A method for extracting glutamic acid using a sequential simulated moving bed

Abstract

The present invention relates to a method for extracting glutamic acid using a sequential simulated moving bed, comprising the following steps: (1) screening an adsorption resin suitable for separating glutamic acid and glucose; (2) determining single column parameters, and determining adsorption isotherm parameters and kinetic parameters; (3) obtaining operating parameters through multi-objective optimization; and (4) performing a sequential simulated moving bed experiment based on the optimized operating parameters. The present invention introduces simulated moving bed technology into an amino acid system, performs separation operations by regulating specific variables in a simulated moving bed, uses a genetic algorithm multi-objective optimization method to explore conditions, takes the final performance of the product as a direct target, uses Fortran programming to perform process simulation, performs operating condition screening and simulates the moving bed experiment design, so that the entire separation process is faster and more efficient, and the process flow is simplified.

Description

Technical Field

[0001] The invention belongs to the technical field of separation and purification, and in particular relates to a process technology for purifying and extracting glutamic acid by a sequential simulated moving bed. Background Art

[0002] Glutamic acid (glutamic acide), chemical name is α-aminoglutaric acid. Glutamic acid has levorotatory, dextrorotatory and racemic isomers. The levorotatory isomer, i.e. L-glutamic acid, is collectively referred to as glutamic acid, which is an acidic amino acid with two carboxyl groups in the molecule. It is a flaky or powdery crystal, slightly acidic and non-toxic. L-glutamic acid is slightly soluble in cold water, easily soluble in hot water, almost insoluble in ether, acetone and cold acetic acid, and insoluble in ethanol and methanol. It sublimates at 200℃ and decomposes at 247℃-249℃. L-glutamic acid has a wide range of uses. Its sodium salt has a special umami taste and is used as a flavor enhancer for food. Its molecular formula is C5H9O4. In recent years, with the development of economy and the advancement of science and technology, the chemical industry has higher requirements for the energy consumption, solvent consumption, separation efficiency, complexity of separation products and flexibility of separation process control in the separation process. The vast majority of amino acid products are produced through fermentation. Therefore, the composition of the crude liquid contains many impurities, such as medium components, biological enzymes and various amino acids. This leads to the possibility of producing impurities such as various other amino acids, sugars, pigments, etc. during the separation of amino acids by crystallization.

[0003] At present, the method for separating and extracting glutamic acid by mainstream manufacturers in China has an isoelectric point-ion exchange method. The isoelectric point-ion exchange method is that after the fermentation liquid is separated and extracted with glutamic acid by isoelectric method, the isoelectric point mother liquor is exchanged and adsorbed by an ion exchange resin column, and then the glutamic acid on the resin is eluted with alkali solution, and the high flow fraction is collected, and it is mixed with the next batch of fermentation liquid, and then the glutamic acid is extracted by isoelectric point method. This extraction process consumes relatively high acid and alkali, and the separation process is cumbersome and complicated, and the process control is relatively difficult. CN101402584A utilizes a metal membrane separation technology and a concentrated continuous isoelectric method to extract and purify glutamic acid from the fermentation liquid. The quality of the refined glutamic acid solution is improved, and the glutamic acid product obtained after concentrated continuous isoelectric extraction can reach more than 99% by amino acid analyzer analysis purity, but it is necessary to use acid and alkali to adjust the pH value in the process, which will produce a large amount of polluted wastewater and cause certain pressure on the environment. CN101293848B discloses a glutamic acid extraction process, which first heats the glutamic acid fermentation liquid, performs ultrafiltration, and then filters it again with a nanofiltration system, heats the fermentation liquid after nanofiltration, adds sulfuric acid to adjust the pH of the fermentation liquid to 3.0-4.0, and then enters a crystallization pot for evaporation and concentration, and after evaporating 50%-80% of the water, discharges the material to a crystallization tank, slowly cools it to below 60°C, and centrifuges to separate glutamic acid. Although the amount of acid and alkali used is small and the method is relatively simple, the purity and extraction yield of the obtained glutamic acid product are relatively low.

[0004] The present invention will focus on studying the adsorption mechanism and parameter change trend of glutamate separation by sequential simulated moving bed, compare its similarities and differences with the traditional simulated moving bed separation process in the glutamate system, use genetic algorithm for multi-objective optimization, summarize the variable control mechanism suitable for glutamate purification, and deeply explore the intrinsic scientific basis for the change of operating conditions with the change of optimization objectives, which is expected to provide new ideas and methods for the separation and purification of glutamate system. Summary of the invention

[0005] The present invention provides a method for extracting glutamic acid using a sequential simulated moving bed (SSMB), comprising the following steps:

[0006] (1) Screening of adsorption resins suitable for separation of glutamate and glucose

[0007] Seven exchange resins were preliminarily selected based on the system characteristics. The resins were thoroughly rinsed, and single pulse experiments of glutamate and glucose were carried out on a single-column device to determine the adsorption resins.

[0008] The single pulse experimental conditions are as follows: glutamate and glucose concentrations of 10-20 g / L, temperature of 40-60° C., flow rate of 4-12 mL / min, and injection volume of 4-10 mL / min.

[0009] Screen out stable chromatographic peaks and compare the resolution, selectivity, porosity and retention time of various resins in the separation system to determine the appropriate stationary phase resin. The higher the selectivity and separation, the better, and the porosity between 0.3-0.4 is more appropriate.

[0010] In one embodiment of the present invention, LX-1850Ca type adsorption resin was screened out, and its selectivity α=1.847, porosity ε=0.373, and glutamic acid retention time t R =40.78min, glucose retention time t R =30.12min, separation R s =0.52.

[0011] (2) Determine the single column parameters, including system dead volume, dead time, half peak width, linear velocity, total column volume, retention time, selectivity, phase ratio and void ratio, and determine the adsorption isotherm parameters and kinetic parameters of glucose and glutamic acid. The single column parameters provide a basis for subsequent sequential simulated moving bed operation and multi-objective optimization.

[0012] The adsorption isotherm parameters are determined by measuring them under a pulse experiment and judging them in combination with a frontal analysis method.

[0013] 2.1 Based on the stationary phase resin determined by the previous screening experiment, pulse experiments were performed, adsorption isotherm parameters were calculated, and the adsorption behavior of glutamate and glucose on the selected chromatographic column was explored: In the liquid chromatography system, pulse experiments were performed on glutamate and glucose at different flow rates, and the elution curves of glutamate and glucose at different flow rates were recorded. According to the elution curves, the retention time, dead time, half peak width, linear velocity and void fraction of glutamate and glucose at different flow rates can be obtained, and the adsorption isotherm parameters can be calculated using formula (2-3).

[0014] (2-3)

[0015] Among them, the linear velocity u is calculated according to formula (2-4):

[0016] (2-4)

[0017] In the formula, H is Henry's constant, i is the component index, including glutamate and glucose. i is the Henry constant of component i, u is the linear velocity in cm / min, ε is the porosity, t R is the retention time, in min, V c is the total volume of the column, in cm 3 , L is the length of the chromatographic column, in cm, r is the inner radius of the chromatographic column, in cm, Q is the flow rate, in mL / min; t0 is the dead time of the chromatographic column t0=t D -t plant , where t D is the system dead time, in min, t plant It is the dead time of the equipment (measured without connecting the chromatographic column), in min.

[0018] In one embodiment of the present invention, the single column parameters also include the half peak width W h / 2 , which is the peak width at half the peak height of the chromatographic elution curve, read from the spectrum; system dead volume V D (measured after connecting the chromatographic column), in mL; and phase ratio .

[0019] In one embodiment of the present invention, the dead volume of the system (measured after connecting to the chromatographic column) V can be obtained by injecting blue dextran that has no adsorption reaction with the stationary phase. D =4.72mL, dead time t D =0.944min, total column volume V c =565.2 cm 3 , phase ratio φ = 1.681, column length L = 100 cm, r = 2.5 cm, equipment dead time t plant =0.420min.

[0020] Substituting into formula (2-4), we get the Henry constant H. Moreover, the basic stability of the Henry constant H of each component at different flow rates indicates that the adsorption characteristics of glutamate and glucose within the stationary phase resin selected in step (1) and the experimental concentration range belong to linear adsorption behavior.

[0021] In one embodiment of the present invention, the flow rates of glutamate and glucose in the pulse experiment are 6 mL / min, 8 mL / min, 10 mL / min, and 12 mL / min, respectively.

[0022] 2.2 Derivation of kinetic parameters using pulse experimental results

[0023] According to the single column parameters and Henry constant obtained from the previous glutamate and glucose pulse experiments, the kinetic parameters were fitted, uλ and N were calculated, and uλ was used as the horizontal axis and 1 / N as the vertical axis to plot the kinetic parameters and fit the linear formulas for glutamate and glucose. The intercept and slope of the fitting formula can be used to solve the mass transfer coefficient k respectively. m and the axial diffusion coefficient N L : Get the parameters of the mass transfer-diffusion (TD) model of glutamate and glucose, the mass transfer coefficient k m and the axial diffusion coefficient N L .

[0024] Specifically, for each component, there is an independent model parameter N (theoretical plate number). N is related to the TD model parameter N L and k m It is related to, expressed by formula (2-5):

[0025] (2-5)

[0026] Among them, D L is the total diffusion coefficient; λ is given by , L and H represent the following expressions:

[0027] (2-6)

[0028] H and N are represented by the following formula:

[0029] (2-3)

[0030] (2-8)

[0031] Where, t R is the retention time, W h / 2 is the half peak width, t plant Device dead time, in minutes, selectivity α ≈1.847.

[0032] According to equation (2-5), the Henry constant measured at different flow rates can be used to calculate the parameter N of the TD model L and k m , when the flow velocity is within a reasonable range, 1 / N is linearly related to u.

[0033] Using uλ as the horizontal coordinate and 1 / N as the vertical coordinate, the kinetic parameters were fitted to obtain the linear fitting formulas for glutamate and glucose. The intercept and slope of the fitting formulas were used to solve the mass transfer coefficient k, respectively. m and the axial diffusion coefficient N L .

[0034] In one embodiment of the present invention, the fitting formula for glutamate is y=3.2474x+0.0197, and the fitting formula for glucose is y=1.3023x+0.0029. The mass transfer coefficient k can be obtained by using the intercept and slope of the fitting formula. m and the axial diffusion coefficient N L .

[0035] 2.3 The adsorption isotherm parameters of glutamate and glucose were determined using the frontal analysis method to verify the accuracy of the pulse experiment and determine the linear adsorption relationship.

[0036] In the mobile phase equilibrium chromatography system, a series of glutamate and glucose samples with different concentrations are prepared, including the concentration of the sample to be tested. The glutamate or glucose solution with the initial concentration is directly and continuously injected as the mobile phase. After the elution curve of the first sample reaches a stable platform and lasts for a fixed period of time, the breakthrough curve at the sample concentration is determined. The next concentration of glutamate and glucose solution is immediately replaced as the injection mobile phase. The experiment is repeated according to the above operation to measure a series of step-shaped breakthrough curves in order of concentration.

[0037] According to the elution curve measured by the frontier analysis method, the corresponding adsorption amounts of glutamate and glucose at different concentrations can be obtained through formula (2-11). The linear model represented by formula (2-12) is used to describe the adsorption equilibrium relationship of glutamate and glucose on the preparation chromatographic column.

[0038] (2-11)

[0039] In the formula, q n is the mass of glutamate or glucose adsorbed per unit volume of resin in the nth step, in g, V c is the total volume of the column, in cm 3 , t D is the system dead time, in min, t n is the breakthrough time of step n, c nis the concentration at step n, and ε is the void fraction of the column, which is dimensionless.

[0040] (2-12)

[0041] In the formula, H i is the Henry constant for component i, q n is the mass of glutamate or glucose adsorbed per unit volume of resin in the nth step, in g, c i It is the concentration of glutamate or glucose, in g / L.

[0042] According to the adsorption isotherm parameter fitting curve determined by pulse experiment, it can be known that glutamate and glucose have linear adsorption characteristics on the selected stationary phase. On this basis, the linear adsorption model can be used to fit the adsorption isotherm. The adsorption isotherm parameter Henry's constant H can be calculated through the slope of the concentration c and adsorption amount q graph. Substituting the parameters into (2-12), the Henry's constants of glutamate and glucose, that is, the adsorption isotherm parameters, can be obtained respectively.

[0043] By comparing the adsorption isotherm parameters obtained by the pulse experiment and the adsorption isotherm parameters (Henry's constant) obtained by the frontier analysis method, if the two values ​​are similar, it can be determined that the adsorption isotherm parameters obtained by the pulse experiment method are correct and meet the linear adsorption model.

[0044] Since the prior art has not studied the adsorption behavior of the two components, glutamate and glucose, in the system of this application, this application is the first to conclude, based on pulse tests and frontier analysis methods, that the adsorption characteristics of glutamate and glucose on the selected resin and within the experimental concentration range belong to linear adsorption behavior, and verifies the accuracy of the adsorption behavior, providing a basis for subsequent modeling.

[0045] (3) Obtaining operating parameters through multi-objective optimization

[0046] According to the single-column experimental results of step (2), the sequential simulated moving bed separation process was modeled using Fortran language based on the mass transfer-diffusion (TD) model and the linear adsorption model. The parameters measured in the single-column experiment and the flow ratio m value of the four separation regions I, II, III, and IV of the sequential simulated moving bed obtained based on the triangle theory were used to select a suitable optimization target. The non-dominated sorting genetic algorithm (NSGA-II) was used to run the program in Compaq Visual Fortran 6.5 software to obtain the injection flow rate and the switching time (t1, t2, and t3) of the three sub-steps of feeding, circulation, and elution.

[0047] The formula of the mass transfer-diffusion (TD) model is:

[0048] (3-1)

[0049] (3-2)

[0050] Linear adsorption model formula: (2-12)

[0051] In the formula, c represents the concentration in the liquid phase (g / L); q represents the concentration in the stationary phase (g / L); i represents the component index; j represents the column index (which can represent the fixed column number 1-8); t represents the time (min); represents phase ratio; ε represents column void ratio; u represents linear velocity; z represents axial coordinate; N L represents the axial diffusion coefficient; k m represents the mass transfer coefficient; H i represents the Henry coefficient of component i.

[0052] According to the m value obtained by trigonometric theory, m is the four regions I, II, III, and IV of SSMB (e.g. Figure 1 The flow ratio (as shown) is dimensionless, and the calculation formula of m is shown in 3-3:

[0053] (3-3)

[0054] Where m is the flow ratio of the four separation regions, x represents the region number (x=I-IV), t s It is the total switching time of the three sub-steps of feeding, circulation and elution, in minutes.

[0055] The optimization objectives may be any two of purity, yield, processing volume, water consumption, etc.

[0056] The process of the non-dominated sorting genetic algorithm (NSGA-II) is shown in Figure 2 .

[0057] Based on the non-dominated sorting genetic algorithm-II (NSGA-II), condition screening is performed, and the corresponding Pareto curve with two optimization objectives as horizontal and vertical coordinates is analyzed and drawn in the "prod.txt" document in the output result folder. At the same time, the four m values ​​corresponding to all points in the Pareto curve are found in turn, and the operating conditions such as the three sub-step time and feed flow rate are calculated.

[0058] The optimized four SSMB variables (three sub-step times and feed flow rate) can be determined sequentially by the following formulas:

[0059] (3-4)

[0060] (3-5)

[0061] (3-6)

[0062] (3-7)

[0063] Where, t1, t2, and t3 are the time of the three sub-steps of feeding, circulation, and elution in SSMB respectively; Q feed is the feed flow rate (mL / min); Q max V is the maximum flow rate suitable for the SSMB system (mL / min); c is the total volume of the column (cm 3 ).

[0064] (4) Based on the optimized operating parameters, a sequential simulated moving bed experiment was carried out. After feeding, circulation, elution and multiple internal cycles, the purity of glutamate and glucose samples was analyzed.

[0065] In one embodiment of the present invention, Figure 1 As shown, the first step is the feeding stage. The prepared solution is sucked into zone III, i.e., columns 5-6 in the chromatographic system, through the feed port by the feed chromatographic pump. At this time, columns 3-4 and 7-8 are isolated. At the same time, the eluent is pulled by the Eluent chromatographic pump into zone I, i.e., columns 1-2. The raffinate port can collect light components (glucose), and the extraction port can collect heavy components (glutamic acid).

[0066] The second step is the circulation stage, where 1-8 columns are all connected through valve switching for internal circulation to achieve the effect of balancing the internal concentration. This step is also a key step for saving mobile phase in the sequential simulated moving bed. No material is collected in the current step.

[0067] The third step is the elution stage. The valve continues to switch to connect the six chromatographic columns 1-6 and isolate the chromatographic columns 7-8. The eluent enters from the chromatographic column 1, flows through the chromatographic columns 2-5, and is discharged from the chromatographic column 6. The raffinate (glucose) can be collected at the port of the chromatographic column 6.

[0068] (5) After the system has completed multiple (10 to 15) cycles, collect the extract and raffinate samples and analyze the sample purity and recovery rate.

[0069] In one embodiment of the present invention, the results of sequential simulated moving bed experiments show that the purity of glutamate can reach 99% and the purity of glucose can reach 94%.

[0070] Beneficial effects of the present invention:

[0071] 1. The present invention uses a sequential simulated moving bed to obtain a high-purity glutamic acid product. Compared with the traditional ion exchange process / isoelectric + ion exchange method, the process experiment only uses pure water solvent, does not require acid and alkali reagents to adjust the pH value, reduces the use of solvents, has low environmental pressure, reduces costs, and reduces the difficulty of subsequent sewage treatment, reduces pollution, is green, energy-saving and environmentally friendly, and has good industrial practical value.

[0072] 2. The adsorption resin used in the present invention is different from the traditional ion resin, does not contain ion exchange groups, and the regeneration process is relatively easy and can be used repeatedly.

[0073] 3. The present invention introduces simulated moving bed technology into the amino acid system. For separation of amino acids, chiral structures, sugars and other systems, the separation operation is performed by regulating specific variables in a simulated moving bed. A more advanced genetic algorithm multi-objective optimization method is used to explore conditions. The final performance of the product (purity, yield, processing volume, solvent consumption, etc.) is taken as a direct target. Fortran programming is used to simulate the process, screen operating conditions and design simulated moving bed experiments, so that the entire separation process is faster and more efficient, does not require a large number of experiments to explore conditions, simplifies the process flow, saves manpower and material resources, and the sequential simulated moving bed can directly obtain a high-purity glutamic acid product (greater than 99%).

[0074] 4. The sequential simulated moving bed of the present invention includes an internal circulation separation step, which can greatly reduce the consumption of water resources and other solvents; at the same time, the sequential simulated moving bed achieves internal concentration stability, and can continuously obtain high-purity, high-recovery glutamic acid products. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 It is a schematic diagram of the sequential simulation moving bed separation process;

[0076] Figure 2 This is the flow chart of the non-dominated sorting genetic algorithm (NSGA-II);

[0077] Figure 3 is the fitting diagram of glutamate, glucose 1 / Ni and uλi;

[0078] Figure 4 is the outflow curve of the frontier analysis method;

[0079] Figure 5 is the glutamate / glucose adsorption isotherm. DETAILED DESCRIPTION

[0080] Example 1

[0081] A method for extracting glutamic acid using a sequential simulated moving bed (SSMB) comprises the following steps:

[0082] (1) Screening of adsorption resins suitable for separation of glutamate and glucose

[0083] According to the relevant prior art, combined with glutamate and glucose functional groups, the present invention specifically selected 6 cation exchange resins (LX-1850 Ca, Amberlite CR99K / 310, Amberlite CR99Ca / 310, ZG SPC106Ca, LX-1850 ammonium type, LX-2850 type), 1 anion resin (Amberlite™ CR5550 type resin), a total of 7 exchange resins, the resins were fully rinsed, and a single pulse experiment of glutamate and glucose (15g / L) was carried out on a single column device to observe the separation degree and determine the adsorption resin.

[0084] The single pulse experiment conditions are: temperature: 45°C, flow rate 4-12ml / min, injection volume: 5ml / min. The selected resins were subjected to pulse experiments at the same flow rate to observe the separation degree and elution curve to further determine the separation performance of the resin.

[0085] Therefore, by comparing and analyzing the advantages and disadvantages of various resins in separation degree, resolution, selectivity, porosity and retention time in the separation system, it is determined that the stationary phase resin LX-1850Ca resin is superior to other resins in terms of separation degree and selectivity, and the resin stability is good after multiple pulse injection experiments. Therefore, LX-1850Ca resin is the selected resin for this experiment. The parameter measurement data of LX-1850Ca resin are shown in Table 1.

[0086] Table 1 Summary of the measured data of LX-1850Ca resin parameters

[0087] .

[0088] (2) Determining single column parameters, including system dead volume, system dead time, half peak width, linear velocity, void fraction and other parameters, and determining adsorption isotherm parameters and kinetic parameters of glucose and glutamic acid.

[0089] According to the pulse experimental conditions in step (1), the single column parameters include: α = 1.847 (selectivity), t D =0.944min (dead time), V c =565.2 cm 3 (Total volume of column), =1.681 (phase ratio), L=100cm (column length), V D =4.72mL (dead volume), ε=0.373 (system porosity), equipment dead time t plant =0.420min.

[0090] 2.1 Determination of adsorption isotherms by pulse experiment

[0091] Based on the stationary phase adsorption resin determined by the screening experiment, the adsorption behavior of glutamate and glucose on the selected chromatographic column was investigated, and 15g / L glutamate and glucose solutions were prepared respectively. The chromatographic column was kept warm by constant temperature water bath heating (45°C). Ultrapure water was used as the mobile phase, and the mobile phase was heated to 45°C after ultrasonic degassing for 30 minutes. Turn on the preparative liquid chromatograph, set the flow rate to 10mL / min to flush the chromatographic column until the baseline of the differential detector is flushed, return the baseline to zero, rotate the manual valve of the six-way valve and press the timer button of the differential detector at the same time to start data collection.

[0092] The elution curves of glutamate and glucose were recorded at four flow rates: 6 mL / min, 8 mL / min, 10 mL / min, and 12 mL / min. Based on the retention time and dead time of the elution curves at different flow rates, the adsorption isotherm parameters were calculated using formula (2-3).

[0093] (2-3)

[0094] Among them, the linear velocity u is calculated according to formula (2-4):

[0095] (2-4)

[0096] The pulse experiment results of glutamate and glucose at different flow rates are shown in Tables 2 and 3, respectively.

[0097] Table 2 Results of glutamate pulse experiment

[0098] .

[0099] Table 3 Glucose pulse test results

[0100] .

[0101] As shown in Tables 2 and 3 above, based on the basic stability of the Henry constants of glutamate and glucose at different flow rates, it can be preliminarily considered that the adsorption characteristics of glutamate and glucose within the resin selected in step (1) and the experimental concentration range belong to linear adsorption behavior.

[0102] 2.2 Using the pulse experiment results to obtain the kinetic parameters:

[0103] Based on the single column parameter results, the kinetic parameters of the separation system were determined and calculated, and the mass transfer coefficients k of glutamate and glucose were obtained. m And the model parameter N L .

[0104] The Henry constant H and single column parameters measured at different flow rates in step 2.1 can be used to calculate the parameter N of the mass transfer-diffusion (TD) model: L and k m .

[0105] Specifically, for each component, there is an independent model parameter N (theoretical plate number) that can measure column efficiency and separation effect. L and k m It is related to, expressed by formula (2-5):

[0106] (2-5)

[0107] Among them, λ is given by , L and H represent the following expressions:

[0108] (2-6)

[0109] H and N are represented by the following formula:

[0110] (2-3)

[0111] (2-8)

[0112] Where, t R is the retention time, W h / 2 is the half-peak width, α ≈ 1.847.

[0113] The 1 / Ni and uλi parameters of glucose and glutamate are shown in Table 4. The kinetic parameters were fitted by plotting uλ as the abscissa and 1 / N as the ordinate. Figure 3 As shown, the fitting formula for glutamate is y=3.2474x+0.0197, and the fitting formula for glucose is y=1.3023x+0.0029. According to formula (2-5), the mass transfer coefficient k can be solved using the intercept and slope of the fitting formula. m and the axial diffusion coefficient N L The final calculation shows that glutamate: k m =50.76, N L =0.308; glucose: k m =344.83, N L =0.768.

[0114] Table 4 Summary of 1 / Ni and uλi parameters

[0115] .

[0116] 2.3 The adsorption isotherm parameters of glutamate and glucose were determined using the frontal analysis method to confirm the accuracy of the pulse experiment and to determine the linear adsorption relationship.

[0117] Use ultrapure water to prepare 1L of 8, 10, 12, 14, and 16g / L glutamic acid and glucose solutions, respectively. Use ultrapure water to flush the chromatographic column at a flow rate of 8mL / min. Turn on the purge button of the differential detector to flush the reference cell until the baseline is stable. Switch to the lowest concentration (8g / L) glutamic acid solution as the mobile phase to start the experiment. The flow rate of the chromatographic pump is 8mL / min. The platform rises to a stable level for more than 15min, then switch to the next concentration sample, and test 10, 12, 14, and 16g / L glutamic acid solutions, respectively. The experimental steps for glucose are the same as those of the above experiments. A series of step-shaped breakthrough curves are measured in order of concentration, such as Figure 4 As shown, each concentration will run out a step (breakthrough curve), and each step (i.e. each flow rate) is a step.

[0118] According to the elution curve measured by the frontier analysis method, the corresponding adsorption amounts of glutamate and glucose at different concentrations can be calculated using formula (2-11), and the linear model represented by formula (2-12) is used to describe the adsorption equilibrium relationship of glutamate and glucose on the chromatographic column.

[0119] (2-11)

[0120] In the formula, q n is the mass of glutamate or glucose adsorbed per unit volume of resin in the nth step, in g; V c is the total volume of the column, in cm 3 ;t D is the system dead time, in min; t n is the breakthrough time of the nth step, in min; ε is the porosity of the chromatographic column, dimensionless.

[0121] (2-12)

[0122] In the formula, H i is the Henry constant for component i, q i is the mass of glutamate and glucose adsorbed per unit volume of resin in step j, in g; c i is the concentration of glutamate and glucose, in g / L.

[0123] By concentration c i and adsorption amount q i The slope of the graph can be used to calculate the Henry constant H, the parameter of the adsorption isotherm. Substituting the parameters into formula (2-12), we can get H1=0.313 for glutamate and H2=0.723 for glucose. The adsorption isotherms of glutamate and glucose are as follows: Figure 5 shown.

[0124] By comparing the adsorption isotherm parameters obtained by the pulse experiment in step 2.2 and the adsorption isotherm parameters H of glutamate and glucose obtained by the frontier analysis method in step 2.3, it can be seen that the values ​​of the two results are similar (Tables 2 and 3, Figure 5 ), the correctness of the adsorption isotherm parameters obtained by the pulse experiment method can be determined, that is, it can be determined to satisfy the linear adsorption model, providing a basis for subsequent modeling.

[0125] (3) Obtaining operating parameters through multi-objective optimization

[0126] According to the single-column experimental results of step (2), the sequential simulated moving bed separation process was modeled using Fortran language based on the mass transfer-diffusion (TD) model and the linear adsorption model. The parameters measured in the single-column experiment and the m value obtained according to the triangle theory were used to select the appropriate optimization target (set as sample purity and processing volume). The non-dominated sorting genetic algorithm (NSGA-II) was used to run the program in Compaq Visual Fortran 6.5 software to obtain the injection flow rate and the switching time (Q) of the three sub-steps of feeding, circulation, and elution. feed , t1, t2 and t3).

[0127] The formula of the mass transfer-diffusion (TD) model is:

[0128] (3-1)

[0129] (3-2)

[0130] Linear adsorption model formula: ;

[0131] In the formula, c represents the concentration in the liquid phase (g / L); q represents the concentration in the stationary phase (g / L); i represents the component index; j represents the column index; t represents the time (min); represents the phase ratio; ε represents the column porosity; u represents the linear velocity; z represents the axial coordinate; N L represents the axial diffusion coefficient; k m represents the mass transfer coefficient; H represents the Henry coefficient.

[0132] The calculation formula of m value obtained according to triangle theory is:

[0133] (3-3)

[0134] m is the flow ratio of the four separation zones of the sequential simulated moving bed, x represents the zone number (x=I-IV), t s is the total switching time.

[0135] Based on the non-dominated sorting genetic algorithm-II (NSGA-II) commonly used in SSMB optimization, condition screening is performed, and the corresponding Pareto curve with two optimization objectives as horizontal and vertical coordinates is analyzed and drawn in the "prod.txt" document in the output result folder. At the same time, the four m values ​​corresponding to all points in the Pareto curve are found in turn, and the operating conditions such as the three sub-step time and feed flow rate are calculated.

[0136] The optimized four SSMB variables (three sub-step times and feed flow rate) can be determined sequentially by the following formulas:

[0137] (3-4)

[0138] (3-5)

[0139] (3-6)

[0140] (3-7)

[0141] Where, t1, t2, and t3 are the time of the three sub-steps of feeding, circulation, and elution in SSMB respectively; Q feed is the feed flow rate unit mL / min; Q max is the maximum flow rate, in mL / min; V c is the total volume of the column, in cm 3 .

[0142] The calculated sequential simulated moving bed feed flow rate Q feed =4mL / min, feeding step (switching) time t1=18min, circulation stage (switching) time t2=25min, elution stage (switching) time t3=8min, in order to facilitate elution, elution flow rate Q D =Q max =10mL / min.

[0143] (4) Based on the optimized operating parameters (such as injection flow rate, elution flow rate and switching time), a sequential simulated moving bed experiment was carried out.

[0144] like Figure 1 As shown, the first step is the feeding stage. The prepared solution is sucked into zone III, i.e., columns 5-6 in the chromatographic system, through the feed port by the feed chromatographic pump. At this time, columns 3-4 and 7-8 are isolated. At the same time, the eluent is pulled by the Eluent chromatographic pump into zone I, i.e., columns 1-2. The raffinate port can collect light components (glucose), and the extraction port can collect heavy components (glutamic acid).

[0145] The second step is the circulation stage, where 1-8 columns are all connected through valve switching for internal circulation to achieve the effect of balancing the internal concentration. This step is also a key step for saving mobile phase in the sequential simulated moving bed. No material is collected in the current step.

[0146] The third step is the elution stage. The valve continues to switch to connect the six chromatographic columns 1-6 and isolate the chromatographic columns 7-8. The eluent enters from the chromatographic column 1, flows through the chromatographic columns 2-5, and is discharged from the chromatographic column 6. The raffinate (glucose) can be collected at the port of the chromatographic column 6.

[0147] (5) After the system has completed 10 to 15 cycles, collect the extract and residue samples and analyze the sample purity and recovery rate.

[0148] SSMB experimental results: According to the optimized parameter Q feed , t1, t2, t3, Q D The SSMB experiment was completed and the experimental process was verified by sequential simulated moving bed process simulation. The simulation results were good and the design parameters could be applied to the sequential simulated moving bed experiment. The purity of glutamate could reach 99%, the purity of glucose could reach 94%, and the processing capacity was 1.41 mL / min, as shown in Table 5.

[0149] The relationship between purity and processing volume, switching time and feed volume is as follows:

[0150] purity: ;

[0151] Processing capacity: U T =Q feed ×t1 / t s

[0152] t s =t1+t2+t3

[0153] Table 5 SSMB separation experimental parameters of Example 1

[0154] .

[0155] Example 2

[0156] According to the operating parameters selected by optimization, the sequential simulated moving bed feed flow rate was set to 4 mL / min, the elution flow rate was set to 10 mL / min, the feed step switching time was set to 18 min, the circulation stage switching time was set to 25 min, and the elution stage switching time was set to 12 min.

[0157] The SSMB separation test was also carried out. After 10 to 15 cycles, the extract and raffinate samples were collected and analyzed for sample purity and recovery, as shown in Table 6. The purity of glutamic acid can reach 95%, the purity of glucose is 90.2%, and the processing capacity is 1.34 mL / min.

[0158] Table 6 Example 2 SSMB separation experimental parameters

[0159] .

[0160] It can be seen from the data in the examples that the SSMB method optimized by the present application can obtain high purity glutamic acid and glucose, and the sequential simulated moving bed can achieve internal concentration stability, and can continuously obtain high-purity and high-recovery glutamic acid products. In addition, no acid or alkali reagent is required to adjust the pH value during the experiment, which reduces damage to the environment and has broad application prospects.

Claims

1. A method for extracting glutamic acid using a sequential simulated moving bed, characterized in that: The steps include: (1) Screening of stationary phase adsorption resins suitable for separation of glutamate and glucose; (2) Determining single column parameters, including system dead volume, dead time, half peak width, linear velocity, total volume of the chromatographic column, retention time, selectivity, phase ratio and void ratio, and determining adsorption isotherm parameters and kinetic parameters; 2.1 Based on the stationary phase adsorption resin determined by screening, pulse experiments are carried out to calculate the adsorption isotherm parameters: Among them, the adsorption isotherm parameters are calculated using formula (2-3) using single column parameters: (2-3) In the formula, Hi is the Henry constant of component i, i is the component index, including glutamate and glucose, u is the linear velocity, unit is cm / min, is the void ratio, t R is the retention time, in min, V c is the total volume of the column, in cm 3 , L is the length of the chromatographic column, in cm, Q is the flow rate, in mL / min, t0 is the dead time of the chromatographic column, in min; 2.2 Using the pulse experiment results to obtain the kinetic parameters: Using the single column parameters and Henry constant H obtained in step 2.1 and formula (2-5), uλ and N are calculated, and uλ is used as the horizontal axis and 1 / N is used as the vertical axis to fit the kinetic parameters. The linear fitting formulas for glutamate and glucose are obtained, and the mass transfer coefficient k is obtained using the intercept and slope of the fitting formula. m and the axial diffusion coefficient N L : （2-5） 2.3 Use the frontier analysis method to determine the adsorption isotherm parameters of glutamate and glucose, verify the pulse experiment, and determine the linear adsorption relationship: Determine the elution curve according to the frontier analysis method, and use formula (2-11) to obtain the corresponding adsorption amount of glutamate and glucose at different concentrations, and use formula (2-12) to calculate the Henry constant H of each component: （2-11） In the formula, q n V is the mass of glutamate or glucose adsorbed per unit volume of resin in the nth step, in g; c is the total volume of the column, in cm 3 ;t D is the system dead time, in min; t n is the breakthrough time of step n, in min; c n is the concentration at step n, in g / L; is the void fraction of the column; （2-12） In the formula, H i is the Henry constant for component i; c i is the concentration of glutamate or glucose, in g / L; (3) Obtain operating parameters through multi-objective optimization: According to the single column experimental results of step (2), the sequential simulated moving bed separation process is modeled using Fortran language based on the mass transfer-diffusion model and the linear adsorption model. The optimization target is selected using the parameters measured by the single column experiment and the flow ratio m of the four separation regions I, II, III, and IV of the sequential simulated moving bed obtained based on the triangle theory. The program is run in Compaq Visual Fortran 6.5 software using a non-dominated sorting genetic algorithm to obtain the injection flow rate and the switching time of the three sub-steps of feeding, circulation, and elution. (4) Based on the optimized operating parameters, a sequential simulated moving bed experiment was carried out.

2. The method according to claim 1, characterized in that The single pulse experimental conditions in step (1) are as follows: glutamate is 10-20 g / L, glucose concentration is 10-20 g / L, temperature is 40-60°C, flow rate is 4-12 mL / min, injection volume is 4-10 mL / min; the adsorption resin is LX-1850Ca.

3. The method according to claim 1, characterized in that In step 2.1, the linear velocity u is calculated according to formula (2-4): (2-4) Where r is the inner radius of the chromatographic column, in cm, and u is the linear velocity, in cm / min. is the porosity, Q is the flow rate, unit is mL / min.

4. The method according to claim 1, characterized in that In step 2.2, λ is given by , L and H represent the calculation formula as follows: （2-6） （2-8） Where W h / 2 is the half-peak width, t plant Device dead time, in min, α is selectivity.

5. The method according to claim 1, characterized in that The formula of the mass transfer-diffusion model is: （3-1） （3-2） Wherein, c represents the concentration in the liquid phase, in g / L; q represents the concentration in the stationary phase, in g / L; i represents the component index; j represents the column index; t represents the time, in min; represents phase ratio; ε represents void ratio; u represents linear velocity; z represents axial coordinate; N L represents the axial diffusion coefficient; k m represents the mass transfer coefficient.

6. The method according to claim 5, characterized in that The calculation formula of m is shown in (3-3): （3-3） m is the flow ratio of the four separation zones of the sequential simulated moving bed, x represents the zone number (x=I-IV), t s It is the total switching time of the three sub-steps of feeding, circulation and elution.

7. The method according to claim 6, characterized in that Based on the non-dominated sorting genetic algorithm-II, conditional screening is performed. In the "prod.txt" document in the output result folder, the corresponding Pareto curve with two optimization objectives as horizontal and vertical coordinates is analyzed and drawn; at the same time, the four m values ​​corresponding to all points in the Pareto curve are found in turn, and the time and feed flow parameters of the three sub-steps of feeding, circulation and elution are calculated according to formulas (3-4), (3-5), (3-6) and (3-7): （3-4） （3-5） （3-6） （3-7） Where, t1, t2, and t3 are the time of the three sub-steps of feeding, circulation, and elution, respectively, in min; Q feed is the feed flow rate, unit is mL / min; Q max It is the maximum flow rate, in mL / min.

8. The method according to claim 1, characterized in that Step (4) The first step of the sequential simulated moving bed experiment is the feeding stage. The prepared solution is sucked into zone III, i.e., column 5-6 in the chromatographic system, through the feed port by the feed chromatographic pump. At this time, columns 3-4 and 7-8 are isolated. At the same time, the eluent is pulled by the Eluent chromatographic pump and enters zone I, i.e., column 1-2. The light component glucose is collected at the raffinate port, and the heavy component glutamic acid is collected at the extraction port. The second step is the circulation stage, where columns 1-8 are all connected through valve switching for internal circulation to achieve the effect of balancing the internal concentration; The third step is the elution stage. The valve continues to switch to connect the six chromatographic columns 1-6 and isolate the chromatographic columns 7-8. The eluent enters from the chromatographic column 1, flows through the chromatographic columns 2-5, and is discharged from the chromatographic column 6. The raffinate is collected at the port of the chromatographic column 6. After the system has been through multiple cycles, the extract and raffinate samples are collected and analyzed for sample purity and recovery.

Images