- Email: [email protected]
In-situ recovery of carboxylic acids from fermentation broths through membrane supported reactive extraction using membrane modules with improved stability
Separation and Purification Technology 241 (2020) 116694
Contents lists available at ScienceDirect
Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur
In-situ recovery of carboxylic acids from fermentation broths through membrane supported reactive extraction using membrane modules with improved stability
T
Angelo Gössia,b, Florian Burgenerb, David Kohlerb, Alessandro Ursob, Boris A. Kolvenbachb, ⁎ Wolfgang Riedlb, Boelo Schuura, a b
University of Twente, Faculty of Science and Technology, Sustainable Process Technology Group, 7500AE Enschede, the Netherlands University of Applied Sciences and Arts Northwestern Switzerland, Institute for Chemistry and Bioanalytics, Hofackerstrasse 30, 4132 Muttenz, Switzerland
A R T I C LE I N FO
A B S T R A C T
Keywords: Reactive extraction Process intensification Product recovery Carboxylic acids Membrane extraction process
Membrane supported reactive extraction (MSE) coupled to back-extraction (MSBE) using a new type of Teflon (PTFE) capillary membrane contactor was studied for the in-situ removal of carboxylic acids from aqueous streams, e.g. fermentation broths. The use of microporous membranes as extraction interface helps avoiding emulsification problems, allows the use of extreme phase ratios, and protects microorganisms, as they are less affected by solvent toxicity during in-situ extractions. The use of PTFE capillary membranes is suitable for longterm use due its high chemical and thermal stability. A simple toxicity screening identified n-decanol with tri n-octyl amine (TOA) as a suitable solvent. MSE experiments were performed using membrane contactors (0.005 m2 to 0.15 m2), working with solvent to feed phase ratios down to 1:40 (mass based). The in-situ removal of lactic acid out of fermentation broths using lactobacillus plantarum led to a glucose conversion rate of 80 mol%. Additionally, a concentration factor up to 7.8 could be shown during back-extraction.
1. Introduction The demand for bio-based chemicals in general and bio-based polymers in particular has tremendously increased in the last few years [1]. Carboxylic acids like lactic acid, mandelic acid, succinic acid and itaconic acid are up-and-coming substances which are used on large scale. For example, a growing interest in bio-based and/or bio-degradable polymers increased the production of lactic acid from 60.000 tons in 2006 to 494.000 tons in 2013 [2]. Many carboxylic acids are currently produced by fermentation of carbohydrates [3,4]. The release of acid during the fermentation leads to a production inhibition, which is why the pH has to be controlled to ensure optimal growth conditions for the microorganisms [5]. Usually, the pH is maintained by addition of calcium hydroxide, leading to carboxylates. After the fermentation has finished, sulfuric acid has to be added to release the acid from its carboxylate salt form into its neutral acid form, resulting in about 1 ton of gypsum per ton carboxylic acid as by-product [3]. With increasing demand of carboxylic acids this costly method of recovery is neither environmentally friendly nor economical [3,6,7]. Several techniques have been investigated in the past to avoid
⁎
gypsum production, mainly by in-situ product removal techniques such as adsorption [8], reactive distillation [9], stripping [10], reactive liquid-liquid extraction [6] and membrane-based processes (electro-dialysis, ultrafiltration, nanofiltration) [11]. Liquid-liquid extraction of carboxylic acids is a widely studied research field [12,13]. Often composite solvents comprising a diluent and a complexing extractant are used, such as tri-octylphosphine oxide [14] or, more frequently, tertiary amines [7,10,15,17–19]. Although ionic liquids have shown great potential [20 21–23], it has been shown that the regeneration from the ionic liquid in dilute systems is challenging [24]. The published method to regenerate the acid from the ionic liquid involves thermal treatment challenging [24], which is not feasible for lactic acid production. For this application the more traditional amine based extraction systems appear more suitable. However, the operational limitations such as the formation of stable emulsions and limited flexibility in solvent/feed (S/F) ratios remain points of attention. By applying a membrane supported extraction (MSE), both emulsification can be prevented, and an unlimited range of (S/F) becomes accessible. The in-situ removal of substances inhibiting production out of fermentation broths has been studied previously ([4,25–27]. However, the
Corresponding author. E-mail address: [email protected] (B. Schuur).
https://doi.org/10.1016/j.seppur.2020.116694 Received 8 November 2019; Received in revised form 23 January 2020; Accepted 8 February 2020 1383-5866/ © 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
conventional liquid-liquid extraction systems suffer from the negative effect of the direct contact of solvent with microorganisms [28]. Several technologies have been described to minimize or even avoid the contact between microorganisms and organic phases. These include for example membrane pertraction [20,29,30] or solvent impregnated resins containing ionic liquids [31]. The use of membranes as extraction interface represents a promising way to reduce solvent toxicity on microorganisms [6,32], since the immobilized extraction interface protects microorganism from long-term contact. Most authors use commercially available hollow fiber membrane contactors, made out of polypropylene and polyurethane [25–27]. While these contactors provide remarkable contact areas based on their size, their solvent stability is limited and they are difficult to clean due to the embedded materials. Disadvantages of membrane contactors include the increased mass transfer resistance, limited membrane lifetime and the risk of fouling. To reduce the mass transfer resistance, the membrane should be thin, while the pore size should be as large as possible. [33,34] Alternatively, dense (=non-porous) membranes can be used, which are less affected by fouling. However, they suffer from reduced permeability compared to porous membranes [35]. This paper aims to characterize a new type of capillary membrane contactor based on commercially available (porous) PTFE fibers, which allow for a long-term use due to its superior chemical stability and good cleanability.
2.3. Experimental methods
2. Material and methods
D=
2.3.1. Solvent and toxicity screening For the extraction efficiency tests, 20 mL of a 0.68 M lactic acid solution was extracted with 20 mL organic phase containing 20% TOA. The occurrence of stable emulsions was examined by vigorous shaking of the samples and observation of the phase separation. Phase separation was reached with centrifugation for 20 min at 4500 × g. The concentration of the lactic acid bound to the reactive agent in the organic phase was determined by mass balance. A simple toxicity screening approach was used regarding dissolved organic solvents (molecular level toxicity) and excess solvent (phase level toxicity) [32,36-38]. For the molecular level toxicity, 30 mL of water was saturated with organic solvent. The excess organic solvent was then removed and the aqueous phase was used for a batch fermentation after sterile filtration (culture medium see chapter 2.3). The phase level toxicity was studied by adding 5 mL of organic solvent to 30 mL of a batch fermentation. The resulting lactic acid concentration for both tests was measured after 72 h by direct titration. The extraction efficiency of a solvent was determined by its distribution coefficient as given in Eq. (1). In this equation, all forms of lactic acid are lumped, thus aqueous lactate and lactic acid concentrations are summed, as well as any free lactic acid in the organic phase, and lactic acid bound to an extractant in the organic phase.
[LA]equil, org [LA]equil, aq
(1)
2.1. Chemicals All chemicals used were purchased from Sigma Aldrich Switzerland. Tri-n-octylamine (TOA) (98%), sodium hydroxide standard solution (volumetric for titration 0.1 M NaOH), n-heptane (> 99%) and n-decanol (> 98%), 2-ethyl-1-hexanol (> 99%), n-octanol (> 99%), tributylphosphate (> 99%) , methyl isobutyl ketone (MIBK, > 98.5%), ndodecanol (98%), isobutanol (> 99%), oleyl alcohol (85%), dodecane (> 99%), itaconic acid (> 99%), mandelic acid (99%), succinic acid (> 99%) were used. The aqueous lactic acid solutions were prepared from an 85 wt% lactic acid solution.
2.3.2. Membrane supported extraction experiments Different starting concentrations of carboxylic acids ranging from 0.3 to 3 wt% were used. If not stated differently, the experiments were performed at room temperature, with 20 wt% TOA in n-decanol, at a flow rate of the aqueous shell side of 12 kg/h and an organic flow rate of 6 kg/h (inside the fibers/lumen). All experiments were started by pumping of the aqueous phase and applying an overpressure of about 100 mbar to ambient pressure. Subsequently, the circulation of the organic phase was started, keeping the pressure about 80 mbar below the aqueous phase. This avoids breakthrough of organic phase into the aqueous phase. Temperature, flow rates and pressure were monitored. In order to work with higher solvent-to-feed ratios (S/F), an external 20 L glass vessel could be connected. Throughout all experiments, the organic phase was pumped through the fibers (lumen side), and the aqueous phase through the shell side of the membrane contactor. Extraction experiments were performed using different capillary membrane contactor sizes (0.005 m2 to 0.15 m2) and phase ratios (up to 1:40 S/F) and different substrate/ reactive agent concentrations. Both the aqueous and the organic phase were circulated loop-wise until reaching the equilibrium, this operation mode allowed adequate and accurate analysis of the process even at the small membrane sized applied. The concentration in the organic phase was calculated via mass balance.
2.2. Experimental setup Fig. 1 shows the applied membrane extraction setup. This setup was used for either extraction or back-extraction experiments. It consists of two temperature controlled 2 L vessels for the aqueous and organic phase, respectively. Two gear pumps are used to pump the phases through the membrane and back into the container. The pressure of both phases can be adjusted by needle valves. Coriolis flow meters record the flow as well as the density of both phases. For continuous fermentation experiments, a glucose-feeding pump is connected to the aqueous phase vessel. Phase equilibrium and back-extraction experiments were conducted in 50 mL Falcon tubes. Intensive phase contact was ensured by regular shaking of the tubes, for a total of one hour. Different temperatures were applied using a water bath. Phase separation was reached with an Eppendorf Centrifuge 5804 R, A-4–44 Rotor, at 4500 × g for 20 min). Table 1 shows characteristics of the used membrane fiber. All parts of the modules are either from stainless steel, glass, polytetrafluoroethene (PTFE) or polyvinylidene fluoride (PVDF). The MSE setup was cleaned after every experiment using acetone and then dried using oil-free compressed air. When working with microorganisms, the setup was disinfected with 70 vol% ethanol for > 20 min. Afterwards, the setup was purged with ethanol and dried using oil-free compressed air. Due to the chemically inert materials used in the membrane contactor, the modules can also be autoclaved, steam sterilized or ozonized.
2.3.3. Back-extraction experiments Lactic acid was added to a solution of 20 wt% TOA in n-decanol to prepare a saturated solution. This solution was then contacted with different volumes of water (the receiving phase) and heptane (antisolvent) in Falcon tubes. The tubes were temperature regulated using a water bath. Intensive phase contact was ensured by regular shaking of the tubes, for a total of one hour. The phases were separated in the temperature regulated centrifuge and the samples of the aqueous phases were taken. The concentration of lactic acid in the aqueous phase was measured by titration (see chapter 2.4). 2
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
Fig. 1. Used MSE-Setup, gear pumps range 0 L/h – 60 L/h, temperature variable 20–60 °C, all phase-contacting parts made out of PVDF, PTFE, glass or stainless steel (1.4404). Additional glucose feeding required for fermentation experiments only.
of carboxylic acids [16,25-27,40,41]. However, the octanol-trioctylamine-water-carboxylic acid system is known to form stable emulsions. Additionally, octanol shows some water solubility and a strong solvent toxicity [6,27]. Generally speaking, with an increasing octanol–water partition coefficient 10log(Poctanol) of a solvent, the probability increases that it is non-toxic to microorganisms [42]. Solvents having a 10log (Poctanol) > 5 are generally biocompatible for cellular biocatalysis [43]. Solvents that can form hydrogen bonds stabilize the carboxylic acidtrioctylamine complex [44,12]. Thus, alcohols are the most widely used species of solvents for this kind of reactive extraction and are expected to result in high extraction efficiencies.
2.3.4. Fermentation experiment: In-situ lactic acid extraction Lactobacillus plantarum (DSM 2648) was used for fermentation experiments. The used culture medium contained the following ingredients, per liter: 20 g Glucose, 5 g Yeast extract, 0.4 g, magnesiumsulfate heptahydrate, 0.1 g mangansulfate monohydrate, 5 g sodiumacetate, 2 g triammonium citrate, 10 g casein protein, 10 g meat extract, 2 g dipotassium phosphate. The pre-culture was developed using the same culture medium at 37 °C for 16 h (150 RPM, Infors HAT Multitron). The pH of the pre-cultures was in the range of 4.5–5, indicating that production of lactic acid had already commenced. Inoculum was added to sterile culture medium in a phase ratio of 1/500 vol/vol. 2.4. Analytical techniques
4. Results and discussion The analysis of the lactic acid concentration in the aqueous phase was performed either by titration or by capillary electrophoresis (CE). A Mettler Toledo G20 device was used for the potentiometric direct titration method (0.1 M NaOH). For low concentrations, capillary electrophoresis was used with the Agilent organic acids solutions kit method. The systematic error was below 3% for both methods.
4.1. Solvent screening A set of nine solvents had been selected for a solvent screening. The selection was based on literature research as well as on the octanol–water-distribution coefficients 10log(Poctanol), which can be an indicator for solvent toxicity. [6,43,45–49]. Key criteria for solvent selection were low toxicity and high extraction efficiency (see Table 2). Due to their low distribution coefficients, oleyl alcohol and dodecane were not studied in further tests. Subsequently, the remaining solvents were investigated regarding their molecular and phase level toxicity. Fig. 2 shows the resulting lactic acid concentration after 72 h of fermentation stressed by either molecular or phase level toxicity of the applied solvents. A significant difference between the molecular and phase level toxicity results can be seen. All solvents strongly inhibit the lactic acid production when applied in direct contact with microorganisms.
3. Theory The reactive extraction using tertiary amines extracts only the undissociated form of the carboxylic acid, meaning that the process is more efficient at low pH. However, when working with fermentation broths, the pH has to be kept > 5 for optimal culture conditions [39]. The trioctylamine-carboxylic acid complex is stabilized by hydrogen bonds with the solvent, leading to much higher complexation constants in the presence of hydrogen bond acceptors or donors (e.g. alcohols). n-octanol is the most widely used solvent for the reactive extraction
Table 1 Used PTFE membrane. The membrane porosity was achieved by stretching; the tortuosity is estimated at 2–3, based on Field Emission Scanning Electron Microscopy (FESEM). Membrane purchased from Memo3, Switzerland. Membrane Material
Porosity [%]
Pore diameter [µm]
Inner Diameter [mm]
Outer diameter [mm]
Wall thickness [µm]
PTFE
54
0.47
2.97
3.49
260
3
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
However, the toxicity on the molecular level is much lower compared to the phase level toxicity. The use of a microporous membrane as extraction interface prevents direct contact with the solvent, helping against phase level toxicity. The negative effects of molecular level toxicity however cannot be reduced. Due to its good extraction efficiency and its low molecular level toxicity decanol is a very promising solvent for membrane supported extractions. Octanol, the most widely used solvent for reactive extraction of carboxylic acids with tri-n-octylamine acids [16,25-27,40,41] shows good extraction efficiency, but its molecular level toxicity is too high for in-situ extraction with efficient fermentation. Thus, a separation of the microorganisms from the extraction process would be required to avoid molecular level toxicity. Therefore, we propose using decanol as solvent in the in-situ extraction. Decanol offers a good compromise between extraction efficiency and toxicity. Apart from the toxicity issue, the membrane helps to overcome the emulsification problems occurring with all investigated solvents and to achieve the desired extreme phase ratios.
available membrane contactors in terms of membrane material, pore diameter and wall thickness. The use of these innovative types of membranes might provide significant improvement regarding process stability and cost efficiency. The MSE set-up described under 2.2 was used for all membrane experiments. Experiments were conducted at 25 °C. Fig. 3 shows typical membrane extraction diagrams. Depending on the membrane area and the process parameters, the extraction process took between 8 h and 7 days. Phase ratios (S/F) between 1:1 and 1:40 were tested. Membrane extractions were performed with following carboxylic acids: lactic-, succinic-, mandelic- and itaconic acid, with initial concentrations of 3%. The differences in the resulting final concentrations can be explained due to different physical distribution coefficients which are highest for itaconic and mandelic acid. The physical distribution coefficients resulting from the extraction of 3% carboxyilic acid solutions with n-decanol are 0.14, 0.15, 0.50 and 0.56 for lactic-, mandelic, itaconic- and succinic acid. Noticeable is the difference between the two dicarboxylic acids succinic- and itaconic acid, which can be explained due to the lower pKa’s of itaconic acid, increasing the extraction efficiency. Additionally, itaconic acid could tend to undergo further reactions (e.g. polymerization) [50]. The used membranes are non-selective, meaning that they do not influence the phase equilibrium in any way. All MSE distribution coefficients were in excellent accordance with distribution coefficients measured in batch liquid-liquid equilibrium experiments. As expected, the formation of emulsions could be avoided during all MSE experiments. The used PTFE membrane allows the immobilization of the extraction interface at the aqueous side of the membrane, filling the pores of the membrane with 20 wt% TOA in n-decanol. Extraction experiments with fermentation broths using n-octanol as solvent led to a rapid decrease in living microorganisms in less than 24 h. However, working with n-decanol as solvent, the number of live cells remained constant for 24 h.
4.2. Membrane supported extraction experiments
4.3. In-situ lactic acid extraction out of fermentation broths
Table 2 Distribution coefficients of lactic acid and octanol–water distribution coefficients of solvents. Extraction of 20 mL of a 0.68 mol/dm3 lactic acid solution with 20 mL organic phase containing 20 wt% TOA at 25 °C, *source: SigmaAldrich. Solvent
Solvent
2-ethyl-1-hexanol n-octanol n-decanol Tributylphosphate MIBK n-dodecanol Isobutanol oleyl alcohol Dodecane
2.9 – 4.5 4.0 1.3 4.8 1.0 7.5 7.0
10
logPoctanol*
Distribution coefficient of lactic acid [-] 15.1 14.6 12.7 10.7 8.8 7.1 5.7 5.4 0.5
The types of membranes used in this study differ from commercially
An in-situ extraction of lactic acid from a batch fermentation with
Fig. 2. Resulting lactic acid concentrations (wt%) from toxicity tests after 72 h of fermentation. The Blank experiment represents the fermentation without any solvents. 4
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
Fig. 3. Extraction of carboxylic acids out of an aqueous stream. Initial carboxylic acid conc. = 3.0 wt%, S/F 1:1 vol/vol, A = 0.059 m2, 20 wt% TOA in n-decanol. Dotted lines are the fitted extraction curves for each carboxylic acid. The last values for each acid are equilibrium values derived from batch extraction experiments.
presence should initiate a salting out of acids and therefore increase the distribution coefficient. Contrarily, studies showed that the presence of salts reduces the extraction of carboxylic acids. As an example, the presence of sodium chloride results in the formation of hydrochloric acid from Cl- and the proton from the carboxylic acid. Hydrochloric acid then competes with lactic acid for the extraction [6,52,53], similar effects were observed for amine-functionalized resins, which were found to prefer sulfuric acid, phosphoric acid and hydrochloric acid over volatile fatty acids [54]. The use of membranes as extraction interface slows down the process due to the limited contact area available. Since tertiary amines only extract undissociated acid, fermentation at low pH using (genetically modified) acid-tolerant organisms promises significantly increased extraction rates [4]. It has to be noted that the production rate of lactic acid had to be kept rather low due to the limited contact area available. Since the organic phase eventually is saturated, the continuous removal of carboxylic acids requires an additional back-extraction step. Fig. 5 shows the desired process on the example of a fermentative production of carboxylic acid. After the in-situ extraction, the loaded organic phase is regenerated in a (membrane supported) back-extraction step, in which the carboxylic acid is transferred into water. The back-extraction is described more in detail in the following chapter.
lactobacillus plantarum (DSMZ 2648) was set up. The goal was to maintain the fermentation process as long as possible and to achieve a high yield of lactic acid. A pH set point of 5 was chosen as compromise between microorganism activity and extraction efficiency. As the size of the membrane contact area (0.059 m2) was prone to limit the mass transfer, the product formation had to be controlled by a constant feed rate of medium containing 20 g/L glucose with a rate of 2 mL/hour. The initial medium contained 5 g/L glucose. Additional glucose was added when the pH of the fermentation rate increased over 5. First extraction trials using n-octanol led to no production of lactic acid at all. In fact, almost no live cells could be detected after 24 h of fermentation. However, the in-situ extraction with n-decanol led to promising results. The fermentation could be kept up for more than 7 days. The pH in the fermenter was solely controlled by the dosage of culture medium, without the addition of any base. The results of the fermentation run with in-situ extraction using MSE are shown in Fig. 4. The fermentation process was stable for more than 7 days. The microorganisms in the fermenter were still active. However, the productivity after 7 days was reduced. After the addition of 5 g of glucose, the pH only dropped slightly. Nevertheless a sample of the fermentation broth after 8 days could be used to start a new pre-culture, resulting in a pH of 4.8 after 16 h, which fits well the pH of other pre-cultures. In total, 25.8 g of glucose was added to the fermentation broth, resulting in 21.0 g of lactic acid. This corresponds to a conversion rate of 81 mol % of the fed glucose. This is remarkable, considering the achievable lactic acid concentrations without neutralization or membrane extraction, as shown in Fig. 5. The in-situ removal of lactic acid increased the yield by factor 4.3 when compared to a fermentation without neutralization. Without a membrane protecting the microorganisms from n-decanol phase toxicity, only less than 0.1% of lactic acid was formed. There was no biomass removed during the experiment. However, during a continuous fermentation process, dead microorganisms need to be removed eventually from fermentation broth e.g. by cross flow filtration [51]. Nevertheless some issues remain to be solved. Culture media typically contain different salts, like NaCl, Na2SO4 or K2HPO4. Their
4.4. Back-extraction of lactic acid In general, the fermentative production of platform chemicals is done in large scales. The use of a membrane supported extraction allows an extraction with big phase ratios. Thus, the cost intensive organic phase can be kept at a minimum. This, on the other hand, only allows the extraction of small amounts of carboxylic acid in a single extraction step. Nevertheless, by introducing a continuous back-extraction, it is still possible to extract sufficient amounts of acid. In order to back-extract the bound carboxylic acid into water, the tri-n-octylamine-carboxylic acid complexes need to be split. As discussed in the theory section, the complexes are stabilized by hydrogen bonds with the solvent n-decanol or n-octanol. By adding a solvent that 5
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
Fig. 4. In-situ extraction of lactic acid out of fermentation broths. Vferm = 0.7 L, Vorg = 0.7 L, dashed green line represents the goal value of pH = 5, red arrows represent additional dosage of glucose. The dashed red lines represent the lactic acid concentrations derived without neutralization and extraction without membrane, respectively.
(wash phase/extract phase, i.e. aqueous back-extraction phase to heptane in n-decanol ratio) between 0:1 and 3:1 were used. A concentration factor was calculated, assuming a feeding concentration at position 1 in Fig. 5 of 0.8% LA. This incoming feed concentration was chosen as a compromise between bacteria activity and extraction efficiency (see Table 3). A good compromise between concentration factor and recovery rate can be achieved at 70 °C, with a phase ratio S/F of 1:2. As shown in Fig. 6, a recovery rate of 31% results in a single step extraction, with an outgoing aqueous-solution containing 4.3%. Higher recovery rates could be achieved by working at even higher temperatures than 70 °C. The PTFE membrane modules can endure
cannot donate or accept hydrogen bonds (like heptane), the nature of the complexes and the stoichiometry changes [12,55], and the distribution can be reduced [16,37]. Additionally, an increase in temperature of the organic phase reduces the complexation constant even more, increasing the recovery rate [10,16,37,47]. However, the added heptane needs to be removed after back-extraction, e.g. by a falling- or thin-film evaporation. Back-extraction experiments were performed using a 20 wt% TOA with a lactic acid concentration of 54 g/L in n-decanol (the concentration was based on lactic acid uptake in previous extraction experiments, extracting from 2.5 wt% lactic acid at S/F of 1:4). Temperatures between 20 and 70 °C, S/F (n-decanol phase to water) ratios between 1:1 and 1:8, and W/E
Fig. 5. Shows the conceptual process design for a fermentation with in-situ removal of lactic acid. The lactic acid produced in the fermenter at ① extracted into 20 wt % TOA in n-decanol and then back-extracted in a temperature- and solvent-swing process into a pure water stream. The resulting product stream is at ②. 6
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
Table 3 Back-extraction experiments with the developed temperature- and solvent-swing process on the example of lactic acid and 20% TOA in n-decanol. Resulting outgoing concentrations of lactic acid (LA) and concentration factor reachable. LA conc. [wt %] ①
Temperature [°C]
W/E [-]
S/F[-]
LA conc.[wt %] ②
Recovery [%]
Concentration Factor[-]
0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8
20 20 20 20 20 20 20 20 20 20 20 20 50 50 50 50 50 50 50 50 50 50 50 50 70 70 70 70 70 70 70 70 70 70 70 70
0 0 0 0 1 1 1 1 3 3 3 3 0 0 0 0 1 1 1 1 3 3 3 3 0 0 0 0 1 1 1 1 3 3 3 3
1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8
0.34 0.43 0.50 0.58 0.89 1.26 1.62 2.04 1.40 2.12 2.86 3.62 0.66 0.82 1.02 1.27 1.39 1.92 2.63 3.36 1.97 3.04 4.44 5.31 0.70 0.97 1.60 1.88 1.46 2.21 3.19 3.84 2.00 3.42 5.03 6.21
6.2% 3.9% 2.3% 1.3% 16.3% 11.6% 7.5% 4.7% 25.7% 19.5% 13.2% 8.3% 12.1% 7.5% 4.7% 2.9% 25.6% 17.6% 12.1% 7.7% 36.2% 28.0% 20.4% 12.2% 12.8% 8.9% 7.3% 4.3% 26.9% 20.3% 14.6% 8.8% 36.7% 31.4% 23.1% 14.3%
0.4 0.5 0.6 0.7 1.1 1.6 2.0 2.6 1.7 2.7 3.6 4.5 0.8 1.0 1.3 1.6 1.7 2.4 3.3 4.2 2.5 3.8 5.6 6.6 0.9 1.2 2.0 2.3 1.8 2.8 4.0 4.8 2.5 4.3 6.3 7.8
removed out of the organic phase by evaporation after the back-extraction. By doing so, the negative effects of heptane on extraction rate efficiency and the toxicity towards the microorganisms can be avoided. 5. Conclusions The here proposed membrane supported extraction (MSE) process using PTFE-based membranes showed promising results for the in-situ removal of carboxylic acids from aqueous streams at superior chemical membrane stability, being compatible with the investigated solvents (tri-n-octylamine in n-decanol showed best results), as well as with the fermentation broth. Reduced solvent toxicity due to the membrane interface was observed in continuous removal of lactic acid from a fermentation, allowing the fermentation to run for five days yielding a conversion of 80% without any neutralization required. Extraction followed by backextraction allowed concentration factors up to 7.8. CRediT authorship contribution statement Angelo Gössi: Investigation, Writing – original draft. Florian Burgener: Investitation. David Kohler: Investitation. Alessandro Urso: Investitation. Boris A. Kolvenbach: Investitation. Prof. Wolfgang Riedl: Supervision, Writing – review and editing. Prof. Boelo Schuur: Supervisor, Writing – review and editing.
Fig. 6. Back-extraction efficiency comparison at different S/F ratios. Incoming LA concentration of 0.8%, 70 °C, W/E 1:3. The dots represent the concentration factor.
temperatures > 120 °C (b.p. heptane 98 °C at ambient pressure). An increased temperature also benefits the removal by means of distillation. Alternatively, other antisolvent (gaseous or liquid) could be used [56]. Due to the significant boiling point difference of heptane and ndecanol (98 °C resp. 230 °C under ambient pressure), heptane can be
Declarations of Competing Interest The authors declare that they have no known competing financial 7
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
interests or personal relationships that could have appeared to influence the work reported in this paper.
[27] K. De Sitter, L. Garcia-Gonzalez, C. Matassa, L. Bertin, H. De Wever, The use of membrane based reactive extraction for the recovery of carboxylic acids from thin stillage, Sep. Purif. Technol. (2018) 177–185, https://doi.org/10.1016/j.seppur. 2018.06.001. [28] M.D. Waghmare, K.L. Wasewar, S.S. Sonawane, D.Z. Shende, Natural nontoxic solvents for recovery of picolinic acid by reactive extraction, Ind. Eng. Chem. Res. 50 (2011) 13526–13537, https://doi.org/10.1021/ie201228u. [29] S. Schlosser, R. Kertész, J. Marták, Recovery and separation of organic acids by membrane-based solvent extraction and pertraction: an overview with a case study on recovery of MPCA, Sep. Purif. Technol. (2005) 237–266, https://doi.org/10. 1016/j.seppur.2004.07.019. [30] N.D. Patil, A.W. Patwardhan, A.V. Patwardhan, Carboxylic acids separation using hollow fiber supported liquid membrane, Indian J. Chem. Technol. 24 (2017) 20–31. [31] M. Blahušiak, S. Schlosser, J. Marták, Extraction of butyric acid by a solvent impregnated resin containing ionic liquid, React. Funct. Polym. 71 (2011) 736–744, https://doi.org/10.1016/j.reactfunctpolym.2011.04.002. [32] K.L. Wasewar, Reactive extraction: an intensifying approach for carboxylic acid separation, Int. J. Chem. Eng. Appl. 3 (2012) 249–255, https://doi.org/10.7763/ ijcea.2012.v3.195. [33] Basic Equations of Mass Transport Through a Membrane Layer, Elsevier, 2019. doi:10.1016/c2016-0-04043-3. [34] E. Drioli, A. Criscuoli, E. Curcio, Membrane Contactors: Fundamentals, Applications and Potentialities, Elsevier Science, vol. 11, 2005. [35] Handbook of Membrane Reactors | ScienceDirect, (n.d.). (accessed January 18, 2020). https://www.sciencedirect.com/book/9780857094148/handbook-ofmembrane-reactors. [36] R. Bar, J.L. Gainer, Acid fermentation in water-organic solvent two-liquid phase systems, Biotechnol. Prog. 3 (1987) 109–114, https://doi.org/10.1002/btpr. 5420030208. [37] Y.K. Hong, W.H. Hong, D.H. Han, Application of reactive extraction to recovery of carboxylic acids, Biotechnol. Bioprocess Eng. 6 (2001) 386–394, https://doi.org/ 10.1007/BF02932319. [38] L. Bassetti, J. Tramper, Organic solvent toxicity in Morinda citrifolia cell suspensions, Enzyme Microb. Technol. (1994) 642–648, https://doi.org/10.1016/01410229(94)90084-1. [39] G. Greafe, C.H. Holten, Lactic acid, properties and chemistry of lactic acid and derivatives, Verlag Chemie GmbH, 1971. [40] A. Schunk, A. Menert, G. Maurer, On the influence of some inorganic salts on the partitioning of citric acid between water and organic solutions of tri-n-octylamine: Part I: Methyl isobutyl ketone as organic solvent, Fluid Phase Equilib. (2004) 55–72, https://doi.org/10.1016/j.fluid.2004.04.010. [41] K.L. Wasewar, V.G. Pangarkar, A.B.M. Heesink, G.F. Versteeg, Intensification of enzymatic conversion of glucose to lactic acid by reactive extraction, Chem. Eng. Sci. 58 (2003) 3385–3393, https://doi.org/10.1016/S0009-2509(03)00221-5. [42] M. Matsumoto, Toxicity of organic solvents and ionic liquids to lactic acid-producing microbes, Org. Solvents Prop. Toxicity, Ind. Eff. (2011) 105–114. [43] M. Vermuë, J. Sikkema, A. Verheul, R. Bakker, J. Tramper, Toxicity of homologous series of organic solvents for the gram-positive bacteria Arthrobacter and Nocardia Sp. and the gram-negative bacteria Acinetobacter and Pseudomonas Sp, Biotechnol. Bioeng. 42 (1993) 747–758, https://doi.org/10.1002/bit.260420610. [44] F. Chemarin, M. Moussa, M. Chadni, B. Pollet, P. Lieben, F. Allais, I.C. Trelea, V. Athès, New insights in reactive extraction mechanisms of organic acids: An experimental approach for 3-hydroxypropionic acid extraction with tri-n-octylamine, Sep. Purificat. Technol. 179 (2017) 523–532, https://doi.org/10.1016/j.seppur. 2017.02.018. [45] N.A. Marinova, D.S. Yankov, Toxicity of some solvents and extractants towards Lactobacillus casei cells, Bulg. Chem. Commun. 41 (2009) 368–373. [46] M. Aimer, Reaktivextraktion von Milchsäure (2015) 163, https://doi.org/10. 18419/opus-1463. [47] I.S. Udachan, A.K. Sahoo, A study of parameters affecting the solvent extraction of lactic acid from fermentation broth, Brazilian J. Chem. Eng. 31 (2014) 821–827, https://doi.org/10.1590/0104-6632.20140313s00002495. [48] D. Yankov, J. Molinier, J. Albet, G. Malmary, G. Kyuchoukov, Lactic acid extraction from aqueous solutions with tri-n-octylamine dissolved in decanol and dodecane, Biochem. Eng. J. 21 (2004) 63–71, https://doi.org/10.1016/j.bej.2004.03.006. [49] S. Isken, H.J. Heipieper, Toxicity of organic solvents to microorganisms, Encycl. Environ. Microbiol, John Wiley & Sons Inc, 2003 10.1002/0471263397.env264. [50] B.E. Tate, Polymerization of itaconic acid and derivatives, Fortschritte Der Hochpolym, Springer-Verlag, 2006, pp. 214–232, , https://doi.org/10.1007/ bfb0051282. [51] E.K. Lindskog, The upstream process: principal modes of operation, Biopharm. Process. Dev. Des. Implement. Manuf. Process, Elsevier, 2018, pp. 625–635 10.1016/B978-0-08-100623-8.00031-1. [52] M. San-Martín, C. Pazos, J. Coca, Reactive extraction of lactic acid with Alamine 336 in the presence of salts and lactose, J. Chem. Technol. Biotechnol. 54 (1992) 1–6, https://doi.org/10.1002/jctb.280540102. [53] A. Keshav, K.L. Wasewar, S. Chand, Extraction of propionic acid from model solutions: effect of ph, salts, substrate, and temperature, AIChE J. 55 (2009) 1705–1711, https://doi.org/10.1002/aic.11780. [54] E. Reyhanitash, S.R.A. Kersten, B. Schuur, Recovery of volatile fatty acids from fermented wastewater by adsorption, ACS Sustain. Chem. Eng. 5 (2017) 9176–9184, https://doi.org/10.1021/acssuschemeng.7b02095. [55] L.M.J. Sprakel, B. Schuur, Improving understanding of solvent effects on intermolecular interactions in reactive liquid–liquid extraction with Isothermal Titration Calorimetry and molecular modeling, J. Ind. Eng. Chem. 72 (2019) 364–373, https://doi.org/10.1016/j.jiec.2018.12.038. [56] L.M.J. Sprakel, A.F.M. Holtkamp, R. Bassa, B. Schuur, Swing processes for solvent regeneration in liquid-liquid extraction of succinic acid, Chem. Eng. Process. Process Intensif. 143 (2019) 107600, , https://doi.org/10.1016/j.cep.2019.107600.
References [1] A. Kuznetsov, A. Beloded, A. Derunets, V. Grosheva, L. Vakar, R. Kozlovskiy, V. Shvets, Biosynthesis of lactic acid in a membrane bioreactor for cleaner technology of polylactide production, Clean Technol. Environ. Policy. 19 (2017) 869–882, https://doi.org/10.1007/s10098-016-1275-z. [2] J. Vijayakumar, R. Aravindan, T. Viruthagiri, Recent trends in the production, purification and application of lactic acid, Chem. Biochem. Eng. Q. 22 (2008) 245–264. [3] R. Datta, M. Henry, Lactic acid: Recent advances in products, processes and technologies - a review, J. Chem. Technol. Biotechnol. 81 (2006) 1119–1129, https:// doi.org/10.1002/jctb.1486. [4] C.S. López-Garzón, A.J.J. Straathof, Recovery of carboxylic acids produced by fermentation, Biotechnol. Adv. 32 (2014) 873–904, https://doi.org/10.1016/j. biotechadv.2014.04.002. [5] C.V. Stevens, The bio-based economy: Step by step from demand push to market pull, in, Biofuels, Bioprod. Biorefining (2012) 495–496, https://doi.org/10.1002/ bbb.1361. [6] K.L. Wasewar, A.A. Yawalkar, J.A. Moulijn, V.G. Pangarkar, Fermentation of glucose to lactic acid coupled with reactive extraction: A review, Ind. Eng. Chem. Res. 43 (2004) 5969–5982. [7] C. King, Amine-based systems for carboxylic acid recovery, Chemtech (United States). 22 (5) (1992) 285–291. [8] I.H. Aljundi, J.M. Belovich, O. Talu, Adsorption of lactic acid from fermentation broth and aqueous solutions on Zeolite molecular sieves, Chem. Eng. Sci. 60 (2005) 5004–5009, https://doi.org/10.1016/j.ces.2005.04.034. [9] R. Kumar, H. Nanavati, S.B. Noronha, S.M. Mahajani, A continuous process for the recovery of lactic acid by reactive distillation, J. Chem. Technol. Biotechnol. 81 (2006) 1767–1777, https://doi.org/10.1002/jctb.1603. [10] D.H. Han, W.H. Hong, Reactive extraction of lactic acid with trioctylamine/methylene chloride/n-hexane, Sep. Sci. Technol. 31 (1996) 1123–1135, https://doi. org/10.1080/01496399608001338. [11] P. Pal, J. Sikder, S. Roy, L. Giorno, Process intensification in lactic acid production: a review of membrane based processes, Chem. Eng. Process. Process Intensif. (2009) 1549–1559, https://doi.org/10.1016/j.cep.2009.09.003. [12] L.M.J. Sprakel, B. Schuur, Solvent developments for liquid-liquid extraction of carboxylic acids in perspective, Sep. Purif. Technol. (2019) 935–957, https://doi. org/10.1016/j.seppur.2018.10.023. [13] E. Reyhanitash, T. Brouwer, S.R.A. Kersten, A.G.J. van der Ham, B. Schuur, Liquid–liquid extraction-based process concepts for recovery of carboxylic acids from aqueous streams evaluated for dilute streams, Chem. Eng. Res. Des. 137 (2018) 510–533, https://doi.org/10.1016/j.cherd.2018.07.038. [14] T. Brouwer, M. Blahusiak, K. Babic, B. Schuur, Reactive extraction and recovery of levulinic acid, formic acid and furfural from aqueous solutions containing sulphuric acid, Sep. Purif. Technol. 185 (2017) 186–195, https://doi.org/10.1016/j.seppur. 2017.05.036. [15] A. Krzyzaniak, M. Leeman, F. Vossebeld, T.J. Visser, B. Schuur, A.B. De Haan, Novel extractants for the recovery of fermentation derived lactic acid, Sep. Purif. Technol. 111 (2013) 82–89, https://doi.org/10.1016/j.seppur.2013.03.031. [16] A. Krzyzaniak, B. Schuur, A.B. De Haan, Equilibrium studies on lactic acid extraction with N, N-didodecylpyridin-4-amine (DDAP) extractant, Chem. Eng. Sci. 109 (2014) 236–243, https://doi.org/10.1016/j.ces.2014.01.030. [17] J. Procházka, A. Heyberger, V. Bízek, M. Koušová, E. Volaufová, Amine extraction of hydroxycarboxylic acids. 2. comparison of equilibria for lactic, malic, and citric acids, Ind. Eng. Chem. Res. 33 (1994) 1565–1573, https://doi.org/10.1021/ ie00030a016. [18] J.G.M. Susanti, B. Winkelman, H.J. Schuur, J. Heeres Yue, Lactic acid extraction and mass transfer characteristics in slug flow capillary microreactors, Ind. Eng. Chem. Res. 55 (2016) 4691–4702, https://doi.org/10.1021/acs.iecr.5b04917. [19] W. Qin, Z. Li, Y. Dai, Extraction of monocarboxylic acids with trioctylamine: equilibria and correlation of apparent reactive equilibrium constant, Ind. Eng. Chem. Res. 42 (2003) 6196–6204, https://doi.org/10.1021/ie021049b. [20] J. Marták, Š. Schlosser, S. Vlčková, Pertraction of lactic acid through supported liquid membranes containing phosphonium ionic liquid, J. Memb. Sci. 318 (2008) 298–310, https://doi.org/10.1016/j.memsci.2008.02.064. [21] J. Marták, Š. Schlosser, M. Blahušiak, Mass-transfer in pertraction of butyric acid by phosphonium ionic liquids and dodecane, Chem. Pap. (2011) 608–619, https://doi. org/10.2478/s11696-011-0069-3. [22] E. Reyhanitash, B. Zaalberg, H.M. Ijmker, S.R.A. Kersten, B. Schuur, CO2-enhanced extraction of acetic acid from fermented wastewater, Green Chem. 17 (2015) 4393–4400, https://doi.org/10.1039/c5gc01061f. [23] K. Tonova, I. Svinyarov, M.G. Bogdanov, Hydrophobic 3-alkyl-1-methylimidazolium saccharinates as extractants for l-lactic acid recovery, Sep. Purif. Technol. 125 (2014) 239–246, https://doi.org/10.1016/j.seppur.2014.02.001. [24] E. Reyhanitash, E. Fufachev, K.D. Van Munster, M.B.M. Van Beek, L.M.J. Sprakel, C.N. Edelijn, B.M. Weckhuysen, S.R.A. Kersten, P.C.A. Bruijnincx, B. Schuur, Recovery and conversion of acetic acid from a phosphonium phosphinate ionic liquid to enable valorization of fermented wastewater, Green Chem. 21 (2019) 2023–2034, https://doi.org/10.1039/c9gc00725c. [25] D.L. Grzenia, D.J. Schell, S. Ranil Wickramsinghe, Detoxification of biomass hydrolysates by reactive membrane extraction, J. Memb. Sci. 348 (2010) 6–12, https://doi.org/10.1016/j.memsci.2009.10.035. [26] D.L. Grzenia, D.J. Schell, S. Ranil Wickramasinghe, Membrane extraction for detoxification of biomass hydrolysates, Bioresour. Technol. 111 (2012) 248–254, https://doi.org/10.1016/j.biortech.2012.01.169.
8
Contents lists available at ScienceDirect
Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur
In-situ recovery of carboxylic acids from fermentation broths through membrane supported reactive extraction using membrane modules with improved stability
T
Angelo Gössia,b, Florian Burgenerb, David Kohlerb, Alessandro Ursob, Boris A. Kolvenbachb, ⁎ Wolfgang Riedlb, Boelo Schuura, a b
University of Twente, Faculty of Science and Technology, Sustainable Process Technology Group, 7500AE Enschede, the Netherlands University of Applied Sciences and Arts Northwestern Switzerland, Institute for Chemistry and Bioanalytics, Hofackerstrasse 30, 4132 Muttenz, Switzerland
A R T I C LE I N FO
A B S T R A C T
Keywords: Reactive extraction Process intensification Product recovery Carboxylic acids Membrane extraction process
Membrane supported reactive extraction (MSE) coupled to back-extraction (MSBE) using a new type of Teflon (PTFE) capillary membrane contactor was studied for the in-situ removal of carboxylic acids from aqueous streams, e.g. fermentation broths. The use of microporous membranes as extraction interface helps avoiding emulsification problems, allows the use of extreme phase ratios, and protects microorganisms, as they are less affected by solvent toxicity during in-situ extractions. The use of PTFE capillary membranes is suitable for longterm use due its high chemical and thermal stability. A simple toxicity screening identified n-decanol with tri n-octyl amine (TOA) as a suitable solvent. MSE experiments were performed using membrane contactors (0.005 m2 to 0.15 m2), working with solvent to feed phase ratios down to 1:40 (mass based). The in-situ removal of lactic acid out of fermentation broths using lactobacillus plantarum led to a glucose conversion rate of 80 mol%. Additionally, a concentration factor up to 7.8 could be shown during back-extraction.
1. Introduction The demand for bio-based chemicals in general and bio-based polymers in particular has tremendously increased in the last few years [1]. Carboxylic acids like lactic acid, mandelic acid, succinic acid and itaconic acid are up-and-coming substances which are used on large scale. For example, a growing interest in bio-based and/or bio-degradable polymers increased the production of lactic acid from 60.000 tons in 2006 to 494.000 tons in 2013 [2]. Many carboxylic acids are currently produced by fermentation of carbohydrates [3,4]. The release of acid during the fermentation leads to a production inhibition, which is why the pH has to be controlled to ensure optimal growth conditions for the microorganisms [5]. Usually, the pH is maintained by addition of calcium hydroxide, leading to carboxylates. After the fermentation has finished, sulfuric acid has to be added to release the acid from its carboxylate salt form into its neutral acid form, resulting in about 1 ton of gypsum per ton carboxylic acid as by-product [3]. With increasing demand of carboxylic acids this costly method of recovery is neither environmentally friendly nor economical [3,6,7]. Several techniques have been investigated in the past to avoid
⁎
gypsum production, mainly by in-situ product removal techniques such as adsorption [8], reactive distillation [9], stripping [10], reactive liquid-liquid extraction [6] and membrane-based processes (electro-dialysis, ultrafiltration, nanofiltration) [11]. Liquid-liquid extraction of carboxylic acids is a widely studied research field [12,13]. Often composite solvents comprising a diluent and a complexing extractant are used, such as tri-octylphosphine oxide [14] or, more frequently, tertiary amines [7,10,15,17–19]. Although ionic liquids have shown great potential [20 21–23], it has been shown that the regeneration from the ionic liquid in dilute systems is challenging [24]. The published method to regenerate the acid from the ionic liquid involves thermal treatment challenging [24], which is not feasible for lactic acid production. For this application the more traditional amine based extraction systems appear more suitable. However, the operational limitations such as the formation of stable emulsions and limited flexibility in solvent/feed (S/F) ratios remain points of attention. By applying a membrane supported extraction (MSE), both emulsification can be prevented, and an unlimited range of (S/F) becomes accessible. The in-situ removal of substances inhibiting production out of fermentation broths has been studied previously ([4,25–27]. However, the
Corresponding author. E-mail address: [email protected] (B. Schuur).
https://doi.org/10.1016/j.seppur.2020.116694 Received 8 November 2019; Received in revised form 23 January 2020; Accepted 8 February 2020 1383-5866/ © 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
conventional liquid-liquid extraction systems suffer from the negative effect of the direct contact of solvent with microorganisms [28]. Several technologies have been described to minimize or even avoid the contact between microorganisms and organic phases. These include for example membrane pertraction [20,29,30] or solvent impregnated resins containing ionic liquids [31]. The use of membranes as extraction interface represents a promising way to reduce solvent toxicity on microorganisms [6,32], since the immobilized extraction interface protects microorganism from long-term contact. Most authors use commercially available hollow fiber membrane contactors, made out of polypropylene and polyurethane [25–27]. While these contactors provide remarkable contact areas based on their size, their solvent stability is limited and they are difficult to clean due to the embedded materials. Disadvantages of membrane contactors include the increased mass transfer resistance, limited membrane lifetime and the risk of fouling. To reduce the mass transfer resistance, the membrane should be thin, while the pore size should be as large as possible. [33,34] Alternatively, dense (=non-porous) membranes can be used, which are less affected by fouling. However, they suffer from reduced permeability compared to porous membranes [35]. This paper aims to characterize a new type of capillary membrane contactor based on commercially available (porous) PTFE fibers, which allow for a long-term use due to its superior chemical stability and good cleanability.
2.3. Experimental methods
2. Material and methods
D=
2.3.1. Solvent and toxicity screening For the extraction efficiency tests, 20 mL of a 0.68 M lactic acid solution was extracted with 20 mL organic phase containing 20% TOA. The occurrence of stable emulsions was examined by vigorous shaking of the samples and observation of the phase separation. Phase separation was reached with centrifugation for 20 min at 4500 × g. The concentration of the lactic acid bound to the reactive agent in the organic phase was determined by mass balance. A simple toxicity screening approach was used regarding dissolved organic solvents (molecular level toxicity) and excess solvent (phase level toxicity) [32,36-38]. For the molecular level toxicity, 30 mL of water was saturated with organic solvent. The excess organic solvent was then removed and the aqueous phase was used for a batch fermentation after sterile filtration (culture medium see chapter 2.3). The phase level toxicity was studied by adding 5 mL of organic solvent to 30 mL of a batch fermentation. The resulting lactic acid concentration for both tests was measured after 72 h by direct titration. The extraction efficiency of a solvent was determined by its distribution coefficient as given in Eq. (1). In this equation, all forms of lactic acid are lumped, thus aqueous lactate and lactic acid concentrations are summed, as well as any free lactic acid in the organic phase, and lactic acid bound to an extractant in the organic phase.
[LA]equil, org [LA]equil, aq
(1)
2.1. Chemicals All chemicals used were purchased from Sigma Aldrich Switzerland. Tri-n-octylamine (TOA) (98%), sodium hydroxide standard solution (volumetric for titration 0.1 M NaOH), n-heptane (> 99%) and n-decanol (> 98%), 2-ethyl-1-hexanol (> 99%), n-octanol (> 99%), tributylphosphate (> 99%) , methyl isobutyl ketone (MIBK, > 98.5%), ndodecanol (98%), isobutanol (> 99%), oleyl alcohol (85%), dodecane (> 99%), itaconic acid (> 99%), mandelic acid (99%), succinic acid (> 99%) were used. The aqueous lactic acid solutions were prepared from an 85 wt% lactic acid solution.
2.3.2. Membrane supported extraction experiments Different starting concentrations of carboxylic acids ranging from 0.3 to 3 wt% were used. If not stated differently, the experiments were performed at room temperature, with 20 wt% TOA in n-decanol, at a flow rate of the aqueous shell side of 12 kg/h and an organic flow rate of 6 kg/h (inside the fibers/lumen). All experiments were started by pumping of the aqueous phase and applying an overpressure of about 100 mbar to ambient pressure. Subsequently, the circulation of the organic phase was started, keeping the pressure about 80 mbar below the aqueous phase. This avoids breakthrough of organic phase into the aqueous phase. Temperature, flow rates and pressure were monitored. In order to work with higher solvent-to-feed ratios (S/F), an external 20 L glass vessel could be connected. Throughout all experiments, the organic phase was pumped through the fibers (lumen side), and the aqueous phase through the shell side of the membrane contactor. Extraction experiments were performed using different capillary membrane contactor sizes (0.005 m2 to 0.15 m2) and phase ratios (up to 1:40 S/F) and different substrate/ reactive agent concentrations. Both the aqueous and the organic phase were circulated loop-wise until reaching the equilibrium, this operation mode allowed adequate and accurate analysis of the process even at the small membrane sized applied. The concentration in the organic phase was calculated via mass balance.
2.2. Experimental setup Fig. 1 shows the applied membrane extraction setup. This setup was used for either extraction or back-extraction experiments. It consists of two temperature controlled 2 L vessels for the aqueous and organic phase, respectively. Two gear pumps are used to pump the phases through the membrane and back into the container. The pressure of both phases can be adjusted by needle valves. Coriolis flow meters record the flow as well as the density of both phases. For continuous fermentation experiments, a glucose-feeding pump is connected to the aqueous phase vessel. Phase equilibrium and back-extraction experiments were conducted in 50 mL Falcon tubes. Intensive phase contact was ensured by regular shaking of the tubes, for a total of one hour. Different temperatures were applied using a water bath. Phase separation was reached with an Eppendorf Centrifuge 5804 R, A-4–44 Rotor, at 4500 × g for 20 min). Table 1 shows characteristics of the used membrane fiber. All parts of the modules are either from stainless steel, glass, polytetrafluoroethene (PTFE) or polyvinylidene fluoride (PVDF). The MSE setup was cleaned after every experiment using acetone and then dried using oil-free compressed air. When working with microorganisms, the setup was disinfected with 70 vol% ethanol for > 20 min. Afterwards, the setup was purged with ethanol and dried using oil-free compressed air. Due to the chemically inert materials used in the membrane contactor, the modules can also be autoclaved, steam sterilized or ozonized.
2.3.3. Back-extraction experiments Lactic acid was added to a solution of 20 wt% TOA in n-decanol to prepare a saturated solution. This solution was then contacted with different volumes of water (the receiving phase) and heptane (antisolvent) in Falcon tubes. The tubes were temperature regulated using a water bath. Intensive phase contact was ensured by regular shaking of the tubes, for a total of one hour. The phases were separated in the temperature regulated centrifuge and the samples of the aqueous phases were taken. The concentration of lactic acid in the aqueous phase was measured by titration (see chapter 2.4). 2
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
Fig. 1. Used MSE-Setup, gear pumps range 0 L/h – 60 L/h, temperature variable 20–60 °C, all phase-contacting parts made out of PVDF, PTFE, glass or stainless steel (1.4404). Additional glucose feeding required for fermentation experiments only.
of carboxylic acids [16,25-27,40,41]. However, the octanol-trioctylamine-water-carboxylic acid system is known to form stable emulsions. Additionally, octanol shows some water solubility and a strong solvent toxicity [6,27]. Generally speaking, with an increasing octanol–water partition coefficient 10log(Poctanol) of a solvent, the probability increases that it is non-toxic to microorganisms [42]. Solvents having a 10log (Poctanol) > 5 are generally biocompatible for cellular biocatalysis [43]. Solvents that can form hydrogen bonds stabilize the carboxylic acidtrioctylamine complex [44,12]. Thus, alcohols are the most widely used species of solvents for this kind of reactive extraction and are expected to result in high extraction efficiencies.
2.3.4. Fermentation experiment: In-situ lactic acid extraction Lactobacillus plantarum (DSM 2648) was used for fermentation experiments. The used culture medium contained the following ingredients, per liter: 20 g Glucose, 5 g Yeast extract, 0.4 g, magnesiumsulfate heptahydrate, 0.1 g mangansulfate monohydrate, 5 g sodiumacetate, 2 g triammonium citrate, 10 g casein protein, 10 g meat extract, 2 g dipotassium phosphate. The pre-culture was developed using the same culture medium at 37 °C for 16 h (150 RPM, Infors HAT Multitron). The pH of the pre-cultures was in the range of 4.5–5, indicating that production of lactic acid had already commenced. Inoculum was added to sterile culture medium in a phase ratio of 1/500 vol/vol. 2.4. Analytical techniques
4. Results and discussion The analysis of the lactic acid concentration in the aqueous phase was performed either by titration or by capillary electrophoresis (CE). A Mettler Toledo G20 device was used for the potentiometric direct titration method (0.1 M NaOH). For low concentrations, capillary electrophoresis was used with the Agilent organic acids solutions kit method. The systematic error was below 3% for both methods.
4.1. Solvent screening A set of nine solvents had been selected for a solvent screening. The selection was based on literature research as well as on the octanol–water-distribution coefficients 10log(Poctanol), which can be an indicator for solvent toxicity. [6,43,45–49]. Key criteria for solvent selection were low toxicity and high extraction efficiency (see Table 2). Due to their low distribution coefficients, oleyl alcohol and dodecane were not studied in further tests. Subsequently, the remaining solvents were investigated regarding their molecular and phase level toxicity. Fig. 2 shows the resulting lactic acid concentration after 72 h of fermentation stressed by either molecular or phase level toxicity of the applied solvents. A significant difference between the molecular and phase level toxicity results can be seen. All solvents strongly inhibit the lactic acid production when applied in direct contact with microorganisms.
3. Theory The reactive extraction using tertiary amines extracts only the undissociated form of the carboxylic acid, meaning that the process is more efficient at low pH. However, when working with fermentation broths, the pH has to be kept > 5 for optimal culture conditions [39]. The trioctylamine-carboxylic acid complex is stabilized by hydrogen bonds with the solvent, leading to much higher complexation constants in the presence of hydrogen bond acceptors or donors (e.g. alcohols). n-octanol is the most widely used solvent for the reactive extraction
Table 1 Used PTFE membrane. The membrane porosity was achieved by stretching; the tortuosity is estimated at 2–3, based on Field Emission Scanning Electron Microscopy (FESEM). Membrane purchased from Memo3, Switzerland. Membrane Material
Porosity [%]
Pore diameter [µm]
Inner Diameter [mm]
Outer diameter [mm]
Wall thickness [µm]
PTFE
54
0.47
2.97
3.49
260
3
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
However, the toxicity on the molecular level is much lower compared to the phase level toxicity. The use of a microporous membrane as extraction interface prevents direct contact with the solvent, helping against phase level toxicity. The negative effects of molecular level toxicity however cannot be reduced. Due to its good extraction efficiency and its low molecular level toxicity decanol is a very promising solvent for membrane supported extractions. Octanol, the most widely used solvent for reactive extraction of carboxylic acids with tri-n-octylamine acids [16,25-27,40,41] shows good extraction efficiency, but its molecular level toxicity is too high for in-situ extraction with efficient fermentation. Thus, a separation of the microorganisms from the extraction process would be required to avoid molecular level toxicity. Therefore, we propose using decanol as solvent in the in-situ extraction. Decanol offers a good compromise between extraction efficiency and toxicity. Apart from the toxicity issue, the membrane helps to overcome the emulsification problems occurring with all investigated solvents and to achieve the desired extreme phase ratios.
available membrane contactors in terms of membrane material, pore diameter and wall thickness. The use of these innovative types of membranes might provide significant improvement regarding process stability and cost efficiency. The MSE set-up described under 2.2 was used for all membrane experiments. Experiments were conducted at 25 °C. Fig. 3 shows typical membrane extraction diagrams. Depending on the membrane area and the process parameters, the extraction process took between 8 h and 7 days. Phase ratios (S/F) between 1:1 and 1:40 were tested. Membrane extractions were performed with following carboxylic acids: lactic-, succinic-, mandelic- and itaconic acid, with initial concentrations of 3%. The differences in the resulting final concentrations can be explained due to different physical distribution coefficients which are highest for itaconic and mandelic acid. The physical distribution coefficients resulting from the extraction of 3% carboxyilic acid solutions with n-decanol are 0.14, 0.15, 0.50 and 0.56 for lactic-, mandelic, itaconic- and succinic acid. Noticeable is the difference between the two dicarboxylic acids succinic- and itaconic acid, which can be explained due to the lower pKa’s of itaconic acid, increasing the extraction efficiency. Additionally, itaconic acid could tend to undergo further reactions (e.g. polymerization) [50]. The used membranes are non-selective, meaning that they do not influence the phase equilibrium in any way. All MSE distribution coefficients were in excellent accordance with distribution coefficients measured in batch liquid-liquid equilibrium experiments. As expected, the formation of emulsions could be avoided during all MSE experiments. The used PTFE membrane allows the immobilization of the extraction interface at the aqueous side of the membrane, filling the pores of the membrane with 20 wt% TOA in n-decanol. Extraction experiments with fermentation broths using n-octanol as solvent led to a rapid decrease in living microorganisms in less than 24 h. However, working with n-decanol as solvent, the number of live cells remained constant for 24 h.
4.2. Membrane supported extraction experiments
4.3. In-situ lactic acid extraction out of fermentation broths
Table 2 Distribution coefficients of lactic acid and octanol–water distribution coefficients of solvents. Extraction of 20 mL of a 0.68 mol/dm3 lactic acid solution with 20 mL organic phase containing 20 wt% TOA at 25 °C, *source: SigmaAldrich. Solvent
Solvent
2-ethyl-1-hexanol n-octanol n-decanol Tributylphosphate MIBK n-dodecanol Isobutanol oleyl alcohol Dodecane
2.9 – 4.5 4.0 1.3 4.8 1.0 7.5 7.0
10
logPoctanol*
Distribution coefficient of lactic acid [-] 15.1 14.6 12.7 10.7 8.8 7.1 5.7 5.4 0.5
The types of membranes used in this study differ from commercially
An in-situ extraction of lactic acid from a batch fermentation with
Fig. 2. Resulting lactic acid concentrations (wt%) from toxicity tests after 72 h of fermentation. The Blank experiment represents the fermentation without any solvents. 4
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
Fig. 3. Extraction of carboxylic acids out of an aqueous stream. Initial carboxylic acid conc. = 3.0 wt%, S/F 1:1 vol/vol, A = 0.059 m2, 20 wt% TOA in n-decanol. Dotted lines are the fitted extraction curves for each carboxylic acid. The last values for each acid are equilibrium values derived from batch extraction experiments.
presence should initiate a salting out of acids and therefore increase the distribution coefficient. Contrarily, studies showed that the presence of salts reduces the extraction of carboxylic acids. As an example, the presence of sodium chloride results in the formation of hydrochloric acid from Cl- and the proton from the carboxylic acid. Hydrochloric acid then competes with lactic acid for the extraction [6,52,53], similar effects were observed for amine-functionalized resins, which were found to prefer sulfuric acid, phosphoric acid and hydrochloric acid over volatile fatty acids [54]. The use of membranes as extraction interface slows down the process due to the limited contact area available. Since tertiary amines only extract undissociated acid, fermentation at low pH using (genetically modified) acid-tolerant organisms promises significantly increased extraction rates [4]. It has to be noted that the production rate of lactic acid had to be kept rather low due to the limited contact area available. Since the organic phase eventually is saturated, the continuous removal of carboxylic acids requires an additional back-extraction step. Fig. 5 shows the desired process on the example of a fermentative production of carboxylic acid. After the in-situ extraction, the loaded organic phase is regenerated in a (membrane supported) back-extraction step, in which the carboxylic acid is transferred into water. The back-extraction is described more in detail in the following chapter.
lactobacillus plantarum (DSMZ 2648) was set up. The goal was to maintain the fermentation process as long as possible and to achieve a high yield of lactic acid. A pH set point of 5 was chosen as compromise between microorganism activity and extraction efficiency. As the size of the membrane contact area (0.059 m2) was prone to limit the mass transfer, the product formation had to be controlled by a constant feed rate of medium containing 20 g/L glucose with a rate of 2 mL/hour. The initial medium contained 5 g/L glucose. Additional glucose was added when the pH of the fermentation rate increased over 5. First extraction trials using n-octanol led to no production of lactic acid at all. In fact, almost no live cells could be detected after 24 h of fermentation. However, the in-situ extraction with n-decanol led to promising results. The fermentation could be kept up for more than 7 days. The pH in the fermenter was solely controlled by the dosage of culture medium, without the addition of any base. The results of the fermentation run with in-situ extraction using MSE are shown in Fig. 4. The fermentation process was stable for more than 7 days. The microorganisms in the fermenter were still active. However, the productivity after 7 days was reduced. After the addition of 5 g of glucose, the pH only dropped slightly. Nevertheless a sample of the fermentation broth after 8 days could be used to start a new pre-culture, resulting in a pH of 4.8 after 16 h, which fits well the pH of other pre-cultures. In total, 25.8 g of glucose was added to the fermentation broth, resulting in 21.0 g of lactic acid. This corresponds to a conversion rate of 81 mol % of the fed glucose. This is remarkable, considering the achievable lactic acid concentrations without neutralization or membrane extraction, as shown in Fig. 5. The in-situ removal of lactic acid increased the yield by factor 4.3 when compared to a fermentation without neutralization. Without a membrane protecting the microorganisms from n-decanol phase toxicity, only less than 0.1% of lactic acid was formed. There was no biomass removed during the experiment. However, during a continuous fermentation process, dead microorganisms need to be removed eventually from fermentation broth e.g. by cross flow filtration [51]. Nevertheless some issues remain to be solved. Culture media typically contain different salts, like NaCl, Na2SO4 or K2HPO4. Their
4.4. Back-extraction of lactic acid In general, the fermentative production of platform chemicals is done in large scales. The use of a membrane supported extraction allows an extraction with big phase ratios. Thus, the cost intensive organic phase can be kept at a minimum. This, on the other hand, only allows the extraction of small amounts of carboxylic acid in a single extraction step. Nevertheless, by introducing a continuous back-extraction, it is still possible to extract sufficient amounts of acid. In order to back-extract the bound carboxylic acid into water, the tri-n-octylamine-carboxylic acid complexes need to be split. As discussed in the theory section, the complexes are stabilized by hydrogen bonds with the solvent n-decanol or n-octanol. By adding a solvent that 5
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
Fig. 4. In-situ extraction of lactic acid out of fermentation broths. Vferm = 0.7 L, Vorg = 0.7 L, dashed green line represents the goal value of pH = 5, red arrows represent additional dosage of glucose. The dashed red lines represent the lactic acid concentrations derived without neutralization and extraction without membrane, respectively.
(wash phase/extract phase, i.e. aqueous back-extraction phase to heptane in n-decanol ratio) between 0:1 and 3:1 were used. A concentration factor was calculated, assuming a feeding concentration at position 1 in Fig. 5 of 0.8% LA. This incoming feed concentration was chosen as a compromise between bacteria activity and extraction efficiency (see Table 3). A good compromise between concentration factor and recovery rate can be achieved at 70 °C, with a phase ratio S/F of 1:2. As shown in Fig. 6, a recovery rate of 31% results in a single step extraction, with an outgoing aqueous-solution containing 4.3%. Higher recovery rates could be achieved by working at even higher temperatures than 70 °C. The PTFE membrane modules can endure
cannot donate or accept hydrogen bonds (like heptane), the nature of the complexes and the stoichiometry changes [12,55], and the distribution can be reduced [16,37]. Additionally, an increase in temperature of the organic phase reduces the complexation constant even more, increasing the recovery rate [10,16,37,47]. However, the added heptane needs to be removed after back-extraction, e.g. by a falling- or thin-film evaporation. Back-extraction experiments were performed using a 20 wt% TOA with a lactic acid concentration of 54 g/L in n-decanol (the concentration was based on lactic acid uptake in previous extraction experiments, extracting from 2.5 wt% lactic acid at S/F of 1:4). Temperatures between 20 and 70 °C, S/F (n-decanol phase to water) ratios between 1:1 and 1:8, and W/E
Fig. 5. Shows the conceptual process design for a fermentation with in-situ removal of lactic acid. The lactic acid produced in the fermenter at ① extracted into 20 wt % TOA in n-decanol and then back-extracted in a temperature- and solvent-swing process into a pure water stream. The resulting product stream is at ②. 6
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
Table 3 Back-extraction experiments with the developed temperature- and solvent-swing process on the example of lactic acid and 20% TOA in n-decanol. Resulting outgoing concentrations of lactic acid (LA) and concentration factor reachable. LA conc. [wt %] ①
Temperature [°C]
W/E [-]
S/F[-]
LA conc.[wt %] ②
Recovery [%]
Concentration Factor[-]
0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8
20 20 20 20 20 20 20 20 20 20 20 20 50 50 50 50 50 50 50 50 50 50 50 50 70 70 70 70 70 70 70 70 70 70 70 70
0 0 0 0 1 1 1 1 3 3 3 3 0 0 0 0 1 1 1 1 3 3 3 3 0 0 0 0 1 1 1 1 3 3 3 3
1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8
0.34 0.43 0.50 0.58 0.89 1.26 1.62 2.04 1.40 2.12 2.86 3.62 0.66 0.82 1.02 1.27 1.39 1.92 2.63 3.36 1.97 3.04 4.44 5.31 0.70 0.97 1.60 1.88 1.46 2.21 3.19 3.84 2.00 3.42 5.03 6.21
6.2% 3.9% 2.3% 1.3% 16.3% 11.6% 7.5% 4.7% 25.7% 19.5% 13.2% 8.3% 12.1% 7.5% 4.7% 2.9% 25.6% 17.6% 12.1% 7.7% 36.2% 28.0% 20.4% 12.2% 12.8% 8.9% 7.3% 4.3% 26.9% 20.3% 14.6% 8.8% 36.7% 31.4% 23.1% 14.3%
0.4 0.5 0.6 0.7 1.1 1.6 2.0 2.6 1.7 2.7 3.6 4.5 0.8 1.0 1.3 1.6 1.7 2.4 3.3 4.2 2.5 3.8 5.6 6.6 0.9 1.2 2.0 2.3 1.8 2.8 4.0 4.8 2.5 4.3 6.3 7.8
removed out of the organic phase by evaporation after the back-extraction. By doing so, the negative effects of heptane on extraction rate efficiency and the toxicity towards the microorganisms can be avoided. 5. Conclusions The here proposed membrane supported extraction (MSE) process using PTFE-based membranes showed promising results for the in-situ removal of carboxylic acids from aqueous streams at superior chemical membrane stability, being compatible with the investigated solvents (tri-n-octylamine in n-decanol showed best results), as well as with the fermentation broth. Reduced solvent toxicity due to the membrane interface was observed in continuous removal of lactic acid from a fermentation, allowing the fermentation to run for five days yielding a conversion of 80% without any neutralization required. Extraction followed by backextraction allowed concentration factors up to 7.8. CRediT authorship contribution statement Angelo Gössi: Investigation, Writing – original draft. Florian Burgener: Investitation. David Kohler: Investitation. Alessandro Urso: Investitation. Boris A. Kolvenbach: Investitation. Prof. Wolfgang Riedl: Supervision, Writing – review and editing. Prof. Boelo Schuur: Supervisor, Writing – review and editing.
Fig. 6. Back-extraction efficiency comparison at different S/F ratios. Incoming LA concentration of 0.8%, 70 °C, W/E 1:3. The dots represent the concentration factor.
temperatures > 120 °C (b.p. heptane 98 °C at ambient pressure). An increased temperature also benefits the removal by means of distillation. Alternatively, other antisolvent (gaseous or liquid) could be used [56]. Due to the significant boiling point difference of heptane and ndecanol (98 °C resp. 230 °C under ambient pressure), heptane can be
Declarations of Competing Interest The authors declare that they have no known competing financial 7
Separation and Purification Technology 241 (2020) 116694
A. Gössi, et al.
interests or personal relationships that could have appeared to influence the work reported in this paper.
[27] K. De Sitter, L. Garcia-Gonzalez, C. Matassa, L. Bertin, H. De Wever, The use of membrane based reactive extraction for the recovery of carboxylic acids from thin stillage, Sep. Purif. Technol. (2018) 177–185, https://doi.org/10.1016/j.seppur. 2018.06.001. [28] M.D. Waghmare, K.L. Wasewar, S.S. Sonawane, D.Z. Shende, Natural nontoxic solvents for recovery of picolinic acid by reactive extraction, Ind. Eng. Chem. Res. 50 (2011) 13526–13537, https://doi.org/10.1021/ie201228u. [29] S. Schlosser, R. Kertész, J. Marták, Recovery and separation of organic acids by membrane-based solvent extraction and pertraction: an overview with a case study on recovery of MPCA, Sep. Purif. Technol. (2005) 237–266, https://doi.org/10. 1016/j.seppur.2004.07.019. [30] N.D. Patil, A.W. Patwardhan, A.V. Patwardhan, Carboxylic acids separation using hollow fiber supported liquid membrane, Indian J. Chem. Technol. 24 (2017) 20–31. [31] M. Blahušiak, S. Schlosser, J. Marták, Extraction of butyric acid by a solvent impregnated resin containing ionic liquid, React. Funct. Polym. 71 (2011) 736–744, https://doi.org/10.1016/j.reactfunctpolym.2011.04.002. [32] K.L. Wasewar, Reactive extraction: an intensifying approach for carboxylic acid separation, Int. J. Chem. Eng. Appl. 3 (2012) 249–255, https://doi.org/10.7763/ ijcea.2012.v3.195. [33] Basic Equations of Mass Transport Through a Membrane Layer, Elsevier, 2019. doi:10.1016/c2016-0-04043-3. [34] E. Drioli, A. Criscuoli, E. Curcio, Membrane Contactors: Fundamentals, Applications and Potentialities, Elsevier Science, vol. 11, 2005. [35] Handbook of Membrane Reactors | ScienceDirect, (n.d.). (accessed January 18, 2020). https://www.sciencedirect.com/book/9780857094148/handbook-ofmembrane-reactors. [36] R. Bar, J.L. Gainer, Acid fermentation in water-organic solvent two-liquid phase systems, Biotechnol. Prog. 3 (1987) 109–114, https://doi.org/10.1002/btpr. 5420030208. [37] Y.K. Hong, W.H. Hong, D.H. Han, Application of reactive extraction to recovery of carboxylic acids, Biotechnol. Bioprocess Eng. 6 (2001) 386–394, https://doi.org/ 10.1007/BF02932319. [38] L. Bassetti, J. Tramper, Organic solvent toxicity in Morinda citrifolia cell suspensions, Enzyme Microb. Technol. (1994) 642–648, https://doi.org/10.1016/01410229(94)90084-1. [39] G. Greafe, C.H. Holten, Lactic acid, properties and chemistry of lactic acid and derivatives, Verlag Chemie GmbH, 1971. [40] A. Schunk, A. Menert, G. Maurer, On the influence of some inorganic salts on the partitioning of citric acid between water and organic solutions of tri-n-octylamine: Part I: Methyl isobutyl ketone as organic solvent, Fluid Phase Equilib. (2004) 55–72, https://doi.org/10.1016/j.fluid.2004.04.010. [41] K.L. Wasewar, V.G. Pangarkar, A.B.M. Heesink, G.F. Versteeg, Intensification of enzymatic conversion of glucose to lactic acid by reactive extraction, Chem. Eng. Sci. 58 (2003) 3385–3393, https://doi.org/10.1016/S0009-2509(03)00221-5. [42] M. Matsumoto, Toxicity of organic solvents and ionic liquids to lactic acid-producing microbes, Org. Solvents Prop. Toxicity, Ind. Eff. (2011) 105–114. [43] M. Vermuë, J. Sikkema, A. Verheul, R. Bakker, J. Tramper, Toxicity of homologous series of organic solvents for the gram-positive bacteria Arthrobacter and Nocardia Sp. and the gram-negative bacteria Acinetobacter and Pseudomonas Sp, Biotechnol. Bioeng. 42 (1993) 747–758, https://doi.org/10.1002/bit.260420610. [44] F. Chemarin, M. Moussa, M. Chadni, B. Pollet, P. Lieben, F. Allais, I.C. Trelea, V. Athès, New insights in reactive extraction mechanisms of organic acids: An experimental approach for 3-hydroxypropionic acid extraction with tri-n-octylamine, Sep. Purificat. Technol. 179 (2017) 523–532, https://doi.org/10.1016/j.seppur. 2017.02.018. [45] N.A. Marinova, D.S. Yankov, Toxicity of some solvents and extractants towards Lactobacillus casei cells, Bulg. Chem. Commun. 41 (2009) 368–373. [46] M. Aimer, Reaktivextraktion von Milchsäure (2015) 163, https://doi.org/10. 18419/opus-1463. [47] I.S. Udachan, A.K. Sahoo, A study of parameters affecting the solvent extraction of lactic acid from fermentation broth, Brazilian J. Chem. Eng. 31 (2014) 821–827, https://doi.org/10.1590/0104-6632.20140313s00002495. [48] D. Yankov, J. Molinier, J. Albet, G. Malmary, G. Kyuchoukov, Lactic acid extraction from aqueous solutions with tri-n-octylamine dissolved in decanol and dodecane, Biochem. Eng. J. 21 (2004) 63–71, https://doi.org/10.1016/j.bej.2004.03.006. [49] S. Isken, H.J. Heipieper, Toxicity of organic solvents to microorganisms, Encycl. Environ. Microbiol, John Wiley & Sons Inc, 2003 10.1002/0471263397.env264. [50] B.E. Tate, Polymerization of itaconic acid and derivatives, Fortschritte Der Hochpolym, Springer-Verlag, 2006, pp. 214–232, , https://doi.org/10.1007/ bfb0051282. [51] E.K. Lindskog, The upstream process: principal modes of operation, Biopharm. Process. Dev. Des. Implement. Manuf. Process, Elsevier, 2018, pp. 625–635 10.1016/B978-0-08-100623-8.00031-1. [52] M. San-Martín, C. Pazos, J. Coca, Reactive extraction of lactic acid with Alamine 336 in the presence of salts and lactose, J. Chem. Technol. Biotechnol. 54 (1992) 1–6, https://doi.org/10.1002/jctb.280540102. [53] A. Keshav, K.L. Wasewar, S. Chand, Extraction of propionic acid from model solutions: effect of ph, salts, substrate, and temperature, AIChE J. 55 (2009) 1705–1711, https://doi.org/10.1002/aic.11780. [54] E. Reyhanitash, S.R.A. Kersten, B. Schuur, Recovery of volatile fatty acids from fermented wastewater by adsorption, ACS Sustain. Chem. Eng. 5 (2017) 9176–9184, https://doi.org/10.1021/acssuschemeng.7b02095. [55] L.M.J. Sprakel, B. Schuur, Improving understanding of solvent effects on intermolecular interactions in reactive liquid–liquid extraction with Isothermal Titration Calorimetry and molecular modeling, J. Ind. Eng. Chem. 72 (2019) 364–373, https://doi.org/10.1016/j.jiec.2018.12.038. [56] L.M.J. Sprakel, A.F.M. Holtkamp, R. Bassa, B. Schuur, Swing processes for solvent regeneration in liquid-liquid extraction of succinic acid, Chem. Eng. Process. Process Intensif. 143 (2019) 107600, , https://doi.org/10.1016/j.cep.2019.107600.
References [1] A. Kuznetsov, A. Beloded, A. Derunets, V. Grosheva, L. Vakar, R. Kozlovskiy, V. Shvets, Biosynthesis of lactic acid in a membrane bioreactor for cleaner technology of polylactide production, Clean Technol. Environ. Policy. 19 (2017) 869–882, https://doi.org/10.1007/s10098-016-1275-z. [2] J. Vijayakumar, R. Aravindan, T. Viruthagiri, Recent trends in the production, purification and application of lactic acid, Chem. Biochem. Eng. Q. 22 (2008) 245–264. [3] R. Datta, M. Henry, Lactic acid: Recent advances in products, processes and technologies - a review, J. Chem. Technol. Biotechnol. 81 (2006) 1119–1129, https:// doi.org/10.1002/jctb.1486. [4] C.S. López-Garzón, A.J.J. Straathof, Recovery of carboxylic acids produced by fermentation, Biotechnol. Adv. 32 (2014) 873–904, https://doi.org/10.1016/j. biotechadv.2014.04.002. [5] C.V. Stevens, The bio-based economy: Step by step from demand push to market pull, in, Biofuels, Bioprod. Biorefining (2012) 495–496, https://doi.org/10.1002/ bbb.1361. [6] K.L. Wasewar, A.A. Yawalkar, J.A. Moulijn, V.G. Pangarkar, Fermentation of glucose to lactic acid coupled with reactive extraction: A review, Ind. Eng. Chem. Res. 43 (2004) 5969–5982. [7] C. King, Amine-based systems for carboxylic acid recovery, Chemtech (United States). 22 (5) (1992) 285–291. [8] I.H. Aljundi, J.M. Belovich, O. Talu, Adsorption of lactic acid from fermentation broth and aqueous solutions on Zeolite molecular sieves, Chem. Eng. Sci. 60 (2005) 5004–5009, https://doi.org/10.1016/j.ces.2005.04.034. [9] R. Kumar, H. Nanavati, S.B. Noronha, S.M. Mahajani, A continuous process for the recovery of lactic acid by reactive distillation, J. Chem. Technol. Biotechnol. 81 (2006) 1767–1777, https://doi.org/10.1002/jctb.1603. [10] D.H. Han, W.H. Hong, Reactive extraction of lactic acid with trioctylamine/methylene chloride/n-hexane, Sep. Sci. Technol. 31 (1996) 1123–1135, https://doi. org/10.1080/01496399608001338. [11] P. Pal, J. Sikder, S. Roy, L. Giorno, Process intensification in lactic acid production: a review of membrane based processes, Chem. Eng. Process. Process Intensif. (2009) 1549–1559, https://doi.org/10.1016/j.cep.2009.09.003. [12] L.M.J. Sprakel, B. Schuur, Solvent developments for liquid-liquid extraction of carboxylic acids in perspective, Sep. Purif. Technol. (2019) 935–957, https://doi. org/10.1016/j.seppur.2018.10.023. [13] E. Reyhanitash, T. Brouwer, S.R.A. Kersten, A.G.J. van der Ham, B. Schuur, Liquid–liquid extraction-based process concepts for recovery of carboxylic acids from aqueous streams evaluated for dilute streams, Chem. Eng. Res. Des. 137 (2018) 510–533, https://doi.org/10.1016/j.cherd.2018.07.038. [14] T. Brouwer, M. Blahusiak, K. Babic, B. Schuur, Reactive extraction and recovery of levulinic acid, formic acid and furfural from aqueous solutions containing sulphuric acid, Sep. Purif. Technol. 185 (2017) 186–195, https://doi.org/10.1016/j.seppur. 2017.05.036. [15] A. Krzyzaniak, M. Leeman, F. Vossebeld, T.J. Visser, B. Schuur, A.B. De Haan, Novel extractants for the recovery of fermentation derived lactic acid, Sep. Purif. Technol. 111 (2013) 82–89, https://doi.org/10.1016/j.seppur.2013.03.031. [16] A. Krzyzaniak, B. Schuur, A.B. De Haan, Equilibrium studies on lactic acid extraction with N, N-didodecylpyridin-4-amine (DDAP) extractant, Chem. Eng. Sci. 109 (2014) 236–243, https://doi.org/10.1016/j.ces.2014.01.030. [17] J. Procházka, A. Heyberger, V. Bízek, M. Koušová, E. Volaufová, Amine extraction of hydroxycarboxylic acids. 2. comparison of equilibria for lactic, malic, and citric acids, Ind. Eng. Chem. Res. 33 (1994) 1565–1573, https://doi.org/10.1021/ ie00030a016. [18] J.G.M. Susanti, B. Winkelman, H.J. Schuur, J. Heeres Yue, Lactic acid extraction and mass transfer characteristics in slug flow capillary microreactors, Ind. Eng. Chem. Res. 55 (2016) 4691–4702, https://doi.org/10.1021/acs.iecr.5b04917. [19] W. Qin, Z. Li, Y. Dai, Extraction of monocarboxylic acids with trioctylamine: equilibria and correlation of apparent reactive equilibrium constant, Ind. Eng. Chem. Res. 42 (2003) 6196–6204, https://doi.org/10.1021/ie021049b. [20] J. Marták, Š. Schlosser, S. Vlčková, Pertraction of lactic acid through supported liquid membranes containing phosphonium ionic liquid, J. Memb. Sci. 318 (2008) 298–310, https://doi.org/10.1016/j.memsci.2008.02.064. [21] J. Marták, Š. Schlosser, M. Blahušiak, Mass-transfer in pertraction of butyric acid by phosphonium ionic liquids and dodecane, Chem. Pap. (2011) 608–619, https://doi. org/10.2478/s11696-011-0069-3. [22] E. Reyhanitash, B. Zaalberg, H.M. Ijmker, S.R.A. Kersten, B. Schuur, CO2-enhanced extraction of acetic acid from fermented wastewater, Green Chem. 17 (2015) 4393–4400, https://doi.org/10.1039/c5gc01061f. [23] K. Tonova, I. Svinyarov, M.G. Bogdanov, Hydrophobic 3-alkyl-1-methylimidazolium saccharinates as extractants for l-lactic acid recovery, Sep. Purif. Technol. 125 (2014) 239–246, https://doi.org/10.1016/j.seppur.2014.02.001. [24] E. Reyhanitash, E. Fufachev, K.D. Van Munster, M.B.M. Van Beek, L.M.J. Sprakel, C.N. Edelijn, B.M. Weckhuysen, S.R.A. Kersten, P.C.A. Bruijnincx, B. Schuur, Recovery and conversion of acetic acid from a phosphonium phosphinate ionic liquid to enable valorization of fermented wastewater, Green Chem. 21 (2019) 2023–2034, https://doi.org/10.1039/c9gc00725c. [25] D.L. Grzenia, D.J. Schell, S. Ranil Wickramsinghe, Detoxification of biomass hydrolysates by reactive membrane extraction, J. Memb. Sci. 348 (2010) 6–12, https://doi.org/10.1016/j.memsci.2009.10.035. [26] D.L. Grzenia, D.J. Schell, S. Ranil Wickramasinghe, Membrane extraction for detoxification of biomass hydrolysates, Bioresour. Technol. 111 (2012) 248–254, https://doi.org/10.1016/j.biortech.2012.01.169.
8