preferably 7.0 ^{m 3} /m ² /d or more and 10.0 ^{m 3} /m ² /d or less. In steps b and c, if a liquid is flowed from the outside to the inside of the hollow fiber membrane at a high flux, the hollow fiber membrane will be compressed and crushed, but by flowing a liquid from the inside to the outside of the hollow fiber membrane, washing at a high flux can be achieved and a high washing effect can be obtained.

(Step d)
<Water rinse>
In step d, the membrane module 2 is rinsed with water, the purpose of which is to wash away the chemical solution remaining after the chemical solution cleaning in step c. The water rinsing method may be the same as that in step b.

(Step e)
<Acid chemical cleaning>
In step e, an acidic chemical solution c is injected from the secondary side to the primary side of the membrane of the membrane module 2 to remove fouling components remaining in the separation membrane that could not be washed in step c.
The pH of the chemical solution c is preferably 0 or more and 4.0 or less, more preferably 1.0 or more and 3.0 or less. The pH is adjusted with at least one compound selected from the group consisting of hydrochloric acid, nitric acid, sulfuric acid, phosphoric acid, formic acid, acetic acid, propionic acid, butyric acid, citric acid, oxalic acid, ascorbic acid, and lactic acid. The lower the pH, the higher the cleaning power, but under conditions where the pH is less than 0, the membrane, membrane module components, piping, etc. may deteriorate. On the other hand, under conditions where the pH is more than 4, the acidity is insufficient, and the cleaning effect is insufficient.

The temperature of the chemical solution is preferably 20°C or higher and 70°C or lower, and more preferably 40°C or higher and 60°C or lower. If the temperature of the chemical solution is below 20°C, it may not provide sufficient cleaning effect. On the other hand, if it exceeds 70°C, the membrane, membrane module components, piping, etc. may deteriorate, which is not preferable.

The injection flux of the chemical solution from the secondary side to the primary side may be set appropriately according to the properties of the raw liquid and the fouling state of the separation membrane, but is preferably 1.5 times or more and 10 times or less, and more preferably 2.0 times or more and 4.0 times or less, of the filtration flux during cross-flow filtration in step a. If the injection flux of the chemical solution from the secondary side to the primary side is less than 1.5 times, the cleaning effect will be reduced, which is not preferable. Furthermore, if the injection flux of the chemical solution from the secondary side to the primary side exceeds 10 times, the power cost will be high and a large amount of chemical solution will be required, which is not preferable. Furthermore, since the initial acid chemical solution injected from the secondary side to the primary side of the membrane of the membrane module contains a large amount of contaminants, by disposing of it without reusing it, it is possible to prevent re-adhesion to the membrane and obtain a higher cleaning effect. It is preferable to dispose of it for about 5 to 10 minutes.

If the separation membrane is a hollow fiber membrane, the filtration in step a is preferably an external pressure filtration in which filtration is performed from the outer periphery side to the inner periphery side of the hollow fiber membrane, and the water rinse in step b and the chemical washing in step c are preferably performed by flowing water or a chemical solution from the inner periphery side to the outer periphery side of the hollow fiber membrane. In this case, the backwash flux, calculated per inner area of the hollow fiber membrane, is preferably 4.0 m ³ /m ² /d or more and 10.0 ^{m 3} /m ² /d or less, and more preferably 7.0 m 3 /m 2 /d or more and 10.0 ^{m 3} /m ² /d or less. In steps b and c, if a liquid is flowed from the outside to the inside of the hollow fiber membrane at a high flux, the hollow fiber membrane will be compressed and crushed, but by flowing a liquid from the inside to the outside of the hollow fiber membrane, washing at a high flux can be achieved and a high washing effect can be obtained.

(Step f)
<Water rinse>
In step f, the membrane module 2 is rinsed with water, the purpose of which is to wash away the chemical solution remaining after the chemical solution cleaning in step e. The water rinsing method may be the same as that in step b.

<Cleaning aid>
It is preferable that the above-mentioned chemical solutions A and B contain, in addition to an alkali and an oxidizing agent, a surfactant, an enzyme, a chelating agent, or a polyphosphate as a cleaning auxiliary, and the above-mentioned chemical solution C contains, in addition to an acid, a surfactant such as a nonionic surfactant, an anionic surfactant, or a cationic surfactant, with nonionic surfactants being particularly preferable, EDTA being the chelating agent, and protease, mannanase, or β-glucanase being particularly preferable as the enzyme.

Examples of nonionic surfactants include polyoxyethylene alkyl ethers, polyoxyethylene alkylphenyl ethers, polysorbates (polyoxyethylene sorbitan fatty acid esters), polyoxyethylene polystyrylphenyl ethers, polyoxyethylene-polyoxypropylene glycols, polyoxyethylene-polyoxypropylene alkyl ethers, polyhydric alcohol fatty acid partial esters, polyoxyethylene polyhydric alcohol fatty acid partial esters, polyoxyethylene fatty acid esters, polyglycerin fatty acid esters, polyoxyethylated castor oil, etc. Nonionic surfactants can be used alone or in combination of two or more. It is believed that the hydrophobic parts of the molecules of fouling substances such as proteins, polysaccharides, and lipids adhere to the separation membrane. Therefore, the interaction between the hydrophobic part (lipophilic part) of the surfactant and the hydrophobic part of the fouling substance, and the solubility of the surfactant in water affect the cleaning properties. The HLB, which represents the hydrophilic-lipophilic balance of the nonionic surfactant, is preferably 12 to 18, and more preferably 13 to 17. When the HLB is in this range, the surfactant has an excellent balance of hydrophilicity and lipophilicity, enhancing cleaning effectiveness.

The concentration of the cleaning aid used in chemical cleaning is preferably 0.05% to 3%, and more preferably 0.1% to 1%. In addition, the critical micelle concentration of surfactants varies depending on the type, but it is preferable to use a concentration equal to or higher than the critical micelle concentration in order to achieve a sufficient cleaning effect.

<Membrane module>
The type of membrane module used in the present invention is not particularly limited, but a hollow fiber membrane module is preferred because it generally has a large specific surface area, can filter a large amount of liquid per unit time, and can be backwashed.

<Separation membrane>
The structure of the separation membrane is not particularly limited, but examples include a symmetric membrane with a uniform pore size overall, an asymmetric membrane with a pore size that varies in the thickness direction of the membrane, and a composite membrane having a support layer for maintaining strength and a separation functional layer for separating the target substance.

The separation membrane preferably satisfies one or more of the following (A) to (D), and more preferably satisfies all of them.
(A) The separation membrane contains a hydrophobic polymer, the pore volume V2 of pores with a pore size of less than 100 nm is 0.02 ^cm3 /g or more and less than 0.1 ^cm3 /g, the pore volume V1 of pores with a pore size of 100 nm or more is 0.1 ^cm3 /g or more, and the ratio V1/V2 of the pore volume V2 to the pore volume V1 is 3 or more and 60 or less.
(B) The separation membrane has a breaking elongation of 38% or more.
(C) The shrinkage rate of the hollow fiber membrane length when the separation membrane is heat-treated in a water vapor atmosphere at 125° C. for 20 hours is 0.5% or more.
(D) The separation membrane contains a hydrophilic polymer, the hydrophilic polymer contains a copolymer composed of N types of monomer units 1, 2, ... i ... N, and the hydration energy density of the hydrophilic polymer calculated based on the following formula (1) is 40 to 70 cal mol ^-1 Å ^-3 .

[In formula (1), the hydration energy of any monomer unit i is the absolute value of the energy of the monomer unit i in water minus the energy of the monomer unit i in a vacuum, N is the total number of monomer species constituting the copolymer, i is an integer of 1 to N, and N is an integer of 2 or more.]
Details are explained below.

(1) Hydrophobic polymer In the present invention, the hydrophobic polymer is not limited as long as it dissolves in the hydrophobic organic solvent used in the film formation. Examples of hydrophobic polymers include polyesters such as polytetrafluoroethylene, polyvinylidene fluoride, polylactic acid, polyhydroxyacetic acid, polycaprolactone, and polyethylene adipate, polyurethanes, poly(meth)acrylic acid esters, polyvinyl acetals, polyamides, polystyrenes, polysulfones, cellulose derivatives, polyphenylene ethers, and polycarbonates, and other single components, polymer alloys and blends of two or more selected from these, or copolymers of monomers forming the above polymers, but are not limited to the above examples. Among these, resin components with excellent heat resistance, chemical resistance, etc. include fluorine-based resins such as polytetrafluoroethylene and polyvinylidene fluoride, or sulfone-based resins such as polysulfone and polyethersulfone. Among these, vinylidene fluoride resins are particularly preferred because they have high compatibility with solvents and can easily produce a uniform manufacturing stock solution. The vinylidene fluoride resin means a resin containing at least one of a vinylidene fluoride homopolymer and a vinylidene fluoride copolymer. The vinylidene fluoride resin may contain a plurality of kinds of vinylidene fluoride copolymers.

A vinylidene fluoride copolymer is a polymer having a vinylidene fluoride residue structure, and is typically a copolymer of vinylidene fluoride monomer and other fluorine-based monomers. Examples of such copolymers include copolymers of vinylidene fluoride and one or more monomers selected from vinyl fluoride, tetrafluoroethylene, hexafluoropropylene, and trifluorochloroethylene.

The weight-average molecular weight of the vinylidene fluoride resin may be appropriately selected depending on the required strength and water permeability of the separation membrane, but as the weight-average molecular weight increases, the water permeability decreases, and as the weight-average molecular weight decreases, the strength decreases. For this reason, the weight-average molecular weight is preferably 50,000 to 1,000,000. For water treatment applications in which the separation membrane is exposed to chemical cleaning, the weight-average molecular weight is preferably 100,000 to 700,000, and more preferably 150,000 to 600,000.

The separation membrane preferably contains vinylidene fluoride resin as a main component, and the proportion of vinylidene fluoride resin in the hollow fiber membrane is preferably 80% by weight or more, more preferably 90% by weight or more, and even more preferably 95% by weight or more.

(2) Average pore size The separation membrane of the present invention preferably has pores with an average pore size in the range of 10 to 10,000 nm. The average pore size of the pores is more preferably in the range of 50 to 5,000 nm, and even more preferably in the range of 100 to 3,000 nm. When the average pore size is 100 nm or more, the permeation resistance is unlikely to increase, and the pressure required for filtration can be prevented from increasing. For example, when filtering a liquid containing microbial particles, the destruction of microbial particles, clogging of the membrane surface due to deformation, and a decrease in filtration efficiency can be prevented. Furthermore, when the average pore size is 3,000 nm or less, sufficient fractionation can be obtained.

The pore size can be measured by mercury intrusion porosimetry. In mercury intrusion porosimetry, pressure p is applied to the mercury so that it is forced into the pores of the separation membrane, and the volume change dV of the mercury in the cell relative to the pressure increment dp is measured, and the pore distribution function F(r) is calculated from the following formula (A):

(Here, r is the pore radius, σ is the surface tension of mercury (0.480 N/m), and θ is the contact angle (140°).)
The average pore size can be calculated by the following formula (B).

(3) Pore volume Pores with a pore size of less than 100 nm are mainly voids within the solid parts that form the separation membrane, and also include fine voids between the solid parts. Pores with a pore size of 100 nm or more are large voids between the solid parts that form the separation membrane.
The fact that the voids within the solid portion are small means that the polymer molecules that make up the solid portion are more densely arranged, and a separation membrane formed from such a solid portion has high breaking strength and breaking elongation.
In addition, pores with a diameter of less than 100 nm can become starting points for the adhesion of contaminating components such as proteins in the fermentation liquid, resulting in clogging, so the small voids in the solid part provide high filtering properties for the fermentation liquid.In addition, the reduced adhesion of flavor components means that there is less odor absorption when switching between types of fermentation liquid to be filtered.

On the other hand, if the pore volume of pores with diameters of less than 100 nm is extremely small, the flow paths in the separation membrane will be poorly branched. If the fine gaps between solids disappear due to partial blockage, the pore volume of pores with diameters of less than 100 nm will often become too small.

Here, in the filtration of liquids that are prone to clogging, such as fermentation liquid, separation membranes have the characteristic that the membrane area of the permeable flow path, i.e. the effective membrane area, is significantly reduced due to localized clogging, and the effective membrane area decreases at an accelerated rate as filtration progresses. Therefore, high filtering properties are imparted by leaving a certain amount of fine voids.

In addition, if the volume of the gaps between the solid parts, i.e., the volume of pores with a pore size of 100 nm or more, is large, the gaps between the solid parts that form the separation membrane are large, and the membrane exhibits high pure water permeability. In addition, since there is little inclusion of chemical contact surfaces during cleaning, the effect of chemical surface modification on the inside of the membrane structure is minimal, and the membrane has excellent chemical resistance.

On the other hand, when the pore volume of pores having a diameter of 100 nm or more is equal to or less than a certain value, good mechanical strength can be obtained. From the viewpoint of mechanical strength, the pore volume of pores having a diameter of 100 nm or more is preferably 0.1 cm ³ /g or more and less than 1.0 cm ³ /g.

That is, in a separation membrane, when the pore volume V1 of pores with a pore diameter of less than 100 nm is 0.02 cm ³ /g or more and less than 0.1 cm ³ /g, the pore volume V2 of pores with a pore diameter of 100 nm or more is 0.1 cm ³ /g or more, and the ratio V1/V2 of the pore volume V2 to the pore volume V1 is 3 or more and 60 or less, the separation membrane has excellent mechanical strength and durability against chemical cleaning, and is less likely to clog during filtration of a fermentation broth and has little adhesion of flavor components during filtration.

The pore volume V1 of pores having a diameter of 100 nm or more is preferably 0.3 cm ³ /g or more, and more preferably 0.35 cm ³ /g or more. The pore volume V1 of pores having a diameter of 100 nm or more is preferably 0.5 cm ³ /g or less, and more preferably 0.45 cm ³ /g or less.

The pore volume V2 of pores with a pore diameter of less than 100 nm is preferably 0.020 cm ³ /g or more, more preferably 0.025 cm ³ /g or more, and even more preferably 0.030 cm ³ /g or more. The pore volume V2 of pores with a pore diameter of less than 100 nm is preferably 0.1 cm ³ /g or less, more preferably 0.075 cm ³ /g or less, and even more preferably 0.050 cm ³ /g or less.

The ratio V1/V2 of the pore volume V2 to the pore volume V1 is preferably 3 or more. In addition, V1/V2 is preferably 60 or less, more preferably 20 or less, and even more preferably 15 or less.

The pore volume of pores with a pore diameter of 100 nm or more may be measured by the above-mentioned mercury intrusion method, and the pore volume of pores with a pore diameter of less than 100 nm may be measured by a gas adsorption method. In the gas adsorption method, for example, after removing water and the like adsorbed to the separation membrane, nitrogen gas adsorption measurement is performed at liquid nitrogen temperature, and the pore volume V ( ^cm3 /g) of the separation membrane may be measured from the amount of nitrogen gas adsorbed (in liquid equivalent) at a relative pressure (adsorption equilibrium pressure/saturated vapor pressure at liquid nitrogen temperature) of 0.99.

Within the above range, separation membranes in which the pore volume of pores with pore sizes smaller than the average pore size accounts for 60% or more of the total pore volume, as measured by mercury intrusion porosimetry, have many branched flow paths and distribute the load caused by localized clogging to other flow paths, making them suitable for filtering liquids that are prone to clogging, such as fermentation liquids.

(4) Spherical structure In the separation membrane of the embodiment of the present invention, the solid portion preferably has a spherical structure. The spherical structure is a structure in which spheres are continuous. By having such a structure inside the separation membrane, the gaps between the solid portions (spheres) are less likely to shrink, and high pure water permeability can be maintained. In addition, the spherical structure can achieve higher strength and higher water permeability than conventional mesh-like structures.
The spherical structure preferably contains vinylidene fluoride resin, and the separation membrane preferably has a spherical structure and is a hollow fiber membrane. A part of a hollow fiber membrane is illustrated in Fig. 4. When the separation membrane is a hollow fiber membrane 30 shown in Fig. 4, the inside of the separation membrane refers to the part excluding the outer surface 31, that is, the substantial inside and/or inner surface 32 of the hollow fiber membrane.

A portion of a spherical structure is shown diagrammatically in FIG. 5. In the spherical structure 40 of FIG. 5, multiple spherical bodies 41 are connected. The spherical bodies 41 are approximately spherical or approximately ellipsoidal. As shown in FIG. 5, since the spherical body 241 is connected to other spherical bodies 41, it is not possible to observe the entire spherical or ellipsoidal surface. However, it is possible to estimate the diameter, major axis, and minor axis from the shape that appears in the external shape.

The connections between the spheres may be formed by direct bonding between the spheres, or by non-spherical portions (also called constrictions) 42 between the spheres. The gaps 43 between the spheres are the gaps between the solid portions described above. As described above for the solid portions, minute gaps also exist within the spheres.

The average diameter of the spherical bodies is in the range of 0.5 to 15 μm, preferably in the range of 0.6 to 10 μm, and more preferably in the range of 0.8 to 8 μm. The diameter of the spherical bodies is determined by taking a photograph of the cross section and/or inner surface of the hollow fiber membrane using a scanning electron microscope or the like at a magnification at which the spherical structure can be clearly confirmed, measuring the diameters of 10 or more, preferably 20 or more, arbitrary spherical bodies, and averaging them. It is also preferable to analyze the photograph with an image processing device and determine the average of the equivalent circle diameter. The density of the spherical bodies is preferably in the range of 10 ³ to 10 ⁸ pieces/mm ² , more preferably in the range of 10 ⁴ to 10 ⁶ pieces/mm ^2. When the density of the spherical bodies is 10 ³ pieces/mm ² or more, the high strength required for fermentation liquid treatment can be realized, and when it is 10 ⁸ pieces/mm ² or less, high pure water permeability can be obtained.

The density of the spheres is measured by taking a photograph in the same manner as for measuring the diameter, and counting the number of spheres per unit area. The circularity ratio (minor diameter/major diameter) is preferably 0.5 or more, more preferably 0.6 or more, and even more preferably 0.7 or more.

(5) Hydrophilic Polymer The hollow fiber membrane preferably contains a copolymer composed of N types of monomer units, and the hydration energy density of the hydrophilic polymer, calculated based on the above formula (1), is 40 to 70 cal mol ^-1 Å ^-3 .

Furthermore, the volume fraction of the monomer unit having the highest hydration energy density is 35 to 90%, and the difference in hydration energy density is 10 to 100 cal·mol ⁻¹ ·Å ⁻³ .

"Monomer unit" refers to a repeating unit in a homopolymer or copolymer obtained by polymerizing a monomer. For example, a hydrophobic monomer unit refers to a repeating unit in a homopolymer or copolymer obtained by polymerizing a hydrophobic monomer.

"Containing N types of monomer units" means that the copolymer contains N types of repeating units (i.e., monomer units). N is an integer of 2 or more, and monomer unit i refers to any one of the N types of monomer units. For example, a vinylpyrrolidone/vinyl decanoate random copolymer contains two types of monomer units, vinylpyrrolidone and vinyl decanoate.

"Copolymer" means a polymer composed of two or more types of monomer units.

"Hydration energy" refers to the energy change that a system undergoes when a solute is placed in an aqueous solution.

"Hydration energy of a monomer unit" means the absolute value of the energy of the monomer unit in water minus the energy of the monomer unit in a vacuum.

"Hydration energy density" refers to the hydration energy per unit volume. For example, in the case of a monomer, it is the numerical value defined by the following formula (2).

Hydration energy density (cal·mol ⁻¹ ·Å ⁻³ ) of any monomer unit i=(hydration energy of the monomer unit i)/(volume of the monomer unit i) (2)
The "difference in hydration energy density" refers to a numerical value defined by the following formula (3).

Difference in hydration energy density (cal·mol ⁻¹ ·Å ⁻³ )=(hydration energy density of monomer unit having the largest hydration energy density)−(hydration energy density of monomer unit having the smallest hydration energy density) (Equation (3))
The term "monomer unit j having the highest hydration energy density" refers to the monomer unit having the highest hydration energy density defined by the formula (2) among the monomer units constituting the hydrophilic polymer.

"Monomer unit k with the smallest hydration energy density" refers to the monomer unit that has the smallest hydration energy density defined by formula (2) among the monomer units that make up the hydrophilic polymer.

For the molecular model of the monomer unit, for example, if the monomer unit has a structure represented by the chemical formula (I) below, the structure represented by the chemical formula (II) below is used as the calculation target. That is, a structure is used in which the carbon terminal on the side to which the side chain R is bonded is terminated with a methyl group ((a) in the following formula (II)), and the carbon terminal on the side to which the side chain R is not bonded is terminated with a hydrogen atom ((b) in the following formula (II)).

The energy of the monomer unit in the above formula (1) in vacuum and in water can be calculated by the following method.

First, the molecular model of the monomer unit is structurally optimized. Density functional theory is used for structural optimization. B3LYP is used as the functional, and 6-31G(d,p) is used as the basis function. In addition, opt is set as a keyword to be written in the input file.

Next, the energy in vacuum and the energy in water are calculated for the optimized structure. The energy in vacuum is calculated using density functional theory. B3LYP is used as the functional and 6-31G(d,p) is used as the basis function. The energy in water is calculated using density functional theory. B3LYP is used as the functional and 6-31G(d,p) is used as the basis function. Furthermore, to calculate the energy in water, a continuous dielectric model is used, and the following keywords are used:
SCRF=(PCM, G03Defaults, Read, Solvent=Water)
Radii=UAHF
Alpha=1.20
The hydration energy of the monomer unit is determined by calculating the SCF energy in vacuum and in water, where the SCF energy is the value of E written on the line labeled "SCF Done:".

The energy calculations are performed using Gaussian09 (registered trademark), a quantum chemistry calculation software made by Gaussian.

In the hydrophilic polymer, the hydration energy density of the hydrophilic polymer is defined based on the above formula (1).

The volume of the monomer unit in the formula (1) is the volume of the structure optimized. The volume of the monomer unit can be calculated, for example, by using the Connollysurface method of MaterialsStudio (registered trademark) manufactured by BIOVIA. In this case, the parameters to be set are as follows.
Gridresolution=Coarse
Grid interval=0.75Å
vdWfactor=1.0
Connolly radius = 1.0 Å
The hydration energy density of the hydrophilic polymer is 40 to 70 cal·mol ⁻¹ ·Å ⁻³ , preferably 43 to 60 cal·mol ⁻¹ ·Å ⁻³ , more preferably 45 to 55 cal·mol ⁻¹ ·Å ⁻³ . Any of the preferred lower limits can be combined with any of the preferred upper limits.

The total number N of monomer species constituting the hydrophilic polymer is not particularly limited as an upper limit, but is preferably 2 to 5, more preferably 2 to 3, and most preferably 2.

It is believed that the hydration energy density of the entire hydrophilic polymer is 40 cal·mol ^-1 ·Å ^-3 or more and 70 cal·mol ^-1 ·Å ^-3 or less, so that the structures of the hydrophilic polymer, the adsorbed water of the hydrophilic polymer, and the impurities in the fermentation liquid and the adsorbed water of the impurities in the fermentation liquid are stabilized. As a result, the electrostatic interaction or hydrophobic interaction between the hydrophilic polymer present on the surface of the hollow fiber membrane and the impurities in the fermentation liquid is reduced, and the adhesion of the impurities in the fermentation liquid to the hollow fiber membrane is suppressed.

The molar fraction of the monomer units of formula (1) above and formula (4) below is the ratio of the number of each type of monomer unit to the total number of monomer units constituting the hydrophilic polymer. As described below, it is measured with a nuclear magnetic resonance (NMR) device and calculated from the peak area. If the molar fraction cannot be calculated by NMR measurement due to overlapping peaks, etc., the molar fraction may be calculated by elemental analysis.

[In the formula (4), N and i are defined as above.]
As described below, the hollow fiber membrane is formed by phase separation of a hydrophobic polymer and the subsequent introduction of a hydrophilic polymer, and has an aggregate of spherical structures composed of hydrophobic polymers and hydrophilic polymers attached to the surfaces of the spherical structures.

The number average molecular weight of the hydrophilic polymer is preferably 2,000 or more, more preferably 3,000 or more, in order to suppress adhesion of impurities in the fermentation liquid. In addition, for high incorporation efficiency into the membrane, the number average molecular weight of the hydrophilic polymer is preferably 1,000,000 or less, more preferably 200,000 or less, and even more preferably 100,000 or less.

In the hydrophilic polymer, the volume fraction of the monomer unit having the highest hydration energy density calculated based on the formula (2) (referred to as monomer unit j for convenience of explanation) is 35% to 90%, preferably 40% to 80%, more preferably 40% to 75%, and even more preferably 40% to 70%. Any of the preferred lower limits can be combined with any of the preferred upper limits.

When the volume fraction is within the above range, it is believed that the interaction between the hydrophilic polymer present on the surface of the hollow fiber membrane and the adsorbed water of the hydrophilic polymer, on the impurities in the fermentation liquid and the adsorbed water of the impurities in the fermentation liquid, is of an appropriate magnitude due to the effects of both the hydrophilic monomer units and the hydrophobic monomer units, and as a result, adhesion of impurities in the fermentation liquid is suppressed.

In addition, in the hydrophilic polymer, the difference in hydration energy density is calculated by the following formula (5).
(difference in hydration energy density) cal·mol ^-1 ·Å ^-3 =
(hydration energy density of monomer unit j having the highest hydration energy density)−(hydration energy density of monomer unit k having the lowest hydration energy density) Equation (5)
The difference in hydration energy density is from 10 to 100 cal·mol ⁻¹ ·Å ⁻³ , preferably from 10 to 80 cal·mol ⁻¹ ·Å ⁻³ , and more preferably from 10 to 60 cal·mol ⁻¹ ·Å ⁻³ .

When the difference in hydration energy density is within the above range, it is believed that the hydrophilic monomer units of the hydrophilic polymer present on the surface of the hollow fiber membrane play a role in retaining the adsorbed water, and the hydrophobic monomer units play a role in controlling the mobility of the adsorbed water. As a result, it is believed that the interaction between the hydrophilic polymer present on the surface of the material and the adsorbed water of the hydrophilic polymer on the impurities in the fermentation liquid and the adsorbed water of the impurities in the fermentation liquid is of an appropriate magnitude, and as a result, adhesion of impurities in the fermentation liquid is suppressed.

The two or more types of monomer units are preferably hydrophobic monomer units and hydrophilic monomer units.

"Hydrophobic monomer unit" means a monomer unit having a smaller hydration energy density than a hydrophilic monomer unit, and for example, a repeating unit in a homopolymer or copolymer obtained by polymerizing a monomer selected from the group consisting of vinyl carboxylate, methacrylic acid ester, acrylic acid ester, and styrene derivative is preferably used. Among these, a repeating unit in a homopolymer obtained by polymerizing vinyl carboxylate or a copolymer obtained by copolymerizing vinyl carboxylate is more preferable, since it is easy to balance with the hydrophilic monomer unit and it is easy to control the mobility of the adsorbed water present on the material surface, and a repeating unit in a homopolymer obtained by polymerizing vinyl carboxylate is even more preferable.

"Hydrophilic monomer unit" means a monomer unit having a larger hydration energy density than a hydrophobic monomer unit, and for example, a repeating unit in a homopolymer or copolymer obtained by polymerizing a monomer selected from the group consisting of allylamine, vinylamine, N-vinylamide, N-vinyllactam, and N-acryloylmorpholine is preferably used. Among these, a repeating unit in a homopolymer obtained by polymerizing N-vinyllactam or a copolymer obtained by copolymerizing N-vinyllactam is preferred, since it does not have too strong an interaction with the adsorbed water present on the material surface and is easy to balance with the hydrophobic monomer unit, and a repeating unit in a homopolymer obtained by polymerizing N-vinyllactam is more preferred. Among these, a repeating unit in a homopolymer obtained by polymerizing vinylpyrrolidone or a copolymer obtained by copolymerizing vinylpyrrolidone is even more preferred, and a homopolymer obtained by polymerizing vinylpyrrolidone is most preferred.

In addition, other monomers, for example, monomers containing reactive groups such as glycidyl groups, may be copolymerized to the extent that the action and function of the hydrophilic polymer is not inhibited, i.e., to the extent that the above [8] is satisfied.

Examples of the arrangement of the hydrophilic monomer units and the hydrophobic monomer units in the hydrophilic polymer include graft copolymers, block copolymers, alternating copolymers, random copolymers, etc. Among these, block copolymers, alternating copolymers, and random copolymers are preferred in terms of their high adhesion suppression function against impurities in the fermentation liquid, and random copolymers and alternating copolymers are more preferred in terms of having an appropriate balance between hydrophilicity and hydrophobicity in one molecule. The reason why block copolymers, alternating copolymers, and random copolymers have a higher adhesion suppression function against impurities in the fermentation liquid than graft copolymers, for example, graft copolymers consisting of hydrophilic monomer units in the main chain and hydrophobic monomer units in the side chain, is thought to be because in graft copolymers, the monomer unit portion grafted to the main chain has many opportunities to come into contact with proteins, etc., and therefore the characteristics of the graft chain portion have a greater influence than the characteristics of the copolymer. In addition, alternating copolymers and random copolymers are more preferred than block copolymers in terms of an appropriate balance between hydrophilicity and hydrophobicity, probably because the characteristics of each monomer unit are clearly separated in block copolymers.

The hydrophilic polymer can be synthesized, for example, by a chain polymerization method such as a radical polymerization method using an azo-based initiator, but the synthesis method is not limited to this.

The hydrophilic polymer is produced, for example, by the following production method, but is not limited to this method.

A predetermined amount of hydrophilic monomer and hydrophobic monomer are mixed with a polymerization solvent and a polymerization initiator, and the mixture is stirred under a nitrogen atmosphere at a predetermined temperature for a predetermined time to cause a polymerization reaction. The ratio of the hydrophilic monomer and the hydrophobic monomer can be determined according to the molar fraction of the hydrophilic monomer units in the copolymer. The reaction liquid is cooled to room temperature to stop the polymerization reaction, and then poured into a solvent such as hexane. The precipitate that has formed is collected and dried under reduced pressure to obtain a hydrophilic polymer.

The reaction temperature for the above polymerization reaction is preferably 30 to 150°C, more preferably 50 to 100°C, and even more preferably 70 to 80°C.

The pressure for the polymerization reaction is preferably normal pressure.

The reaction time of the above polymerization reaction is appropriately selected depending on the reaction temperature and other conditions, but is preferably 1 hour or more, more preferably 3 hours or more, and even more preferably 5 hours or more. If the reaction time is short, a large amount of unreacted monomer may easily remain in the copolymer. On the other hand, the reaction time is preferably 24 hours or less, and more preferably 12 hours or less. If the reaction time is long, side reactions such as the formation of dimers may easily occur, making it difficult to control the molecular weight.

The polymerization solvent used in the above polymerization reaction is not particularly limited as long as it is a solvent that is compatible with the monomer. For example, ether solvents such as dioxane or tetrahydrofuran, amide solvents such as N,N-dimethylformamide, sulfoxide solvents such as dimethyl sulfoxide, aromatic hydrocarbon solvents such as benzene or toluene, alcohol solvents such as methanol, ethanol, isopropyl alcohol, amyl alcohol or hexanol, or water may be used. However, from the viewpoint of toxicity, it is preferable to use an alcohol solvent or water.

As the polymerization initiator for the above polymerization reaction, for example, a photopolymerization initiator or a thermal polymerization initiator is used. Although a polymerization initiator that generates any of radicals, cations, or anions may be used, a radical polymerization initiator is preferably used since it does not cause decomposition of the monomer. As the radical polymerization initiator, for example, an azo initiator such as azobisisobutyronitrile, azobisdimethylvaleronitrile, or azobis(isobutyrate)dimethyl, or a peroxide initiator such as hydrogen peroxide, benzoyl peroxide, di-tert-butyl peroxide, or dicumyl peroxide is used.

After the polymerization reaction has stopped, the solvent into which the polymerization reaction solution is poured is not particularly limited as long as it is a solvent in which the copolymer precipitates. For example, a hydrocarbon solvent such as pentane, hexane, heptane, octane, nonane, or decane, or an ether solvent such as dimethyl ether, ethyl methyl ether, diethyl ether, or diphenyl ether can be used.

(6) Others When the separation membrane of the present invention is in the form of a hollow fiber membrane, the outer diameter and membrane thickness of the hollow fiber may be determined so that the water permeation rate of the membrane module is a target value, taking into consideration the pressure loss in the longitudinal direction inside the hollow fiber membrane, within a range that does not impair the strength of the membrane. That is, a thicker outer diameter is advantageous in terms of pressure loss, but the number of fibers packed is reduced, which is disadvantageous in terms of membrane area. On the other hand, a thinner outer diameter is advantageous in terms of membrane area, since the number of fibers packed can be increased, but is disadvantageous in terms of pressure loss. In addition, the membrane thickness is preferably thin within a range that does not impair the strength. Therefore, as a rough guide, the outer diameter of the hollow fiber membrane is preferably 0.3 to 3 mm, more preferably 0.4 to 2.5 mm, and even more preferably 0.5 to 2.0 mm. In addition, the membrane thickness is preferably 0.08 to 0.4 times the outer diameter, more preferably 0.1 to 0.35 times, and even more preferably 0.12 to 0.3 times.

The separation membrane of the present invention preferably has substantially no macrovoids. Here, the term "macrovoids" refers to voids having a major axis of 50 μm or more observed in the membrane substance portion in the cross section of the separation membrane. "Substantially no macrovoids" means that the number of voids is 10 or ^less per mm2 in the cross section, more preferably 5 or ^less per mm2, and most preferably no macrovoids at all.

The separation membrane of the present invention has a water permeability at 10 kPa and 25°C in the range of 0.1 to 10 ^m3 / ^m2 ·h, preferably 0.2 to 5 ^m3 / ^m2 ·h, and more preferably 0.4 to 2 ^m3 / ^m2 ·h, and when the separation membrane of the present invention is in the form of a hollow fiber membrane, the breaking strength of the hollow fiber is in the range of 0.3 to 3 kg/fiber, preferably 0.4 to 2.5 kg/fiber, and more preferably 0.5 to 2 kg/fiber, and the breaking elongation is preferably in the range of 20 to 1000%, preferably 40 to 800%, and more preferably 60 to 500%. By being in these ranges, sufficient water permeability is exhibited under normal use conditions and the hollow fiber membrane does not break.

<Method of manufacturing separation membrane>
Next, a method for obtaining a hollow fiber membrane from a polyvinylidene fluoride resin as a hydrophobic polymer among the separation membranes of the present invention will be described, but the present invention is not limited to these manufacturing methods. An example of the manufacturing method for the hollow fiber membrane described above will be described below. The manufacturing method described below includes the following steps:
The method comprises: (a) forming hollow fibers by subjecting a solution containing a hydrophobic polymer to phase separation, and then post-treating the hollow fibers by immersion in a solvent; and (b) introducing a hydrophilic polymer into the hollow fibers.

(a) Step of forming hollow fibers and post-treating them by solvent immersion Methods for producing hollow fiber membranes from polyvinylidene fluoride resins include thermally induced phase separation, non-solvent induced phase separation, melt extraction, and stretching and opening methods. Of these, it is preferable to use the thermally induced phase separation or non-solvent induced phase separation method.

Thermally induced phase separation is a phase separation that occurs when a resin solution dissolved at high temperatures is cooled and solidified, while non-solvent induced phase separation is a phase separation that occurs when a resin solution is brought into contact with a non-solvent and solidified.

When manufacturing hollow fiber membranes using the thermally induced phase separation method, the solvent for the polyvinylidene fluoride resin solution is preferably a poor solvent for the resin, and particularly preferably a poor solvent in which the resin has a relatively high solubility, such as cyclohexanone, isophorone, γ-butyrolactone, alkyl ketones such as dimethyl sulfoxide, or esters.

When a hollow fiber membrane is manufactured using the non-solvent induced phase separation method, the solvent for the polyvinylidene fluoride resin solution is preferably a good solvent for the resin, and examples of such good solvents include lower alkyl ketones such as N-methyl-2-pyrrolidone, dimethylacetamide, dimethylformamide, methyl ethyl ketone, acetone, and tetrahydrofuran, esters, amides, and mixtures thereof. On the other hand, the non-solvent is a non-solvent for the resin, and examples of such non-solvents include water, hexane, pentane, benzene, toluene, methanol, ethanol, carbon tetrachloride, o-dichlorobenzene, trichloroethylene, ethylene glycol, diethylene glycol, triethylene glycol, propylene glycol, butylene glycol, pentanediol, hexanediol, and low molecular weight polyethylene glycol, as well as aliphatic hydrocarbons, aromatic hydrocarbons, aliphatic polyhydric alcohols, aromatic polyhydric alcohols, chlorinated hydrocarbons, or other chlorinated organic liquids and mixtures thereof.

In the thermally induced phase separation method and the non-solvent induced phase separation method, a good solvent or a poor solvent in which the resin has high solubility is used, so that the resin and the solvent can be mixed at the molecular level, and when solidified, the solvent molecules are interposed between the resin molecules. This results in a large amount of inter-molecular voids being formed. Therefore, in order to control the pore volume range to a range that provides excellent filterability for the fermentation liquid, it is preferable to carry out a post-treatment to reduce the inter-molecular voids in the separation membrane obtained by the phase separation method.

A preferred post-treatment for reducing the gaps between molecules is a solvent immersion treatment, in which the resin is immersed for a short time at a constant temperature in a poor solvent in which the solubility of polyvinylidene fluoride resin is relatively high, and only the outer surface structure of the solid portion is dissolved, and then the solution is replaced with a non-solvent for polyvinylidene fluoride resin. The solvent immersion treatment closes the gaps within the solid portion, but is thought to have almost no effect on the gaps between the solid portions, resulting in a pore structure suitable for filtering liquids that are prone to clogging, such as fermentation liquids.

On the other hand, when a separation membrane obtained by thermally induced phase separation or non-solvent induced phase separation is heat-treated at a high temperature near the melting point Tm of the separation membrane, micro-Brownian motion of the polymer becomes active in the amorphous part of the solid part that forms the separation membrane, and then some of it crystallizes, or the crystalline part that melts at a lower temperature than the melting point of the separation membrane in the solid part melts once and then recrystallizes into a crystalline part that melts at a higher temperature, causing the solid part to shrink. Since such crystalline and amorphous parts have tiny voids, the voids are filled by the shrinkage, which is expected to have the effect of reducing the starting points for adhesion of fouling components such as proteins in the fermentation liquid.

However, it is thought that the shrinkage of the entire membrane during heat treatment causes the microvoids between the solid parts to become clogged, resulting in less branching of the flow paths, making it unsuitable for filtering liquids that are prone to clogging, such as fermentation liquid. One method for quantifying the microvoids between the solid parts is to measure the shrinkage rate of the hollow fiber membrane when it is heat treated at a temperature T that satisfies Tm-60°C≦T<Tm-40°C, where Tm is the melting point of the separation membrane.

The melting point Tm of the separation membrane is the peak top temperature when a dry separation membrane is heated at a rate of 10°C/min using a differential scanning calorimetry (DSC) device, and is around 175°C in the case of a hollow fiber membrane made of polyvinylidene fluoride. In the case of a hollow fiber membrane made of vinylidene fluoride resin, if the shrinkage rate of the hollow fiber membrane length when heat-treated for 20 hours in a water vapor atmosphere at 125°C is 0.5% or more, it has sufficient microvoids between solid parts. A more preferable shrinkage rate is 1.0% or more, and even more preferable is 2.0% or more. On the other hand, excessive heat shrinkage can induce thread breakage and deformation inside the module, so it is preferable that it is 25% or less.

There are two main types of phase separation mechanisms in thermally induced phase separation: one is liquid-liquid phase separation, in which a resin solution that is uniformly dissolved at high temperatures separates into a thick phase and a thin phase of the resin as the temperature decreases due to a decrease in the dissolving ability of the solution, and the other is solid-liquid phase separation, in which a resin solution that is uniformly dissolved at high temperatures separates into a polymer solid phase and a polymer dilute solution phase as the temperature decreases due to crystallization of the resin.
In the former method, a three-dimensional network structure is mainly formed, and in the latter method, a spherical structure is formed. In the present invention, for the reasons described above, it is more preferable to form a spherical structure by the latter phase separation mechanism. According to this phase separation mechanism, a spherical structure having spherulites (spherical bodies) containing a lamellar structure and polymer chains connecting one lamellar structure to another lamellar structure is formed. In the case of a spherical structure, since the solid part is bulky, the voids between the solid parts are less likely to shrink, and high pure water permeability can be maintained, which is also a reason why it is preferable. For this reason, it is preferable to select a resin concentration and a solvent that induce solid-liquid phase separation. In addition, it is preferable to use a cooling bath when cooling, and it is preferable to contain the same solvent as the resin solution or a low concentration of a non-solvent for the resin to speed up solidification, etc.

For the above reasons, the method for producing the separation membrane of the embodiment of the present invention preferably employs a solid-liquid phase separation method that uses thermally induced phase separation to form a spherical structure, and the resulting separation membrane is preferably immersed in a solvent.

Here, when the separation membrane of the present invention is in the form of a hollow fiber membrane, the hollow fiber membrane produced using the non-solvent induced phase separation method or the heat induced phase separation method is preferably stretched to expand the voids and improve the pure water permeability. The stretching conditions are preferably a temperature range of 50 to 120°C, more preferably 60 to 100°C, and a stretching ratio of preferably 1.1 to 4 times.

At temperatures below 50°C, it is difficult to achieve stable and uniform stretching, and at temperatures above 120°C, the hollow fiber membrane may soften and the hollow portion may collapse. Stretching is preferably performed in a liquid because it is easy to control the temperature, but it may also be performed in a gas such as steam. Water is the most convenient liquid, but when stretching at temperatures above 90°C, it is also preferable to use low molecular weight polyethylene glycol, etc.

Heat treatment of separation membranes is generally known, but in conventional technology it is used for dimensional stability below the heat treatment temperature. Therefore, the heat treatment temperature is lower than in the present invention, and does not impart excellent filtration performance to the separation membrane for fermentation liquid.

(b) Hydrophilic polymer introduction step The types of hydrophilic polymers are as explained in the section on hollow fiber membranes. Methods for introducing hydrophilic polymers include a method in which an aqueous solution in which a hydrophilic polymer is dissolved is passed through or immersed in the hollow fiber membrane, followed by irradiation or heat treatment to insolubilize the hydrophilic polymer, or a method in which a covalent bond is formed by a chemical reaction with a reactive group present in the hollow fiber membrane.

If the concentration of the hydrophilic polymer aqueous solution is too low, a sufficient amount of hydrophilic polymer will not be introduced to the surface. Therefore, the copolymer concentration in the aqueous solution is preferably 10 ppm or more, more preferably 100 ppm or more, and most preferably 500 ppm or more. However, if the concentration of the hydrophilic polymer in the aqueous solution is too high, there is a concern that the amount of elution from the module will increase, so the copolymer concentration in the aqueous solution is preferably 100,000 ppm or less, and more preferably 10,000 ppm or less. The number average molecular weight of the hydrophilic polymer is measured by gel permeation chromatography (GPC) as described below.

When the hydrophilic polymer is poorly soluble or insoluble in water, the hydrophilic polymer may be dissolved in an organic solvent that does not dissolve the hollow fiber or in a mixed solvent of water and an organic solvent that is compatible with water and does not dissolve the hollow fiber. Specific examples of organic solvents that can be used in the organic solvent or mixed solvent include, but are not limited to, alcohol-based solvents such as methanol, ethanol, and propanol.

In addition, if the proportion of organic solvent in the mixed solvent is high, the hollow fibers may swell, which may cause the pore size of the hollow fiber membrane to change. Therefore, the weight fraction of the organic solvent in the mixed solvent is preferably 60% or less, more preferably 10% or less, and most preferably 1% or less.

For the radiation irradiation, alpha rays, beta rays, gamma rays, X-rays, ultraviolet rays, electron beams, etc. can be used. Here, from the viewpoints of safety and simplicity, radiation methods using gamma rays or electron beams are preferred. The radiation exposure dose is preferably 15 kGy or more, more preferably 25 kGy or more. By using a dose of 15 kGy or more, the hydrophilic polymer can be effectively introduced. In addition, the above-mentioned exposure dose is preferably 100 kGy or less. If the exposure dose exceeds 100 kGy, the copolymer may be prone to three-dimensional crosslinking or decomposition of the ester group portion of the vinyl carboxylate monomer unit.

Antioxidants may be used to suppress crosslinking reactions during radiation exposure. Antioxidants refer to substances that have the property of readily donating electrons to other molecules, and examples of such antioxidants include, but are not limited to, water-soluble vitamins such as vitamin C, polyphenols, and alcohol-based solvents such as methanol, ethanol, or propanol. These antioxidants may be used alone or in combination of two or more types. When using antioxidants, safety must be considered, so antioxidants with low toxicity such as ethanol or propanol are preferably used.

In addition, a method of forming a covalent bond by chemical reaction with a reactive group is specifically achieved by reacting a reactive group such as an amino group, a sulfonic acid group, or a halogenated alkyl group on the surface of the substrate of the material with a reactive group introduced at the end of the main chain or in the side chain of the copolymer.

Methods for introducing reactive groups onto the surface of a material include, for example, a method in which a monomer having a reactive group is polymerized to obtain a substrate having a reactive group on its surface, and a method in which reactive groups are introduced by ozone treatment or plasma treatment after polymerization.

Examples of methods for introducing reactive groups into the ends of the main chains of the above copolymers include methods that use initiators having reactive groups, such as 2,2'-azobis[2-methyl-N-(2-hydroxyethyl)propionamide] and 4,4'-azobis(4-cyanovaleric acid).

Methods for introducing reactive groups into the side chains of the copolymer include copolymerizing a monomer having a reactive group, such as glycidyl methacrylate, to an extent that does not inhibit the action and function of the copolymer.

As mentioned above, the amount of the hydrophilic polymer introduced into the hollow fiber membrane can be quantified by attenuated total reflection infrared spectroscopy (ATR-IR). If necessary, it can also be quantified by X-ray photoelectron spectroscopy (XPS) or the like.

Furthermore, the hollow fiber membrane produced by the above-mentioned production method can be used as a hollow fiber membrane module. A module is a device in which multiple hollow fiber membranes are bundled together and placed in a cylindrical container, and both ends or one end are fixed with polyurethane, epoxy resin, etc. to enable collection of permeate water, or both ends of the hollow fiber membrane are fixed to a flat plate shape to enable collection of permeate water. By providing a pressurizing means such as a pump or water level difference on the raw liquid side of this hollow fiber membrane module, or a suction means such as a pump or siphon on the permeate side, it can be used as a separation device that performs membrane filtration of the raw liquid, which is the fermentation liquid. Using this separation device, a purified permeate can be produced from the fermentation liquid.

<Use of hollow fiber membrane>
The hollow fiber membrane described above is preferably used in the process of filtering the fermentation liquid of the present invention using the hollow fiber membrane. The fermentation liquid is not particularly limited by the type of microorganism or fermentation product, but examples thereof include beer, wine, sake, vinegar, soy sauce, amino acid fermentation liquid, organic acid fermentation liquid, and fermentation liquid of biopharmaceuticals. By filtering, unnecessary components for the product, such as microbial particles and turbidity in the fermentation liquid, can be removed. The hollow fiber membrane described above is suitable for these applications because it is less likely to clog during filtration and its performance recovery by chemical cleaning after clogging is maintained for a long period of time.

The present invention is not limited to the above-described embodiment, and can be modified as desired. In addition, the material, shape, dimensions, values, form, number, location, etc. of each component in the above-described embodiment are arbitrary and not limited as long as they can achieve the present invention.

The present invention will be described in more detail below with reference to examples, but the present invention is not limited to these examples in any way.

The following examples will explain the case where the separation membrane of the present invention takes the form of a hollow fiber membrane.

<Measuring pore size distribution of hollow fiber membranes using mercury intrusion method>
The hollow fiber membrane was dried by the following method. The hollow fiber membrane wetted with water was freeze-dried at -20°C for about 50 hours, and then vacuum-dried at room temperature for about 8 hours. The hollow fiber membrane not wetted with water was vacuum-dried at room temperature for about 8 hours. The dried hollow fiber membrane was cut to a length of about 5 mm, and the sample weight was weighed with an electronic balance (AW220 manufactured by Shimadzu Corporation). The pore size distribution was measured with a Pore Sizer 9320 manufactured by Micromeritics. The test piece was sealed in a glass cell of about 5 ^cm3 attached to the device, mercury was injected under reduced pressure, and then the pressure was increased in the range of about 4 kPa to 207 MPa (corresponding to a pore diameter of about 7 nm to 350 μm) through oil in a pressure-resistant container attached to the device. The surface tension of mercury was calculated as 484 dyn/cm, and the contact angle of mercury was calculated as 141.3°. The volume of pores with a pore diameter of 100 nm or more was defined as V1.

<Measurement of pore volume of hollow fiber membrane by gas adsorption method>
The hollow fiber membrane dried under reduced pressure was cut to a length of about 10 cm, and the sample weight was weighed using an electronic balance (AW220, manufactured by Shimadzu Corporation). The amount of nitrogen gas adsorption was measured using an automatic specific surface area/pore size distribution measuring device (BELSORP-mini II, manufactured by Microtrack-Bell). The BET specific surface area was calculated using the attached analysis program. The pore volume of pores with a pore size of less than 100 nm was designated as V2.

<Preparation of membrane module>
A plurality of hollow fiber membranes were cut to a length of about 30 cm and wrapped in a polyethylene film to form a hollow fiber membrane bundle. This hollow fiber membrane bundle was inserted into a cylindrical polycarbonate module case, and both ends were fixed with an epoxy potting agent. The ends were cut to obtain a module with both ends open. The number of hollow fiber membranes was appropriately set so that the membrane area based on the inner diameter of the hollow fiber membrane was 100 to 200 cm ² in Examples 1 to 7, 10, and 11 and Comparative Examples 1 to 3, and 1000 to 2000 cm ² in Examples 8, 9, and Comparative Examples 4 to 9. Note that the cylindrical module case was provided with ports at two locations near both ends to allow fluid to perfuse the outer surface of the hollow fiber membrane, and end caps having inlets and outlets for liquid were attached to both ends to allow fluid to perfuse the hollow part of the hollow fiber membrane.

<Measurement of total sugar concentration>
The total sugar in the sample solution was quantified by the phenol-sulfuric acid method. That is, the sample and 5% phenol solution were mixed in equal amounts, and then concentrated sulfuric acid, 1.5 times the amount of the mixture, was quickly added and stirred. The temperature was kept constant (20-30°C) for 20-30 minutes. Then, the absorbance of the mixed solution at a wavelength of 480 nm was measured using a UV-Visible Spectrophotometer (Shimadzu Corporation, UV2450). On the other hand, a calibration curve was created using a glucose standard solution, and the total sugar concentration (glucose equivalent value) in the sample solution was calculated using the obtained calibration curve.

<Measurement of protein concentration>
150 μL of the sample solution was mixed with 1350 μL of Coomassie (Bradford) Protein Assay Reagent (manufactured by Thermo Fisher Scientific), which is a Bradford reagent, and allowed to stand at room temperature for 10 minutes to react. After the reaction, the absorbance of the mixed solution at a wavelength of 470 nm was measured using an ultraviolet-visible spectrophotometer (Shimadzu Corporation, UV2450). Meanwhile, a bovine serum albumin standard sample was reacted with the Bradford reagent in the same manner as above, a calibration curve was created, and the protein concentration in the sample solution was calculated using the obtained calibration curve.

<Measurement of polyphenol concentration>
500 μL of the sample solution was mixed with 2500 μL of 1% phenol reagent solution, stirred, and allowed to stand at room temperature for 5 minutes. Then, 2 mL of 7.5% aqueous sodium carbonate solution was mixed, stirred, and allowed to stand at room temperature for 60 minutes. The absorbance of the mixed solution at a wavelength of 765 nm was measured using a UV-Vis spectrophotometer (U-2001, manufactured by Hitachi, Ltd.).

<Shrinkage rate during sterilization>
The hollow fiber membrane was sterilized by the following method: The hollow fiber membrane was cut to a length of 40 cm, and sterilized for 20 hours using an autoclave (Tomy Seiko Co., Ltd., LSX-300) at a set temperature of 125° C. The length of the hollow fiber membrane was measured, and the shrinkage rate was calculated by dividing the shrinkage length by the length of the hollow fiber membrane before sterilization.

<Measurement of maximum stress and breaking elongation>
The measurement was performed using a tensile tester (TENSILON/RTM100 manufactured by Toyo Baldwin Co., Ltd.) with a full-scale load of 5 kg and a crosshead speed of 50 mm/min. A test piece having a test length of 50 mm was used for the measurement in a wet state.

<Measurement of β-glucan>
The measurement of β-glucan was carried out using a β-glucan measurement kit, Mixed-linkage β-glucan, manufactured by Megazyme, Inc. The measurement was carried out according to the attached protocol.

(Reference Example 1)
28% by weight of vinylidene fluoride homopolymer having a weight average molecular weight of 417,000 and 72% by weight of dimethyl sulfoxide were dissolved at 120° C. This vinylidene fluoride homopolymer solution was discharged from the outer tube of a double-tube type spinneret, and at the same time, an aqueous solution of 90% by weight of dimethyl sulfoxide was discharged from the inner tube of the double-tube type spinneret and solidified in a bath of an aqueous solution of 85% by weight of dimethyl sulfoxide at a temperature of 10° C., and then washed with water and stretched to 1.4 times in water at 90° C. The pore volume (V1) of the obtained hollow fiber membrane having a pore diameter of 100 nm or more was 0.213 (cm ³ /g), and the pore volume (V2) of the pores having a pore diameter of less than 100 nm was 0.229 (cm ³ /g).

(Reference Example 2)
The hollow fiber membrane obtained in Reference Example 1 was immersed in dimethyl sulfoxide at 10°C for 15 minutes to replace the poor solvent with water. The pore volume (V1) of the obtained hollow fiber membrane with a pore diameter of 100 nm or more was 0.310 ( ^cm3 /g), the pore volume (V2) of the pore diameter of less than 100 nm was 0.030 ( ^cm3 /g), and V1/V2 was 10.3. In addition, the shrinkage rate during sterilization at 125°C was 5.1%, and the breaking elongation was 35%.
(Reference Example 3)
Vinylpyrrolidone/vinyl acetate random copolymer (Kollidon VA64; manufactured by BASF)
The hollow fiber membrane prepared in Reference Example 2 was immersed for 2 hours in a 0.1 wt% aqueous ethanol solution in which 1000 ppm of a hydrophilic polymer consisting of was dissolved, and then irradiated with 25 kGy of gamma rays to introduce a vinylpyrrolidone/vinyl acetate random copolymer into the hollow fiber membrane. The ratio of monomer units of the vinylpyrrolidone/vinyl acetate random copolymer was 6/4, and the hydration energy density of the hydrophilic polymer/volume fraction of the monomer unit (vinylpyrrolidone) with the largest hydration energy density/hydration energy density difference between vinylpyrrolidone and vinyl acetate was 45.2/60/14.

Example 1
Using the hollow fiber membrane module prepared using the hollow fiber membrane of Reference Example 1, cross-flow filtration of beer was carried out with the filtration device of FIG. 1 (step a). The beer used was a commercially available unfiltered beer containing brewer's yeast, "Ginga Kogen Beer" (hereinafter referred to as evaluation beer). After filling the module with RO water and leaving it for more than one hour, the RO water outside the hollow fiber membrane was discharged, and the water present in the hollow part was replaced with the evaluation beer. 2 L of evaluation beer maintained at 0°C was prepared in a container, and the evaluation beer was perfused from this container through a pump on the outer surface of the hollow fiber membrane and returned to the container, and at the same time, a circuit was set up so that the filtrate filtered by the hollow fiber membrane was collected in a container different from the container containing the evaluation beer. At that time, the inlet pressure and outlet pressure of the evaluation beer to the module and the pressure on the filtration side could be measured. The evaluation beer was introduced into the module inlet so that the evaluation beer flowed at a membrane surface linear velocity of 1.5 m/sec. The filtration flux was adjusted to 100 [L/ ^m2 /h]. In this state, the beer for evaluation was perfused on the outer surface of the hollow fiber membrane at 5±3°C, and cross-flow filtration was continued in which a portion of the beer was filtered. The inlet pressure, outlet pressure, and pressure on the filtration side were measured at predetermined time intervals, and filtration was continued until the transmembrane pressure difference (TMP) rose to 150 kPa, and the beer throughput was calculated. The beer for evaluation was appropriately added so that 2 L was maintained in the container.
The formula was TMP=(Pi+Po)/2-Pf.
where Pi is the inlet pressure, Po is the outlet pressure and Pf is the pressure on the filtration side.
Beer throughput [L/m ² ]=filtration flux 100 [L/m ² /h]×time required for pressure to increase to 150 kPa [h].
After the TMP reached 150 kPa, the filtration was stopped and the liquid was drained, and rinsing with pure water was performed (step b). The rinsing with pure water was performed for 5 minutes at a membrane surface linear velocity of 1.5 m/s and a filtration flux of 100 [L/m ² /h], and then backwashing was performed at 100 [L/m ² /h] for 3 minutes.
Subsequently, chemical washing (step c) was performed. Chemical washing was performed by back-filtration with a 0.3% aqueous solution of sodium hypochlorite (pH 12) at 50° C. from the inside (secondary side) of the hollow fiber membrane toward the outside (primary side) at 140 [L/m ² /h], and the membrane was washed until the beer throughput until the TMP rose to 150 kPa in the second beer filtration was 90% or more of the first time. Finally, the chemical solution was drained, followed by rinsing with pure water (step d). Rinsing with pure water was performed for 5 minutes at a membrane surface linear velocity of 1.5 m/s and a filtration flux of 100 [L/m ² /h]. After that, the water was drained and rinsed with pure water again, and the same operation was repeated a total of three times.
The recovery rate of the beer throughput was calculated from the amount of beer throughput after repeating the above steps a to d 20 times. Note that the time for the first chemical cleaning was used for the time for the second and subsequent chemical cleanings.
Beer processing volume recovery rate (%)=beer processing volume (20th time)/beer processing volume (1st time)×100
The results of each filtration (beer throughput of the first filtration, first cleaning time, beer throughput of the 20th filtration, and beer throughput recovery rate) are shown in Table 1.

Example 2
The chemical cleaning (step c) was carried out in the same manner as in Example 1, except that the first step was carried out using a 0.04% NaOH aqueous solution at 50° C., and the second step was carried out using a 0.3% Na hypochlorite aqueous solution (pH 12) at 50° C. The results in Table 1 were obtained. Note that the first and second cleaning steps were carried out for the same time.

Example 3
In Example 2, after step d, an additional chemical cleaning was performed for 15 minutes using a 3.0% citric acid aqueous solution at 50° C. in the same manner as in step b (step e). Furthermore, the chemical solution was drained, followed by rinsing with pure water (step f). Rinsing with pure water was performed for 5 minutes at a membrane surface linear velocity of 1.5 m/s and a filtration flux of 100 [L/m ² /h]. Thereafter, the water was drained and rinsing with pure water was performed again, and the same operation was repeated three times in total. The same method as in Example 2 was performed except that the above-mentioned steps a to f were repeated, and the results in Table 1 were obtained.

Example 4
The same procedure as in Example 2 was repeated except that in step c of Example 2, the flux of backfiltration from the inside (secondary side) of the hollow fiber membrane to the outside (primary side) during chemical cleaning was changed to 250 [L/ ^m2 /h], and the results in Table 1 were obtained.

Example 5
The same procedure as in Example 2 was repeated except that in step c of Example 2, 0.2% of Tween 20 (polyoxyethylene sorbitan monolaurate, manufactured by Wako Pure Chemical Industries, Ltd.) was added as a nonionic surfactant to the chemical solution in the first stage, and the results in Table 1 were obtained.

Example 6
The same procedure was carried out as in Example 2, except that the hollow fiber membrane of Reference Example 2 was used, and the results were obtained as shown in Table 1. In addition, a sample of the membrane used was cut out and subjected to stress measurement, and the maximum stress was found to be 1,800 gf/ ^mm2 .

(Example 7)
The same procedure was carried out as in Example 2, except that the hollow fiber membrane of Reference Example 3 was used, and the results in Table 1 were obtained.

(Comparative Example 1)
The same procedure was carried out as in Example 2, except that in step c, the sample was immersed in a chemical solution, and the results in Table 1 were obtained.

(Comparative Example 2)
The same procedure was carried out as in Example 2, except that in step c, the chemical solution was caused to flow from the outside (primary side) to the inside (secondary side), and the results in Table 1 were obtained.

(Comparative Example 3)
The same procedure was carried out as in Example 7, except that in step c, the chemical solution was caused to flow from the outside (primary side) to the inside (secondary side), and the results in Table 1 were obtained.

(Example 8)
In Example 2, the filtration in step a was performed in the same manner, except that the amount of evaluation beer held in the container was 500 mL and the filtration was performed until the β-glucan concentration in the raw water reached 1.3 times the initial concentration, and the results in Table 1 were obtained. Note that the density of the beer was calculated by arranging an electromagnetic flowmeter and a Coriolis flowmeter in series in the circulation section of the evaluation beer, and when the beer density reached a value corresponding to the previously measured β-glucan concentration, it was determined that the β-glucan concentration had reached 1.3 times, and washing was performed.

(Comparative Example 4)
The same procedure was carried out as in Example 8, except that the chemical solution in step c was flowed from the outside (primary side) to the inside (secondary side), and the results in Table 1 were obtained.

(Comparative Example 5)
The same method was used as in Example 8, except that the filtration in step a was carried out until the β-glucan concentration in the raw water reached 1.2 times the initial concentration, and the results shown in Table 1 were obtained.

(Comparative Example 6)
In Example 8, the filtration in step a was carried out until the β-glucan concentration in the raw water reached 1.2 times the initial concentration, and the chemical solution in step c was flowed from the outside (primary side) to the inside (secondary side). The same method was used, and the results in Table 1 were obtained.

(Example 9)
In Example 6, the filtration in step a was carried out in the same manner except that the amount of evaluation beer held in the container was 50 mL and the filtration was carried out until the β-glucan concentration in the raw water reached 15 times the initial concentration, and the results in Table 1 were obtained.

(Comparative Example 7)
The same procedure was carried out as in Example 9, except that the chemical solution in step c was flowed from the outside (primary side) to the inside (secondary side), and the results in Table 1 were obtained.

(Comparative Example 8)
In Example 6, the same method was used except that the amount of evaluation beer held in the container was 20 mL and the filtration of step a was carried out until the β-glucan concentration in the raw water reached 20 times the initial concentration, and the results in Table 1 were obtained.

(Comparative Example 9)
In Comparative Example 8, the same procedure was carried out except that the chemical solution in step c was flowed from the outside (primary side) to the inside (secondary side), and the results in Table 1 were obtained.

(Example 10)
The same procedure was carried out as in Example 6, except that the first step of step c was carried out using a 0.4% NaOH aqueous solution at 50° C., and the results in Table 1 were obtained. The membrane used was cut out and subjected to stress measurement, and the maximum point stress was 1800 gf/ ^mm2 , and no decrease in strength was observed compared to Example 6.

Example 11
The same procedure was carried out as in Example 6, except that the second step of step c was carried out using a 1.0% aqueous solution of sodium hypochlorite (pH 12) at 50° C., and the results in Table 1 were obtained. The membrane used was cut out and subjected to stress measurement, and the maximum point stress was 1800 gf/ ^mm2 , and no decrease in strength was observed compared to Example 6.

(Comparative Example 10)
The same procedure was carried out as in Example 2, except that the first step of step c was carried out using a 0.4% NaOH aqueous solution at 50° C., and the results in Table 1 were obtained. The membrane used was cut out and subjected to stress measurement, and the maximum point stress was 1600 gf/mm ^2. The recovery rate was the same as in Example 10, but the maximum point stress was lower.

Compared to Comparative Examples 1 to 3, Examples 1 to 7 were superior in terms of beer processing volume, chemical cleaning time, and beer processing volume recovery rate. These results demonstrate that the filtration method of the present invention allows fermentation broth to be stably filtered over a long period of time.

The fermentation liquor filtration method of the present invention can be used to filter fermentation liquors in the fermentation industry, food industry, and other fields.

1 Raw liquid tank 2 Membrane module 3 Raw liquid pump 4 Circulation control valve 5 Filtration control valve 6 Filtrate tank 7 Backwash liquid tank 8 Backwash pump 9 Backwash control valve 10 Washing liquid tank 11-20 Valves 21-23 Pressure gauges 24-26 Flowmeter 30 Hollow fiber membrane 31 Outer surface 32 Inner surface 40 Spherical structure 41 Spherical body 42 Non-spherical portion (also called constriction)
43 Void

Claims (13)

A method for filtering a fermentation liquid , comprising the steps of: (a) filtering the fermentation liquid through a membrane module having a separation membrane;
(b) rinsing the separation membrane with water;
(c ) permeating a chemical solution having a pH of 10 to 14 from the secondary side to the primary side of the separation membrane;
(d) rinsing the separation membrane with water;
After step d,
(e) passing a chemical solution C having a pH of 0 to 4 from the secondary side to the primary side of the separation membrane;
(f) after step e, rinsing the membrane module with water;
Including,
The cross-flow membrane permeate in the step (a) has a total sugar concentration of 1,000 mg/L or more and 100,000 mg/L or less, a protein concentration of 50 mg/L or more and 10,000 mg/L or less, and a polyphenol concentration of 50 mg/L or more and 10,000 mg/L or less,
The step c
(c-1) A method for filtering a fermentation liquid, comprising a step of passing a chemical solution A having a pH of 10 or more and 14 or less and an oxidant concentration of 0.01% or more from the secondary side to the primary side of a membrane module. The step c
(c-2) Before the step c-1, a step of permeating a chemical solution B having an oxidant concentration of less than 0.01% and a pH of 10 to 14 from the secondary side of the separation membrane to the primary side is further provided. The filtration method for the fermentation liquid according to claim 1. The method for filtering a fermentation broth according to claim 1 or 2, wherein in the steps c and e, the flux ( ^m3 / ^m2 /d) of the chemical solution passing from the secondary side to the primary side is 1.5 to 10 times the cross-flow flux in the step a . The fermentation liquor filtration method according to any one of claims 1 to 3, wherein the chemical solution C is an aqueous solution containing at least one compound selected from the group consisting of hydrochloric acid, nitric acid, sulfuric acid, phosphoric acid, formic acid, acetic acid, propionic acid, butyric acid, citric acid, oxalic acid, ascorbic acid and lactic acid . 3. The method for filtering a fermentation liquid using a membrane module according to claim 2 , wherein the chemical solutions A, B and C contain at least one cleaning auxiliary selected from the group consisting of a surfactant, an enzyme and a chelating agent. 6. The method for filtering a fermentation broth according to claim 1 , wherein the polysaccharide concentration _Ct in the raw water is in the range of 1.3 to 15 times the concentration _C0 at the start of the filtration when the process is transferred from the process a to the process b. The method for filtering a fermentation liquid according to any one of claims 1 to 6 , wherein the separation membrane is a hollow fiber membrane. The separation membrane contains a hydrophobic polymer resin,
The pore volume V2 of pores having a pore diameter of less than 100 nm as measured by a gas adsorption method is 0.02 cm ³ /g or more and less than 0.1 cm ³ /g,
the pore volume V1 of pores having a diameter of 100 nm or more as measured by mercury intrusion porosimetry is 0.1 cm ³ /g or more;
8. The method for filtering a fermentation broth according to claim 7 , wherein the ratio V1/V2 of the pore volume V2 to the pore volume V1 is 3 or more and 60 or less.
V1: pore volume with a pore diameter of 100 nm or more measured by mercury intrusion porosimetry V2: pore volume with a pore diameter of less than 100 nm measured by gas adsorption method 9. The method for filtering a fermentation broth according to claim 8 , wherein the separation membrane has a breaking elongation of 38% or more. 10. The separation membrane according to claim 8 , wherein the shrinkage rate of the hollow fiber membrane length when the membrane is heat-treated in a water vapor atmosphere at 125°C for 20 hours is 0.5% or more. The separation membrane contains a hydrophilic polymer,
The method for filtering a fermentation broth according to any one of claims 1 to 10, wherein the hydrophilic polymer contains a copolymer composed of N types of monomer units 1 , 2, ... i ... N, and the hydration energy density of the hydrophilic polymer calculated based on the following formula (1) is 40 to 70 cal·mol ^-1 ·Å ^-3 .

[In formula (1), the hydration energy of any monomer unit i is the absolute value of the energy of the monomer unit i in water minus the energy of the monomer unit i in a vacuum, N is the total number of monomer species constituting the copolymer, i is an integer of 1 to N, and N is an integer of 2 or more.] The method for filtering a fermentation broth according to any one of claims 8 to 10 , wherein the hydrophobic polymer resin is a fluorine-based resin or a sulfone-based resin. The method for filtering a fermentation liquid according to any one of claims 1 to 12, wherein the filtration in step a is an external pressure filtration in which filtration is performed from the outer periphery side to the inner periphery side of the hollow fiber membrane, and the flux of the chemical solution to be passed through in steps c and e is 7.0 ^m3 / ^m2 /d or more and 10 ^m3 / ^m2 /d or less per inner area of the hollow fiber membrane.

Images