Off-gas capture: A promising strategy for removal and recovery of toxic bioproducts in aerobic fermentation

doi:10.21203/rs.3.rs-7745535/v1

Download PDF

Research Article

Off-gas capture: A promising strategy for removal and recovery of toxic bioproducts in aerobic fermentation

https://doi.org/10.21203/rs.3.rs-7745535/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

In many bioprocesses, maximum achievable titers are limited below economically viable levels by toxic accumulation of the primary end-product. To combat end-product inhibition, a variety of in situ product removal technologies have been developed to selectively remove or partition toxic bioproducts, thereby prolonging fermentation and improving overall process efficiency. Use of an in situ organic overlay to partition toxic hydrophobic products is a commonly employed approach, but this technique occupies valuable space in the fermentor, imposes replacement costs for unrecovered solvent, and increases downstream separation due to formation of stable emulsions. In addition, for many volatile hydrophobic products produced under aerobic conditions – including medium-chain alcohols, esters, monoterpenes, and other aviation fuel precursors - a significant fraction of the product is volatilized to the fermentor off-gas and must be recovered separately to maximize product yield. To address these challenges, we explore the viability of leveraging existing aeration energy to fully strip and recover volatile products from the fermentor off-gas. We compare two strategies of in situ product removal - liquid-liquid extraction and direct recovery from fermentation off-gas – for the production and recovery of intermediates used to generate isoprene and DMCO (1,4-dimethylcyclooctane), a high-performance jet fuel. We evaluate product toxicity, solvent toxicity, solvent partitioning, and the impact of aeration and internal overlay configurations on product volatilization rates. We then optimize product recovery from fermentor off-gas via condensation in chilled solvent, achieving 84% capture efficiency. In addition to greatly simplifying downstream processing, relying on aeration for product volatilization in the absence of an internal overlay enables continuous removal of toxic fermentation products up to maximum isoprenol titers of 20.4 g/L, the highest reported to date.

isoprenol

isoprene

off-gas capture

SAF

in situ product recovery

solvent extraction

product toxicity

downstream processing

In contrast to the projected moderation of gasoline fuel demand driven by continuing advances in electrification of light-duty transportation, global aviation fuel demand is projected to rise sharply in the coming decades. Aviation is expected to account for roughly 15% of the growth in global oil consumption by 2030, with fuel use increasing from 106 billion gallons in 2019 to 230 billion gallons by 2050 (1, 2). Biobased aviation fuels offer a practical pathway to meet this rising demand while enhancing energy security and creating new, distributed manufacturing opportunities in rural areas. Given the long service life of aircraft and the absence of alternative propulsion systems with comparable energy density, liquid hydrocarbon fuels will remain essential for powering commercial air travel for decades to come. Consequently, alongside continued improvements in aircraft efficiency, synthetic aviation fuels (SAF) - including bio-based options - will play a central role in meeting the sector’s future energy needs. These high–energy-density fuels can also serve hard-to-electrify segments such as off-road, heavy-duty, and marine transport (3).

In addition to the contributions to energy security, the development of new fuel technologies offers opportunities to enhance performance compared to conventional petrochemical alternatives. One example is 1,4-dimethylcyclooctane (DMCO), a cyclic alkane with a volumetric net heat of combustion up to 9.2% higher than Jet A, which can be produced from biologically sourced isoprene via selective dimerization and hydrotreating (4). The weight and volume savings achieved by incorporating DMCO into jet fuel can improve aircraft range and fuel efficiency, providing greater operational flexibility (5). As a SAF blendstock, DMCO can provide seal-swelling capacity currently provided by petrochemical aromatics, enabling deployment of 100% SAF blends (6). DMCO also has a lower kinematic viscosity and freezing point than Jet A, alongside an acceptable flash point.

Beyond its potential value as a jet fuel precursor, the global market for isoprene itself is projected to reach $4.39 billion by 2027, with a compound annual growth rate of over 7% (7). While commercial isoprene production is currently entirely fossil-based, it can be readily expressed in biological systems, with Whited et al. demonstrating equivalent titers of up to 60 g/L of isoprene from sugar with engineered E. coli (8). Bio-isoprene production is aerobic, and the co-occurrence of highly volatile isoprene and oxygen in aerobic production raises significant safety concerns for production and recovery. Avoiding development of explosive conditions requires investment in new fermentation infrastructure with robust and redundant safety mechanisms, hindering commercialization progress to date. The low concentration of bio-isoprene in the vapor phase presents an additional challenge, with off-gas concentrations typically below 2% by volume, well below that of petroleum-derived isoprene obtained through extractive distillation processes (10–20%). Because the low boiling point and high vapor pressure of isoprene precludes recovery via condensation above cryogenic temperatures, recovery and purification from the off-gas stream is uneconomical with currently available technologies, generating additional barriers to deployment (9).

Bio-isoprenol presents an attractive alternative route to isoprene, as it converts readily to isoprene via dehydration. Isoprenol possesses significant safety and recovery advantages in bioproduction, driven by its reduced volatility, lower vapor pressure, and lower toxicity. These properties also make isoprenol easier to capture, handle, store and transport in liquid form. In storage, isoprenol exhibits enhanced stability and is less susceptible to spontaneous polymerization or degradation. Additionally, when compared to isoprene, isoprenol demonstrates lower reactivity in the atmosphere, mitigating release of volatile organic compounds (VOCs) and reducing potential contributions to air pollution and smog formation. In addition to its value as an isoprene precursor, isoprenol is an intriguing spark-ignition fuel additive due to its high octane number, its synergistic blending characteristics as a gasoline additive, and its higher energy density and lower hygroscopicity in comparison to ethanol (10, 11).

Two classes of metabolic pathways have been engineered to produce isoprenol in microbial hosts: amino acid production pathways and isoprenoid biosynthesis pathways, including both the mevalonate (MVA) and non-mevalonate pathways. To overcome intrinsic limitations of the conventional MVA pathway, including lower pathway efficiency and toxicity of the essential metabolic intermediate isopentenyl diphosphate (IPP), Kang et al. (12) developed and optimized an alternative IPP-bypass MVA pathway, preventing IPP toxicity and thereby enhancing flux to isoprenol. Isoprenol is water soluble and toxic at higher concentrations, leading to product inhibition in aqueous culture. To overcome this inhibition, a two-phase fermentation process was utilized to partition a fraction of the isoprenol and thereby mitigate product toxicity, achieving an isoprenol titer of 10.8 g/L. In addition to metabolic engineering efforts with E. coli, isoprenol production has been demonstrated in a variety of alternative microbial hosts including Corynebacterium glutamicum (1.1 g/L), Pseudomonas putida (3.5 g/L), Saccharomyces cerevisiae (383 mg/L), and Bacillus subtilis (2.1 mg/L) (13–16)

Product toxicity towards microorganisms is a common and significant bottleneck for production of fuels and chemicals at titers required for efficient recovery. For a number of biofuel precursors including isoprenol, butanol, isobutanol, and monoterpenes, product toxicity presents a primary barrier to commercial deployment (3). This toxicity is commonly mitigated via a combination of microbial tolerance engineering and process engineering, including implementation of in situ product recovery (ISPR) strategies to continuously maintain aqueous concentrations below the toxicity threshold. Tolerance engineering via both rational and adaptive approaches can significantly improve the toxicity threshold, including for isoprenol (17), but biological limits to this strategy often necessitate a combined approach. Extractive two-phase fermentation, in which a hydrophobic solvent is used to continuously partition a hydrophobic product, is commonly utilized for production of toxic, hydrophobic products (18). Two-phase systems rely on several crucial factors for their effectiveness. The extraction efficiency from water to solvent is dictated by the partition coefficient of the target molecule between the aqueous and organic phases. The solvent used must be biocompatible, exhibit low solubility in the aqueous phase, and maintain a density different from that the broth to ensure phase separation by gravity. Emulsion formation can significantly challenge separation of two-phase fermentations (19); the chosen solvent must therefore maintain low viscosity, large interfacial tension, and low tendency to emulsify in the broth. Additionally, the solvent should demonstrate high stability and be cost-effective (20). Finally, the product must be readily separable from the recovered solvent.

For volatile products, continuous gas stripping represents an alternative approach to in situ extraction, and is typically applied via vacuum extraction in anaerobic systems (21). Gas stripping in anaerobic systems typically relies on application of vacuum in the reactor headspace, however this approach conflicts directly with high oxygen transfer and gas sparging requirements in aerobic systems. As an alternative approach in aerobic cultivation, we propose removal of the product using existing aeration to strip isoprenol from the fermentor, with product recovered from the off-gas stream via continuous condensation in chilled solvent with high affinity for the product. The high vapor pressure, high toxicity to microbial hosts, and relatively poor partitioning of isoprenol make it an ideal candidate to evaluate recovery opportunities leveraging aerobic gas stripping. Stripping can effectively reduce the concentration of the product in the aqueous phase, mitigating toxic effects on the microorganism while selectively removing only volatile components, thereby facilitating recovery of the product with high purity (22). In a comparative study by Sun et al. (23) on the production costs of limonene using liquid-liquid extraction and gas stripping strategies, it was concluded that the gas stripping-based process offer significant advantages in terms of fixed costs compared to the liquid-liquid extraction process, though this process has not yet been validated experimentally. This cost advantage arises from reduced cost of downstream purification in the gas stripping approach, as no further recovery is required from the aqueous broth. In the case of the liquid-liquid extraction unit, transportation, mixing, and centrifugation of liquid streams with significantly higher flow rates than in the gas stripping process are required, necessitating larger vessel sizes and more powerful pumps. Consequently, the average inside battery limit (ISBL) cost of the liquid-liquid extraction unit is 40% higher than that of the gas-stripping recovery unit. Furthermore, the gas-stripping recovery unit excels at concentrating limonene in the organic mixture to levels exceeding 50 wt.%, far surpassing the 2 wt.% achieved in the liquid-liquid extraction unit. This substantial concentration capability directly translates into a reduced burden for downstream separation and subsequently a 30% capital cost reduction for the distillation unit.

In this study, we conduct a comprehensive comparison between liquid-liquid extraction and off-gas capture strategies for production and recovery of isoprenol. We evaluate key factors such as product toxicity, solvent toxicity, solvent partitioning, and the impact of aeration and internal overlay configurations on product volatilization rates. We then optimize product recovery from fermentor off-gas via condensation in chilled solvent, achieving 84% capture efficiency. This approach not only simplifies downstream processing but also demonstrates the potential of utilizing off-gas recovery as a viable alternative to two-phase fermentation for the production of volatile products. Continuous removal of toxic fermentation products led to a substantial 2-fold increase in effective product titers, the highest ever reported for a hemiterpene alcohol.

Host and product toxicity

Product accumulation in fermentation broth inhibits the growth of the producer organisms and decreases overall productivity. To evaluate the toxic effects of C5 alcohols and relative tolerance across variable microbial hosts, different concentrations of isopentanol, prenol and isoprenol were added into the culture medium, with optical density used as response variable. The experiments were completed in 24-well plates with 5 mL working volume and an inoculum size of 10% (v/v). The isoprenol-producing strain, E. coli AK30, and wild type strains (Corynebacterium glutamicum ATCC 13032, Saccharomyces cerevisiae CEN.PK113-7D, Escherichia coli DH1, Pseudomonas putida KT2440 and Rhodosporidium toruloides IFO0880) were used to compare relative toxicity across common biomanufacturing hosts. LB medium supplemented with 20 g/L glucose was used in the experiments using bacterial hosts, with standard YPD medium used for the yeast strains. Temperature was maintained at 30°C for 24 h and optical density (OD) was normalized by the control experiments (i.e. no C5 alcohols added to the medium).

Solvent toxicity and solvent partitioning

One approach to overcome product toxicity is selective removal of the inhibitor via addition of an immiscible solvent, generating a two-phase extractive fermentation. The most important factors to consider during the selection of a suitable solvent are extraction capacity towards the product of interest and biocompatibility of the solvent towards the microorganism. Selection of a suitable extractant was accomplished via screening of various solvents to evaluate biocompatibility with the E. coli AK30 strain alongside analysis of isoprenol liquid-liquid partitioning for each solvent when mixed with water. Solvent biocompatibility experiments were completed in 24-well plates with 5 mL working volume (3.6 mL LB media + 0.4 mL inoculum + 1 mL solvent) at 30°C and 900 rpm. Optical density at 600 nm was used as the response variable and compared with control experiments (no solvent). To determine the partition coefficient, a 20 g/L isoprenol aqueous solution was mixed to equal volumes of each solvent in microcentrifuge tubes with O-ring seal screw caps. All the water-solvent mixtures were vortexed for 15 min and then centrifuged for 5 min to separate both fractions. The partitioned layers were quantitatively analyzed by gas chromatography with a flame ionization detector (GC-FID), and the results were expressed as the ratio of the isoprenol concentration in the organic phase divided by the concentration of isoprenol in the aqueous phase.

Solvent selection for isoprenol capture

An alternative approach used to overcome product toxicity during fermentation involves use of required aeration to continuously remove the product from the fermentation broth and capture the off-gas stream in a column filled with chilled solvent. To evaluate the capture efficiency of different solvents, mock fermentations were performed in 2 L benchtop glass fermentors (Biostat B, Sartorius Stedim, Germany) with 0.6 L working volume of an aqueous solution containing 2% (w/v) isoprenol. A jacketed glass column with 0.4 L working volume was connected to the exhaust outlet of the fermentors and the gas stream was dispersed through a sparger in the bottom of the column filled with chilled solvent (Fig. 1B). In all experiments, temperature, agitation and aeration were kept constant at 37°C, 400 rpm, and 2 vvm, respectively, and samples from the fermentor and the recovery solvent were taken in regular intervals and analyzed by gas chromatography.

Volatilization rates

The effect of different fermentation conditions on the evaporation rate of isoprenol were investigated using 2 L bench top glass fermentors (Biostat B, Sartorius Stedim, Germany). In all experiments, agitation was kept constant at 400 rpm and the initial concentration of isoprenol was 2% (w/v). The following variables were studied: working volume (0.6 L and 1.2 L), air flow (0.5 vvm and 1 vvm), temperature (30°C and 37°C) and internal overlay. Samples were taken in regular intervals and analyzed by gas chromatography.

Inoculum cultivation, media and fermentation set up (strategies)

The isoprenol-producing strain used in this study, E. coli AK30, was optimized by Kang et al. (12) and maintained at -80°C after a four-stage adaptation step. First, a single colony was inoculated in LB media overnight and diluted 10-fold (v/v) in M9-MOPS minimal medium. Cells were grown at 30°C, 200 rpm for 24 h, and re-diluted 10-fold (v/v) in fresh M9-MOPS minimal medium. The last step was repeated three more times and the final cell cultures adapted to M9-MOPS medium were stored in 25% (w/v) glycerol stock solution at -80°C. Prior to fermentation, an aliquot of the frozen M9-MOPS adapted glycerol stock was used to inoculate 5 mL of M9-MOPS seed culture medium in a test tube to cultivate an overnight starter culture. The primary inoculum culture was grown in 250 mL baffled flasks containing 45 mL of the M9-MOPS media and 5 mL of the starter culture. The flasks were incubated at 30°C, at 200 rpm for approximately 6 h.

The medium used for seed cultures is described by Kang et al. (12). It contains M9-MOPS minimal medium supplement with 2 mM MgSO₄, 1 mg/L thiamine, 10 µM FeSO₄, 0.1mM CaCl₂, 2 g/L NH₄Cl, 20 g/L glucose, 10 g/L yeast extract, 30 mg/L carbenicillin, 30 mg/L chloramphenicol, and micronutrients including 3×10^− 8 M (NH₄)₆Mo₇O2₄, 4×10^− 6 M boric acid, 3×10^− 7 M CoCl₂, 1.5×10^− 8 M CuSO₄, 8×10^− 7 M MnCl₂, and 1×10^− 7 M ZnSO₄. Isoprenol production was performed in the same media without MOPS.

Fed-batch bioreactor experiments were performed in 2 L benchtop glass fermentors (Biostat B, Sartorius Stedim, Germany) equipped with two 6-blade Rushton impellers. The working volume of all tanks was 600 mL aqueous media with a 10% (v/v) inoculum size. Temperature was maintained at 30°C and pH was controlled at 6.8 with 14% NH₄OH. Air flow was fixed at 0.5 vvm for the liquid-liquid extractive process and 1 vvm for the off-gas capture process. Dissolved oxygen was cascade-controlled at 30% via agitation (300–1000 rpm) and expression of the recombinant isoprenol production pathway was induced with 0.5 mM isopropyl β-D-1-thiogalactopyranoside. Glucose feeding was initiated when the initial amount of glucose was depleted with a feed solution containing 600 g/ L glucose, 30 mg/L carbenicillin, 30 mg/L chloramphenicol.

To limit the inhibition caused by the end-product accumulation during fermentation, two different strategies were evaluated. First, 40% (v/v) of oleyl alcohol was added to the fermentor at the time of induction recover isoprenol in situ (Fig. 1A). The fermentor was equipped with a condenser maintained at 4°C to minimize isoprenol loss. In the second strategy, isoprenol was continuously stripped from the fermentor and the off-gas stream was captured in a chilled column (4°C) filled with 400 mL of octanol (Fig. 1B); during this process, the condenser was turned off. The dimensions and schematic of the columns used in the off-gas recovery system are provided in Supplementary Materials (Fig. S1)

Isoprenol and sugar analysis

Isoprenol concentrations in the aqueous and organic phase were quantified by gas chromatography with flame-ionization detection (GC-FID). An aliquot of the aqueous phase (250 µL) was combined with an equal volume of ethyl acetate (250 µL) containing 1-butanol (30 mg/L) as an internal standard. The mixture was vortexed for 15 min and then centrifuged at 20,000 x g for 5 min. The ethyl acetate fraction was diluted 10-fold in ethyl acetate containing 1-butanol (30 mg/L). The organic phase was diluted 100-fold in ethyl acetate containing 1-butanol (30 mg/L). An aliquot (1 µL) of each of the diluted samples was analyzed by gas chromatography – flame ionization detection (Thermo Focus GC) equipped with a DB-WAX column (30 m, 0.32 mm inner diameter, 0.25-µm film thickness, Agilent, USA). The initial oven temperature was 40°C, followed by a ramp of 15°C/min to 100°C, then a ramp of 40°C/min to 230°C and held at 230°C for 3 min. All isoprenol titers are normalized to the aqueous volume of the fermentation. The off-gas stream exiting the solvent column was continuously monitored to quantify isoprenol losses during fermentation. Off-gas samples were analyzed using a Prima BT Bench Top Process Mass Spectrometer (Thermo Fisher Scientific, UK), and the instrument was calibrated with known concentrations of isoprenol prior to the experiment. Gas component concentrations, including isoprenol, were recorded at 5-minute intervals throughout the fermentation process using Thermo Scientific GasWorks Software. Glucose was quantified by High Performance Liquid Chromatography (Thermo Fisher Scientific, Ultimate 3000, Waltham, MA, USA) with an Aminex HPX-87H column (Bio-Rad Laboratories, Inc) equipped with a refractive index detector. The HPLC was operated using 4 mM H₂SO₄ mobile phase at 0.6 mL.min-1 and column oven temperature of 65°C. The samples were filtered through 0.22 µm filter membranes and diluted with 4 mM H₂SO₄ for injection.

Host tolerance and product toxicity

The selection of a microbial production host for an industrial process is primarily determined by its potential to efficiently convert substrate to the product of interest. However, the accumulation of some compounds can be inhibitory, affecting growth and limiting maximum production titers. At present, data on toxic effects of C5 alcohols on microbial growth are limited. Improved availability of this data can help inform both host selection for industrial production, as well as process design strategies to manage product accumulation below the toxicity threshold.

To better quantify these toxicity thresholds, we investigated toxic effects of three C5 alcohols (isopentanol, isoprenol and prenol) across variable concentrations and across a suite of common microbial hosts. We assessed growth of the isoprenol producing strain (E. coli AK30) and five wild type strains (C. glutamicum, S. cerevisiae, P. putida, R. toruloides, and E. coli), and observed that toxicity varied among alcohols and particularly across hosts, generating diverse responses in different species. Specifically, isoprenol exhibited lower toxicity than prenol and isopentanol in all strains. Moreover, the tolerance for concentrations above 3 g/L of all alcohols declined considerably for all hosts, except for Corynebacterium, which displayed superior tolerance to all three alcohols. This finding suggests that this microorganism has significant potential as a product-tolerant host for production of hemiterpene alcohols. Kang et al. (12) observed that product toxicity was a limiting factor for isoprenol production using E. coli and that significant cell growth ceased at concentrations above 3 g/L. Engineered biological systems can result in decreased growth and global physiological changes due to the unnatural burden caused by extra expression of construct genes (24). This metabolic burden likely explains the reduced performance of the isoprenol producing strain when compared with the wild type strain under the same conditions, as shown in Fig. 2.

Overall tolerance to isoprenol can be improved via evolutionary engineering, enabling tolerance of up to 8 g/L isoprenol for P. putida and increasing tolerance by up to 47% in E. coli (17, 25). Huffer et al. (26) suggest that while membrane fluidity is the primary factor in determining alcohol tolerance, it is not the sole determinant of a microorganism's response to alcohol exposure. Therefore, continued development of more tolerant microorganisms will require modifications that address multiple aspects, including the regulation of redox, osmotic, and ion homeostasis, as well as preservation of protein structure and function, in addition to maintaining membrane fluidity. While such improvements remain a possibility for enhanced production, given the low toxicity threshold (< 10g/L) across a range of wild-type strains, achieving production titers required for economically viable biofuel production will require a combination of tolerance engineering and in situ product removal to minimize product exposure and inhibition.

Solvent toxicity and solvent partitioning

To overcome challenges with product toxicity and enable production at titers beyond the threshold of feedback inhibition, the toxic compound can be removed while it is being produced. Several integrated systems coupling fermentation and separation, also known as in situ product recovery (ISPR), have been developed to selectively remove the final product while it is being produced to limit end-product inhibition and improve the overall process efficiency. Methods for continuous product removal include pervaporation, adsorption, gas stripping, vacuum fermentation and extractive fermentation (27). In extractive fermentation, also known as liquid-liquid extractive fermentation, the product can be recovered by contacting the broth with a suitable immiscible organic solvent. The solvent of choice must be biocompatible (nontoxic to the fermenting microorganism) and have a high product partitioning. It is also desirable to choose a solvent with low solubility in the aqueous phase, density different from that of the broth to ensure phase separation by gravity, low viscosity, large interfacial tension, and low tendency to emulsify in the broth, high stability, and low cost (28).

To evaluate optimal strategies for extractive fermentation with isoprenol, eleven solvents (dodecane, oleyl alcohol, 1-octanol, PPG-725, PPG-1,000, PPG-1,200, PPG- 2,000, 2-ethyl-1,3-hexanediol, durasyn 164, pentadecane and dioctyl phthalate) from different chemical families were evaluated for their ability to extract isoprenol from the aqueous phase (Fig. 3A). Among the solvents evaluated, 1-octanol had the highest partition coefficient (6.48), followed by 2-ethyl-1,3-hexanediol (6.0). PPGs (725, 1,000, 1,200 and 2,000) presented a partition coefficient ranging from 4 to 5.1, however their high viscosity make the use of these compounds less attractive due to low mass transfer and high energy input required for mixing. Oleyl alcohol had a partition coefficient of 2.8; the non-polar solvents dodecane, durasyn 164, pentadecane and dioctyl phthalate had partition coefficients below 1, indicating that most of the product remained in the aqueous phase.

As described above, besides a high extractant capability, a suitable solvent must be non-toxic to the fermenting microorganism to enable in situ extraction. The effects of ten solvents on microbial growth were therefore evaluated to assess biocompatibility, with relative optical density of the isoprenol-producing strain used as a response variable (Fig. 3B). Eight of the ten solvents tested were found to have a high degree of biocompatibility, with relative optical densities equal or greater than 90% of the control. 1-octanol and 2-ethyl-1,3-hexanediol were found to be inhibitory, with ODs 76% and 89% lower than the control condition, respectively.

According to Marinova and Yankov (29), the mechanism of toxicity caused by the organic solvents is still not well understood, but it is likely that the membrane functions are disturbed by the interaction of the solvents with membrane lipids. These disturbances include inactivation or denaturation of membrane-bound enzymes, breakdown of transport mechanisms, and - at high concentrations - solvolysis of the cells. The interaction of solvent with cells via direct contact of the cells with the water-immiscible organic phase is known as phase toxicity, and the mechanism in which the solvent molecules dissolve in fermentation broth is known as molecular toxicity. For phase-separating organic solvents, there appears to be a trend between increasing relative molecular mass and biocompatibility. According to Barton and Daugulis (30), this could be explained by the relationship between decreasing solvent polarity with increasing chain length, as the polar effect of alcohol, ether or other active groups is diluted by the nonpolar hydrocarbon portion. The best-partitioning solvents, 1-octanol and 2-ethyl-1,3-hexanediol, were also the most toxic, due to their relatively high polarity and low molar mass. The biocompatible solvents with the highest partition coefficient were the propylene glycols, however these appeared to react with the fermentation media to form a stable emulsion, complicating downstream separation. In addition, the density of the PPGs exceeds that of water, further complicating the recovery process when separating solid, aqueous, and organic phases. The non-polar solvents dodecane, durasyn and pentadecane were biocompatible but unsuitable as extractants due to low partition coefficients. For this reason, oleyl alcohol was selected as the optimal extraction solvent for evaluation of in situ liquid-liquid extraction.

Quantifying volatilization rates

The two strategies evaluated in this study, liquid-liquid extraction and gas stripping, have distinct goals. Liquid-liquid extraction aims to minimize product losses to the off-gas while maximizing retention in the extractive solvent, while gas stripping seeks to maximize volatilization of the product and thereby consolidate downstream purification to the off-gas stream. By understanding the effect of the volatilization rate of isoprenol under different conditions, it is possible to optimize the fermentation conditions to achieve maximum yield and productivity using these different approaches. To evaluate the impact of variable fermentation conditions on volatilization rates, volatilization of isoprenol was assessed based on variable temperature, working volume, air flow, and overlay conditions, providing insight into the optimal operating conditions for each strategy. All experiments were conducted with an initial concentration of 2% isoprenol (w/v) and results from this assessment are presented in Fig. 4.

Experimental condition A, operated without an overlay, with the highest temperature (37°C), highest air flow rate (1 vvm), and lowest working volume (0.6L), demonstrated the highest evaporation rate, resulting in 94% isoprenol loss at the end of a 3-day period. When comparing conditions across different working volumes (A, B, C, and D), higher evaporation rates are consistently observed in setups with lower working volumes reflecting accelerated evaporation due to the larger surface area-to-volume ratio. At 37°C and without an internal overlay (B and D), similar levels of isoprenol evaporation were attained for both 0.6 L and 1.2 L. However, when decreasing the temperature to 30°C with an internal overlay, the difference became more pronounced, with the lower volume condition exhibiting a notably faster evaporation rate.

In examining the impact of an internal overlay on retaining isoprenol over a 5-day period, at a fixed temperature of 30°C and airflow rate of 1 vvm (A, C, E and G), distinct trends emerged. In conditions without an overlay, including condition G with a smaller volume and condition E with a larger volume, the smaller volume (G) showed a 98% loss of isoprenol, while the larger volume (E) resulted in a 92% loss. However, when an internal overlay was added to the setup under the same conditions (A and C), the presence of the overlay led to a significant reduction in volatilization rate, resulting in only 79% and 68% of isoprenol evaporating for the 0.6 L and 1.2 L volumes, respectively. These results indicate that the addition of an overlay has a greater impact on isoprenol retention than working volume. Increasing the airflow from 0.5 vvm to 1 vvm (E and F) resulted in nearly complete evaporation of the product at the end of 6 days, while the lower airflow achieved only 76% evaporation. The lowest evaporation rates were observed in the processes with an overlay at 30°C, 1 vvm, and 1.2 L, with 30% of isoprenol remaining in the fermentor after six days of the process. In sum, these results highlight the high sensitivity of solvent volatility to bioreactor operating conditions, including air flow, temperature, operating volume, and overlay selection. Back pressure plays a significant role as well, driving the equilibrium away from the vapor phase, and should be accounted for in pressurizable or larger-scale reactor systems. This sensitivity can be leveraged as part of the overall process design, to push the capture equilibrium towards either in situ extraction or off-gas capture. It should be noted that full removal of product to the off-gas is possible at 2% w/v under conditions tuned for evaporation, but substantial evaporation was observed in every condition, including operation with internal overlays. At these titers, it is therefore feasible to drive isoprenol production towards full off-gas capture, but even in the extractive fermentation case a substantial fraction of the product will still be lost to the off-gas, necessitating a hybrid capture approach. Moreover, driving the reaction towards full retention via reflux condensation, lower aeration rates, and low-temperature operation would be at odds with toxicity mitigation, particularly given the high toxicity of isoprenol and its relatively poor partitioning into extractive solvents.

Enhancing product capture efficiency

Effective and efficient product capture is crucial to maximize recovery yields and optimize the overall production process. Solvent selection plays a critical role in capturing target products, as different solvents have varying solubility properties, costs, and viscosities when cooled. In this study, we evaluated four solvents with known affinity for isoprenol - octanol, oleyl alcohol, 2-ethyl-1,3-hexanediol, and water - to identify the factors driving efficient and effective capture of isoprenol from fermentation off-gas. To generate a representative off-gas stream, 2% isoprenol was added to 2 L fermentors with a working volume of 1 L. The fermentors were maintained at a temperature of 37°C, with agitation of 400 rpm and an air flow rate of 1 vvm. The exhaust gas from each fermentor was directed into a recovery column containing 400 mL of solvent maintained at 4°C. A sparger positioned at the bottom of the solvent column dispersed the incoming gas stream, allowing the volatilized isoprenol to pass through the chilled solvent for absorption. As shown in Fig. 5, after 52 h all of the isoprenol had evaporated completely. Octanol and oleyl alcohol exhibited the highest capture efficiencies of the four solvents tested, retaining 69% and 65% of isoprenol, respectively, after 52 h. The capture efficiency of 2-ethyl-1,3-hexanediol was 61%, followed by water, which had the lowest efficiency of 35%.

Overall capture efficiency mirrored water-solvent partitioning, with capture rates generally tracking polarity and chemical properties of both the solvent and solute. While isoprenol is soluble in water up to 170 g/L (31), it contains a hydrophobic olefin functionality and partitions favorably into hydrophobic alcohols from water. The long hydrocarbon chains and polar hydroxyl (-OH) groups of oleyl alcohol and octanol create higher miscibility for isoprenol, improving capture. Moreover, unlike water, both solvents possess significantly higher boiling points than isoprenol, facilitating recovery via distillation from the capture solvent and subsequent solvent reuse. While 2-ethyl-1,3-hexanediol exhibited improved solvent partitioning in liquid-liquid extraction, at cold temperatures its viscosity increases dramatically, resulting in larger bubble formation and poor gas-to-liquid mass transfer. While oleyl alcohol was favored over octanol for extractive fermentation due to the poor biocompatibility of octanol, for vapor-phase capture biocompatibility is no longer required. Octanol was therefore utilized as the preferred solvent for further evaluation of off-gas capture under true fermentation conditions due to its high affinity for isoprenol and low viscosity when chilled.

Liquid-liquid extractive fermentation (in situ overlay)

To assess the suitability of extractive fermentation approaches to minimize product inhibition during isoprenol fermentation, a liquid-liquid extraction approach using 40% oleyl alcohol as an extractant was evaluated in a fed-batch fermentation configuration. Figure 6 shows the data obtained from a six-day fermentation, where the maximum isoprenol concentration achieved was 10.5 g/L at 100 h fermentation, representing normalized titers of 4.2 g/L isoprenol in the aqueous fraction and 6.3 g/L in the overlay. The production rate of isoprenol showed an increasing trend until day 3 of fermentation, but slowed down as the concentration in the aqueous phase surpassed 3 g/L, a concentration previously shown to inhibit E. coli replication. There was no further cell growth beyond this point, and after four days, the concentration of isoprenol began to decline due to evaporative losses. On the second day of fermentation, continuous agitation of the solvent led to the formation of a stable emulsion, resulting in a three-phase system. Emulsion formation is a substantial drawback of extractive fermentation, particularly with amphipathic solvents such as oleyl alcohol. In a study by Kang et al. (13), the production of isoprenol was compared with and without the use of an internal overlay, obtaining similar results to the findings in this study with a maximum concentration of 10.8 g/L when oleyl alcohol was used as the overlay. While liquid-liquid extractive fermentation is a promising strategy for enhancing product yield for toxic and hydrophobic molecules, product titers for isoprenol are effectively limited by partition efficiency. In addition, the downstream process is challenged by the even distribution of product in both aqueous and solvent phases, increasing the volume to be processed and the complexity of the recovery step. Successful deployment of such a strategy to reach industrial titers would require substantial improvements in strain tolerance to isoprenol, coupled with either substantial advancements in solvent partitioning, and/or continuous recycling of the extractive solvent to maintain low concentrations in the fermentor.

Off-gas product capture

To overcome the limitations of the liquid-liquid extraction strategy, we then evaluated an alternative approach that combines gas stripping and an ex situ solvent recovery to continuously remove and recover isoprenol from the fermentor and prevent it from accumulating and inhibiting further production. In this integrated method, aeration used for oxygen supply to fermentation is used to strip isoprenol to fermentor off-gas, removing it from the fermentation medium. Isoprenol is continuously captured from the off-gas stream via gas stripping in chilled octanol, with the off-gas sparged through a chilled column prior to final analysis via off-gas mass spectroscopy. In this configuration, bioreactor operating conditions are optimized for volatilization by eliminating the condenser, operating with low working volume to maximize volatility, and eliminating the extractive overlay – as shown in Fig. 4. Although higher temperatures were also found to increase the volatilization rate of isoprenol, the recombinant enzymes engineered into this strain do not function as effectively at 37°C, and the experiments were therefore conducted at 30°C to ensure optimal enzyme functionality.

Utilizing this off-gas capture approach, a maximum titer of 20.4 g/L isoprenol was obtained at 122 h, with 17.1 g/L captured in the solvent phase, 2.2 g/L in the aqueous phase and 1.1 g/L lost in the off-gas (Fig. 7). By continuously removing the product from the fermentor, we were able to capture 84% of the product in the solvent while maintaining the concentration of isoprenol in the aqueous phase below toxic levels (3 g/L), resulting in an extended production phase and a 94% increase in product titer. Additionally, by mitigating product toxicity, this process led to improvements in production rate, product yield on substrate (YP/S), and productivity—15%, 75%, and 60% higher, respectively, compared to the liquid–liquid extraction strategy (Table 1).

This strategy provides an effective way to improve yield and productivity, separating isoprenol from non-volatile solubles present in the fermentation and avoiding limitations of the liquid-liquid extraction. In addition, the product is recovered in a single phase and in a smaller volume, simplifying the downstream process by eliminating any product recovery from the fermentor. In a techno-economic assessment conducted by Sun et al., the authors compared the efficiency of a liquid-liquid extraction strategy with a gas-stripping method combined with gas scrubbing for the production and recovery of limonene. The study revealed that the cost of the liquid-liquid extraction unit was 40% higher than the gas-stripping recovery method. Furthermore, by simplifying the downstream process and achieving a more concentrated stream, a significant 30% reduction in capital costs for the distillation unit was estimated. Elimination of internal overlays also reduces solvent losses and increases the effective aqueous volume of the fermentor.

Table 1
Fermentation performance metrics for solvent extraction and off-gas capture processes
Process	Isoprenol (g/L)	Y_P/S (g/g)	Yield (%)	Productivity (g/L.h)	Production rate (g/h)
Solvent extraction	10.5	0.075	23.284	0.104	0.097
Off-gas capture	20.4	0.130	40.706	0.166	0.111

This fermentation approach effectively doubled the maximum achievable production of isoprenol in aerobic fermentation, creating a strategy that may be applied to a variety of other volatile bioproducts including medium-chain alcohols, esters, aromatics, and monoterpenes. Fully realizing these benefits at industrial intensity will require continued advancements in both product volatilization and product capture efficiency. Operation at higher temperatures enhances product volatility, and may therefore support development of thermophilic processes suitable for maintenance of rapid product stripping without costly increases in aeration rate. Host development should also include tolerance engineering, which can accelerate volatilization rates by enabling operation with higher product concentrations in the fermentor. For less volatile products, hybrid approaches may be necessary, with off-gas stripping supplementing more conventional extractive fermentation approaches. While the solvent stripping approach demonstrated here was highly effective for isoprenol, achieving 84% recovery, it utilized relatively large quantities of capture solvent in comparison to the reactor volume. New innovations in volatile product capture could substantially improve this performance, either via continuous solvent stripping or more advanced adsorptive approaches.

Product toxicity towards microbial hosts constitutes a significant challenge for development of economically viable bioprocesses. While toxicity of hydrophobic products can often be mitigated via use of internal overlays, the high cost of extractive fermentation systems limits their deployment for commodity bioproducts. Here, we demonstrate a new approach for mitigation of product toxicity towards microorganisms by utilizing the existing aeration stream to continuously remove product from the fermentor and recapture it in the off-gas stream. This process resulted in a 94% increase in isoprenol titer when compared to liquid-liquid extraction, far in excess of the product toxicity threshold. Using a capture system at above-freezing temperatures, we were able to achieve an 84% recovery efficiency from the fermentor off-gas. Additionally, this process eliminates non-volatile contaminants present in the fermentation media and streamlines purification to a single phase outside the fermentor itself, dramatically simplifying the downstream purification process. Isoprenol is a viable precursor to isoprene via direct dehydration; this strategy therefore represents a compelling pathway to bypass the drawbacks of direct isoprene biosynthesis and generate biologically-sourced polymers, rubbers, and fuels.

DMCO: 1,4-dimethylcyclooctane

GC-FID: Gas chromatography - flame ionized detector

HPLC: High performance liquid chromatography

IP: Isoprenol

IPA: Isoprenyl acetate

IPP: Isopentenyl diphosphate

ISBL: Inside battery limit

ISPR: In situ product recovery

MVA: Mevalonate

OD: Optical density

PPG: Propylene glycol

SAF: Synthetic aviation fuels

VOCs: Volatile organic compounds

Funding

This work was funded by U.S. Department of Energy, Energy Efficiency and Renewable Energy Office, Bioenergy Technologies Office (BETO) under the BETO Bioprocessing Separations Consortium through Contract DE-AC02-05CH11231 between Lawrence Berkeley National Laboratory and the U.S. Department of Energy. The portion of
the work conducted at the Joint BioEnergy Institute was supported by the U.S.
Department of Energy, Office of Science, Biological and Environmental Research
Program, through contract DE-AC02-05CH11231 between Lawrence Berkeley National
Laboratory and the U.S. Department of Energy. The United States Government
retains and the publisher, by accepting the article for publication,
acknowledges that the United States Government retains a nonexclusive, paid-up,
irrevocable, worldwide license to publish or reproduce the published form of
this manuscript, or allow others to do so, for United States Government
purposes. Any subjective views or opinions that might be expressed in this
paper do not necessarily represent the views of the U.S. Department of Energy
or the United States Government.

Availability of data and materials

No data sets were generated or analyzed during the current study.

Contributions

CAB: conceptualization, methodology, investigation, data analysis, visualization, writing (original draft), writing (review and editing). ES: funding acquisition, supervision, conceptualization, writing (review and editing). DMP: resources (provided the strain), writing (review and editing). TLS: writing (review and editing)

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests

Le Feuvre P. Are aviation biofuels ready for take off?. International Energy Agency. 2019. https://www.iea.org/commentaries/are-aviation-biofuels-ready-for-take-off. Accessed 19 Apr 2023.
Annual Energy Outlook 2020. U.S. Energy Information Administration (EIA). 2020. https://www.eia.gov/outlooks/aeo/data/browser/#/?id=57-AEO2020&cases=ref2020&sourcekey=0. Accessed 19 Apr 2023.
Keasling J, Garcia Martin H, Lee TS, Mukhopadhyay A, Singer SW, Sundstrom E. Microbial production of advanced biofuels. Nat Rev Microbiol. 2021;11:701–15. https://doi.org/10.1038/s41579-021-00577-w.
Rosenkoetter KE, Kennedy CR, Chirik PJ, Harvey BG. [4 + 4]-cycloaddition of isoprene for the production of high-performance bio-based jet fuel. Green Chem. 2019;21(20):5616–23. https://doi.org/10.1039/C9GC02404B.
Baral NR, Yang M, Harvey BG, Simmons BA, Mukhopadhyay A, Lee TS, et al. Production cost and carbon footprint of biomass-derived dimethylcyclooctane as a high-performance jet fuel blendstock. ACS Sustainable Chem Eng. 2021,9(35):11872–11882. https://doi.org/10.1021/acssuschemeng.1c03772.
Holladay J, Abdullah Z, Heyne J. Sustainable Aviation Fuel - Review of Technical Pathways. US Department of Energy - Office of Energy Efficiency& Renewable Energy. 2020 Sep. https://www.osti.gov/biblio/1660415. Accessed 20 Apr 2023. https://doi.org/10.2172/1660415.
Nagrale P. Isoprene global market report. Market Research Future. 2023. https://www.marketresearchfuture.com/reports/isoprene-industry-4799. Accessed 19 Apr 2023.
Whited GM, Feher FJ, Benko DA, Cervin MA, Chotani GK, McAuliffe JC, et al. Technology update: Development of a gas-phase bioprocess for isoprene-monomer production using metabolic pathway engineering. Ind Biotechnol. 2010;6(3):152–63. https://doi.org/10.1089/ind.2010.6.152.
Zou H, Liu H, Aboulnaga E, Liu H, Cheng T, Xian M. Microbial Production of Isoprene: Opportunities and Challenges. Wittmann C, Liao JC. Liao, editors. Industrial Biotechnology, Wiley; 2017, p. 473–504. https://doi.org/10.1002/9783527807833.ch16.
Mack JH, Rapp VH, Broeckelmann M, Lee TS, Dibble RW. Investigation of biofuels from microorganism metabolism for use as anti-knock additives. Fuel. 2014;117:939–43. https://doi.org/10.31224/osf.io/e6qdr.
Demain AL, Báez-Vásquez MA. Biofuels of the present and the future. In: Suib SL, editor. New and future developments in catalysis. Elsevier; 2013. p. 325–370. https://doi.org/10.1016/B978-0-444-53878-9.00016-3.
Kang A, Mendez-Perez D, Goh E-B, Baidoo EEK, Benites VT, Beller HR, et al. Optimization of the IPP-bypass mevalonate pathway and fed-batch fermentation for the production of isoprenol in Escherichia coli. Metab Eng. 2019;56:85–96. https://doi.org/10.1016/j.ymben.2019.09.003.
Sasaki Y, Eng T, Herbert RA, Trinh J, Chen Y, Rodriguez A, et al. Engineering Corynebacterium glutamicum to produce the biogasoline isopentenol from plant biomass hydrolysates. Biotechnol Biofuels. 2019;27:12–41. https://doi.org/10.1186/s13068-019-1381-3.
Banerjee D, Yunus IS, Wang X, Kim J, Srinivasan A, Menchavez R, et al. Genome-scale and pathway engineering for the sustainable aviation fuel precursor isoprenol production in Pseudomonas putida. Metab Eng. 2024;82:157–70. https://doi.org/10.1016/j.ymben.2024.02.004.
Kim J, Baidoo EEK, Amer B, Mukhopadhyay A, Adams PD, Simmons BA, et al. Engineering Saccharomyces cerevisiae for isoprenol production. Metab Eng. 2021;64:154–66. https://doi.org/10.1016/j.ymben.2021.02.002.
Phulara SC, Chaurasia D, Diwan B, Chaturvedi P, Gupta P. In-situ isopentenol production from Bacillus subtilis through genetic and culture condition modulation. Process Biochem. 2018;72:47–54. https://doi.org/10.1016/j.procbio.2018.06.019.
Babel H, Krömer JO. Evolutionary engineering of E. coli MG1655 for tolerance against isoprenol. Biotechnol Biofuels. 2020;13(1):183. https://doi.org/10.1186/s13068-020-01825-6.
Kim JK, Iannott EL, Bajpai R. Extractive recovery of products from fermentation broths. Biotechnol Bioprocess Eng. 1999;4:1–11. https://doi.org/10.1007/BF02931905.
Li T, Liu X, Xiang H, Zhu H, Lu X, Feng B. Two-Phase Fermentation Systems for Microbial Production of Plant-Derived Terpenes. Molecules. 2024;29(5). https://doi.org/10.3390/molecules29051127.
Huang HJ, Ramaswamy S, Tschirner UW, Ramarao BV. Separation and purification processes for lignocellulose-to-bioalcohol production. In: Bioalcohol Production - Biochemical Conversion of Lignocellulosic Biomass. Elsevier Inc.; 2010. p. 246–77. https://doi.org/10.1533/9781845699611.3.246.
Xue C, Zhao J, Liu F, Lu C, Yang S-T, Bai F-W. Two-stage in situ gas stripping for enhanced butanol fermentation and energy-saving product recovery. Bioresour Technol. 2013;135:396–402. https://doi.org/10.1016/j.biortech.2012.07.062.
Schewe H, Mirata MA, Schrader J. Bioprocess engineering for microbial synthesis and conversion of isoprenoids. Adv Biochem Eng Biotechnol. 2015;148:251–86. https://doi.org/10.1007/10_2015_321.
Sun C, Theodoropoulos C, Scrutton NS. Techno-economic assessment of microbial limonene production. Bioresour Technol. 2020;300:122666. https://doi.org/10.1016/j.biortech.2019.122666.
Ceroni F, Boo A, Furini S, Gorochowski TE, Borkowski O, Ladak YN, et al. Burden-driven feedback control of gene expression. Nat Methods. 2018;15(5):387–93. https://doi.org/10.1038/nmeth.4635.
Lim HG, Srinivasan A, Menchavez R, Yunus IS, Noh MH, White M, et al. Evolution-guided tolerance engineering of Pseudomonas putida KT2440 for production of the aviation fuel precursor isoprenol. Metab Eng. 2025;91:322–35. https://doi.org/10.1016/j.ymben.2025.05.007.
Huffer S, Clark ME, Ning JC, Blanch HW, Clark DS. Role of alcohols in growth, lipid composition, and membrane fluidity of yeasts, bacteria, and archaea. Appl Environ Microbiol. 2011;77(18):6400–8. https://doi.org/10.1128/AEM.00694-11.
Zentou H, Abidin Z, Yunus R, Awang Biak D, Korelskiy D. Overview of Alternative Ethanol Removal Techniques for Enhancing Bioethanol Recovery from Fermentation Broth. Processes. 2019;7(7):458. https://doi.org/10.3390/pr7070458.
Shuler ML, Steinmeyer DE, Togna AP et al. Economic evaluation of a multicompartment bioreactor for ethanol production using in Situ extraction of ethanol. Appl Biochem Biotechnol, 2992;28:571–588. https://doi.org/10.1007/BF02922634.
Marinova NA, Yankov DS. Toxicity of some solvents and extractants towards Lactobacillus casei cells. Bulgarian Chemical Communications. 2009;41:368–373.
Barton WE, Daugulis AndrewJ. Evaluation of solvents for extractive butanol fermentation with Clostridium acetobutylicum and the use of poly(propylene glycol) 1200. Appl Microbiol Biotechnol. 1992;36(5). http://dx.doi.org/10.1007/BF00183241.
National Center for Biotechnology Information. PubChem Compound Summary for CID 12988, 3-Methyl-3-buten-1-ol. https://pubchem.ncbi.nlm.nih.gov/compound/3-Methyl-3-buten-1-ol. Accessed 5 Aug 2025.

No competing interests reported.

Download PDF

Reviews received at journal
18 Oct, 2025
Reviewers agreed at journal
18 Oct, 2025
Reviewers agreed at journal
16 Oct, 2025
Reviewers invited by journal
14 Oct, 2025
Editor assigned by journal
30 Sep, 2025
Submission checks completed at journal
30 Sep, 2025
First submitted to journal
29 Sep, 2025

You are reading this latest preprint version

Off-gas capture: A promising strategy for removal and recovery of toxic bioproducts in aerobic fermentation

Status:

Version 1

Abstract

Figures

Background

Materials and Methods

Host and product toxicity

Solvent toxicity and solvent partitioning

Solvent selection for isoprenol capture

Volatilization rates

Inoculum cultivation, media and fermentation set up (strategies)

Isoprenol and sugar analysis

Results and discussion

Host tolerance and product toxicity

Solvent toxicity and solvent partitioning

Quantifying volatilization rates

Enhancing product capture efficiency

Off-gas product capture

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1