Effects of Organic Phase, Fermentation Media, and Operating Conditions on Lactic Acid Extraction

Abstract

Lactic acid has extensive uses in the food, pharmaceutical, cosmetic and chemical industry. Lately, its use in producing biodegradable polymeric materials (polylactate) makes the production of lactic acid from fermentation broths very important. The major part of the production cost accounts for the cost of separation from very dilute reaction media where productivity is low as a result of the inhibitory nature of lactic acid. The current method of extraction/separation is both expensive and unsustainable. Therefore, there is great scope for development of alternative technology that will offer efficiency, economic, and environmental benefits. One of the promising technologies for recovery of lactic acid from fermentation broth is reactive liquid-liquid extraction. In this paper the extraction and recovery of lactic acid based on reactive processes is examined and the performance of a hydrophobic microporous hollow-fiber membrane module (HFMM) is evaluated. First, equilibrium experiments were conducted using organic solutions consisting of Aliquat 336/trioctylamine (as a carrier) and tri-butyl phosphate (TBP)/sunflower oil (as a solvent) The values of the distribution coefficient were obtained as a function of feed pH, composition of the organic phase (ratio of carrier to solvent), and temperature (range 8–40 °C). The optimum extraction was obtained with the organic phase consisting of a mixture of 15wt % tri-octylamine (TOA) and 15% Aliquat 336 and 70% solvent. The organic phase with TBP performed best but is less suitable because of its damaging properties (toxicity and environmental impact) and cost. Sunflower oil, which performed moderately, can be regarded as a better option as it has many desirable characteristics (nontoxic, environment- and operator-friendly) and it costs much less. The percentage extraction was approximately 33% at pH 6 and at room temperature (can be enhanced by operating at higher temperatures) at a feed flow rate of 15–20 L/h. These results suggest that the hollow-fiber membrane process yields good percentage extraction at the fermentation conditions and its in situ application could improve the process productivity by suppressing the inhibitory effect of lactic acid.

1. Introduction

Lactic acid (2-hydroxypropanoic acid, CH₃CHOHCOOH) is a colorless, organic liquid. It has a variety of applications in the food, chemical, pharmaceutical and cosmetic industries (1). The Food and Drug Administration (FDA) have approved lactic acid and its salts to be GRAS (Generally Recognized as Safe) (2). It can be converted to a polylactic acid used for the synthesis of biodegradable materials (3). As well as being environmentally friendly, there is a growing demand due to strict environmental laws being legislated for biodegradable polymers as a substitute for conventional plastic materials. Biodegradable copolymers are also used for the production of new materials with biomedical applications such as drug delivery systems (4).

Lactic acid is typically produced via either chemical synthesis or the fermentation of whey or another inexpensive carbon source (2). Due to the increasing cost of crude oil, a common raw material for the chemical synthesis of lactic acid, the efficient production of lactic acid through fermentation has become increasingly important (5−9). An economical and efficient method for the recovery of lactic acid from fermentation broth is vital as the overall cost of production is dominated by its extraction and recovery from the broth (5, 7).

The production of most organic acids from fermentation media are subject to product inhibition as the reaction proceeds (8). Hence, the separation of the organic acid as it is being produced is highly desirable. Extractive fermentation, in situ application of the solvent extraction technique, keeps the product concentration in the broth at a low level and suppresses the product inhibition by continuously removing them from a fermentation broth (9).

Various methods for the extraction of lactic acid have been reported, such as precipitation, ion exchange process, adsorption, diffusion dialysis, microcapsules, esterification, and hydrolysis reactive extraction, as well as a simulated moving bed process (1, 10−13). These methods have several disadvantages including high cost, and they produce large volumes of waste, require multiple steps, and operate with low efficiency under practical conditions. The reactive liquid−liquid extraction (RLLE) method using microporous hollow-fiber membrane module (HFMM) may potentially overcome many of the disadvantages and has been evaluated for the production of lactic acid (14−17). In a recent review, a process based on RLLE in HFMM has been found to be competitive from the process, economic, and environmental points of view (18−20). To understand and explore more of this process the main aims of this research were set as

• to determine a less toxic, environmentally friendly solvent and a carrier or a mixture of carriers for extraction of lactic acid at conditions similar to the fermentation,

• to examine the effects of various operating variables, especially temperature and pH on the distribution coefficient of lactic acid in the organic phase

• to upgrade the process to a small pilot-scale module and evaluate the performance of the less toxic solvent for extraction under fermentation conditions (i.e., in presence of salts and lactose),

2. Modeling of Equilibrium and Mass Transfer

2.1. Extraction. The reaction of undissociated lactic acid (HLA) with a carrier (T) dissolved in the solvent gives a reaction complex (THLA) that remains largely in the organic phase and may be represented by (21, 22)

• (1)

2.2. Mass Transfer in the Extraction Process. A schematic of the transport mechanism of lactic acid from an aqueous feed side to the organic side through hollow-fiber wall is shown in Figure 1. The lactic acid molecules are transported from the bulk feed solution to the feed−membrane interface and can be expressed by eq 2. At the interface the reaction takes place (eq 1) to form a lactic acid−carrier complex. The concentrations at the equilibrium in the aqueous and organic phases can be related by an apparent distribution coefficient (DE), given by eq 3. The mass transfer of the complex through the hollow fiber is the next step and can be expressed by eq 4. The final step is the transport through the organic interface and can be expressed as in eq 5.

• (2)

• (3)

• (4)

• (5)

where k_f, k_mf, and k_o are the mass transfer coefficient in the feed side, on the membrane, and in the organic side, respectively. The concentrations C_LAf and C_LAfi are the total lactic acid concentrations in the bulk and at the interface, respectively. The concentrations C_LAOi and C_LAO are the concentration of lactic acid at the membrane−organic interface and in the bulk organic phase, respectively.

Figure Figure 1.. (a) Schematic of the mass transfer and extraction process in the membrane module. (b) Concentration profiles of lactic acid in the feed phase (fiber side), in membrane wall, and in the organic phase (shell side).

Combining the above equations the flux in the system at the steady state is obtained as

• (6)

where K_of is the overall mass transfer coefficient of the process. K_of is related to the individual transfer coefficients by the following equation:

• (7)

In order to calculate the overall mass transfer coefficient from the above equation the individual mass transfer coefficients have to be known in addition to the distribution ratio of lactic acid between the aqueous−organic solutions. There are many correlations available in the literature for calculating the individual mass transfer coefficients (17). Each is based on the specific experimental conditions and equipment setup used to develop the correlation. So the assumption of the correlation needs to be matched for its appropriateness before applying to any other system.

In the section below an approximate solution is presented for hollow-fiber membrane modules to evaluate the overall mass transfer coefficient from an analysis of concentration versus time data.

2.3. Approximate Analytical Solution for Extraction in the Hollow-FiberModule. The membrane modules are operated in recycling mode as the percentage extraction in once-through operation is small. In the recycling mode, it is considered that the feed and the organic solutions are circulated through the fiber side and shell side of the module, respectively.

The mathematical model consists of two mass balance equations, eqs 8 and 9, which define the change in lactic acid concentration (i) in the module and (ii) in the feed tank, where aqueous solution is continuously circulated.

• (8)

Mass balance in the tank:

• (9)

where (V/A)_in is the ratio of the volume to inner area of mass transfer in the fibers, L is the length of the fiber, u_f is the linear velocity, q_f is the feed flow rate, and v^T is the tank volume. The superscripts m and T refer to the membrane module and tank, respectively.

The ratio (V/A)_in for the feed solution circulating along the inside of the fiber is equal to di/4, where di is the inner radius of each fiber. The factor (V/A)_in 1/L is a small number for this type of contactor, i.e., approximately 4 × 10⁻⁴. Using this and assuming a slow rate of change of solute concentration in the module, eq 8 can be simplified to

• (10)

An approximate solution of the model equations form is given by

• (11)

where B has been defined by the following equation:

• (12)

The overall mass transfer coefficient (K_of) can be determined from the value of the slope of the linear plots of the left-hand side (LHS) of eq 11 versus t (time). The LHS of eq 11 requires (i) the experimental values of the concentrations and (ii) the partition coefficient of the solute. We also need to calculate B from eq 12, which requires the velocity inside the fiber, module characteristics (volume, mass transfer area), and the partition coefficient.

The degree of extraction for the HFM experiments is measured by the extraction efficiency which is defined by the following equation:

• (13)

where C_LAi and C_LAfare the initial and final concentration values of the feed solution in the recirculation system, respectively.

3. Materials and Methods

3.1. Chemicals.l-(+)-Lactic acid (85% by mass), lactose monohydrate, magnesium sulfate, heptahydrate, potassium phosphate monobasic anhydrous, and tri-butyl phosphate (97%) were purchased from Sigma Aldrich, USA. Octanol and sodium carbonate were from BDH laboratory Reagents, Poole, England. Sodium chloride and sodium hydroxide pellets were from Scharlau Ltd., Spain. Tri-octylamine (98%), hexane (99%), dodecane (99%), oleic acid, and 2-octylamine (98%) were from Acros Organics, USA. Oleyl alcohol (85%), Aliquat 336 and decanol were from Lancaster synthesis, England. Shellsol TK was from Specialty Chemicals, New Zealand.

3.2. Hollow-Fiber Membrane Module. A microporous hollow-fiber membrane contactor, 5PCM-218, was purchased from Separation Products Division, Hoechst Celanese Corporation, Charlotte, NC. The contactor had a shell-and-tube configuration with a total of 10,000 polypropylene hollow fibers (Celgard X-30, 240 μm i.d., 300 μm o.d., length 15 cm) potted in polyethylene in a polypropylene case of 6 cm i.d., The surface area of the contactor was 1.4 m². The hollow-fiber module was set up as shown in Figure 2.

Figure Figure 2.. Schematic of hollow-fiber membrane experimental setup.

3.3. Feed and Synthetic Fermentation Solutions. Feed solutions were prepared by dissolving a weighed amount of lactic acid (as supplied) into distilled water; the natural pH was 2.4 and any pH adjustment was done using 0.1 M HCl or 0.1 M NaOH. The feed concentration of lactic acid was 0.2 M (unless otherwise mentioned). For the physical extraction, straight solvent with no carrier was used. For the reactive extraction the organic solution was prepared with the desired amount of carrier (or carriers) into a solvent (or a mixture of solvents).

The procedure described in the above paragraph was followed to examine the effect of the fermentation media, but the aqueous solution was made up of 0.2 M lactic acid, 11.36 mM lactose, 31.3 mM sodium chloride (NaCl), 6.68 mM potassium sulfate (KH₂PO₄), and 0.368 mM magnesium sulfate (MgSO₄·7H₂O) (12, 15).

3.4. Equilibrium Experiments.3.4.1. Extraction and Analytical Procedures. A 5 mL portion of lactic acid solution at the desired pH was mixed with 5 mL of the organic phase which contained the organic solvent in a 15 mL centrifuge tube. The mixture was stirred for 2 h at 1000 rpm. After that, the mixture was placed in the centrifuge for 15 min at 4000 rpm to separate the two phases. After the separation, the aqueous lactic acid concentration at pH 2.4 was determined at a wavelength of 192 nm by a UV/vis Spectrophotometer (Perkin-Elmer 35 UV/vis spectrophotometer, USA).

For pH values greater than 2.4, the acid concentration was calculated by titrating aqueous solutions before and after equilibrium with 0.05 M NaOH. For lactic acid concentration in the organic phase, 4 mL of sample solution was placed in a vial and 2−3 drops of phenolphthalein indicator was added to the sample. The sample was then titrated with NaOH. When the color changed from white to pink, this indicated that the end-point was reached and the volume of NaOH used was recorded. This was used to calculate the amount of lactic acid transferred to the organic solution. The amount of lactic acid in the organic phase was determined by a mass balance taking in to account the volumes of the organic and aqueous phases, respectively. This is an approximate method but has been reported for the determination of lactic acid (15, 16). The measurements of lactic acid by this method (although less accurate) give reproducible results with an error of 11−14%. For more accurate determination especially in producing analytical grade lactic acid the HPLC method or enzymatic method should be used (17, 18).

3.4.2. Extraction with Organic Solvents and Carriers. The effect of solvent was examined by using these solvents (used alone): tri-butyl phosphate, oleic acid, decanol, oleyl alcohol, hexane, dodecane, shellsol TK and sunflower oil.

The effect of adding carrier in the solvent was examined on the extraction of lactic acid at its natural pH (2.4) and at room temperature. With the carrier the solutes are extracted only in the dissociated form whereas with the solvent only undissociated form is extracted. Three different amines, 2-octyl amine a secondary amine, tri-octyl amine (TOA), a tertiary amine and Aliquat 336 a quaternary amine were chosen because of their insolubility and effectiveness in lactic acid extraction (2, 5).

The equilibrium experiments were carried out for extraction of 0.2M lactic acid within the pH range 3−6 with sunflower oil and Aliquat 336/TOA. The effect of temperature (range 8−40 °C) was studied for extraction at pH 2.4 using an organic phase of 20% TOA and 80% sunflower oil.

3.5. Hollow-Fiber Module Experiments. Prior to use, the shell side and fiber side of the hollow-fiber module was cleaned using 25% ethanol solution and then rinsed with water; 500 mL of the organic solution was prepared and placed in a 1 L conical flask. The organic solution used for extraction in the HFMM was made up of 10 wt % TOA in TBP, unless otherwise stated. It was prepared by measuring out 59.2 mL of TOA and 440.8 mL of TBP in a conical flask. The aqueous solution was 0.2 M lactic acid solution at pH 2.4.

To saturate the pores of the hollow-fiber walls, the organic solution was passed through the shell side of the membrane for an hour. Initially, the organic solution was pumped at a high flow rate to remove air bubbles and then at a low flow rate of approximately 30 mL/min. After 1 h, the tube side was cleaned with RO water and lactic acid was allowed to pass through the tube side. During this time, both the pressure and flow rate were adjusted. A higher pressure of 30 kPa on the aqueous phase was applied to stabilize the organic-aqueous interface at the pore mouth. The aqueous and the organic solutions were pumped and circulated through the membrane for approximately 6 h. Samples were taken at every 15 min for the first hour and then at every hour.

4. Results and Discussions

4.1. Equilibrium Experiments.4.1.1. Effect of Organic Solvent. The physical extraction of lactic acid with pure solvent (no carrier added) in the equilibrium experiments is assessed by the values of the distribution coefficient and these values for various organic phases are listed in Table 1.

Table Table 1.. Distribution Coefficient of Lactic Acid at pH 2.4 in Various Solvents

solvent	distribution coefficient, DE
tri-butyl phosphate (TBP)	0.82
oleyl alcohol (OlA)	0.25
decanol (D)	0.43
hexane (H)	0.25
dodecane (DD)	0.25
shellsol TK (STK)	0.11
oleic acid (OA)	0.67
sunflower oil	0.11

The data in Table 1 suggest that tri-butyl phosphate is the optimum solvent for extraction of lactic acid. Tri-butyl phosphate is followed by oleic acid, decanol, oleyl alcohol, hexane, dodecane. Shellsol TK and sunflower oil gave the lowest distribution ratio of lactic acid into the organic phase. It is evident that active solvents (with functional groups), e.g., tri-butyl phosphate and oleic acid, have greater extraction power than the inactive solvents, e.g., shellsol TK and dodecane. The active solvent can assist the solvation of the lactic acid molecules and enhance the solubility of the lactic acid complex (23). The complex formed with the functional group of the active solvent may be more stable and soluble in the organic phase and thus allow greater extraction compared to those solvents without any functional groups (24). Although this result with active solvent is good, other points like solvent loss (due to solubility in the aqueous phase) and environmental effects should also be considered before the final selection of the solvent. It has been reported that the solvents containing phosphorus-bonded oxygen atoms (like tributyl phosphate) are not favorable from the points of water solubility and environmental considerations (25).

4.1.2. Effect of Carrier at Natural pH of 2.4. The values of DE increased when extraction was performed with adding carrier in the solvent phase at natural pH (2.4) of lactic acid and at room temperature. The results are presented in Table 2 for extraction of lactic acid at pH 2.4 using tributyl phosphate as solvent. At this low pH the values of the distribution coefficient were moderate to high for all the carriers. The organic system of 10 wt % TOA−90 wt % TBP gave a distribution coefficient as 4.88. This value increased with the increase in the concentration of TOA for a fixed feed concentration.

Table Table 2.. Effect of Carrier in Tri-butyl Phosphate on Extraction of Lactic Acid at pH 2.4

10 wt % carrier in TBP	distribution coefficient, DE
2-octyl amine	4.26
trioctyl amine	4.88
Aliquat 336	1.22

The values of the distribution coefficients at these conditions were obtained using various solvents with 20% TOA and 80% solvent. The solvents were sunflower oil, oleic acid, and tributyl phosphate. The effect of the solvent had a large influence on the distribution coefficient. Again the best organic system was TOA−TBP which achieved a distribution ratio of 15. This value is almost two times the value of the second best performer, sunflower oil, which had a distribution ratio of 8.8. It is observed that oleic acid did not perform as well (DE was less than 4) when carrier was added to it. However, the best organic solvent, TBP, is expensive and is not recognized as a GRAS (generally regarded as safe) solvent. Sunflower oil is a nontoxic, cheap and can be regarded as an environmentally friendly solvent. The hollow-fiber experimental runs are provided by using both the solvents individually or in combination (except otherwise mentioned) as organic phase. It is worth mentioning that the solvents used in this study are less toxic to the fermentation media compared to those reported in the study of organic acid extraction (23, 26).

4.1.3. Effect of Aqueous Feed pH (Range 4−6). The equilibrium experiments were carried out for extraction of 0.2M lactic acid with sunflower oil (as diluent) and Aliquat 336/TOA (as carriers). As shown in Figure 3 the distribution ratio decreased considerably with the increase in aqueous solution pH range 4−6 (26). The magnitude of the distribution ratio was greater with Aliquat 336 than for TOA at higher pH. The two extractants when combined also gave considerably higher distribution ratios than the individual extractant. The synergistic effect found in the literature (22, 29,30) was also apparent in these results, with the combination of these carriers. However, this synergistic effect is much more apparent at lower pH (pH 4) than at higher pH (pH 6). The organic phase which comprised of 15% Aliquat 336 and 15% TOA gave the highest distribution ratios at all pH values. Since higher pH is used for fermentative production of lactic acid, the effect of the distribution coefficients at around pH 6 will provide useful information for their in situ applicability. Although the values of the distribution coefficient increased slightly these small change may allow continuous extraction (without any pH adjustments of the fermentation media) and reduce the inhibitory effect of lactic acid to a reasonable extent.

Figure Figure 3.. Effect of feed pH on distribution ratio of different carriers in sunflower oil for extraction of 0.2 M lactic acid at room temperature.

4.1.4. Effect of Temperature. As shown in Figure 4 increasing the temperature of the equilibrium experiment (both phases at the same temperature) increased the distribution coefficient at the natural pH of 2.4. This means that better extraction could be expected at temperatures higher than the room temperature (27). This trend is a positive result, as the optimal temperature for lactic acid fermentation could be in the range 30−38 °C although this optimal varies with the type of micro-organism. Therefore, by performing the extraction at temperatures higher than the room temperature a greater amount of lactic acid could be removed from the broth, thus enhancing the productivity further.

Figure Figure 4.. Effect of temperature on distribution coefficient for extraction with 20% TOA and 80% sunflower oil at pH 2.4 and 0.2 M lactic acid.

4.2. Hollow-Fiber Experiments.4.2.1. Effect of Flow Rate. The experiments were conducted with recirculation flow rate in the range 4.16−12.5 mL/s, solution pH at 2.4, and lactic acid feed concentration of 0.2 M. The same flow rates were maintained for both the aqueous and organic solutions. The maximum extraction efficiency of approximately 93% was obtained with 10% TOA−TBP organic phase within a processing time of 2−3 h (Figure 5). The extraction profiles for the flow rates were identical except for the profile of 8.33 mL/s that attained lower than the other two organic systems. This could be due to the experimental error (as is observed in Figure 5) that occurred during the initial monitoring and the initial difference was carried throughout with final extraction being less by that percentage. The effect of flow rate is small and suggests that the overall extraction is not significantly affected by the external mass transfer of lactic acid (i.e., the aqueous side resistance).

Figure Figure 5.. Effect of flow rate on percentage extraction of lactic acid (0.2 M and pH 2.4) with 10% TOA and 90% TBP.

Although TBP gave very good extraction efficiency and attained equilibrium the fastest, it is expensive and more importantly is not recognized as a food grade solvent. Oleyl alcohol (OA) is nontoxic to acid-producing bacteria and hence was also examined as a solvent. The percentage extraction increased with time and then attained a plateau at approximately 93% within 2−3 h (Figure 6). With OA the extraction rate was initially slower, increased with time and finally attained approximately 84% (which is within 10% of that for TBP). The lower efficiency and slower rate could be attributed to the high viscosity of oleyl alcohol resulting in a decreased mass transfer rate and requiring longer time to complete the process compared to the less viscous solvent systems viscous.

Figure Figure 6.. Extraction (%) of lactic acid (0.2 M and pH 2.4) with 10% TOA and 90% solvent (TBP or oleyl alcohol).

4.2.2. Effect of Fermentation Media. The synthetic fermentation broths (containing lactose and salts) and the pure lactic acid solution gave nearly the same extraction efficiency of about 90% as shown in Figure 7. The variation of the solution pH, i.e., a pH of 4.5 for the synthetic fermentation broth (compared to pH 2.4 for pure lactic acid) and presence of salts did not significantly affect the extraction efficiency. These results compare very well with those in the literature (10, 30).

Figure Figure 7.. Extraction (%) of lactic acid from solutions with no salts (at pH 2.4), with salts and lactose (0.2 M and pH 2.4) and from synthetic broth at pH 4.5 with 10% TOA and 90% TBP.

Because the above-mentioned solvents are not considered suitable (although many studies in the literature have used them) for large scale applications due to cost and health, safety, and environmental reasons, further hollow-fiber experiments were conducted with sunflower oil.

4.3. Hollow-Fiber Extraction with Sunflower Oil.4.3.1. Effect of Carrier. The presence of carrier (either singly or combination with others) in the solvent had a large affect on the efficiency of the extraction. All of the trials reached steady state after a period of 45−60 min. The run with no carrier (sunflower oil only) performed poorly with only 9% extraction efficiency (Figure 8). The experimental run with 30% carrier, which consisted of half Aliquat 336 and half TOA, performed the best, and reached an extraction efficiency of 33%. This percentage extraction is low but was obtained at conditions similar to the fermentation conditions; so separation of lactic acid can be achieved without addition of any chemicals to adjust the pH of the broth. Thus in situ application would be possible without interruption of the fermentation and this continuous removal of lactic acid is expected to reduce its inhibitory effect of lactic acid.

Figure Figure 8.. Effect of carrier (and a mixture of carriers) on extraction (E %) of lactic acid (0.2 M and pH 6) with sunflower oil (only), a 10% mixture of Aliquat 336 and TOA (ratio 3:1) and 30% mixture of Aliquat 336 and TOA (1:1) in sunflower oil.

4.3.2. Comparison with Other Solvent (TBP). The effect the solvent used on the extraction efficiency was compared by running a trial using sunflower oil and TBP as solvents at room temperature (Figure 9). The carrier used was a mixture of 15% Aliquat 336 and 15% TOA and the lactic acid solution was at pH 6. As expected TBP performed better; TBP achieved 50% extraction efficiency compared to 33% with sunflower oil. It is surprising to see that sunflower oil, having a much lower distribution coefficient than TBP, achieved a moderately good performance. Comparing the cost, industrial acceptability and environmental benefits sunflower oil could be a very good candidate as a single solvent or in combination with other high performing solvents.

Figure Figure 9.. Extraction (%) of lactic acid (0.2 M and pH 6) with 10% TOA and 90% solvent (TBP or sunflower oil).

4.3.3. Effect of Temperature on Extraction with Sunflower Oil. The increase in extraction with an increase in temperature found in the equilibrium extraction experiments was also observed in the HFMM experiments. As can be seen in Figure 10, the experiment performed at room temperature reached 24%, whereas the run at 30.5 °C reached 28% extraction using a 10% TOA in sunflower oil (lactic acid feed pH 6). The increase in temperature increased the distribution and transport of lactic acid into the organic phase. The increase in temperature would have increased the diffusivity (as the viscosity of the two phases decreased) and allowed faster mass transport through the pores (31, 33). Therefore, the percentage extraction can be further increased by operating (i) at higher temperature (the maximum temperature of the membrane used in this study is limited to 40 °C), (ii) at lower aqueous solution pH, (iii) at increased TOA concentration and (iv) by mixing sunflower oil with other active solvents like TBP. The effects of these changes are demonstrated in a trial performed at 35 °C, pH 5.0, with a different carrier and solvent system reached 70% extraction (Figure 11).

Figure Figure 10.. Effect of temperature on extraction (E %) of lactic acid (0.2 M and pH 6) with 10% TOA and 90% sunflower oil.

Figure Figure 11.. Extraction (%) of lactic acid in a binary mixture of sunflower oil and TBP using a mixture of carriers at pH 5 and at a temperature of 35 °C.

4.4. Optimal Operating Conditions. The optimal operating parameters discussed in the preceding paragraphs were tested to determine the percentage extraction that could be achieved under the optimum conditions. The extraction was performed at pH 5.0, at 35 °C, with an organic phase, a mixture of 15% Aliquat 336 and 15% TOA and a binary solvent system of 35% TBP and 35% sunflower oil. It is seen in Figure 11 that approximately 70% of lactic acid was extracted within 4 h of contact time. The extraction rate was still increasing gradually and if the experiment was continued for a longer time the percentage extraction would have reached a higher value.

4.5. Overall Mass Transfer Coefficient. The overall transfer coefficient (K_of) was calculated from the slope of the plot of the left-hand side of eq 11 versus time. The value of the distribution coefficient obtained experimentally for each organic phase was used in the calculation. The experimental data during the initial period (approximately 60−70 min) showed good correlation (R² = 0.9 at least) and were considered for the calculation. A plot of the LHS of eq 11 versus the extraction time with various organic phases is shown in Figure 12. From the best fit line of the plot of LHS versus time, the slope was calculated and with the value of B from eq 12, the overall mass transfer coefficient has been calculated. The values of these coefficients are in the range (0.4−2.3) × 10⁻⁵ cm/s (shown in Table 3). This is in consistent with the values of the distribution coefficient and the extraction data (Figure 9), i.e., the mass transfer rate is much faster in tributyl phosphate than in the other two organic phases. The values of overall mass transfer coefficient reported are in the range: (0.2−5) × 10⁻⁵ cm/s (31−33). Considering the fact that our analysis is based on the extractions only (no simultaneous re-extraction and no optimization of operating conditions) the values obtained in this study compare very well. The other advantage of the method used in the present analysis is that the estimation of the mass transfer coefficient is based on semianalytical equation and requires no correlation and additional data from any other experiments or literature.

Table Table 3.. Values of Overall Mass Transfer Coefficient (K_of) for Various Organic Phases

organic phase	feed solution	overall mass transfer coefficient (Kof, cm/s × 10^5)
10% TOA in TBP (250 mL/min)	aqueous lactic acid (0.2 M)	1.82
10% TOA in TBP (750 mL/min)	aqueous lactic acid (0.2 M)	2.30
10% TOA in OA (250 mL/min)	aqueous lactic acid (0.2 M)	0.52
10% TOA in TBP (250 mL/min)	synthetic fermentation broth	1.57
15% TOA + 15% Aliquat 336 in sunflower oil	aqueous lactic acid (0.2 M)	0.4

Figure Figure 12.. Plot for the overall mass transfer coefficient of lactic acid; effect of various organic phases for extraction of lactic acid from aqueous solution and synthetic fermentation media.

The knowledge of the analysis of results in terms of mass transfer coefficient could be useful in estimating the membrane area for a large-scale fermentative production process where lactic acid can be removed as it is produced and thus enhancing the productivity.

5. Conclusions

The values of the distribution coefficient of lactic acid with solvents such as tri-butyl phosphate (TBP, the best), oleic acid, oleyl alcohol, shellsol TK, and sunflower oil are small when used for physical extraction (solvent only).

Higher values of the distribution coefficient can be achieved by reactive extraction with tri-octylamine (TOA/Aliquat 336 as the carrier) dissolved in any of the above solvents. The best organic system TOA−TBP gave a high distribution ratio at the natural pH 2.4. A maximum extraction of 93% was achieved in the hollow-fiber module within 2−3 h using 10% (w/w) TOA−TBP.

The use of less toxic solvents such as oleyl alcohol (instead of TBP or its mixture with TBP) gave a lower percentage extraction with the final value around 83%.

The presence of the components of the fermentation media, such as, lactose and salts did not significantly affect the percentage extraction in the membrane module.

Extractions with these organic systems are considered less suitable for large scale production because of cost, toxicity and potentially harmful environmental effects.

For industrial application an organic phase with sunflower oil, is more attractive because of its cost (cheaper by an order of magnitude), nontoxicity and environmental benefits. It has proven to be effective at operating conditions (pH and temperature) similar to fermentation and shows stability with the commercially available large processing device.

With sunflower oil−TOA/Aliquat 336 as the organic phase the percentage extraction is lower (30−35%) at room temperature. The extraction can be enhanced to approximately 70% by operating at higher temperature.

The percentage removal of lactic acid at this rate (at feed flow rate of approximately 8.3 mL/s) from the fermentation media should reduce the product inhibition effects and improve the productivity by allowing fermentation over a longer period of time.

Notation

A

surface area (m²)

B

defined in eq 12

C

concentration of lactic acid (mmol/L)

di, do

inner diameter, outer diameter of a hollow fiber (cm)

D_E

distribution coefficient of lactic acid defined in eq 3

E

percent of extraction of lactic acid, defined in eq 13

H+

hydrogen ion

HLA

undissociated lactic acid

K_of

mass transfer coefficient in eq 6, (cm/s)

L

length of the fiber (cm)

n_f

number of fibers

N

steady-state flux, defined in eq 6

t

time, s

v

flow rate, L/s

u

linear velocity in the hollow fibers (cm/s)

V

volume (L)

Z

axial distance in the module (cm)

Subscripts

 

aq

aqueous phase

f

feed solution

fi

feed side at aqueous−organic interface

fo

organic side at aqueous−organic interface

in

based on inside fiber diameter

LAf

final lactic acid concentration

LAi

initial lactic acid concentration

oi

organic side interface

o, org

organic phase

Superscripts

 

m

module

T

tank

Acknowledgements

The financial assistance of the University of Auckland Staff Research Grant, Auckland, New Zealand, is gratefully acknowledged. The first author is also thankful to the United Arab Emirates University, Al Ain, UAE, for their overall support.