1-s2.0-s1383586614003608-main

Separation and Purification Technology 132 (2014) 513–540

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier .com/ locate /seppur

Review

Separation and purification of biobutanol during bioconversionof biomass

http://dx.doi.org/10.1016/j.seppur.2014.06.0131383-5866/� 2014 Elsevier B.V. All rights reserved.

Abbreviations: [BMIM]+, 1-butyl-3-methylimidazolium; BTESE, 1,2-bis(triethoxysilyl)ethane; [DCA], dicyanamide; [DHSS]�, dihexylsulfosuccinate; [DMIM]+, 1methylimidazolium; [FAP]�, tris(pentafluoroethyl)trifluorophosphate; ETES, ethyl triethoxysilane; [Hmim]+, 1-hexyl-3-methylimidazolium; [HOhmim], 1-(6-hydroxymethylimidazolium; IPA, isopropanol; [N1,8,8,8]�, methyltrioctylammonium; [NTf2]�, bis(trifluoromethylsulfonyl)imide; OA, oleyl alcohol; [OMIM]+, 1-octyl-3-methylimid[P6,6,6,14]�, trihexyltetradecylphosphonium; PDMS, poly(dimethyl siloxane); PE, polyethylene; PEBA, poly(ether-block-amide); PES-DVB, poly(4-ethylstyrene-co-DVBhexafluorophosphate; [Phosph], bis(2,4,4-trimethylpentyl) phosphinate; PS-DVB, poly(styrene-co-divinylbenzene); PTMSP, poly(1-trimethylsilyl-1-propyne), membranesbutyl phosphate; [TCB]�, tetracyanoborate; [THA]+, tetrahexylammonium; [TOA MNaph], tetraoctylammonium 2-methyl-1-naphthoate; b-CD, b-cyclodextrin.⇑ Corresponding authors. Address: Department of Bioproducts and Biosystems Engineering, University of Minnesota, 2004 Folwell Ave, St. Paul, MN 55108, USA

6126248797.E-mail addresses: [email protected] (H.-J. Huang), [email protected] (S. Ramaswamy).

Hua-Jiang Huang a,⇑, Shri Ramaswamy a,⇑, Youyan Liu b

a Department of Bioproducts and Biosystems Engineering, University of Minnesota, St. Paul, USAb School of Chemistry & Chemical Engineering, Guangxi University, Nanning, Guangxi, PR China

a r t i c l e i n f o a b s t r a c t

Article history:Received 27 January 2014Received in revised form 5 June 2014Accepted 7 June 2014Available online 17 June 2014

Keywords:Butanol separationGas strippingVacuum flashSolvent extractionMembrane separationAdsorption

Biofuels from biomass are becoming increasingly more important, due to the need for reduction in green-house gas (GHG) emissions, energy independence, and limited global availability and increasing demandand costs of fossil fuels. Butanol has several advantages over ethanol as a drop-in biofuel such as higherenergy content, potential for higher blending percentage with gasoline, lower vapor pressure, and lowerhygroscopy. It can be used in existing transportation fuel distribution infrastructure. Butanol can be pro-duced by fermentation of carbohydrates derived from biomass using Clostridium acetobutylicum or C. bei-jerinckii under anaerobic conditions. There are still many unsolved challenges for making biobutanoltechnically, and economically viable. The unsolved challenges lie in severe product (especially butanol)inhibition during bioprocessing, which leads to low butanol yield and productivity, and very low finalproduct concentration (<3 wt%), causing expensive downstream processing (product separation) costs.There are two ways for solving these problems. One is the modification of microorganisms for ABE (Ace-tone, Butanol and Ethanol) fermentation by genetic engineering, which could keep the microorganismsalive and active under higher concentration of products in the broth. This could significantly increasethe product yield, productivity, and concentration and hence reduce the production costs. However, thisis still an unrealized long term goal. Another approach is the development of efficient separation andpurification processes for product recovery. And, even if the modification of microorganisms becomesa reality, product separation and purification will still remain a major critical challenge. In this study,an extensive review of separation and purification of butanol from fermentation broth is provided,including gas stripping, vacuum flash, liquid–liquid extraction, membrane solvent extraction or perstrac-tion, membrane pervaporation, membrane distillation, thermopervaporation, reverse osmosis, adsorp-tion, and integrated bioprocessing with various separation methods. It is concluded that membranepervaporation, solvent extraction, and adsorption are the most energy-efficient approaches for removalof butanol from the ABE fermentation broths. It should be noted that this is not a strict comparisonand it is suggested that each separation process should be optimized before comparison. Integration ofbioreactors with these energy-efficient separation methods could significantly increase the product yield,productivity and concentration and hence lower the production cost. Butanol dehydration is also dis-cussed. This review could be helpful in the research and development and commercialization of biobut-anol as renewable drop-in biofuels and biochemicals.

� 2014 Elsevier B.V. All rights reserved.

-decyl-3-hexyl)-3-azolium;); [PF6]�,; TBP, tri-

. Tel.: +1

http://crossmark.crossref.org/dialog/?doi=10.1016/j.seppur.2014.06.013&domain=pdf

http://dx.doi.org/10.1016/j.seppur.2014.06.013

mailto:[email protected]

mailto:[email protected]

http://dx.doi.org/10.1016/j.seppur.2014.06.013

http://www.sciencedirect.com/science/journal/13835866

http://www.elsevier.com/locate/seppur

Nomenclature

C concentration (g L�1 or mol L�1)J pervaporation permeate flux (kg m�2 h�1)K distribution coefficientK0 distribution coefficientK00 distribution coefficientm mass (g) or mole numbers (mol)P pressure (Pa)S selectivity of an extracting agent (extractant) for the

solute (butanol/isobutanol) over waterT temperaturew weight percent (wt%) or mass fraction (g g�1)x mole fraction

Greek lettersq density (kg m�3)

l viscosity (mPa s)c solvent–air interfacial tension (i.e., surface tension) or

solvent–water interfacial tension (mN m�1)b separation factor of butanol/isobutanol (over water) for

membrane pervaporationd thickness of membrane (lm)

SubscriptAP aqueous phaseD distributioni componentOP organic phasePerm permeate side

514 H.-J. Huang et al. / Separation and Purification Technology 132 (2014) 513–540

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5142. Separation of butanol from dilute solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516

2.1. Gas stripping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
2.1.1. Batch fermentation–gas stripping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5162.1.2. Fed-batch fermentation–gas stripping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5172.1.3. Continuous fermentation–gas stripping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5172.1.4. Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 517
2.2. Vacuum flash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5172.3. Liquid–liquid extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 518

2.3.1. Extractive fermentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5192.3.2. External solvent extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5232.3.3. Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524

2.4. Membrane-assisted solvent extraction (perstraction). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5242.5. Membrane pervaporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524

2.5.1. Pervaporation with hydrophobic polymeric membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5252.5.2. Pervaporation with polymeric composite membranes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5252.5.3. Pervaporation with supported liquid membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5262.5.4. Integrated fermentation–membrane pervaporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5272.5.5. Energy requirement for membrane pervaporation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5282.5.6. Advantages and disadvantages of membrane pervaporation, and summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529

2.6. Other membrane technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529
2.6.1. Membrane distillation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5292.6.2. Thermopervaporation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5302.6.3. Reverse osmosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 530
2.7. Adsorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5312.8. Comparison of separation methods for product removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534

3. Further separation and purification/dehydration of butanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5344. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536

1. Introduction

Biofuels from biomass are becoming increasingly more impor-tant, due to the need for reduction in GHG emissions, energy inde-pendence, and limited global availability and increasing demandand costs of fossil. Biobutanol as biofuel has several advantagesover bioethanol such as higher energy content, lower vapor pres-sure, higher flash point (37 �C vs. 15 �C), lower hygroscopy, andbetter miscibility with gasoline [1–4]. Specifically, butanol con-tains 30% more energy content on a unit volume basis than ethanol(27.83 MJ (L butanol)�1 vs. 21.27 MJ (L ethanol)�1 [5]. Life-cycleassessments by Swana et al. [4] show that the net energy return

associated with corn-to-biobutanol conversion is greater than thatof the corn-derived bioethanol (6.53 MJ L�1 vs. 0.40 MJ L�1). Buta-nol’s lower vapor pressure and higher flash point represents thatit is safer than ethanol, and being less hygroscopic means less cor-rosion to the fuel pipelines and equipments. Besides, butanol canbe blended with gasoline at a higher percentage than ethanol. Cur-rent US regulations allow biobutanol to be blended at up to 16% byvolume vs. 10% for ethanol (butamax.com). In addition, butanol iscompatible with the current automobile engine design and can beused as a drop-in fuel and used in existing transportation fuel dis-tribution infrastructure, making it an ideal candidate to replacegasoline [6,7]. In addition to being used as biofuel, butanol can

http://butamax.com

H.-J. Huang et al. / Separation and Purification Technology 132 (2014) 513–540 515

be used as solvent in food and pharmaceutical industries, a build-ing block for producing a variety of chemical products, such aspaints, coatings, bio-based polymers and plastics, and other chem-icals. Butanol is even more valuable than ethanol as a chemical [8].

Butanol has four isomers: n-butanol (1-butanol), sec-butanol(2-butanol), isobutanol (2-methyl-1-propanol), and tert-butanol(2-methyl-2-propanol). They have different physicochemical prop-erties due to their different structures. Among these four isomers,tert-butanol is miscible with water and it has a low melting pointof about 25–26 �C. Sec-butanol has a high solubility in water (29 g/100 ml at 25 �C). Thus, both tert-butanol and sec-butanol are notsuitable for use as gasoline alternatives, though tert-butanol canbe used as gasoline octane booster and oxygenate. The other twoisomers n-butanol and isobutanol have limited solubility in water,and they are potential gasoline alternatives. In this review, n-buta-nol and isobutanol will be mainly discussed.

Currently, n-butanol is mainly produced from petroleum, with aglobal capacity of 2.8 million tons of petro-butanol, equivalent to$5 billion. Using low-cost biomass feedstock and pretreatmenttechnology, advanced biocatalyst, faster fermentation system andenergy-efficient product separation process, production cost ofbio-based n-butanol can be 30–60% less than the cost of petroleumbased butanol (cobalttech.com). Also, butanol has the potential tosubstitute both bioethanol and biodiesel in the biofuels marketestimated to be worth $247 billion by 2020 [9].

Recently, due to the significant advantages of biobutanol overbioethanol as alternative fuels as described above, and its higherprofitability than ethanol in the chemicals market, there have beenincreasing interests in butanol bioconversion and significantefforts have been made in commercial production of biobutanolor retrofitting some existing bioethanol plants for butanol produc-tion. In China, for instance, the total ABE solvent production capac-ity was about 0.21 million TPA (tons per annum) in 2008 andexpected to be expanded to 1 million TPA [10]. In USA, for example,Gevo, a Denver-based company, and Butamax, a joint venturebetween BP and DuPont, vigorously conducted research on isobu-tanol and n-butanol bioconversion, respectively, aiming at thecommercialization of the solvent production, with current on-going retrofit of existing bioethanol plants for isobutanol/n-buta-nol production. Other companies around the world such as CobaltBiofuels, Cathay Industrial Biotech, and Green Biologics are alsodedicated to the research and development of butanol bioconver-sion [11].

Biochemically, n-butanol can be produced by ABE fermentationof biomass carbohydrates. Current commercial biobutanol pro-cesses are based on fermentation of starch or sugar-based feed-stocks such as corn [12], molasses [13], whey permeate [13], anddried distillers’ grains and solubles [14]. Most existing biobutanolplants use corn, which competes with food and animal feed. Therelatively high cost of corn leads to higher butanol production cost.For this reason, there has been a growing research interest indeveloping technologies for producing drop-in biofuels such asbutanol from non-food cellulosic biomass including corn fiber[13,15], corn stover [16], corn stalk [13], rice bran [17], rice straw[18], barley straw [19], wheat straw [20–22], wheat bran [23],switchgrass [16], and cassava bagasse [24] as well as other non-food, renewable biomass such as starch-containing microalgae;for example, Chlorella, which contains up to 30–65% starch ondry basis, making it a good substrate for butanol fermentation[25–27]. In addition to starch, sugar, and cellulosic biomass ascarbohydrate sources, stillage [28] and glycerol [28–32], a majorby-product of biodiesel plants, can also be applied for butanolproduction by fermentation.

There are four species of ABE producing clostridia for ABEfermentation of various carbohydrates (glucose, xylose, etc.):

C. Acetobutylicum, C. beijerinckii, saccharobutylicum, and C.Saccharoperbutylacetonicum [33], which are based on the Clostridiabutanol metabolic pathway [34]. Among these, the most com-monly researched clostridia are C. acetobutylicum [10,18,24,35–42] and C. beijerinckii [12,20–23,43–49]. Due to the toxicity or inhi-bition of butanol to the cell, the obtained ABE fermentation broth isvery dilute, usually with total ABE concentration of less than20 g L�1 (the common weight ratio A:B:E = 3:6:1, dependent onthe substrate and the type of bacteria used), or the butanol concen-tration of less than 12 g L�1. Based on the best solventogenic clos-tridia strains in the literature summarized by Green [9], the besttotal ABE solvent titer and butanol titer are about 28 g L�1 and20 g L�1, respectively. Such low concentrations of butanol makethe product separation highly costly. Furthermore, the butanolyield and productivity are also low. Thus, the overall biobutanolproduction cost is relatively high. Recent reviews on biobutanolproduction have been published [11,34,50–55].

The ABE fermentation could be switched to the isopropanol,butanol and ethanol (IBE) fermentation with engineered C. acet-obutylicum through introducing a single secondary alcohol dehy-drogenase, where acetone is converted to isopropanol [56,57]. Asisopropanol can be used as a fuel additive for increasing high-octane value of gasoline, the IBE mixture produced can be directlyused as biofuel. Also, isopropanol can be used as an industrial sol-vent. The IBE fermentation is equivalent to the ABE fermentationplus further conversion of acetone to isopropanol. Thus, IBE con-centration is also dilute, similar to ABE.

Isobutanol, which has similar energy content to that of n-buta-nol, can also be produced biochemically from glucose and cellu-losic biomass, using the recombinant organisms such asEscherichia coli, Saccharomyces cerevisiae, Bascillus subtilis, Coryne-bacterium glutamicum, and Clostridium cellulolyticum, based onthe valine pathway. The metabolic pathways for producing higheralcohols from various resources are reviewed [53,55]. Among theseorganisms, E. coli was engineered to produce 22 g L�1 isobutanol ata yield of 0.35 g isobutanol (g glucose)�1, corresponding to 86% ofthe theoretical maximum yield of isobutanol [53]. In addition,the integration of fermenters with in situ product removal (ISPR)by gas stripping could increase isobutanol titer to 50 g L�1 [54],which represents the highest concentration among the literaturereviewed.

In this review, separation of n-butanol and isobutanol fromtheir dilute fermentation broths will be mainly discussed. Whilemost references reported are related to separation of n-butanolfrom ABE fermentation broth, usually, the separation technologiesare also suitable for the separation of isobutanol from its fermen-tation broth.

In general, challenges for biobutanol production from biomass[9] include high feedstock cost, low butanol titers (ABE < 20 g L�1),increasing separation costs, low butanol yield, low volumetric pro-ductivity increasing capital and operating costs, and high waterusage [9]. In order to make biomass-based butanol economicallyand commercially viable, two major challenges need to be solved:(i) severe product (especially butanol) inhibition, i.e., solvent toxic-ity to the microorganisms used for fermentation, which leads tolow butanol yield and productivity, and (ii) expensive downstreamprocessing (product separation and purification) [6,58]. Thoughgenetic modification of the microorganisms for ABE productioncould increase their tolerance to the toxicity of products, leadingto increase in butanol yield, broth concentration, and productivityas described previously, a high toxicity-tolerant microorganism forABE that can tolerate ABE titer of well over 3% is still years away. Amore near term plausible solution is to eliminate or reduce theproduct inhibition through continuous product removal from thebroth during fermentation, i.e., integration of in situ product sepa-

http://cobalttech.com

Fig. 1. Simplified block diagram of the integrated FBF–gas stripping system [60].


ration with ABE fermentation so that the ABE titer remains belowthe inhibition concentration (usually <2 wt% butanol).

Recent review on separation technologies for biofuels withmore focus on separation of bioethanol from fermentation brothincludes Huang et al. [59], and Vane [60]. There are similaritiesbetween separation of bioethanol from fermentation broth andbiobutanol from ABE fermentation broth. For example, both brothsare very dilute and hence their separation from dilute solutions areenergy intensive and costly; both have larger molecular size thanwater, so a small fraction of water can be removed from alcohol-rich mixture by using molecular sieve for alcohol dehydration;both have a –OH group, so they can be extracted from aqueoussolutions using some similar type of extraction solvents, for exam-ple, oleyl alcohol, n-dodecanol, butyl acetate, and [P6,6,6,14][DCA],just to name a few. However, there exists significant differencesbetween them. Specifically, the concentration of ethanol fermenta-tion broth is significantly higher than that of ABE fermentationbroth (5–12 wt% ethanol vs. <2 wt% butanol). Also, at atmosphericpressure, the ethanol forms azeotropes with water at 95.5% ethanolat 78.2 �C that is significantly lower than the water’s boiling point(100 �C), while butanol forms azeotropes with water at 55.5 wt%butanol at 93 �C that is only slightly lower than that of water. Thus,separation or pre-concentration of butanol from initial ABE brothsby using ordinary distillation directly should consume much moreenergy compared to the case for ethanol separation from ethanolfermentation broth. In fact, the separation of butanol from ABEbroths by using direct ordinary distillation is infeasible as itrequires more energy than the energy content in the recoveredbutanol. Thus it is necessary to explore some other separationmethods for pre-concentration of butanol from ABE broths. Onthe other hand, ethanol is completely miscible with water andthe azeotropes formed by ethanol with water are homogeneous.In contrast, at 20 �C butanol has a lower solubility limit in water(7.7 wt% butanol) and an upper solubility limit in water(79.9 wt% butanol) [61], i.e., the water solubility in butanol being20.1 wt% water. Thus, the azeotropes formed by butanol withwater are heterogeneous. This is favorable for butanol–water sys-tem for further separation of butanol from pre-concentrated solu-tions containing more than 7.7 wt% butanol through phaseseparation. The pre-concentrated solutions having butanol concen-trations between 7.7% and 79.9% will be separated into two phases:one aqueous phase containing 7.7wt% butanol, and another organicphase containing butanol of more than 7.7wt% at 20 �C and101.325 kPa. This suggests that special separation approaches suchas gas stripping, vacuum flash, and membrane pervaporation canbe first used for in situ product removal and/or pre-concentrateABE solutions to significantly higher than 7.7 wt% of butanol con-centration. After phase separation of the pre-concentrated solu-tions, a solution of higher concentration of significantly over 7.7%butanol will be achieved, along with another (aqueous) solutionof 7.7%.

In order to enhance substrate utilization and butanol produc-tivity, many researchers investigated in situ simultaneous removalof butanol from fermentation broth to reduce product inhibitionusing different separation techniques. To explore butanol recov-ery technologies for reducing its overall production cost, in thisstudy, an extensive review of separation and purification of buta-nol from fermentation broth is provided, including gas stripping,vacuum flash, liquid–liquid extraction, membrane solvent extrac-tion or perstraction, membrane pervaporation, other membranetechnologies such as membrane distillation, thermopervaporationand reverse osmosis, adsorption, as well as different integratedfermentation–separation processes. The comparison between dif-ferent separation methods in terms of energy consumption andthe further separation and purification of butanol are also dis-cussed. This could be helpful in the research and development

and commercialization of renewable butanol biofuels andbiochemicals.

2. Separation of butanol from dilute solutions

2.1. Gas stripping

Gas stripping is a simple approach for removal of product inhi-bition from ABE fermentation [8]. The fermentation off-gas (CO2

and H2) or another inert gas can be used as the carrier gas. A num-ber of researchers have studied the use of gas stripping for butanolremoval from fermentation [12,21,22,46–48,62]. Gas stripping canbe integrated into the ABE fermentation system for in situ recoveryof ABE product. The gas stripping mechanism is based on one-stagegas–liquid equilibrium [60]. Batch, fed-batch, or continuous fer-mentation can be integrated with gas stripping for in situ productremoval. Fig. 1 shows an integrated fed-batch fermentation(FBF)–gas stripping product recovery system where the fermenta-tion gas are used as the carrier gases. In this system, substrate inhi-bition and product inhibition are reduced by fed-batch operationand gas stripping, respectively. At the fermentation temperature,ABE is stripped off by the fermentation gas, and condensed intoliquid. After separation from the liquid products, the uncondensedinert gas is recycled back to the fermenter.

2.1.1. Batch fermentation–gas strippingThe combination of gas stripping and simultaneous saccharifi-

cation (enzymatic hydrolysis) and fermentation (SSF) for ABE pro-duction from wheat straw also significantly increases the ABEproductivity. For example, the solvent productivity of the inte-grated SSF–gas stripping process in a batch bioreactor using C. bei-jerinckii P260 for butanol production from wheat straw is up to0.31 g L�1 h�1, 15% higher than the SSF only, while having similaryield (0.41 g g�1 of SSF–gas stripping vs. 0.42 g g�1 of SSF only)[21]. It is also found that the yield of the SSF–gas stripping systemusing wheat straw hydrolyzate is comparable to the batch fermen-tation control using expensive glucose substrate. Xue et al. [63]investigated the integration of the batch ABE fermentation in afibrous bed bioreactor with two-stage gas stripping for butanolrecovery. The first-stage gas stripping was used for in situ productremoval. Its condensate contained 175.6 g L�1 butanol (227.0 g L�1

ABE), which was subjected to phase separation resulting in anorganic phase containing 612.3 g L�1 butanol (660.7 g L�1 ABE)and an aqueous phase containing 101.3 g L�1 butanol (153.2 g L�1

ABE). The second-stage gas stripping was then used to strip offthe product from the aqueous phase, leading to a highly concen-trated product containing 420.3 g L�1 butanol (532.3 g L�1 ABE).The high product concentration implies the significant energy-sav-ing in the subsequent product separation and purification process.Similarly, gas-stripping can be used in the IBE fermentation for


in situ removal of IBE mixture. The recovery of IPA and butanolfrom model solutions (17 g L�1 IBE mixture with I:P:A weight ratioof 6:10:1) by gas stripping at 37 �C are 77.2% and 82.3% respec-tively. For the IBE batch fermentation, the use of gas-strippingcould increase the IBE productivity to 0.29 from 0.16 g L�1 h�1 ofthe control (without gas-stripping) [64]. The comparisons betweensome batch fermentation (BF)–gas stripping ISPR processes andbatch fermentation controls are given in Table 1.

2.1.2. Fed-batch fermentation–gas strippingEzeji et al. [47] studied the ABE production from the glucose

substrate using C. beijerinckii BA101 as biocatalyst in the fed-batchfermentation (FBF) integrated with gas stripping. Results showedthat the product yield and the productivity were significantlyimproved to 0.47 g g�1 and 1.16 g L�1 h�1 respectively, comparedto 0.39 g g�1 and 0.29 g L�1 h�1 of the batch control [47]. Fermen-tation of wheat straw hydrolysate using C. beijerinckii P260 forbutanol production in a fed-batch reactor where a sugar solutionconsisting of glucose, xylose, arabinose, mannose, and galactosewas added as supplement showed significant improvement inABE productivity of 0.36 g L�1 h�1 and ABE yield of 0.44 comparedto the SSF–gas stripping in batch fermentation as described above[22]. The gas stripping method not only continuously separatesABE products from the broth and hence removes the product inhi-bition but also brings agitation, which causes more hydrolysis inthe SSF process and hence increases the ABE productivity [21].Most recently, gas stripping was also integrated with fed-batch fer-mentation for producing butanol from concentrated cassavabagasse hydrolysate (584.4 g L�1) with C. acetobutylicum [24]. Withcontinuous ABE removal by gas stripping, an ABE concentration of108.5 g L�1 (butanol 76.4 g L�1, acetone 27.0 g L�1, and ethanol5.1 g L�1) was obtained, with the ABE yield of 0.32 g g�1 and buta-nol yield of 0.23 g g�1. The butanol concentration could beincreased to about 64% (w/v) after phase separation [24], whichcould result in significant cost reduction in the post separationand purification of the product. Also, high concentration of n-buta-nol was produced from glucose by C. acetobutylicum JB200 in a fed-batch fermentation coupled with intermittent gas stripping [65].Setlhaku et al. [67] investigated a two-stage ABE fermentation pro-cess using C. acetobutylicum ATCC 824 strain, with the 1st stageoperated in a continuous mode and the 2nd stage as a fed-batch.With gas stripping integrated with the 2nd stage and operatedintermittently, additional glucose feeding was allowed and thebroth concentration of up to 59 g L�1 of butanol and 73 g L�1 oftotal ABE was obtained. The comparisons between some fed-batch

Table 1Comparisons of BF–gas stripping, FBF–gas stripping, and continuous fermentation–gas str

BF–gas stripping vs. batch control

Butanol conc. (g L�1) 19.8(16.2)Total ABE titer (g L�1) 75.9(17.7)a 31.8(25.5) 21.42(11.9Cond. conc. (g L�1) 175.6/227.0b

Org. phase conc. (g L�1) 612.3/660.7BuOH yield (g g�1) 0.25(0.20)ABE yield (g g�1) 0.47(0.40) 0.40(0.32) 0.41(0.42)BuOH prod. (g L�1 h�1) 0.41(0.30)ABE prod. (g L�1 h�1) 0.60(0.29) 0.66(0.48) 0.31(0.27)Cells conc. (g L�1) 11.0(3.2) –Initial glucose (g L�1) 161.7(59.2) 80(80) 26.1(25.6)Glucose used (g L�1) – –Glucose utiliz. rate (g L�1 h�1) 1.53(0.76) –Gas flow rate (L min�1) 1.5Condenser temp. (�C) 2Refs. [46] [63] [21]

GS = Gas Stripping; CF = Continuous Fermentation.a Data between parentheses is the data of the control without gas-stripping.b The data over the slash ‘/’ is the data for butanol while the data after ‘/’ is the data

fermentation (FBF)–gas stripping ISPR processes and batch fermen-tation controls are also given in Table 1.

2.1.3. Continuous fermentation–gas strippingEzeji et al. [66] integrated the continuous one-stage ABE fer-

mentation using C. beijerinckii BA101 with gas stripping forin situ product recovery. With semi-continuous removal of inhibi-tory chemicals and products, the integrated system produced461.3 g L�1 ABE from 1125.0 g L�1 sugar as compared to a controlbatch process in which 18.4 g L�1 ABE was produced from 47.3 gsugar. These results demonstrate that ABE fermentation can beoperated in an integrated continuous one-stage fermentation andproduct recovery. The comparisons between the continuous fer-mentation–gas stripping system (CF–GS) and the BF control (with-out gas stripping) are also given in Table 1.

2.1.4. SummaryIn summary, gas stripping has many advantages: simple tech-

nology and operation, selective removal of volatile compoundswithout loss of nutrients, use of fermentation off-gas, withoutrequirement of chemicals or membrane, and no toxicity to cells,efficient removal of product inhibition. This also has additionaladvantages of improvement of hydrolysis for SSF process by intro-ducing agitation, significant increase in reactor productivity andproduct yield, higher product concentration after condensationand phase separation, and hence potential cost reduction.

2.2. Vacuum flash

Flash is a separation process based on a one-stage V–L equilib-rium. The continuous vacuum flashing process can be combinedwith the ABE fermentation for in situ product removal from themicrobial culture, as seen in Fig. 2. In this process, a side streamof the fermentation broth liquor is heated by heat exchanger to aflash temperature. The heated stream then enters the flash tankand is flashed into a vapor phase and a liquid phase. The obtainedliquid phase, mainly containing water, is pumped back to the fer-menter for reuse, while the vapor phase, after combined with thefermentation off-gas, enters the condenser where the volatile sol-vents (ABE) and chemicals are condensed into liquid mixture thatis later fed into the liquid–liquid phase separator connected to avacuum system. The butanol-rich (organic phase) and the water-rich (aqueous phase) solutions are obtained from the phase separa-tor, and the non-condensable gas is pulled out of the system by avacuum pump.

ipping with batch controls without gas stripping.

FBF–gas stripping vs. batch control CF–GS vs. BF

151.7 113.3(19.1) 76.44(9.71)3) 232.8(17.6) 172 108.50(15.41) 461.3(18.4)

150.5/195.9610/660 6400.24(0.21) 0.23(0.22)

0.47(0.39) 0.36 0.32(0.34)0.35(0.24) 0.29(0.24)

1.16(0.29) 0.53 0.41(0.39) 0.41(0.39)15.0100.0(59.9) 600(100) 584.4500.1(45.4) 474.9(86.4) 336.6(44.8) 1125(47.3)2.49(0.76) 1.46(1.11) 1.28(1.12) 0.92(0.28)

1.52

[47] [65] [24] [66]

for ABE.

Fig. 2. Continuous vacuum flash fermentation process for removal of butanol.


A few of studies on the ABE fermentation–vacuum flash sys-tems, simply called vacuum flash fermentations, have been con-ducted [68–72]. For instance, Mariano et al. [69–70] studied anABE fermentation using C. beijerinckii P260 under continuous orintermittent vacuum at normal fermentation temperature. The sol-vents were boiled and flashed out, reducing the butanol inhibition.Evanko et al. [71] used the combined fermentation and continuousvacuum flashing process for butanol recovery. The fermentation isoperated under simultaneous saccharification and fermentation(SSF) mode at 32 �C. A side stream containing 4 wt% butanol is con-tinuously removed from the fermenter and a heat exchanger isused to control the temperature of a flash tank feed at 34 �C. Thevacuum pressure is about 6.7 kPa (50 mmHg). The flashed buta-nol–water azeotrope vapor containing 54 wt% butanol and46 wt% water is condensed into a butanol rich phase having a buta-nol concentration of about 680 g L�1 and a water rich phase havinga butanol concentration of about 86 g L�1. It has been verified byAspen Hysys process simulation that a binary aqueous solutioncontaining 4 wt% isobutanol (rather than n-butanol) under theabsolute (rather than vacuum) pressure of 6.7 kPa could be flashedto obtain a vapor containing 52.36 wt% isobutanol and 47.64 wt%water, i.e., 21.08 mol% isobutanol and 78.92 mol% water. This isin agreement with the T–xy diagram as shown in Fig. 3.

Abdehagh et al. [68] simulated the vacuum flash fermentationto determine the final concentration of butanol recovered. For

Fig. 3. T–xy for water/isobutanol at 6.7

the vacuum fermentation conducted at 37 �C with a compositionof 7 g ABE L�1 with a ratio of A:B:E of 3:6:1, the boiling of the qua-ternary mixture at this temperature would occur at an absolutepressure of around 7.1 kPa as calculated using UniSim� DesignR400 (Honeywell). The vapor exiting the fermenter at equilibriumwith the model fermentation broth was predicted to have massfractions of 0.144, 0.164, 0.008 and 0.684 for butanol, acetone, eth-anol and water, respectively. These corresponds to the enrich-ments of 34, 78, and 11 times for butanol, acetone, and ethanol,respectively. In terms of this, vacuum fermentation appears to bea promising alternative for biobutanol production.

Vacuum flash fermentation has many advantages includingno requirement of additional equipment such as membrane,sparger or agitator, no damage to the microorganism, andsignificant increase in butanol concentration in the productvapor stream [68]. Its major disadvantage is that it requires avacuum system.

2.3. Liquid–liquid extraction

Liquid–liquid (L–L) extraction, or solvent extraction, is a con-ventional separation process used to extract desirable solutes fromdilute solution with one or more mixed extractants that is immis-cible with the solvent of the solution. L–L extraction has been com-mercially used in chemical and pharmaceutical industries. It alsohas great potential in separation of butanol from dilute fermenta-tion broth. Liquid–liquid extraction is even more favorable forbutanol separation from ABE fermentation broth than cellulosicethanol separation from carbohydrates-to-ethanol fermentation.This is because butanol is more hydrophobic hence less misciblewith water than ethanol due to its longer carbon chain, and atthe same time, the concentration of butanol from ABE fermenta-tion is much less than that of ethanol obtained from carbohy-drates-to-ethanol fermentation.

In general, there are two basic categories in separating butanolfrom the ABE fermentation broth with solvent extraction: one isintegration of solvent extraction into the bioreactor, also calledextractive fermentation, where butanol can be recovered in situduring fermentation so that product (especially butanol) inhibitionis significantly reduced and hence the butanol yield and productiv-ity significantly increased. The other one is external solvent extrac-tion of butanol from the broth after the fermentation step.

kPa (plotted by Aspen Hysys 8.0).

Fig. 4. Extractive fermentation.


2.3.1. Extractive fermentationExtractive fermentation, an integration of fermentation and

in situ solvent extraction for butanol removal, has been widelystudied. The simplified diagram of integrated fermentation – L–Lextraction process is shown in Fig. 4 assuming that no water isextracted [59,73]. The distillation column in the process is usuallya multi-stage distillation column where the extractant is recoveredand recycled to the fermenter, while the product ABE is isolatedout for further separation. However, if a nonvolatile extractant,e.g., ionic liquid that is discussed later in this article, is applied,an evaporator, the simplest distillation with only one-stage phaseequilibrium, can be used instead of the distillation column.

For efficient in situ isolation of butanol from ABE fermentation,suitable solvents should be selected based on the following criteria[59,60]:

– Nontoxic to microorganism and environment.– High distribution coefficient.– High selectivity (separation factor).– Low solubility in aqueous solution.– Density significantly different from that of the broth for easy

phase separation.– Low viscosity for less energy consumption in extraction.– Large interfacial tension for easy coalescence of emulsions and

phase separation.– High stability.– Suitable volatility or boiling point.– Commercial availability at low cost.

Distribution coefficient and selectivity are the two majorparameters for determining an extractant. The distribution coeffi-cient of the component i between organic and aqueous phases isdefined as the ratio of the concentration of the component i inthe organic phase to the concentration of the component i in theaqueous phase at extraction equilibrium. That is,

KDi ¼Ci;OP

Ci;APð1aÞ

where the subscript OP is used to represent the organic phase andAP the aqueous phase. i can be butanol, acetone, ethanol, or water,for example, in the ABE broth. The concentration C in g L�1 ormol L�1 (both have the same value of KD).

The distribution coefficient of component i can also beexpressed by

K 0Di ¼wi;OP

wi;APð1bÞ

where w is weight percent (wt%) or mass fraction (g g�1).

To our knowledge, few of the literatures cited defined the distri-bution coefficient of component i by the ratio of the amount (molor g) of the component i in the organic phase to the amount of thecomponent i in the aqueous phase at extraction equilibrium:

K 00Di ¼mi;OP

mi;APð1cÞ

where mi is the mass (g) or mole numbers (mol) of component i inorganic phase (OP) or aqueous phase (AP).

Eq. (1c) is less commonly used.In this paper, the extraction of butanol or isobutanol will be

mostly discussed, so the distribution coefficient of butanol or iso-butanol will be simply expressed by KD, K 0D, or K 00D, depending onthe unit basis used.

The relationship between KD and K 0D can be derived as follows:

KDi ¼qOP

qAP

� �K 0Di ð2Þ

where q is density of organic phase (OP) or aqueous phase (AP).Assuming the density of organic phase being close to the den-

sity of extractant and the density of aqueous phase being close tothe density of water, Eq. (2) can be simplified to:

KDi �qExtractant

qH2O

!K 0Di ð3Þ

Thus, if the density of extractant is not very different from that ofwater, then KD andK 0D should be the same for a given extractionsystem.

The selectivity (Si) for the solute i (to be extracted, e.g., butanol)over water, is defined by the ratio of the distribution coefficient ofthe component i to the distribution coefficient of water:

Si ¼KDi

KD;H2O¼ Ci;OP=CH2O;OP

Ci;AP=CH2O;APð4Þ

Apparently, whether Si is defined as (2), or Si ¼K 0Di

K 0D;H2 O, or Si ¼

K 00DiK 00D;H2 O

,the Si value should be the same.

The potential extractants for butanol recovery include oleylalcohol [35], glyceryl tributyrate [74,75], methylated crude palmoil [76] and other chemically modified plant oils such as hydroxyl-ated corn oil [77], biodiesel [76,78], surfactants [79], and ionic liq-uids [80–83]. These extractants have significantly higher boilingpoints than butanol, thus from an energy consumption standpoint,favoring the subsequent separation of the extractants and butanolby distillation.

2.3.1.1. Conventional extractants for recovery of butanol. Oleyl alco-hol is a nontoxic, most widely researched benchmark extractantfor separating butanol from dilute ABE broth with a butanol distri-bution coefficient (KD) of 3.4 and a selectivity of 208 for initial con-centration of 1 wt% butanol in aqueous solution [84]. Thedistribution coefficients (KD) of butanol, acetone, and ethanol are4.1, 0.45, and 0.22, respectively, using oleyl alcohol as extractantfor the cell-free broth containing 5.2 kg m�3 butanol, 2.9 kg m�3

acetone and 0.07 kg m�3 ethanol at 37 �C and a ratio of organicphase: aqueous phase = 2:5 (v/v) [85]. Davison and Thompson[35] used oleyl alcohol to remove butanol in a fluidized-bed biore-actor with immobilized C. acetobutylicum, and direct simultaneousextraction and fermentation. Their results showed a 50–90%increase in product yield and butanol productivity as organic-to-aqueous ratio increased from 1 to 4, compared to the non-extrac-tive case. The maximum productivity obtained was 1.8 g butanolL�1 h�1, and the extractant had a distribution coefficient (KD) ofaround 3 for butanol at an aqueous concentration of 1 g L�1 (noselectivity was given). The density of oleyl alcohol is 845–855 kg m�3, significantly different from that of water, which is

Table 2Extraction equilibrium compositions for selected solvents at 25 �C [71].

Solvent Organic phase Aqueous phase K 0D Selectivity for i-BuOHa

i-BuOH (wt%) H2O (wt%) i-BuOH (wt%) H2O (wt%)

Butyl acetate 0.83 1.3 0.34 99.7 2.5 187Butyl acetate 3.1 1.9 1.5 98.4 2.0 105TBP 1.2 7.1 0.22 99.8 5.4 76TBP 5.6 6.5 1.2 98.8 4.9 74Decanol 1.11 3.8 0.30 99.7 3.8 99Decanol 7.5 4.4 1.4 98.6 5.2 1162-Heptanone 1.15 1.7 0.44 99.2 2.6 1552-Heptanone 5.4 2.5 1.7 98.0 3.3 125

a The original term used for this column was ‘‘enrichment factor’’, which expressed as ‘‘the ratio of alcohol/water in the condensed vapor divided by the ratio of alcohol/water in the aqueous dilute solution’’ [71]. This is actually equivalent to the term ‘‘Selectivity’’ as described by Eq. (4). The conventional term ‘‘enrichment factor’’ or‘‘concentration factor’’ of a phase (e.g., organic phase) is usually defined as the solute concentration in this phase divided by the solute concentration in the initial feedsolution.


favorable for phase separation after extraction. In addition, a con-tinuous two-stage ABE fermentation using immobilized cells inte-grated with liquid–liquid extraction was reported for removal ofproduct inhibitions and improvement of the fermentation perfor-mance [86]. With a mixture of oleyl alcohol and decanol as extract-ant, high solvent (ABE) productivity of 2.5 g L�1 h�1 and solventyield of 0.35 g g�1 were obtained.

Glyceryl tributyrate is a water-immiscible solvent capable ofextracting both n-butanol and acetone from aqueous solution[75]. Recently, glyceryl tributyrate was studied for the in situ selec-tive extraction of both alcohols and acetone to enable the simpleintegration of ABE fermentation and chemical catalysis, for thepurpose of reducing the overall process energy consumption [74].Compared to oleyl alcohol, the benchmark solvent, glyceryl tribu-tyrate has much lower boiling point (287–288 �C vs. 330–360 �Cat 1 atm). Thus, the subsequent separation of glyceryl tributyrateand butanol could have less energy demand than the separationof oleyl alcohol and butanol. However, the specific gravity of glyc-eryl tributyrate is very close to 1.0 (1.0340–1.0370 at 25.00 �C),leading to more difficulty in phase separation between the organicand aqueous phases, especially in commercial practice.

Non-ionic surfactants were also investigated as butanol extract-ants. For example, five non-ionic surfactants (Triton X 114, L64,L62, L62LF, and L61) were examined as extractants for butanolrecovery from the fermentation broth in extractive acetone–buta-nol (AB) fermentation using Clostridium pasteurianum. Biocompat-ibility tests using 3% (vol) surfactants showed that L62, L62LF, andL61 did not show inhibition to AB production in 72-h fermentationusing C. pasteurianum, while L64 reduced the AB yield and Triton X114 inhibited the AB production. The results showed that L62 is agood extractant for isolating butanol from the fermentation broth,with the distribution coefficient (KD) of 3–4 for butanol, and it sig-nificantly enhanced the butanol production with a butanol yield of225% higher than the control using 6% (vol) L62 [79].

In addition to the above solvents, butyl acetate, TBP, decanol,and 2-Heptanone, etc. were also tested for extracting isobutanol(i-BuOH) from dilute aqueous solution. The related compositionsof two liquid phases under equilibrium, distribution coefficients,

Table 3Extraction performance of selected solvents at 25 �C [71].

Solvent Initial wt% i-BuOH Extractant/aq. sol. (w/w) i-BuOH wt%

Butyl acetate 0.7 0.5 0.8Decanol 0.8 0.5 1.12-Heptanone 1.0 0.5 1.1TBP 1.4 1.0 1.2

a The extraction yield was calculated by dividing the alcohol amount in the organic p

and selectivities, obtained at 25 �C, are summarized in Table 2.The extraction performances of these solvents at 25 �C for specificconditions (initial isobutanol concentration and solvent to aq. sol.vol. ratio) are shown in Table 3 [71]. Furthermore, Groot et al.[87] examined thirty-six conventional extractants for the distribu-tion coefficient for butanol, the selectivity of butanol over water,and the toxicity towards Clostridia.

2.3.1.2. Biodiesel and oil-derived extractants. Oil derived from bio-mass, or plant oil can be chemically modified into an extractantfor in situ butanol removal from a fermentation broth. Ishizakiet al. [76] studied the removal of product inhibition andenhancement of butanol productivity in ABE fermentation usingmethylated crude palm oil (MCPO) as extractant. With thisextractant, 47% of the total butanol generated was recovered,with butanol concentration up to 20.9 g L�1. Results also showedthat MCPO did not influence the fermentation performance of theorganism used. The glycerides in the oil can be chemically con-verted into a product of C7 to C22 carbons, such as fatty acids,fatty alcohols, fatty amides, fatty acid methyl esters, fatty acidglycol esters, and hydroxylated triglycerides, and mixturesthereof, which is water-immiscible and can be utilized asextractant having a greater distribution coefficient for a productalcohol than the biomass oil [77]. The distribution coefficient(KD), one of the most important parameters for the extractionperformance, of representative corn oil derived extractants forthe isobutanol fermentation broth are summarized in Table 4as compared with conventional commercial solvents of oleylalcohol and 2-butyl-1-octanol [77]. It has been shown that cornoil derived extractants have the distribution coefficients of 2.1–3.2 for isobutanol, representing good candidates as extractant,though the coefficients are lower than those of oleyl alcoholand 2-butyl-1-octanol. The selectivities of these extractants forbutanol, though not shown, should be very high due to theimmiscibility between oil and water.

Being non-toxic to species of clostridia [76,88], biodiesel wasstudied as a biocompatible extractant for in situ separation ofbutanol from fermentation [76,78,88]. It has lower distribution

in org. phase i-BuOH wt% in aq. phase K 0D Extraction yield (%)a

0.3 2.67 600.3 3.67 600.4 2.75 600.2 6.0 85

hase by the initial alcohol amount in the aqueous solution.


coefficient for butanol compared to oleyl alcohol (KD = 0.91 vs. 2.8at the biodiesel/aq. sol. vol. ratio of 1:1), but it is much less costlythan oleyl alcohol. The major advantage of biodiesel as extractantis its capability of good blending with butanol as fuel, thus subse-quent separation of biodiesel–butanol could be avoided resultingin cost savings. With biodiesel as solvent in a single-stage extrac-tion, butanol recovery of 45–51% was obtained at a biodiesel/aque-ous phase volume ratio of 1:1 and an initial butanol concentrationof 11.1 g L�1 and 16.7 g L�1 [78]. Biodiesel has a density of about880 kg m�3, making the phase separation relatively easy. Gasolinecan also be used for in situ extraction of butanol from fermentation.The distribution coefficient K 0D

� �of gasoline for butanol is 0.62–

0.66 at 24 �C [71], which is much lower than that of oleyl alcohol.However, gasoline has a very high selectivity for n-butanol (>600)[71]. Yen and Wang [88] studied the fed-batch ABE fermentationusing C. acetobutylicum and the glucose substrate coupled within situ butanol removal using biodiesel extraction for improvementof productivity. The butanol productivity of 0.295 g L�1 h�1 and themaximum total butanol concentration of 31.44 g L�1 wereobtained at an extractant/broth volume ratio of 1:1, as comparedto the productivity of 0.185 g L�1 h�1 and the butanol concentra-tion of 9.85 g L�1 obtained in the control batch (without the addi-tion of biodiesel).

2.3.1.3. Ionic liquids for butanol extraction. Room temperature ionicliquids (RTILs), the organic salts in the liquid state at or belowroom temperature, are the novel extractants having potential foruse in recovering butanol from dilute solution [80,81]. For instance,Simoni et al. [81] studied the LLE of n-butanol from water usingionic liquids (ILs) as solvents. Experimental results show that someILs have high distribution coefficients and selectivities of 25–300.[Hmim][FAP] shows especially good extraction capability withthe distribution coefficient K 00D

� �of 5 and the selectivity of 300

for 5 wt% n-butanol aqueous mixture. Ha et al. [83] investigatedthe extraction behavior of 11 different imidazolium-based ionicliquids for butanol extraction. The butanol distribution coefficientsof ILs are highly dependent on the hydrophobicity of anions of ILsfollowed by the hydrophobicity of cations of ILs. The hydrophobic-ity of anions studied is in the increasing order of [TfO]�

< [BF4]� < [PF6]� < [NTf2]�. Considering extraction efficiency andselectivity, [NTf2]-based ILs among the tested ILs showed to bethe best extractants for recovery of butanol from aqueous solution.Among them, [OMIM][NTf2] showed the highest selectivity (132),

Table 4Distribution coefficients for isobutanol at initial aqueous solution of 3wt% i-BuOH [77].

Extractant i-BuOH conc. in organic phase (g

2-Butyl-1-octanol 20.12-Butyl-1-octanol 20.0Oleyl alcohol 21.5Oleyl alcohol 13.6Oleic acid 12Oleic acid 3Corn oil fatty acid 16.3Corn oil fatty acid 9.8Corn oil ethylene glycol ester 13.4Corn oil ethylene glycol ester 12.2Corn oil fatty alcohols 18.0Corn oil fatty alcohols 14.5Corn oil fatty amides/acid 12.067% hydroxylated corn oil –28% hydroxylated corn oil –Corn oil fatty acid (COFA) –Corn oil fatty acid methyl esters (FAME) –

Note: Values in the 4th column (directly from the literature [77]) deviate a bit from the rarow). The difference might be the fact that the values of the 3rd and the 2nd columnscalculated using the original data having two or more precisions of the 3rd and the 2nd

with high butanol distribution coefficient (KD = 1.939) and extrac-tion efficiency (74%) at 50 �C, respectively, for aqueous solutionhaving initial butanol concentration of 6 wt%.

Other ILs including 1-decyl-3-methylimidazolium tetracyanob-orate ([DMIM][TCB]) [89,90], 1-decyl 3-methylimidazoliumtris(pentafluoroethyl)trifluorophosphate ([DMIM][FAP]) [89], tri-hexyltetradecylphosphonium tetracyanoborate ([P6,6,6,14][TCB])[84,90], and 1-hexyl-3-methylimidazolium tetracyanoborate[HMIM][TCB] [90] were also studied as extractants for extractingbutanol from dilute aqueous solution. For instance, Domanskaand Królikowski [90] determined the L–L equilibrium data for thethree ternary systems [HMIM][TCB], or [DMIM[TCB], or[P6,6,6,14][TCB] + n-butanol + water at 35 �C and ambient pressure.The NRTL model was well in agreement with the experimentalLLE data. In addition, the RTIL with the longer alkyl chain on thecation showed higher selectivity and distribution coefficient. Thedistribution coefficient and the selectivity of these three ionicliquids are in the same order: [P6,6,6,14][TCB] > [DMIM[TCB] >[HMIM][TCB]. The phosphonium cation shows higher distributioncoefficient and selectivity as compared to imidazolium cation.Davis and Morton [91] studied the ternary liquid–liquid equilib-rium data for two ionic liquid + water + butanol systems. The RTILsused are 1-butyl-3-methylimidazolium bis(trifluoromethylsulfo-nyl)imide ([BMIM][NTf2]) and 1-hexyl-3-methylimidazoliumbis(trifluoromethylsulfonyl)imide ([HMIM][NTf2]). The LLE datashow that [BMIM][NTf2] and [HMIM][NTf2] exhibit high selectivityfor butanol for solutions of low initial (feed) butanol concentra-tions. However, the IL-rich phases contain some water as the hydro-phobic properties of [BMIM][NTf2] and [HMIM][NTf2] are not strongenough. The liquid–liquid equilibrium data for the ternary systembutanol + water + ionic liquid such as 1-ethyl-3-methylimidazo-lium bis(trifluoromethylsulfonyl)imide ([EMIM][NTf2]) at tempera-ture ranging from 281 K to 340 K were also published [92].

The liquid–liquid equilibriums in the binary systems have alsobeen investigated. For instance, n-butanol + ionic liquid such as1-alkyl-3-methylimidazolium hexafluorophosphate ([RnMIM][PF6], where Rn = butyl, pentyl, hexyl, heptyl, and octyl) [93], 1-ethyl-3-methylimidazolium hexafluorophosphate ([EMIM][PF6])at 55–116.5 �C [94], 1-ethyl-3-methyl-imidazolium bis(trifluoro-methylsulfonyl)imide ([C2MIM][NTf2]) at 275–345 K and ambientpressure [95], 1-n-butyl-3-methylimidazolium hexafluorophos-phate ([BMIM][PF6]) [96], and 1-butyl-3-methylimidazoliumhexafluorophosphate ([BMIM][PF6]) [97].

L�1) i-BuOH conc. in aqueous phase (g L�1) KD for i-BuOH

4.8 4.225.0 4.056.4 3.383.6 3.774.6 2.611.2 2.515.9 2.763.9 2.525 2.684 3.076.3 2.875.0 2.884.3 2.81– 3.2– 2.1– 2.8– 1.06

tio of the values in the 3rd and the 2nd columns (e.g., 4.19 instead of 4.22 in the firstare data retaining only one precision, while KD values of the 4th column might be

columns.

Table 5Extraction performance of representative ILs.

IL Initial/eq. BuOH conc. of aq. sol. (wt%) Temp (�C) KD (or K 0D;K00D) S Refs.

[HMIM][FAP] 5 22 5 (K 00D) 300 [81][BMIM][PF6] 6 25 0.74a 21 [83][HMIM][PF6] 6 25 0.97 37 [83][OMIM][PF6] 6 25 1.11 49 [83][BMIM][NTf2] 6 25 1.03 39 [83][HMIM][NTf2] 2 22 6 (K 00D) 90 [81][HMIM][NTf2] 6 25 1.25 66 [83][HOHIMI][NTf2] 5 22 1.5 (K 00D) 40 [81][OMIM][NTf2] 6 25 1.37 79 [83][P6,6,6,14][NTf2] 2 25 1.10 – [102][N1,8,8,8][NTf2] 2 25 1.44 – [102][P6,6,6,14][DCA] 2 25 7.49 – [102][THA][DHSS] 2 25 7.99 – [102][DMIM][FAP] 1 25 0.8 420 [89][DMIM][TCB] 1 25 3.2 100 [89][DMIM][TCB] 2 25 3.27 104 [84][OMIM][TCB] 1 25 3.7 97 [84][P6,6,6,14][TCB] 1 25 2.0 500 [84][P6,6,6,14][Phosph] 0.66Eq. 25 41.8 266 [103]

1.09Eq. 25 46.11 219[TOA MNaph] 0.05Eq. 25 21 (K 0D) 274 [101]Oleyl alcohol 1 25 3.42 (K 0D) 194 [101]Oleyl alcohol 1 25 3.4 208 [84]

Eq. Equilibrium mole percent of aqueous phase.a Data without stated as K 0D or K 00D between parentheses are the values of KD.


Table 5 shows the extraction performance of representativeionic liquids in terms of distribution coefficient and selectivity.Density, viscosity and interfacial tension are also important datafor choosing a suitable extractant. The density, viscosity and inter-facial tension for ILs (only those with available data) in Table 5 aretabulated in Table 6. Viscosities and surface tensions of[BMIM][PF6], [HMIM][PF6], [OMIM][PF6], and [BMIM][NTf2] werealso reported elsewhere [98]. The surface tension of ionic liquidsand ionic liquid solutions were recently reviewed [99].

Extractants with good distribution coefficient (KD,BuOH > 2) andgood selectivity (ideally >100), immiscibility with water, andnon-toxicity to the microorganism are considered good candidatesfor butanol recovery [84]. From Table 5 it can be seen that[DMIM][TCB], [OMIM][TCB], [P6,6,6,14][TCB], [P6,6,6,14][Phosph], and[TOA MNaph] have good distribution coefficient (KD,BuOH > 2) andgood selectivity (�100 or >100). Note that given K 0D ¼ 21 for[TOA MNaph], its KD,BuOH is estimated to be larger than 2 basedon Eq. (3). [HMIM][FAP] is not included in the ‘‘good’’ list as itsKD cannot be derived from its K 0D (=5), thus it is not sure whetherits KD is larger or smaller than 2. From Table 6, it is found thatthe interfacial tension of ILs are large enough for coalescence ofemulsions and phase separation; The densities of ILs are often sig-nificantly different from that of water, for example, the densities of[BMIM][PF6] and [OMIM][PF6] are 1369 kg m�3 and 1235 kg m�3,respectively [83]. This makes the aqueous and organic phases eas-ily separated after extraction. However, most ILs except those con-taining the anion [TCB]� have very high viscosity compared toconventional extractants (e.g., OA), which results in low masstransfer and requires high energy consumption for mixing and agi-tation for extraction. Note that the presence of water in ILs signif-icantly changes their viscosities, especially for the lesshydrophobic ILs where the water content is significantly higher[100]. However, even the water-saturated ILs are still much moreviscous than the conventional solvents. To improve the mass trans-fer and extraction performance through decreasing the viscosity ofILs, extraction temperature can be increased and/or appropriatediluents can be added with the extractants. For the extractantswith good distribution coefficient and selectivity as just men-tioned, [DMIM][TCB], [OMIM][TCB] and [P6,6,6,14][TCB] should have

much lower viscosity than those ILs that do not consist of the[TCB]� anion. This can be seen by comparing the viscosity between[HMIM][TCB] and [HMIM][PF6]. The ILs containing the [TCB]�

anion are often called low-viscosity family of ILs. Hence,[DMIM][TCB], [OMIM][TCB] and [P6,6,6,14][TCB] favor mass transferand requires low energy consumption. The densities of them werenot found yet. [P6,6,6,14][Phosph] has a density of slightly more than10% lower than that of water. This is still significantly far awayfrom the good density of extractants that should be ±20% higheror lower than that of aqueous phase. It is worth noting that[P6,6,6,14][Phosph] and [TOA MNaph] have very high selectivityand the best distribution coefficients (from a few to 20 timeshigher than the other ILs listed in Table 5). The conceptual designand simulation showed that the butanol extraction with [TOAM-Naph] requires 73% less energy in comparison with conventionaldistillation (5.65 MJ (kg BuOH)�1 vs. 21.3 MJ kg�1) [101].

Ionic liquids have many advantages as extractants for butanolseparation, including: non-volatility or negligible vapor pressure,which can translate into less energy demand, extremely low sol-vent loss, and much less capital cost in the subsequent ILs-butanolseparation process. Their physical–chemical properties can betailored or tuned for better separation performance [59,81].Especially, its non-volatility makes it easily separated from butanolafter extraction by simply using evaporation or a one-stage V–Lequilibrium operation (flash distillation), resulting in less energyconsumption, compared to ordinary multi-stage distillation.

The binary mixtures of butanol and ILs are non-ideal solutions.The activity coefficients and the VLE of organic solvents in the(organic solvent + IL) binary mixtures can be predicted byCOSMO-RS model. The energy consumptions of the IL recoveryvia evaporation mainly consist of the energy required for the fluidheating and vaporization. In addition, the operating conditions forthe binary mixtures of organic solvents and ILs were determined[121].

With a butanol–water model solution (2.0 wt% butanol, 0.5 wt%ethanol, 0.7 wt% acetone, 0.5 wt% acetic acid and 0.2 wt% butyricacid), the KD values measured at 25 �C were found in the order[110]: [THA][DHSS] (7.99) > [P6,6,6,14][DCA] (7.39) > OA (3.32) >[N1,8,8,8][NTf2] (1.44) > [P6,6,6,14][NTf2] (1.10). No selectivity data

Table 6Density, viscosity and interfacial tension of representative ILs.

IL q (kg m�3) at 25 �C l (mPa s) at 25 �C c (mN m�1), 25 �C Mutual solubilities (wt%) of IL and water at 25 �C

ILs in H2O H2O in ILs

[BMIM][PF6] 1369 [83] 207.2 [82]; 204@30a [104] 43.52@30 [105] 1.70 [83] 2.69 [83][HMIM][PF6] 1301 [83] 363@30 [104] 38.35@30 [105] 0.60 [83] 2.25 [83][OMIM][PF6] 1235 [83] 477.4 [82], 452@30 [104] 34.60@30 [105] 0.15 [83] 1.75 [83][BMIM][NTf2] 1436 [83] 52 [106] 33.09@30 [105] 0.51 [83] 1.75 [83][HMIM][NTf2] 1372 [83] 35.92@30 [107]; 14.49*@30 [108] 0.23 [83] 1.47 [83][OMIM][NTf2] 1325 [83] 22.81*@30 [109];16.47*@30 [108] 0.09 [83] 1.19 [83][P6,6,6,14][NTf2] 1.070 [110] 91/46b@45 [102] 33.08 [111] 0.087 [112][N1,8,8,8][NTf2] 1.105 [110] 83/42@45 [102] 27.93 [113][P6,6,6,14][DCA] 0.898 [110] 123/37@45 [102] 35.04 [111] 0.510 [112] 3.31 � 10�2 [114][THA][DHSS] 0.975 [110] 8.60 � 10�5 mol L [115] 5 [116][OMIM][TCB] 38.6@30 [118][P6,6,6,14][Phosph] 895 [103] 785.77/97.03@35 [100]Oleyl alcohol 0.855 [110] 28.32 [119]; 26@30 [120] 0.0019 1.14

q = density, l = viscosity, c = ionic liquid–air interfacial tension (i.e., surface tension), or ionic liquid–water interfacial tension (labeled with ⁄).a The data right after @ represent the temperature in �C at which the physical properties were measured. Properties without following @ were all measured at 25 �C.b The data right after slash ‘/’ for the l column are the viscosities of the water-saturated ILs; otherwise, they represent the viscosities of the pure (dry) ILs.

* Data with star (⁄) in c column are ionic liquid–water interfacial tension; otherwise, data without ‘‘⁄’’ are ionic liquid–air interfacial tension (i.e., surface tension).


were reported. OA can be used as a possible RTIL diluent [110] as ithas good distribution coefficient and much lower viscosity than theILs discussed here. The biocompatibility tests shows that amongthese extractants only [P6,6,6,14][NTf2] and OA are both completelybiocompatible with C. beijerinckii at saturation levels but inhibitorywith C. acetobutylicum [110]. In addition, [P6,6,6,14][NTf2] [122],[N1,8,8,8][NTf2] [122] are biocompatible with E. coli.

Stoffers and Górak [123] conducted an experimental and model-ing study on continuous multi-stage extraction of n-butanol fromdilute aqueous mixtures using [HMIM][TCB] in a mixer-settler unitat 35 �C. The NRTL model was found suitable for description of the L–L equilibrium of the ternary system [HMIM][TCB]/n-butanol/water.The continuous separation of n-butanol was confirmed feasible.

2.3.1.4. Biocompatibility or toxicity. It is worth noting that biocom-patibility between extractants and the microorganisms used forABE fermentation must be strictly tested before use in extractivefermentation. Though the extractants discussed above have goodextraction capabilities and have large potentials to use in ABEextractive fermentation for in situ product removal, all of their bio-compatibilities have not been tested yet. Most recently, the toxicityof solvents towards C. beijerinckii was discussed by Yang and Lu[124].

The anions [BF4]� and [PF6]� hydrolyze in aqueous media andthe resulting hydrolytic products are toxic to microorganisms[125]. Thus, ILs consisting of these anions might not be suitablefor in situ extraction of butanol from fermentation broth [83]. Inaddition, [OMIM][PF6] suppresses biological activity in the ABE

Fig. 5. Simplified block diagram of hybrid fe

fermenter if present at saturation level [82]; its concentration levelin the broth must be well controlled under the limit if it is used asextractant. These ILs, though discussed under this section of‘‘Extractive Fermentation’’, should be considered for externalextraction or ex situ product removal.

2.3.2. External solvent extractionExtractive fermentation with in situ product removal, as

described above, may not be suitable for large-scale productiondue to slow mass transfer into organic phase, formation of emul-sions through agitation, cell inhibition by solvent (interface toxic-ity) and loss of cells at interface, and difficult process control, etc.[126]. In addition, as described above, extractants for in situ sepa-ration in extractive fermentation must be biocompatible with theorganisms used. This greatly limits the range of suitable solventsand till now only limited biocompatible solvents have been identi-fied. For this reason, fermentation integrated with external productremoval in an extraction column with a recycle of product-leanbroth (Fig. 5) was proposed [126].

Kraemer et al. [126] used the computer-aided molecular designto screen potential solvents for butanol, and identified mesitylene(i.e., 1,3,5-trimethybenzene) as novel solvent with excellent proper-ties for ABE extraction from the fermentation broth. The compari-sons between mesitylene (measured data) and oleyl alcohol areshown in Table 7 [126]. Specifically, mesitylene has much lower dis-tribution coefficient for butanol than oleyl alcohol at lower temper-atures (25–30 �C), but it has relatively large distribution coefficientfor butanol at higher temperatures (�80 �C). Simulation results

rmentation-external LLE process [126].


showed that, the energy demand of the hybrid extraction-distilla-tion process with mesitylene as extractant is much lower than theextraction with oleyl alcohol (4.8 vs. 13.3 MJ kg�1 butanol).

The major advantages of mesitylene over oleyl alcohol are:much higher selectivity (�6 times), much lower (�40 times) vis-cosity, and much lower (�2 times) boiling point. The major disad-vantage of fermentation followed by external solvent extraction isthe requirement of a very large volume external extractor, whichwould greatly increase the capital cost.

2.3.3. SummaryL–L extraction is a potentially energy-efficient approach for

removal of butanol from the ABE fermentation broths. Novel sol-vents – ILs are especially promising extractants due to their prop-erties such as nonvolatility, tunability of properties such ashydrophobility or affinity to butanol, viscosity, and density. Novelextractants are still needed to be developed with the aid of molec-ular modeling technique and chemically modifying its affinity tobutanol or hydrophobicity. The disadvantages of extraction in situcontinuous recovery of butanol are: potential emulsions, high vis-cosity of ILs, and possible toxicity to the biocatalysts,

2.4. Membrane-assisted solvent extraction (perstraction)

Membrane-assisted solvent extraction, also referred as mem-brane solvent extraction, or perstraction, is a special solventextraction for recovery of butanol from the fermentation broth. Itis the combination of membrane separation and L–L extraction inone unit operation. In the perstraction process, the extractantand the fermentation broth are isolated by a membrane wherebutanol diffuses through the membrane and then extracted bythe extractant, while the other components are retained.

Perstraction can be utilized for recovery and separation of buta-nol from ABE fermentation broth [75,127]. Membrane-assistedextractive (MAE) fermentation of ABE by C. saccharoperbutylaceton-icum N1-4 using a polytetrafluoroethylene (PTFE) membrane and1-dodecanol was studied. The membrane separates the aqueous

Table 7Comparisons between mesitylene and oley alcohol as solvents for butanol extraction.

1,3,5-Trimethybenzene [126] Oleyl alcohol [120]

25 �C 80 �C 30 �C

K 0D;BuOH 0.76 2.2 3.8

K 0D;Acetone 0.43 0.83 0.34

K 0D;EtOH 0.03 0.1 0.28

Selectivity 1650 1970 330Viscosity (mPa s) 0.66 26Boiling point (�C) 165 330–360

Fig. 6. ABE fermenter integra

phase from the organic phase allowing the use of the toxic extract-ant 1-dodecanol with high distribution coefficients in extraction ofbutanol without direct contact and toxifying the microorganism.Compared to conventional batch fermentation, MAE–ABE batchfermentation with 1-dodecanol as an extractant decreased butanolinhibition and increased glucose consumption from 59.4 to86.0 g L�1, and the total butanol production increased from 16.0to 20.1 g L�1. The maximum butanol production rate increasedfrom 0.817 to 0.979 g L�1 h�1. The butanol productivity wasremarkably high with this system, i.e., 78.6 g L�1 h�1 m�2 [128].Qureshi and Maddox [38] studied the ABE fermentation with wheypermeate as substrate and lactose as supplement in a batch reactorusing C. acetobutylicum P262, coupled with the ABE removal byperstraction using oleyl alcohol as extractant. ABE of 98.97 g L�1

was produced from lactose of 227 g L�1 at a yield of 0.44 and pro-ductivity of 0.21 g L�1 h�1. Results show that the coupled fermen-tation–perstraction process is superior to the control batchfermentation where ABE concentration is only 9.34 g L�1.

The major advantages of perstraction are that the extractantstoxic to cells having good extraction performance (high distribu-tion coefficient and selectivity) can be chosen for in situ productextraction as the organic phase and the aqueous phase do not con-tact directly. In situ product removal by perstraction could reduceproduct inhibition, enhance cell growth, increase productivity,and save energy in recovering butanol. However, perstraction alsohas the obvious disadvantage in that the membrane barrier limitsthe extraction rate [38], which results in low butanol productivity.

2.5. Membrane pervaporation

Pervaporation is a separation process where a liquid mixture(feed) is in contact with one side of a membrane and permeate isremoved as a low-pressure vapor from the other side connectedto a vacuum system. Membrane pervaporation is highly selective,efficient, and safe. It is the most promising technology in themolecular-scale liquid/liquid separations existing in biorefinery,petrochemical, pharmaceutical industries [129], and has beenwidely studied for removal of inhibitory products from fermenta-tion broth [59]. Fermentation incorporated with membrane per-vaporation has been regarded as an efficient way to remove orreduce the butanol inhibition in ABE fermentation. A typical fer-mentation–pervaporation system for product removal is shownin Fig. 6. In the fermentation–pervaporation system, ultrafiltration(UF) is used to retain and recycle the microorganisms or cells usedback to the fermenter. The permeates without microorganismsthen enter the hydrophobic pervaporation (PV) unit where theABE solvents pass through the pores of the hydrophobic mem-brane, and then evaporated under the vacuum. The recoveredsolvents are then condensed into liquid for further product

ted with pervaporation.


separation. The aqueous retentate is recycled to the fermenter.Hydrophobic membranes are chosen in separating butanol fromdilute broth by pervaporation as butanol is hydrophobic. In gen-eral, there are three categories of hydrophobic membranes: poly-meric, inorganic, and composite membranes.

2.5.1. Pervaporation with hydrophobic polymeric membranesThe commonly investigated hydrophobic polymeric mem-

branes for separating n-butanol from ABE fermentation brothinclude polyether block amide (PEBA) [40,130–133], polyvinyli-dene fluoride (PVDF) [134], polytetrafluoroethylene (PTFE) [135],polypropylene (PP) [43], poly(1-trimethylsilyl-1-propyne) (PTMSP)[136,137], and polydimethylsiloxane (PDMS) [67,133,138–140].Most of single-polymer membranes have either low separation fac-tors or low permeate fluxes. PTFE and PP membranes allow a largeamount of water pass through the membrane [141], leading to lowselectivity. PTMSP shows best performance based on the permeateflux and separation factor. This is similar to the case of ethanolrecovery from ethanol fermentation broth by pervaporation wherepervaporation using PTMSP membranes shows a distinct advan-tage over PDMS membranes in ethanol removal [142]. PDMS wasalso reported to have good performance and promising potentialdue to its highly hydrophobic properties, good thermal, chemicaland mechanical stability, relatively low price, as well as its abilityto be easily fabricated [143]. A thin PDMS membrane has high fluxbut low selectivity for butanol, while a thick PDMS membraneoffers high selectivity but low flux [141]. Thus, its thickness mustbe appropriate in order to achieve appropriate flux and selectivity.

The performance of membrane pervaporation is basicallyexpressed by solute (butanol) permeate flux, total permeate flux,and separation factor. The separation factor of butanol (over water)for membrane pervaporation is defined as:

b ¼ ðwBuOH=wH2OÞPðwBuOH=wH2OÞF

ð5Þ

where wBuOH and wH2O are concentrations of butanol and water(in wt%) in permeate (P) and feed (F), respectively.

The pervaporation performances of representative hydrophobicpolymeric membranes for butanol removal are shown in Table 8.

To improve the overall pervaporation performance, heteroge-neous polymeric membranes consisting of different polymers, suchas PDMS/PE [143], hydroxyterminated polybutadiene-based polyu-

Table 8Pervaporation performance of polymeric membranes for butanol recovery.

Membrane Thickness(lm)

Butanol in feed(wt% or g L�1)

T (�C) PPerm

PDMS 35PDMS 50 1.0 50/70 67–4PDMS 170 1 78 267–PDMS (GFT 1060) 190 1.0 37 107PDMS (Pervap 4060) (58.9)a 37 1000PDMS (Pervap 4060) 5 50 4Thin PDMS on polyimide (Pervatech) 1 0.58–1.8 37 1230PTMSP 0.3–1.5 25PTMSP 0.3–1.5 70PTMSP 2 53PEBA 50 0.8–1.2F 37 5333PDMS/b-CD 1.0 40Polysiloxane (CMX-GF-010-D) 10 0.2 40 307PDMS copolymers (PERTHESE@ 500–1) 0.2 40PEBA 2.34 37 1000

CMX-GF-010-D: commercial hydrophobic membrane by CELFA AG, Switzerland.GFT 1060: membrane of PDMS only, by Deutsche Carbone AG, Germany.PERTHESE@ 500-1: dimethyl and methyl vinyl siloxane copolymers by Perouse Plastie, FPERVAP 4060: Sulzer Chemtech, Switzerland.Pervatech thin PDMS: product of Pervatech, The Netherlands.

F Fermentation broth; others are simulated solutions.a Concentration of butanol in feed and permeate shown in ‘‘()’’ are in the unit g L�1; o

rethaneurea (HTPB-PU) [149], PDMS–PAN [150], PDMS/polyimide[41], and PDMS/PTFE [151] were studied. Li et al. [143] studied anovel composite membrane consisting of three active layers ofPDMS (Sylgard� 184) and dual support layers of high porositypolyethylene (PE) and high mechanical stiffness perforated metalfor separation of butanol from dilute aqueous solution by pervap-oration. Experiments showed that both total flux and separationfactor increased when placing a layer of hydrophobic PE betweenthe PDMS and the metal support. With the PDMS/PE/Brass supportcomposite membrane, a total flux of 132 g h�1 m�2 and a separa-tion factor of 32 were obtained at 37 �C for the feed solution of2% butanol. Tong et al. [149] prepared hydroxyterminated polybu-tadiene-based polyurethaneurea (HTPB-PU) pervaporation mem-branes for recovering n-butanol and acetone from dilute aqueoussolutions. The results demonstrated that the pervaporation perfor-mance for the ternary mixture was superior to that for the binarymixture because of the permeant–permeant and permeant–mem-brane interactions. High selectivity towards n-butanol and acetonewas obtained. Butanol and acetone were concentrated from 3.0 to43.5 wt% and 1.5 to 11.9 wt% at 40 �C for the simulated ternarymixtures, respectively. For fermentation broth at 45 �C, the separa-tion factors of butanol and acetone were 17.6 and 18.9, respec-tively, with concentration of butanol and acetone increased from1.1 to 16.4 wt% and 0.5 to 8.7 wt%, respectively. However, the totalflux was only 9.7 g m�2h�1. Relatively low permeate flux is thecommon weakness of heterogeneous polymeric membranes. Thisis because this category of membrane made of only polymers usu-ally has to be relatively thicker to achieve the necessary strengthand the ability to endure long time operation. This leads to lowpermeate flux, though the increase in thickness could increasethe butanol separation factor.

Overall, unlike composite membranes described later, heteroge-neous polymeric membranes have not shown significant improve-ment of pervaporation performance over single polymericmembranes. Hence, there is an increasing interest towards thestudy of composite membranes.

2.5.2. Pervaporation with polymeric composite membranesThe performance of polymeric membranes is often limited by

the trade-off between permeability and selectivity [40]. For thisreason, the composite membranes consisting of polymers andinorganic fillers such as ceramic, zeolites (especially silicalite),

(Pa) Total permeateflux (kg m�2 h�1)

Sep. factor, BuOH/acetone/EtOH

Butanol in permeate(wt% or g L�1)

Refs.

0.025–0.035 8.8–18.8 [44]00 0.145/0.350 49.6/48.7 [144]667 0.084 44.9 [138]

0.129 27 [145]>0.4 26.4 (167.1) [67]3.4 39 [146]0.561–0.621 16–20/20–33/6–10 (60.4–131.6) [41]0.06–0.124 51–78 13.2–45.6 [136]0.762–1.03 47–70 12.3–51.6 [136]1.75 115 [137]

–6666 0.161 14 [40]0.800 40 [147]0.330 �38 [148]0.033 �52 [148]0.297(BuOH) 11.57 27.08 [133]

rance.

thers are in wt%.


and silica are gaining more interests and have been widely tested[143,144,148,152–156]. Both polymers and inorganic fillers arehydrophobic, favoring separation of butanol. Polymers are easy touse to fabricate membranes. They are relatively cheap, but theylack sufficient strength, and are subject to aging over a long periodof time, especially at a given temperature. The hydrophobic inor-ganic materials possess strong affinity to butanol and have highstrength and long-term stability under high temperature, but thematerials and the production cost for inorganic membrane areexpensive. With incorporating the advantages of both polymersand inorganic materials, the composite membranes can usuallybe prepared to have high permeate flux and separation factor ata moderate cost. The incorporation of hydrophobic inorganic fillersinto polymer could increase the permeate flux without losing theseparation selectivity.

The polymeric composite membranes commonly studied forpervaporation of butanol from aqueous solutions include PEBA-inorganic composite membranes [40,157,168], PDMS-based com-posite membranes such as PDMS/ceramic [154–156,158], PDMS/zeolites [147,159,160], PDMS/silica [143], PDMS/silicalite-1[138,161,162], and PDMS/PE/perforated alloy metal (i.e., PDMScomposite with the dual support consisting of the porous polyeth-ylene sheet and perforated alloy metal) [163], PTMSP/silica[146,164], and PVA/ceramic [165,166]. Marszałek and Kaminski[167] used commercial flat-sheet membrane PERVAP 4060 for per-vaporative separation of butanol from dilute solution. Pervapora-tion allows to concentrate butanol over 10 times. The permeatefluxes were 523, 1494, and 1063 g m�2h�1 for butanol, ethanoland acetone, respectively. The separation factor for butanol iswithin the range of 4.12–19.48.

Some researchers investigated pervaporation using PEBA-inor-ganic composite membranes for butanol separation [40,157,168].Tan et al. [157] prepared and characterized the composite mem-branes incorporating ZSM-5 zeolite into PEBA. Characterizationby TGA, XRD, and SEM showed that the zeolite could distributewell in the polymer matrix. Using the 5% ZSM-5-PEBA membranefor separation of butanol from aqueous solution, the separationfactor of 22–30 and the permeate flux of 250–550 were obtainedfor 2.5 wt% butanol feed at the flow rate of 50 L h�1 and the feedtemperature of 30–45 �C. The PEBA – carbon nanotubes (CNTs)pervaporation membrane, a special kind of composite membranesusing the novel inorganic material – carbon nanotubes, has greatpotential when integrated in the ABE fermentation [40]. The com-bination of a 5-L fermenter with the pervaporation membrane ofPEBA + CNTs (10%) resulted in a 26% increase in butanol productiv-ity and a 18% increase in its yield compared to using PEBA only.Also, the addition of CNTs to PEBA could greatly enhance the sol-vent permeate flux without decreasing the separation factor, andthe mechanical strength of the membrane, which can be used fora longer operation. Most recently, Paradis et al. [169] developedand characterized the hydrophobic hybrid silica membranes con-sisting of different R-triethoxysilanes (R = C1 to C10 alkyl) and1,2-bis(triethoxysilyl)ethane (BTESE). The affinity of these mem-branes could be tailored from hydrophilic to hydrophobic, specifi-cally, their hydrophobicity increases with increasing the length ofthe ending R-group, on which their separation properties arestrongly dependent. Using the membrane of this type with C10

alkyl ending group for pervaporative separation of n-butanol fromaqueous solution containing 6.8% butanol, permeate of more than40 wt% butanol was obtained. In addition, being stable and non-swelling, the membrane separation factor remained constant ataround 15 over a wide range of temperatures (30–90 �C) and buta-nol feed concentrations (0.5–6.8 wt%). The butanol fluxes obtainedare 0.5–0.6 kg m�2 h�1 at 60 �C and 1.2–1.5 kg m�2 h�1 at 90 �C.

Butanol forms azeotropes with water at 55.5 wt% butanol.When butanol concentration of permeate is between about 7.7

and 79.9 wt%, it separates into two layers of liquid phases: a buta-nol-rich solution and a butanol-thin solution. To recover all thebutanol in the permeate, butanol needs to be separated from thesetwo layers of liquid. The butanol-rich solution can be further sep-arated by distillation, and the butanol-thin solution can be recycledto the membrane system. Another option is to develop membraneshaving very high separator factors that can produce high butanolconcentrations of over 80% in the permeate through pervaporationof dilute solution containing less than 1% butanol. The solution ofhigh butanol concentration can be directly separated further bydistillation. Without the need of recycling butanol-thin solution,the burden of membrane process can be relieved. For example,Negishi et al. [170] developed silicalite membranes coated with0.3% silicone-rubber on a porous support tube for pervaporativeseparation of butanol from 1 wt% butanol solution. Using thismembrane, permeates with high butanol concentration of81.8 wt% were obtained, which could simplify the butanol concen-tration process.

The pervaporation performances of representative hydrophobiccomposite membranes for butanol removal are shown in Table 9.

Membrane fouling is a critical issue and should be paid atten-tion to. It was reported that the silicalite–silicone composite mem-brane exposed to fermentation broth for 120 h was not fouled byfermentation broths [37]. Based on another literature [158], thePV membrane exposed to the fermentation broth for 200 h wasfouled by the active fermentation broth, nevertheless the separa-tion performances were partially recovered by offline membranecleaning, and the solvent productivity was increased to0.252 g L�1 h�1, corresponding to 19% higher compared with thatin situ recovery process without membrane cleaning. To avoid foul-ing of PV membrane, UF is often considered to be used prior to thePV module as seen in Fig. 6.

2.5.3. Pervaporation with supported liquid membranesPolymeric membranes for pervaporative separation of butanol

are usually difficult to offer high values in both permeate fluxand butanol selectivity. As described above, composite membranesconsisting of polymers filled with inorganic fillers such as sillicaliteand ceramic are designed for enhancing the permeability and sep-aration factor [144,152,154]. Supported liquid membranes (SLMs)provide another approach for improving the pervaportion perfor-mance (permeability and separation factor) [172]. Generally, thereare several types of SLMs in terms of different methods for immo-bilizing the hydrophobic organic solvent liquid on the membranesupport [173]: covalent binding of ILs, ion exchange membranes,gelation of ILs, inclusion in polymer matrix, and inclusion by sili-cone coating. An immobilized liquid membrane is the membranesystem in which the immobilized liquid (called the membranephase) is held by capillary forces in the pores of microporous poly-meric or inorganic film (the support for the membrane). Usually ahydrophobic organic solvent is immobilized in a polymeric mem-brane through impregnation to form a SLM used for pervaporativeseparation of butanol [174]. In principle, in a SLM system theimmobilized extractant as membrane phase separates two aque-ous solutions: the feed (donor) and the strip (receiving, acceptor)phase. The compounds are separated from the aqueous feed phaseinto the immobilized extractant in a support diffusing through themembrane phase, and then they are continuously stripped to theother side of the membrane into the strip phase [174].

Supported liquid membranes for separating butanol from dilutefermentation broths have been studied using conventional solventssuch as oleyl alcohol [120,175–177] and recently ionic liquids[102,172,178,179] as extractants. Thongsukmak and Sirkar [177]studied SLM using trioctylamine (TOA) as liquid membrane for sep-arating butanol from aqueous solution. High selectivities of 275, 220and 80 for butanol, acetone and ethanol, respectively, were obtained

Table 9Pervaporation performance of representative composite membranes for butanol recovery.

Membrane Thickness (lm) Butanol infeed (wt%)

T (�C) PPerm (Pa) Total flux a

(kg m�2 h�1)Sep. factorfor BuOH


Refs.

PEBA + CNTs 50 0.8–1.2F 37 5333–6666 0.147 18 [40]PEBA/ZIF-71 (20 wt%) 10–20 1 37 <400 0.520 18.8 18.6 [168]PDMS/ceramic 1.26 37 <400 0.951 16 17.1 [154]PDMS/ceramic 1 40 0.457 26.1 [155]PDMS/ceramic 30 1–2 30 <300 0.36–0.47 11–14 26–30 [156]PDMS/ceramic 0.5F 37 <400 0.338–0.847 5–27 (96.2)b [158]PDMS/ZSM-5 zeolite 1.0 40 1.500 45 [147]PDMS/silicalite-1 96 1 78 267–667 0.235 50 [138]PDMS/silicalite-1 306 1 78 267–667 0.090 97 [138]PDMS/silicalite (Pervap 1070) 29 �0.35 35/65 �667 0.170/0.55 16/10 5.5/3 [152]PDMS/silicalite 0.3 0.2–3 80 5.0–11.2 25–42 [161]PDMS-PE/silica (Sylgard� 184) 2 37 <133 0.132 32 [143]Thin-film silicalite-filled silicone 13–15 1.5–2.0F 70 67–400 0.907 49/14/4 45.7 [144]Thin-film silicalite-filled silicone 19 1.0 50/70 67–400 0.191/0.607 111/93 46.2–42.9 [144]PERVAP-1070 29 1.0 50/70 67–400 0.137/0.344 50/48 [144]HybSi@ membrane 0.5 silica 5 60 0.450–1.40 �11 42 [169]Hydrophobic HybSi@ 2 60 0.5–0.6BuOH 15 [169]Hydrophobic HybSi@ 2 90 1.2–1.5BuOH 15 [169]PTMSP/silica (Vito 1) 2.4 5 50 9.5 104 [146]Silicone-rubber-coated silicalite 1.04 45 0.038 465 82.9 [170]ETES/silicalite-1 2 60 0.100 150 [171]

HybSi@ membrane: hybrid organic (BTESE-based)–inorganic (silica) membranes.PERVAP-1070: silicate-filled composite PDMS membrane, by Sulzer Chemtech, Germany.

F Fermentation broth; others are simulated solutions.a For the ‘‘Total Permeate Flux’’ column, all data are total permeate fluxes except those data with ‘‘BuOH’’ representing butanol permeate flux.b Concentration of butanol in feed and permeate shown in ‘‘()’’ are in the unit g L�1; others are in wt%.


at 54 �C for a feed mixture containing 1.5 wt% butanol, 0.8 wt% ace-tone, and 0.5 wt% ethanol, with the permeate mass fluxes of 11, 5and 1.2 g m�2 h�1 for butanol, acetone and ethanol, respectively.Results also showed that selectivities and fluxes increased consider-ably as the temperature of the feed solution was increased from25 �C. The permeate fluxes could be increased by a factor of 5 byreducing the thickness of the TOA layer in the porous wall of thecoated fibers. For instance, with 50 lm of thin TOA-LM containing20 vol% TOA and hexane (80 vol%) as diluent, the butanol flux of53 g m�2 h�1 and the selectivity of 240 were achieved for a feedsolution of 1.5 wt% butanol. More recently, hydrophobic ILs wereselected for SLMs as ILs are novel and mostly ‘‘green’’ solvents withhigh extraction performance as in the extraction section describedabove. With proper selection of IL components, PM-ILs can offer highfluxes and faster separation [178]. Izák et al. [172] prepared a sup-ported ionic liquid membrane (SILM) and integrated it into an ABEfermentation system using C. acetobutylicum at 37 �C. The SILMwas impregnated with 15 wt% of the ionic liquid, tetrapropylammo-nium tetracyanoborate, and 85 wt% of PDMS. The product was con-tinuously removed at 2.34 g L�1 h�1 by in situ pervaporation usingSILM when aqueous feed concentration was 0.9 wt% butanol. Casconand Choudhari [102] studied the SILMs using hydrophobic ammo-nium- and phosphonium-based RTILs: [P6,6,6,14][NTf2],[N1,8,8,8][NTf2], [P6,6,6,14][DCA], and the mixture of [P6,6,6,14][NTf2]and [P6,6,6,14][DCA] for pervaporative recovery of n-butanol fromaqueous solutions containing 0.5–2.5 wt% butanol at temperaturevarying from 35 to 70 �C. Among these four SLMs, [P6,6,6,14][DCA]shows the best evaporation performance. Some of their results areshown in Table 10.

The immobilized SLMs could not offer long-term stability, i.e.,leaking of the impregnated or immobilized ILs from the supportmembrane because of the large pressure difference in the pervap-oration process [180]. For this reason, polymer inclusion mem-branes (PIMs) and gelled liquid membranes have been studied toimprove SLM stability. PIMs are formed by casting the solutioncontaining a liquid extractant and a base polymer, usually cellulosetriacetate (CTA) or poly (vinyl chloride) (PVC), and plasticizer toform a thin, flexible and stable film (self-supporting PIM), while

gelled liquid membranes are form by liquid-phase gelation in thePVC pores of an SLM. Separation by SLMs combines the solventextraction and stripping processes in a single step, with greatpotential for energy saving, low capital and operating cost [174].Recently, Matsumoto and co-workers [180] investigated the sepa-ration of n-butanol from a ternary mixture (butanol/IPA/water) bypervaporation using polymer inclusion membranes containingionic liquids: Cyphos IL-101, 102, 104, and 109, Aliquat 336 and[Bmim][PF6]. PVDF membranes were used as the support mem-brane. The highest n-butanol flux, total flux, and separation factorobtained by the membrane (30 lm in thickness) with the ratio ofAliquat 336/PVC = 70/30 (w/w) were 26.08 g m�2 h, 1193 g m�2 hand 4.5 respectively, for initial butanol concentration of 5 g L�1 inthe feed. Obviously, the separation factor (4.5) obtained is muchlower compared to the other pervaporation membranes. Plazaet al. [181] prepared a SLM by gellation of the ionic liquid,[bmim][PF6], into the pores of PTFE hollow fibers and used theSLM for separating ABE from its fermentation broth by sweep gaspervaporation. The butanol/ethanol selectivity was improved com-pared to the membrane pervaporation process using the same hol-low fiber support with IL. Heitmann et al. [173] studied SILMsusing [DMIM][TCB], [P6,6,6,14][TCB], and [DMIM][FAP] as liquidmembrane and PEBA as host polymer for pervaporation of butanolfrom aqueous solutions. Two types of SILMs were tested: one withIL immobilized by inclusion between silicone layers, and the otherwith IL immobilized by dissolution in PEBA. Results showed thatthe maximum permeate flux was 560 g m�2 h�1 for the initialbutanol concentration of 5 wt% at pervaporation temperature of37 �C. It was also found that the permeate flux increased withincreasing IL content in the membrane.

2.5.4. Integrated fermentation–membrane pervaporationMembrane pervaporation were widely integrated into batch

fermentation [44,182], fed-batch fermentation [37,152,154,182],and continuous fermentation [39,41,42,182] of carbohydrates toABE, for product removal so as to eliminate or reduce the productinhibition and hence increase the product yield, the ABE concentra-tion, and the productivity. In a fed-batch fermentation–pervapora-

Table 10Pervaporation performance of representative supported liquid membranes for butanol recovery.

Solvents asliquid membrane

d (lm) B/A/E infeed (wt%)

T (�C) PPerm (Pa) Total permeateflux (kg m�2 h�1)

BuOH permeateflux (kg m�2 h�1)

Sep. factorfor B/A/E


Refs.

TOA (Full thickness) 1.5/0.8/0.5 54 426.6 – 11.0/5.0/1.2 275/220/80 67 [177]TOA (20 vol%) + Hexane (diluent) 50 1.5/0.8/0.5 54 426.6 – 53 �240/220/

100– [177]

TOA (30 vol%) + Hexane (diluent) 50 1.7/0.8/0.7F

54 426.6 – 31/9/4 197/111/54 – [177]

[P6,6,6,14][NTf2] 25 1.0 45 <310 77 29 63 37 [102][N1,8,8,8][NTf2] 25 1.0 45 <310 82 22 37 26 [102][P6,6,6,14][DCA] 25 1.0 45 <310 298 126 77 42 [102][P6,6,6,14][NTf2]/[DCA] 25 1.0 45 <310 117 72 57 38 [102][DMIM][TCB] 5 37 0.560 55 [173]


tion system, ultrafiltration membrane with 500,000 molecularweight cut-off was used to retain and recycle the cell culture tothe fermenter while allowing its cell-free permeate enter the per-vaporation membrane for product recovery and concentration, assimilar in Fig. 6 but with additional buffering tank placed betweenthe UF membrane unit and the hydrophobic PV membrane unit.Using the glucose substrate and C. acetobutylicum biocatalyst, andthe hydrophobic silicalite–silicone composite membrane for prod-uct recovery and concentration, the ABE concentration is up to154.97 g L�1 while the ABE yield is in the range of 0.31–0.35. Theaverage selectivity of butanol and acetone were about 40 and95–200, and the average flux was about 86 g m�2 h�1. Table 11summarizes the performance of the integrated FBF-pervaporation(PV) vs. the performance of the control batch fermentation (BF)[37].

Most recently, continuous fermentation systems integratedwith membrane pervaporation have been studied. Fig. 7 showsthe general continuous fermentation coupled with membrane per-vaporation. For instance, a continuous two-stage ABE fermentationusing C. acetobutylicum ATCC 824 was coupled with pervaporationseparation using hydrophobic thin PDMS composite membraneswith 1 lm PDMS separating layer on 200 lm porous polyimidesupport for product (butanol) removal, leading to significantdecrease in the product inhibition in the fermenter. Hence, the glu-cose concentration increased from 60 to 126 g L�1, the productivityincreased from 0.13 to 0.30 g L�1 h�1, and the permeate wasenriched to 57–195 g L�1 total solvents depending on the solventconcentrations in the fermenter [39]. Hecke et al. [41] integrateda continuous two-stage ABE fermentation using C. acetobutylicumATCC 824 with pervaporation where PDMS composite pervapora-tion membranes were directly coupled to the second fermenterresulting in decreased solvent titers. With increasing the initialcarbohydrate concentration in the feed from 60 g L�1 to150 g L�1, the overall productivity was increased from 0.45 g L�1 -h�1 to 1.13 g L�1 h�1, even though productivity decreased signifi-cantly in the first fermenter due to substrate inhibition. In the

Fig. 7. Two-stage ABE fermentation integ

last phase that lasted 200 h, the average flux of 0.621 kg m�2 h�1,the separation factor for butanol of 19.8, and the total solvent con-centration in the permeate of 202 g L�1 were achieved.

Chen et al. [42] integrated an ABE fermentation using C. Acet-obutylicum with PDMS composite membrane pervaporation forin situ butanol recovery into a continuous and closed-circulatingfermentation (CCCF) system. Two cycles of experiment were con-ducted for 274 h and 300 h, respectively. At their respective brothconcentrations of 7.6 and 7.0 g L�1, the total permeate fluxes of thetwo cycles were 784 and 566 g m�2 h�1, respectively, the separa-tion factors are 10 and 7, respectively, and the corresponding per-meate butanol concentrations are 71 and 47 g L�1, respectively.The average cell concentration, the glucose consumption rate, thebutanol productivity, and the butanol production of the secondcycle were 1.68 g L�1, 1.12 g L�1 h�1, 0.205 g L�1 h�1 and61.43 g L�1, respectively.

2.5.5. Energy requirement for membrane pervaporationThe evaporation energy requirement for the PV separation of

butanol from dilute aqueous solutions varies with the concentra-tion of butanol in the feed and the membrane separation factorwas estimated by Vane [60] and is shown in Fig. 8. It can be seenfrom this figure that evaporation energy consumption decreaseswith increasing feed butanol concentration. It also decreases withincreasing butanol separation factor. At the feed butanol concen-tration of 2 wt% the evaporation energy required is about 8.8, 4.9,and 2.8 MJ/kg BuOH (data read from figure) for separation factorof 30, 60, and 120, respectively. These energy consumptions corre-spond, respectively, to 24.4, 13.6, and 7.8 wt% of the butanolenergy content. At this 2wt% feed concentration level, it is sug-gested that the membrane butanol separation factor of PV shouldbe at least 30, and 50 or above would be considered as good buta-nol separation factor. At a higher feed butanol concentration, theevaporation energy required drops significantly, especially for thelow separation factor cases. At 4 wt% feed butanol concentration,for example, the evaporation energy required by PV with a

rated with a pervaporation unit [39].


separation factor of 30 is about the same as the energy required bythe PV with a separation factor of 60 at the feed concentration of2 wt% butanol. This suggests that if the ABE fermentation can beimproved via advanced biotechnologies such as genetic engineer-ing and higher broth concentration, e.g., 4 wt% butanol, is achieved,the hydrophobic membrane with high permeate flux but low sep-aration factor (e.g., �30) could still be energy-efficient in the PVprocess. This also gives a hint that improvement of microorgan-isms is also important for enhancing the separation efficiencyand reducing the separation cost.

Matsumura et al. [175] estimated the energy requirement of acombined OA-LM plus distillation system based on the assump-tions: Fermentation broth with a butanol mass fraction of 0.005at 30 �C is fed to the membrane module at a mass flow rate of24 kg/h. The outlet butanol mass fraction: xB,out = 0.004. Membranepervaporation operates at T = 25 and P = 0.133 kPa. The averagebutanol composition of the permeate was 0.439. The condensationtemperature is �13.0 �C. Assuming that 98% vapor permeate wascondensed. The condensed permeate was then fed to a distillationcolumn for purification to the product purity of 99.9%. The sum ofthe energies supplied to the cold trap and vacuum pump, and thespecific energy requirement was 4.67 MJ/kg of butanol, while theenergy requirement for butanol purification was 2.75 MJ/kg ofbutanol. Thus, the overall energy requirement in the combinedOA-LM + distillation system was 7.42 MJ/kg of butanol.

Most recently, Qureshi et al. [183] carried out an economic eval-uation of biological conversion of wheat straw to butanol. Theresults showed that the use of a membrane recovery processincluding membrane pervaporation for butanol removal fromABE fermentation broth significantly reduced the butanol produc-tion cost compared to the distillative recovery process. Negishiet al. [170] estimated the energy requirement for pure butanol pro-duction from 1 wt% butanol solution using silicone-rubber coatedsilicalite membranes. The resulting permeates contained a veryhigh butanol concentration of 81.8 wt%. The energy requirementestimated was 4.3 MJ kg�1 n-butanol, mainly consisting of theenergy required for concentrating from 1 wt% butanol to81.8 wt% butanol by pervaporation using the silicone-rubbercoated silicalite membrane and the energy required for the subse-

Fig. 8. Evaporation energy requirement as a function of separation factor andbutanol feed concentration for PV separation of butanol from water for a fixedbutanol recovery of 80% (permission from John Wiley and Sons Ltd.) [60].

quent product dehydration using a hydrophilic PV membrane. Thisis only about 13% of the energy content of butanol.

2.5.6. Advantages and disadvantages of membrane pervaporation, andsummary

Membrane pervaporation have many advantages [129]:

(1) high separation efficiency in terms of both permeate fluxand separation factor, since separation mechanism is basedon the differential transport/diffusion of permeates throughthe pores of membrane, instead of the vapor–liquidequilibrium;

(2) high energy-efficiency due to the fact that only the perme-ates need to be evaporated and consume the latent heat.Energy efficiency is defined as the ratio of energy gained toenergy consumed, i.e., the ratio of energy content in per unitweight of butanol to energy consumption for producing unitweight of butanol;

(3) no harmful effects on the microorganisms used becausethere is no need for adding any other components;

(4) good flexibility, simplicity in scaling-up, operation, andcontrol.

The major issues of membrane pervaporation are associatedwith the potential membrane fouling especially when dealing withreal fermentation broths, the long-term stability for polymeric andcomposite membranes, as well as the membrane preparation cost.Thus, it is required to study the long-term fouling characteristics offermentation broth on PV membranes. In addition, there are manyfactors influencing the pervaporation separation performance: (1)factors of membrane itself such as thickness of active layer, mem-brane structure, membrane pore size, filler’s adsorption capacityfor butanol, particles of filled inorganic materials of compositemembranes, and membrane surface modification (2) operation fac-tors such as feed temperature and concentration, and pressure onpermeate side. Therefore, further work on optimization of mem-brane preparation and pervaporation operating conditions are stillneeded.

Earlier critical reviews provide additional discussions on mem-brane pervaporation [184,185].

In summary, even though hydrophobic PV membrane has notyet been utilized in industry as hydrophilic PV membrane, it hasgreat potential in efficient separation of butanol from ABE fermen-tation broth and is a very promising technology in biorefineries.

2.6. Other membrane technologies

2.6.1. Membrane distillationMembrane distillation (MD) is a separation process involving

evaporation and transportation of volatile components (vapor)through a porous hydrophobic membrane based on a tempera-ture-induced vapor pressure difference [186–188]. As the volatilecompounds evaporate through the pores of MD membrane, theseparation mechanism of MD is based on the VLE of the liquid mix-ture. There are four types of MD: direct contact membrane distilla-tion (DCMD), air gap membrane distillation (AGMD), sweeping gasmembrane distillation (SGMD), and vacuum membrane distillation

Table 11Performance of integrated fermentation–pervaporation and the control BF [37].

Control BF Integrated FBF-PV

Total ABE conc. (g L�1) 19.2 155.0ABE yield 0.29 0.31–0.35Glucose conc. in feed (g L�1) 70.30 700.0

Feed, TF1

HeatExchanger

TF2 TF3

Retentate , TR

Permeate

CondensingSheet

Membrane

Liqu

id fe

ed

Satu

rate

d va

por

Liqu

id p

erm

eate

Fig. 9. TPV concept [191].


(VMD) [59,186]. Among these types, AGMD and VMD are suitablefor removal of volatile components from an aqueous solution. Sep-aration of ethanol using MD (mainly VMD) from dilute ethanolaqueous solutions were previously reviewed elsewhere [59,60].Besides, a general comprehensive review on membrane distillationwas recently presented [189]. Literature on recovery of butanolfrom aqueous solution by VMD is rare. Banat and Al-Shannag[187] studied the separation of butanol from ABE aqueous solu-tions using AGMD by mathematical simulation based on the mul-ticomponent Stefan–Maxwell model. The effects on the AGMDperformances (permeate flux and concentration factor) by theoperating conditions such as coolant temperature, ABE concentra-tions, and air–gap width were predicted. MD was claimed to be acost and energy efficient process for separation of aqueous mix-tures, which could use low-grade energy such as industrial residualheat and waste heat, and alternative energy sources such as solarand geothermal energy [186,190]. However, this simulation hasnot yet been verified by experimental studies. Also, since MD isbased on VLE and butanol is less volatile (or has low partial pres-sure), the MD technique is not applicable in separating butanolfrom water. It also has several limitations including MD membraneand module design, membrane pore wetting, low permeate flowrate and flux decay, as well as uncertain energetic and economiccosts [190]. In addition, the porous membrane of VMD cannot offersignificant selectivity advantages over what is offered by a singlestage of VLE. Thus, VMD is not a very attractive process comparedto pervaporation [60].

2.6.2. ThermopervaporationThermopervaporation (TPV) is a novel concept of pervaporation

with internal heat recovery [191]. TPV uses similar nonporousmembrane to that of conventional PV, but the former operatesunder the atmospheric pressure while the latter usually underthe vacuum condition. On the other hand, the TPV system is similarto the MD system in that both operates under atmospheric pres-sure and donot require the vacuum system, with difference in thata nonporous membrane is used for the former while a porousmembrane is use for the latter. In the TPV process shown inFig. 9, the feed enters the condenser of the module at a tempera-ture TF1, where it is heated by the heat of condensation of pervapo-rated vapors to a temperature TF2 (heat recovery fromcondensation heat). The feed is further heated to TF3 in an externalheat exchanger using additional heat, and then enters the feed sideof the membrane module where a fraction of feed permeates themembrane as a vapor and condenses on the condenser plate whilethe remaining feed leave as retentate at a temperature TR. As thecondensation occurs inside the membrane module with a smalldistance from the membrane surface on the permeate side to thecondensing sheet and the heat resistance in the condenser couldbe neglected, the internal heat recovery is maximized. As a proofof principle, heat recovery was up to 30% for ethanol separationusing TPV based on the literature [191]. Till now, heat recoveryfor butanol separation, which is equally important, has not yetbeen reported. In addition, low grade heat (80–100 �C) can be usedfor further heating the feed in the heat exchanger shown in Fig. 9 tothe required pervaporation temperature.

Borisov et al. [192] experimentally and theoretically studied theselective TPV of butanol from dilute aqueous solution through ahydrophobic PTMSP membrane in a plate-and-frame module withan air gap, as seen in Fig. 10. In the TPV process, feed of initial fer-mentation broth at 30–40 �C was heated to 40–75 �C in a heatexchanger using low grade heat, and then entered the membranemodule. A fraction of liquid permeated and pervaporated throughthe membrane. The resulting vapor on the membrane permeateside (air–gap channel) was condensed on the condensing sheet(cooling wall) by the cooling liquid (�15 �C) on the other side into

liquid permeate. Note that there exists a liquid film (not shown inFigs. 9 and 10) between the liquid phase on the condensing sheetand the vapor phase.

The effects of feed temperature, coolant temperature, initialfeed concentration, and membrane thickness on the permeate fluxand the permeate composition were observed. Results showed thatthe permeate flux of the TPV is not inferior to that of vacuum per-vaporation at condensation temperatures of 0.5–15.0 �C [192].With the TPV membrane of 19 lm in thickness operated at the feedtemperature (or PV temperature) of 65 �C, coolant temperature of15 �C, and under atmospheric pressure inside the membrane space,increasing the butanol feed concentration from 2 wt% to 5.5 wt%,the separation factor increased from 25 to 75, the total permeateflux increased from 0.18 to 0.55 kg m�2 h�1, and the concentrationof butanol in permeate from 33 to 60 wt% [192]. This means thathigher concentration of feed broths resulted from the future suc-cessful modification of microorganism could significantly improvethe separation performance of TPV. This also means the modifica-tion of microorganism is equally important.

In brief, TPV can operate with low grade heat, e.g., at 80–100 �Cfor separation of butanol from ABE fermentation. TPV configuredwith internal condenser can efficiently recover the condensationheat of pervaporated vapors for increasing the feed temperatureto the pervaporation temperature. In addition, it has a great poten-tial for commercial uses [191], and as it is a novel technology, moreresearch work are required for further study of the TPV process.

2.6.3. Reverse osmosisReverse osmosis (RO), a relatively mature technology, is widely

used for desalination and water purification [193]. RO can also beused to concentrate aqueous butanol solutions. For instance, Itoet al. [194] recently invented a patented butanol separation andpurification process using RO technology to concentrate the fer-mentation broth. The process consists of three steps: nanofiltration(NF) of aqueous butanol solution by a butanol–permeable mem-brane to remove impurities and get a butanol-containing solutionfrom the permeate side; concentration of the butanol permeateusing RO, followed by phase separation of the concentrated buta-nol solution into a butanol phase and aqueous phase; separationof butanol from the butanol phase with distillation. The NF mem-brane used is a composite membrane with a polyamide functional

Heat Exchanger

Coolingfluid

Retentate, TR

Permeate

CondensingSheet

Membrane

Liqu

id fe

ed

Satu

rate

d va

por

Liqu

id p

erm

eate

Feed

Fig. 10. The plate-and-frame flow-through TPV membrane module.


layer placed on a support made of a porous membrane or non-woven fabric. It has high pressure resistance, high permeabilityto water and high solute removal performance. In the RO stepthe butanol concentration in the concentrated solution is not lessthan the butanol saturation solubility of 8 wt%. As the RO processoperates at high pressure, it should cost more energy than thehydrophobic pervaporation process. However, the former is amature technology while the latter has not yet been commerciallyapplied. It would be interesting and important to estimate the effi-ciency and the cost for the RO process to see if the RO process isfeasible for concentrating the butanol solution.

Fig. 11. Integrated fermentation–adsorption process [196].

2.7. Adsorption

Adsorption technology, specifically molecular sieve adsorption,has been widely used in dehydration of bioethanol for productionof anhydrous fuel-grade ethanol in corn-to-ethanol plants. Adsorp-tion also has the potential to be used for removal of butanol fromABE fermentation broth. The fundamentals of adsorption can referto the recent work of Venkatesan [195]. A typical integrated fer-mentation–adsorption process is shown in Fig. 11. In this process,broth liquor enters ultrafiltration (UF) membrane unit where cellsare retained and recycled to the fermenter. The cell-free permeatesthen enter the adsorption column containing adsorbents to adsorbABE. The ABE lean liquid is recycled to the fermenter. The use of UFmembrane for cells retaining and recycling prior to the adsorptioncolumns can not only avoid the possible fouling of adsorbent bycells, but also reduce cells loss and retain high cell concentrationin the fermenter.

As butanol is hydrophobic, based on the similarity rule hydro-phobic adsorbents are utilized for separation of butanol from dilutesolution by adsorption. Compared to ethanol, butanol is morehydrophobic due to its longer carbon chain. Hence, butanol is abetter suited solute for product recovery using hydrophobic resins.Three types of hydrophobic adsorbents have been explored forbutanol recovery, including activated carbon [197–199], zeolitesor molecular sieve [198–203], composite such as calixarene-basedadsorbents [204], and polymer resins such as XAD-4, XAD-16[205], polyvinylpyridine [36,206], poly(styrene-co-divinylbenzene)(PS-DVB)-derived resins Dowex� Optipore L-493 and SD-2, Dow-ex� M43, and Diaion� HP-20 [207], polystyrene diethylbenzene-

derived polymeric resins H-511 and KA-I [208–210], and polyam-ide-derived resin XD-41 [208].

Hydrophobic polymeric resins have been studied for adsorptionof butanol by a number of researchers [36,71,197,204–213]. Poly-meric resins including XAD-2, XAD-4, XAD-7, XAD-8, XAD-16,and Bonopore, the copolymer of divinylbenzene and styrene arehydrophobic, non-ionic, and cross-linked polymeric adsorbents.Experimental studies on adsorption of butanol in dilute modelsolutions show that the butanol adsorption capacity of resinsXAD-2, XAD-7, XAD-8, and Bonopore are far below 100 mg butanol(g adsorbent)�1 [197,205,214], while the maximum equilibriumadsorption capacities of XAD-4 is 0.1 g butanol g�1 adsorbent[197]. Recently, Evanko et al. used Amberlite� XAD-16 to extractbutanol from an aqueous solution. The adsorption capacity ofAmberlite�, XAD16 is dependent on its macromolecular structure,high surface area and the aromatic nature of its surface. The exper-iments demonstrated high efficiency of butanol enrichment withdistribution coefficients ranging from 10 to 35 at 25 �C. The distri-bution coefficient for butanol adsorption is defined as the ratio ofconcentration of butanol adsorbed in the resin to the concentrationof butanol in the aqueous solution [71]. Nielsen and Prather [207]compared and screened 20 commercially available polymeric res-ins derived from poly(styrene-co-DVB), poly(ester), poly(acry-lates), poly(amide), polyurethane, poly(styrene-butadiene), andpoly(ethylene-co-vinyl acetate) at initial n-butanol concentrationof 20 g L�1 in the aqueous solution. Three different PS-DVB-derivedresins Dowex� Optipore L-493 and SD-2, and Diaion� HP-20, aswell as poly(ethylene-co-vinyl acetate) possess high n-butanolaffinity and have high adsorption capacity for n-butanol. Theiradsorption capacities were strongly limited or controlled by theirspecific surface area. The recovery of n-butanol from these resinsby heating is efficient and economical [207]. About 95% of theadsorbates on the adsorbent Dowex� Optipore L-493 can be recov-ered using 140 �C steam at the steam-to-adsorbent mass ratio of 1[215]. The other resins tested have low adsorption capacity (wellbelow 100 mg butanol (g adsorbent)�1) [207].

For adsorption of butanol from fermentation broth, XAD-16showed the highest adsorption capacity of 75 mg g�1 among testedXAD resins [205,211]. Yang et al. [36] used polyvinylpyridine asadsorbent for in situ removal of products from the acetone-butanolbatch fermentation. With the integration of fermentation-in situadsorption, the total final product concentration increased by54% and the productivity by 130%, compared to batch fermentationcontrol without in situ adsorption.

In general, those resins with non-polar monomeric structureand high specific surface area provide the highest loading of buta-nol [212]. For instance, PS-DVB derived resins that are comprisedof non-polar monomeric units, exhibit the highest butanol affinity.Those resins with the highest surface area, such as Dowex�

Optipore L-493 and SD-2, provided the greatest equilibrium


distribution coefficients for n-butanol between resin-aqueoussolution as the total number of sites available for interaction withsolute molecules increases with increasing the specific surface areaof the resin [207].

Activated carbons (Norit-Row 0.8, AC F600 and AC F400) weretested and found to have very high adsorption capacity in therange of 150–280 mg g�1 for butanol from dilute solution[197,198,216]. However, the recovery of butanol from the adsor-bent by desorption is incomplete [214]. In general, the strongerthe adsorption, the more difficult the desorption. Thus, activatedcarbons are the most difficult to be regenerated.

Zeolites also possess high adsorption capacity for butanoladsorption from dilute solutions [138,198,202,217–219]. It wasfound that ZSM-5 and NaY had the butanol adsorption capacitiesof 155 and 116 mg g�1, respectively, at the same concentrationequilibrium [198]. Saravanan et al. [202] examined two commer-cial zeolites ZSM-5 with high Si/Al ratio, CBV28014, and CBV901(faujasite), as adsorbents for recovering butanol from ABE fermen-tation broth. Results showed that these two zeolites adsorbed n-butanol quickly and almost completely from aqueous solutionswith around 1% butanol. Their adsorption capacities were about0.12 g butanol g�1 zeolite, and remained constant until the equilib-rium butanol concentration as low as 0.04 wt% in aqueous solution.Desorption was conducted by carefully choosing the temperatureprogram, i.e., initial stage of desorption was conducted at lowertemperature to firstly remove the more volatile componentsincluding water, acetone, and ethanol, followed by butanol desorp-tion at higher temperature to obtain a high concentration of buta-nol up to 84.3%, which corresponds to a concentration factor of 65.The major disadvantage of this adsorbent is that only 80% butanolrecovery was obtained at 150 �C. At higher temperature butanolcould only be desorbed as butene. More recently, Oudshoornet al. [220] studied the desorption of water and butanol fromCBV28014 and CBV901. The results show that the desorption heatrequired for CBV901 are 2440 J g�1 of water and 1080 J g�1 of buta-nol, while the desorption heat required for CBV28014 are2730 J g�1 (water) and 1160 J g�1 (butanol); CBV28014 showedless water adsorption than CBV901; the desorption rate of butanolfrom CBV28014 was significantly slower than from CBV901, and acatalytic reaction, most probably butanol dehydration, occurs atthe desorption temperature of �200 �C for CBV28014.

Metal–organic framework ZIF-8 is a promising adsorbent forbutanol recovery from the fermentation broth [199,221]. Remiet al. [199] compared the butanol adsorption performances of themetal–organic framework ZIF-8, silicalite zeolite, and active carbonfor recovery of butanol from dilute mixtures by experiments. It wasfound that ZIF-8 has a high adsorption capacity for n-butanol in anaqueous solution, a high selectivity of n-butanol over the by-prod-ucts, and easy desorption ability. In addition, SAPO-34 was testedto remove water and ethanol from n-butanol for purification ofbiobutanol by adsorption. SAPO-34 showed a high affinity andadsorption capacity for water and ethanol, while butanol wasexcluded from the pores of SAPO-34. Thus, the combination ofZIF-8 and SAPO-34 is very attractive for the recovery and purifica-tion of biobutanol by adsorption. Ikegami et al. studied the adsorp-tion properties of butanol on silicalite powder under variousconditions. It was found that the adsorbed amounts of ABE at equi-librium per gram of silicalite from aqueous solutions of binary mix-tures at 30 �C increased in the order: ethanol (95 mg) < acetone(100 mg) < n-butanol (120 mg) [219]. The same order of adsorp-tion capacity of silicalite-1 for ethanol, acetone and butanol wasfound elsewhere [144]. The desorption of butanol from silicaliteoccurred efficiently at 78 �C and 1–3 Torr [138]. As the hydropho-bicity of zeolites increases with increasing the ratio of SiO2/Al2O3,sililcalite-1, the special zeolites having no aluminum, is the mosthydrophobic zeolite. Therefore, sililcalite shows good adsorption

for butanol, with the butanol adsorption capacity of 85–125 mgL�1 for dilute aqueous solution [138,217,218]. With silicalite asadsorbent, high butanol concentration of up to 98% (by weight)was achieved from dilute aqueous solutions (0.5–2% by weight)after desorption [222]. This could save significant energy consump-tion in the further separation and purification costs. The desorptionfor ABE recovery and silicalite regeneration was conducted bysequential heating, i.e., heating to 40 �C to firstly desorb andremove most of water from silicalite, followed by heating to150 �C to recover butanol [214,217]. In addition, desorption at78 �C could recover 100%, 95.5%, and 80% of butanol, acetone,and ethanol, respectively, of the absorbed amount [223]. Thus, heattreatment at 150 �C for silicalite regeneration was suggested [214].The column dynamics of an adsorption–drying–desorption (ADD)process for butanol recovery from aqueous solutions using silica-lite pellets as adsorbent was experimentally studied at the dryingtemperature of 50–70 �C and the desorption temperature of 130–150 �C. Also, the models of adsorption, drying, and desorption wereestablished to simulate the ADD process with heat integration withthe butanol fermentation process for heat recovery from the efflu-ent of the desorption process (to pretreat the feed streams to thefermenter, and to dry the solid residues of fermentation [222]. Sim-ulation showed that the proposed process can be an energy-effi-cient alternative, with high butanol recovery (95%) and lowenergy requirement (3.4 MJ kg�1).

Unlike materials for membrane pervaporation, composites asadsorbents were rarely studied. Thompson et al. [204] synthesizedgrafted calixarenes as adsorbents, characterized them, and investi-gated their adsorption performances for butanol isolation fromdilute solutions. The adsorbents are composed of hydrophobic,cavity-containing calixarenes covalently bound directly to porous,hydrophilic silica supports through a Si linker atom rather than aflexible organic linker. The butanol adsorption capacity of 0.119 gbutanol g�1 adsorbent was obtained at equilibrium concentrationbelow 8.9 g L�1. As a representative adsorbent of this type, tert-butylcalix(4)arene was explored. Butanol could be desorbed asgas phase from tert-butylcalix(4)arene at 150 �C, which is wellbelow the stability temperature of calixarenes (<300 �C).

Table 12 below summarizes the major properties (surface areaand butanol adsorption capacity) of typical adsorbents for butanolremoval by adsorption (usually at room temperature). Additionalproperties see the earlier review by Qureshi et al. [214].

Integration of adsorption with ABE batch, or fed-batch fermen-tation can significantly improve the sugar substrate concentration,ABE yield and productivity. Yang and Tsao [196] made compari-sons between the integrated fermentation–adsorption for productrecovery system, or simply called in situ adsorptive fermentation,with the batch fermentation control without in situ product recov-ery, as shown in Table 13. Compared with the batch fermentation(BF) control without the simultaneous product removal, the in situadsorptive batch process has 130% and 54% higher productivityand solvent concentration, respectively, while the in situ adsorptivefed-batch (FBF) process has 323% and 145% higher productivity andsolvent concentration, respectively.

Comparison of different types of adsorbents in terms of butanoladsorption property Abdehagh et al. [198] compared two activatedcarbons (AC F600 and AC F400) and two zeolites (ZSM-5 and NaY)for butanol adsorption from dilute solutions. It was found that ACF600 and AC F400 showed at least two times faster adsorptionkinetics than zeolites. From Table 12, by rough comparison it isfound that among four categories of adsorbents (activated carbons,zeolites, composites, and polymeric resins), activated carbons ingeneral have the highest butanol adsorption capacity. Zeolites havethe second highest butanol adsorption capacity, which are usuallyabove 100 mg g�1. Polymeric resins have a wide range of butanoladsorption capacities between different resins. PS-DVB, PES-DVB,

Table 12Major properties of typical adsorbents for butanol.

Adsorbent Matrix Surface area(m2 g�1)

BuOH adsorp. capacity(mg g�1)

Butanol conc. in aq. sol(g L�1)

Refs.

Activatedcarbon

Norit-ROW 0.8 Activated carbon 252 [197]AC 400 Activated carbon 1090 283 15EBS [198]AC 400 Activated carbon 1090 258 10IBS [216]AC F600 Activated carbon 710 150 15EBS [198]AC F600 Activated carbon 710 149 10IBS [216]

Zeolites ZSM-5 155 15EBS [198]ZSM-5 425 139 10IBS [216]

CBV28014 ZSM-5 (MFI) 400 120 10IQS [202]CBV901 Y (FAU) 700 120 10IQS [202]

MFI 120 20IQS [224]NaY 116 15EBS [198]NaY 107 10IBS [216]Silicalite 120 [219]Silicalite 85–97 [217]Silicalite 64–85 11.7–16.8BF [225]Silicalite 400 103 10IBS [216]

Composite Calixarene–silica 119 [204]

Polymericresins

Dowex� OptiporeL-493

PS-DVB 1100 175 20IBS [207]

Dowex� OptiporeL-493

PS-DVB 1100 �50–100 2–8EBS [215]

Dowex� OptiporeSD-2

PS-DVB 800 152 20IBS [207]

PES-DVB 650 �110 20IBS [207]Diaion HP-20 PS-DVB 500 �118 20IBS [207]KA-I resin Polystyrene diethylbenzene 850–900 �100–304 4–10EBS [208,209]H-511 Polystyrene diethylbenzene 1000–1100 �100–140 4–10EBS [208]XD-41 Cross-linked polyamide 549–620 �60–90 4–10EBS [208]Amberlite XAD-16 Macroreticular aliphatic cross-linked

polymer900a 75 9.2BF [211]

Amberlite XAD-4 Macroreticular cross-linked aromaticpolymer

725a 27 11.7–16.8BF [205]

EBS: equilibrium butanol concentration for binary solutions of butanol–water.IBS: initial butanol concentration for binary solutions of butanol–water.IQS: initial (not equilibrium) butanol concentration in quaternary (ABE and H2O) model solution.BF: batch fermentation broth.

a As available and reported by the supplier.

Table 13Comparison of performances between integrated batch/fed-batch fermentation–adsorption processes and the control batchfermentation (BF) [196].

BF control Integrated BF–adsorption Integrated FBF–adsorption

Total glucose used (g) 43.8 73.3 1198.5Total ABE produced (g) 13.5 23.2 387.3Productivity (g L�1 h�1) 0.40 0.92 1.69ABE yield (g g�1) 0.31 0.32 0.32ABE concentration (g L�1) 19.3 29.8 47.2


and polystyrene diethylbenzene derived resins have comparable orsimilar butanol adsorption capacities to those of the zeolites. Mostof other resins including those not shown in Table 12, for exam-ples, XAD-2, XAD-7, and XAD-8 have low butanol adsorptioncapacities of far below 100 mg g�1. It should be noted that morestricter comparison of the adsorption capacities should be basedon comparison of the adsorption isotherms.

By comparing the adsorption performance in terms of adsorptioncapacity, kinetics and selectivity, among the adsorbents F-400, F-600, ZSM-5, silicalite and NaY, it was found that F-400 is mostly suit-able for butanol adsorption with a highest adsorption capacity andthe fastest kinetics for butanol adsorption. In addition, the adsorp-tion capacity of F-400 for other broth components was very low, thusit has high selectivity for butanol [216]. The effect of the presence ofother components in the broths on butanol adsorption was alsostudied, and results showed that the presence of ethanol, glucose,and xylose did not affect the butanol adsorption of F-400; the pres-

ence of acetone slightly decreased the adsorption capacity at lowBuOH concentrations, while acids, especially butyric acid, affectedthe adsorbent capacity significantly [216].

In summary, activated carbons are the most strong adsorbentsand have the highest adsorption capacities for butanol, but theyhave disadvantages such as incomplete desorption and difficultyin regeneration. PS-DVB, PES-DVB, and polystyrene diethylbenzenederived resins have good butanol adsorption capacities as zeolites,but they have less long-term stability as zeolites, e.g., aging prob-lems. Zeolites, especially, silicalite-1 is the most promising adsor-bent for butanol recovery, having many advantages such as highadsorption capacity (�100 mg g�1), high selectivity, completedesorption by heating (200 �C), high stability (<1000 �C), and highbutanol concentration obtained (810 g L�1).

Adsorption is one of the most energy-efficient approaches forremoval of butanol from the ABE fermentation broths. Modificationof current high performing adsorbents, study on adsorbent fouling

Fig. 12. Simple comparison of energy consumption between different separationapproaches (percent data on the top of bars, e.g., ‘‘1–99.9%’’ represents recovery,concentration and purification from 1 wt% of dilute butanol solutions to 99.9 wt%purity of butanol product; ‘‘0.78%-pure’’ means recovery, concentration andpurification from 0.78 wt% of dilute butanol solutions to pure butanol, withoutdetailed purity data; Literature numbers are shown as x label ticks. For the figurelegend, the separation method placed after the dash ‘‘–’’ represents the downstreamseparation process, e.g., ‘‘gas stripping – distillation’’ means gas stripping forbutanol recovery from dilute butanol solutions while distillation is used as themajor downstream separation processes).


problems and stability of adsorbents, methods for efficient desorp-tion or adsorbent regeneration with high concentration of butanolare also worthy of further exploration.

2.8. Comparison of separation methods for product removal

Comparisons of different separation methods for butanol recov-ery from dilute ABE fermentation broths in terms of energy con-sumption have been carried out [200,214,226]. Groot et al. [226]compared gas stripping, pervaporation, L–L extraction, perstrac-tion, and adsorption for separating butanol from dilute solutioncontaining 0.78 wt% butanol in terms of energy consumption.Among them, pervaporation, L–L extraction, and perstraction werefound the best. However, this was published many years ago in1992, and the comparison needs to be updated due to the signifi-cant advances in the research and development of novel pervapo-ration membranes, extracting solvents, and adsorbents with highseparation performance. For example, the adsorbent used for com-parison by Groot et al. [226] has low specific surface area and lowbutanol adsorption ability thus leading to high energy consump-tion of up to 33 MJ kg�1 of n-butanol. By contrast, the currentadsorption technology using high adsorption capacity of adsorbentcould be an energy-efficient process. For instance, comparativestudy for separation and concentration from 5 to 810 g L�1 byQureshi et al. [214] showed that the adsorption with silicalite isthe most energy efficient method for n-butanol recovery fromthe ABE dilute solution, with 8.2 MJ kg�1 of n-butanol. L–L extrac-tion and pervaporation are the second and the third best methodsin terms of energy consumption [214]. However, this work did notconsider the product purification. In addition, similar comparisonsbetween gas stripping, pervaporation, L–L extraction and adsorp-tion were made, and pervaporation, L–L extraction and adsorptionwere also found the most energy-efficient. However, only the prod-uct recovery/pre-concentration step and a simple quantitativeapproach were considered, rather than a detailed process modelingand simulation. In this work, an updated comparison is presentedin Fig. 12 using the energy consumption data from the most recentliterature except in the case of gas stripping – distillation wheredata from old literature was used. The energy consumption ofthe gas stripping plus distillation as downstream separation pro-cess under the same conditions did not change over time as it isbased on one stage vapor–liquid equilibrium. From this figure, itcan be seen that pervaporation, L–L extraction and adsorptionare the most energy-efficient for butanol recovery and purificationfrom dilute solution.

For the gas-stripping process, a large fraction of water will bestripped off with butanol, leading to low separation factor. A typicalrange of the gas-stripping separation factor for butanol is 4–19[200]. Similar to gas stripping, one stage distillation or evaporationis based on vapor–liquid equilibrium. Distillation of dilute solutionsalso has low separation factor for butanol. Therefore, the energy con-sumptions of the gas-stripping and distillation for recovery of buta-nol from fermentation broth are larger than the other methods.Some membranes (Table 9 and Table 10) and extractants (Table 2and Table 5) have high separation factors of over 100. Novel adsor-bents such as zeolites [198,216] and PS-DVB [207] also have highseparation factors for butanol due to their high adsorption capacity.Therefore, these three processes are more energy-efficient for recov-ery of butanol from the fermentation broth.

Strictly speaking, data from different literature are not compa-rable as they were not based on the same feed concentration andproduct purity, and their downstream separation processes or theiroperating conditions were different. Thus, Fig. 2 only shows thevery simple (not strict) comparison, though to some extent thetrend is reasonable as the feed concentration and product purityfor different cases are very close.

3. Further separation and purification/dehydration of butanol

The above reviews are focused on separation of butanol fromABE fermentation broths, including integration of fermentationwith different separation processes. In the whole separation pro-cess, ABE is isolated and pre-concentrated from the fermentationbroth by one of the separation approaches described above. Theresulting ABE solution usually has a butanol concentration of above7.7 wt% butanol. This pre-concentrated ABE mixture is then furtherseparated and purified. Acetone (A) can be firstly separated out ofthe ABE mixture by a simple distillation column as it is the mostvolatile among acetone, butanol, ethanol, and water. Ethanol isthen separated from the remaining butanol–ethanol (BE) aqueoussolution in the another distillation column where ethanol–waterhomogeneous azeotrope comes out of the top of the column whilethe butanol–water heterogeneous azeotrope exits at the bottom.The resulting ethanol–water azeotrope can be dehydrated toobtain high purity ethanol by using one of the methods such ashydrophilic membrane pervaporation and adsorption with hydro-philic adsorbents [59,60,129]. The butanol–water heterogeneousazeotrope can be carried out in the two-column-and-decanter sys-tem proposed by Doherty and Malone [227]. Luyben [228] usedthis flowsheet (Fig. 13) for the steady-state simulation, dynamicsimulation and control study.

In the steady-state simulation [228], the feed enters the decan-ter at a flow rate of 1000 kmol h�1 with mole fraction compositionsof 0.40 (butanol) and 0.60 (water). Phase separation occurs in thedecanter at 0.5 atm and 343 K, with aqueous liquid phase contain-ing 97 mol% water and the organic phase containing 58 mol% n-butanol. The butanol product purity was specified as 99.9 mol%.Other operation conditions and the calculated energy consump-tions by two boilers of two distillation columns and the condenserswere also shown in the figure. By simple calculation, the totalenergy consumptions was 1.792 MJ kg�1 butanol, which is 5.0%of the energy content in recovered butanol. The calculationomitted the energy consumed by the four pumps (note that twoother pumps for transporting two bottoms streams are not shownin the figure) as this part of energy is very small compared to the

Fig. 13. Butanol–water separation [228].


energy required by the boilers and the condenser. Simulation fortwo other feeds of different compositions, i.e., 2 mol% (7.75 wt%)butanol and 10 mol% (31.39 wt%) butanol, at the same mole flowrate of 1000 kmol h�1 were also performed. The energy requiredby the boilers and condenser were 9.566 and 2.279 MJ kg�1 buta-nol for these two cases, respectively. These corresponds to 26.6%and 6.3% of the energy content in recovered butanol, respectively.By comparison, it can be seen that the feed butanol concentrationsignificantly influences the energy consumption required for thefurther separation and purification with distillation. The higherthe feed butanol concentration, the lower amount of energyrequired per kg butanol, and also lower capital cost of distillationcolumns required for the same butanol production capacity.

L–L extraction with solvents such as RTILs could also be used fordehydration of butanol. For example, Hu et al. [229] presented theternary L–L equilibrium of 1-(2-hydroxyethyl)-3-methylimidazo-lium tetrafluoroborate ([C2OHmim]BF4) or 1-(2-hydroxyethyl)-2,3-dimethylimidazolium tetrafluoroborate ([C2OHdmim]BF4) +water + 1-butanol at 20 �C. The experimental results show thatthe two ionic liquids are potential candidates to extract water fromn-butanol by liquid–liquid extraction. The experimental tie lineswere correlated with the NTRL equation.

Purification and dehydration of butanol-rich solution for highpurity butanol product is quite similar to that of ethanol in thatboth processes are to remove minor amount of water from the con-centrated solution to get higher purity products. Methods for eth-anol dehydration discussed previously [59,60,129] can usually beused for butanol dehydration. For example, hydrophilic membranepervaporation was used for butanol dehydration in the economicevaluation performed by Qureshi et al. [183].

Furthermore, TPV with water selective membranes is also suit-able for dehydration of butanol [191]. Especially, adsorption usingmolecular sieves that has been widely and commercially used forthe final ethanol purification and dehydration can also be effectivein butanol purification and dehydration.

4. Conclusions

Biobutanol can be widely used as fuel superior to bioethanol,solvent in food and pharmaceutical industrials, and importantbuilding block for producing a variety of chemicals and materials

such as paints, coatings, biopolymers, bioplastics, and otherchemicals. Butanol can be produced through ABE fermentation,but the production cost are still relatively high due to its verydilute concentration. Separation of butanol from ABE fermenta-tion broth is very challenging and critical for successfulcommercialization.

Integration of ABE fermentation with a product recovery pro-cess such as gas stripping, vacuum flash, solvent extraction, per-straction, membrane pervaporation, thermopervaporation, andadsorption, as process intensification, can efficiently eliminateproduct inhibition, enhance cell growth, increase productivityand yield, and hence could reduce energy consumption and down-stream separation cost and make the overall system more viable.

Based on comprehensive review of the literature and compari-son of the various separation and purification technologies, it isconcluded that membrane pervaporation, L–L extraction, andadsorption are the most energy-efficient approaches for removalof butanol from the ABE fermentation products. Improvement ofmembrane technologies, especially development of thin mem-branes with higher flux, higher separation factor, increasedstrength and high stability, as well as anti-fouling characteristicsor better cleaning technologies, etc. still need to be explored. ForL–L extraction, novel extractants are still needed to be developedwith the aid of molecular modeling technique and chemicallymodifying its affinity to butanol or hydrophobicity. For adsorptionmethod, modification of current high performing adsorbents, studyon adsorbent fouling problems and stability of adsorbents, meth-ods for efficient desorption or adsorbent regeneration with highconcentration of butanol are also worthy of further exploration.In addition, these three hybrid separation processes incorporatedwith fermentation as a whole system need to be optimized, andrigorous comparison based on modeling and analysis of the wholebiorefineries are important and need to be considered. It is alsocritical to understand the effects of different separation processesintegrated with fermentation on the overall production costs andthe capital costs. Process synthesis and modeling of the wholehybrid reaction–separation system integrated with the whole bior-efinery with heat integration, combined with experimental studyand pilot scale analysis would enable us to gain a deeper insightand aid in the successful commercial implementation of the buta-nol biorefinery.


TPV can operate with low grade heat, e.g., at 80–100 �C forbutanol separation from ABE broths. TPV with internal condensercan efficiently recover the condensation heat of pervaporatedvapors for increasing the feed temperature to the required PVoperation temperature. In addition, it has a great potential forcommercial success. As TPV is a novel promising technology,more research work is needed further exploring its capabilitiesand potentials.

References

[1] P. Dürre, Biobutanol: an attractive biofuel, Biotechnol. J. 2 (2007) 1525–1534.[2] S.Y. Lee, J.H. Park, S.H. Jang, L.K. Nielsen, J. Kim, K.S. Jung, Fermentative

butanol production by Clostridia, Biotechnol. Bioeng. 101 (2008) 209–228.[3] V. García, J. Päkkilä, H. Ojamo, E. Muurinen, R.L. Keiski, Challenges in

biobutanol production: how to improve the efficiency?, Renew Sust. EnergyRev. 15 (2011) 964–980.

[4] J. Swana, Y. Yang, M. Behnam, R. Thompson, An analysis of net energyproduction and feedstock availability for biobutanol and bioethanol,Bioresour. Technol. 102 (2011) 2112–2117.

[5] B. Boundy, S.W. Diegel, L. Wright, S.C. Davis, Biomass Energy Data Book,fourth ed., 2011. <http://cta.ornl.gov/bedb>.

[6] M. Kumar, K. Gayen, Developments in biobutanol production: new insights,Appl. Energy 88 (2011) 1999–2012.

[7] P.S. Nigam, A. Singh, Production of liquid biofuels from renewable resources,Prog. Energy Combust. Sci. 37 (2011) 52–68.

[8] C.M. Daza Montaño, Process design and evaluation of butanol productionfrom lignocellulosic biomass, in: Bioenergy 2009 4th International BioenergyConference and Exhibition, Jyväskylä, Finland, 31st August–4th September,2009. <http://www.ecn.nl/docs/library/report/2009/l09117.pdf>.

[9] E.M. Green, Fermentative production of butanol – the industrial perspective,Curr. Opin. Biotechnol. 22 (2011) 1–7.

[10] Y. Ni, Z. Sun, Recent progress on industrial fermentative production ofacetone–butanol–ethanol by Clostridium acetobutylicum in China, Appl.Microbiol. Biotechnol. 83 (2009) 415–423.

[11] C. Jin, M. Yao, H. Liu, C.-F. Lee, J. Ji, Progress in the production and applicationof n-butanol as a biofuel, Renew. Sust. Energy Rev. 15 (2011) 4080–4106.

[12] T.C. Ezeji, N. Qureshi, H.P. Blaschek, Production of acetone–butanol (AB) fromliquefied corn starch, a commercial substrate, using Clostridium beijerinckiicoupled with product recovery by gas stripping, J. Ind. Microbiol. Biotechnol.34 (2007) 771–777.

[13] N. Qureshi, H.P. Blaschek, Butanol production from agricultural biomass, in:K. Shetty, G. Paliyath, A. Pometto, R.E. Levin (Eds.), Food Biotechnology,Taylor & Francis, Boca Raton, 2006, pp. 525–549.

[14] T.C. Ezeji, H.P. Blaschek, Fermentation of dried distillers’ grains and solubles(DDGS) hydrolysates to solvents and value-added products by solventogenicclostridia, Bioresour. Technol. 99 (2008) 5232–5242.

[15] N. Qureshi, T.C. Ezeji, J. Ebener, B.S. Dien, M.A. Cotta, H.P. Blaschek, Butanolproduction by Clostridium beijerinckii. Part I: Use of acid and enzymehydrolyzed corn fiber, Bioresour. Technol. 99 (2008) 5915–5922.

[16] N. Qureshi, B.C. Saha, R.E. Hector, B. Dien, S. Hughes, S. Liu, L. Iten, M.J.Bowman, G. Sarath, M.A. Cotta, Production of butanol (a biofuel) fromagricultural residues: Part II – Use of corn stover and switchgrasshydrolysates, Biomass Bioenergy 34 (2010) 566–571.

[17] J. Lee, E. Seo, D.H. Kweon, K. Park, Y.S. Jin, Fermentation of rice bran anddefatted rice bran for butanol production using Clostridium beijerinckii NCIMB8052, J. Microbiol. Biotechnol. 19 (2009) 482–490.

[18] A. Ranjan, V.S. Moholkar, Comparative study of various pretreatmenttechniques for rice straw saccharification for the production of alcoholicbiofuels, Fuel 112 (2013) 567–571.

[19] N. Qureshi, B.C. Saha, B. Dien, R.E. Hector, M.A. Cotta, Production of butanol (abiofuel) from agricultural residues: Part I – Use of barley straw hydrolysate,Biomass Bioenergy 34 (2010) 559–565.

[20] N. Qureshi, B.C. Saha, M.A. Cotta, Butanol production from wheat strawhydrolysate using Clostridium beijerinckii, Bioproc. Biosyst. Eng. 30 (2007)419–427.

[21] N. Qureshi, B.C. Saha, R.E. Hector, S.R. Hughes, M.A. Cotta, Butanol productionfrom wheat straw by simultaneous saccharification and fermentation usingClostridium beijerinckii: Part I—Batch fermentation, Biomass Bioenergy 32(2008) 168–175.

[22] N. Qureshi, B.C. Saha, M.A. Cotta, Butanol production from wheat straw bysimultaneous saccharification and fermentation using Clostridium beijerinckii:Part II—Fed-batch fermentation, Biomass Bioenergy 32 (2008) 176–183.

[23] Z. Liu, Y. Ying, F. Li, C. Ma, P. Xu, Butanol production by Clostridium beijerinckiiATCC 55025 from wheat bran, J. Ind. Microbiol. Biotechnol. 37 (2010) 495–501.

[24] C. Lu, J. Zhao, S.-T. Yang, D. Wei, Fed-batch fermentation for n-butanolproduction from cassava bagasse hydrolysate in a fibrous bed bioreactor withcontinuous gas stripping, Bioresour. Technol. 104 (2012) 380–387.

[25] P.W. Behrens, S.E. Bingham, S.D. Hoeksema, D.L. Cohoon, Studies on theincorporation of CO2 into starch by Chlorella vulgaris, J. Appl. Phycol. 1 (1989)123–130.

[26] G. Dragone, B.D. Fernandes, A.P. Abreu, A.A. Vicente, J.A. Teixeira, Nutrientlimitation as a strategy for increasing starch accumulation in microalgae,Appl. Energy 88 (2011) 3331–3335.

[27] B. Maršálková, M. Širmerová, M. Kurec, T. Brányik, I. Brányiková, K. Melzoch,V. Zachleder, Microalgae chlorella sp. as an alternative source of fermentablesugars, Chem. Eng. Trans. 21 (2010) 1279–1284, http://dx.doi.org/10.3303/CET1021214.

[28] J.H. Ahn, B.I. Sang, Y. Um, Butanol production from thin stillage usingClostridium pasteurianum, Bioresour. Technol. 102 (7) (2011) 4934–4937.

[29] H. Biebl, Fermentation of glycerol by Clostridium pasteurianum – batch andcontinuous culture studies, J. Ind. Microbiol. Biotechnol. 27 (1) (2001) 18–26.

[30] K.A. Taconi, K.P. Venkataramanan, D.T. Johnson, Growth and solventproduction by Clostridium pasteurianum ATCC� 6013™ utilizing biodiesel-derived crude glycerol as the sole carbon source, Environ. Prog. Sustain.Energy 28 (1) (2009) 100–110.

[31] T.O. Jensen, T. Kvist, M.J. Mikkelsen, P.V. Christensen, P. Westermann,Fermentation of crude glycerol from biodiesel production by Clostridiumpasteurianum, J. Ind. Microbiol. Biotechnol. 39 (5) (2012) 709–717.

[32] A. Malaviya, Y.S. Jang, S.Y. Lee, Continuous butanol production with reducedbyproducts formation from glycerol by a hyper producing mutant of Clostridiumpasteurianum, Appl. Microbiol. Biotechnol. 93 (4) (2012) 1485–1494.

[33] Y. Tashiro, K. Sonomoto, Advances in butanol production by clostridia, in:Current Research, Technology and Education Topics in Applied Microbiologyand Microbial Biotechnology, A. Mendez-Vilas (Ed.), Microbiology Book Series– Number 2, vol. 2, Formatex Research Center, 2010. <http://www.formatex.info/microbiology2/1383-1394.pdf>.

[34] P.P. Peralta-Yahya, F. Zhang, S.B. del Cardayre, J.D. Keasling, Microbialengineering for the production of advanced biofuels, Nature 320–328 (488)(2012).

[35] B.H. Davison, J.E. Thompson, Continuous direct solvent extraction of butanolin a fluidized-bed bioreactor with immobilized Clostridium acetobutylicum,Appl. Biochem. Biotechnol. 39 (40) (1993) 415–426.

[36] X. Yang, G.-J. Tsai, G.T. Tsao, Enhancement of in situ adsorption on theacetone–butanol fermentation by Clostridium acetobutylicum, Sep. Technol. 4(2) (1994) 81–92.

[37] N. Qureshi, M.M. Meagher, J. Huang, R.W. Hutkins, Acetone butanol ethanol(ABE) recovery by pervaporation using silicalite–silicone compositemembrane from fed-batch reactor of Clostridium acetobutylicum, J. Membr.Sci. 187 (1–2) (2001) 93–102.

[38] N. Qureshi, I.S. Maddox, Reduction in butanol inhibition by perstraction:utilization of concentrated lactose/whey permeate by Clostridiumacetobutylicum to enhance butanol fermentation economics, Trans IChemEPart C. Food Bioprod. Process. 83 (C I) (2005) 43–52.

[39] W.V. Hecke, P. Vandezande, S. Claes, S. Vangeel, H. Beckers, L. Diels, H. DeWever, Integrated bioprocess for long-term continuous cultivation ofClostridium acetobutylicum coupled to pervaporation with PDMS compositemembranes, Bioresour. Technol. 111 (2012) 368–377.

[40] H.-W. Yen, Z.-H. Chen, I.-K. Yang, Use of the composite membrane ofpoly(ether-block-amide) and carbon nanotubes (CNTs) in a pervaporationsystem incorporated with fermentation for butanol production by Clostridiumacetobutylicum, Bioresour. Technol. 109 (2012) 105–109.

[41] W.V. Hecke, T. Hofmann, H.D. Wever, Pervaporative recovery of ABE duringcontinuous cultivation: enhancement of performance, Bioresour. Technol.129 (2013) 421–429.

[42] C. Chen, Z. Xiao, X. Tang, H. Cui, J. Zhang, W. Li, C. Ying, Acetone–butanol–ethanol fermentation in a continuous and closed-circulating fermentationsystem with PDMS membrane bioreactor, Bioresour. Technol. 128 (2013)246–251.

[43] J.R. Gapes, D. Nimcevic, A. Friedl, Long-term continuous cultivation ofClostridium beijerinckii in a two-stage chemostat with on-line solventremoval, Appl. Environ. Microbiol. 62 (1996) 3210.

[44] N. Qureshi, H.P. Blaschek, Production of acetone–butanol–ethanol (ABE) by ahyper-producing mutant strain of Clostridium beijerinckii BA101 and recoveryby pervaporation, Biotechnol. Prog. 15 (1999) 594–602.

[45] N. Qureshi, H.P. Blaschek, Butanol production using Clostridium beijerinckiiBA101 hyper-butanol producing mutant strain and recovery bypervaporation, Appl. Biochem. Biotechnol. 84 (2000) 225–235.

[46] T.C. Ezeji, N. Qureshi, H.P. Blaschek, Production of acetone, butanol andethanol by Clostridium beijerinckii BA101 and in situ recovery by gas stripping,World J. Microbiol. Biotechnol. 19 (2003) 595–603.

[47] T.C. Ezeji, N. Qureshi, H.P. Blaschek, Acetone butanol ethanol (ABE)production from concentrated substrate: reduction in substrate inhibitionby fed-batch technique and product inhibition by gas stripping, Appl.Microbiol. Biotechnol. 63 (2004) 653–658.

[48] T.C. Ezeji, P.M. Karcher, N. Qureshi, H.P. Blaschek, Improving performance of agas stripping-based recovery system to remove butanol from Clostridiumbeijerinckii fermentation, Bioprocess Biosyst. Eng. 27 (2005) 207–214.

[49] T.C. Ezeji, N. Qureshi, H.P. Blaschek, Butanol production from agriculturalresidues: impact of degradation products on Clostridium beijerinckii growthand butanol fermentation, Biotechnol. Bioeng. 97 (2007) 1460–1469.

[50] T. Lütke-Eversloh, H. Bahl, Metabolic engineering of Clostridiumacetobutylicum: recent advances to improve butanol production, Curr. Opin.Biotechnol. 22 (2011) 1–44.

[51] G. Jurgens, S. Survase, O. Berezina, E. Sklavounos, J. Linnekoski, A. Kurkijärvi,M. Väkevä, A. van Heiningen, T. Granström, Butanol production fromlignocellulosics, Biotechnol. Lett. 34 (2012) 1415–1434.

http://refhub.elsevier.com/S1383-5866(14)00360-8/h0005









http://cta.ornl.gov/bedb





http://www.ecn.nl/docs/library/report/2009/l09117.pdf





























































http://dx.doi.org/10.3303/CET1021214

http://dx.doi.org/10.3303/CET1021214

















http://www.formatex.info/microbiology2/1383-1394.pdf

http://www.formatex.info/microbiology2/1383-1394.pdf






























































[52] Y.-S. Jang, A. Malaviya, C. Cho, J. Lee, S.Y. Lee, Butanol production fromrenewable biomass by clostridia, Bioresour. Technol. 123 (2012) 653–663.

[53] S. Atsumi, T. Hanai, J.C. Liao, Non-fermentative pathways for synthesis ofbranched-chain higher alcohols as biofuels, Nature 451 (2008) 86–89.

[54] A. Baez, K.M. Cho, J.C. Liao, High-flux isobutanol production using engineeredEscherichia coli: a bioreactor study with in situ product removal, Appl.Microbiol. Biotechnol. 90 (2011) 1681–1690.

[55] E.I. Lan, J.C. Liao, Microbial synthesis of n-butanol, isobutanol, and otherhigher alcohols from diverse resources, Bioresour. Technol. 135 (2013) 339–349.

[56] Z. Dai, H. Dong, Y. Zhu, Y. Zhang, Y. Li, Y. Ma, Introducing a single secondaryalcohol dehydrogenase into butanol-tolerant Clostridium acetobutylicum Rh8switches ABE fermentation to high level IBE fermentation, Biotechnol.Biofuels 5 (2012) 44, http://dx.doi.org/10.1186/1754-6834-5-44. <http://www.biotechnologyforbiofuels.com/content/5/1/44>.

[57] J. Lee, Y.-S. Jang, S.J. Choi, J.A. Im, H. Song, J.H. Cho, D.Y. Seung, E.T.Papoutsakis, G.N. Bennett, S. Yup, Metabolic engineering of Clostridiumacetobutylicum ATCC 824 for isopropanol–butanol–ethanol fermentation,Appl. Environ. Microbiol. 78 (5) (2012) 1416–1423.

[58] K. Kraemer, A. Harwardt, B. Bronneberg, W. Marquardt, Separation of butanolfrom acetone–butanol–ethanol fermentation by a hybrid extraction–distillation process, in: S. Pierucci, G. Buzzi Ferraris (Eds.), 20th EuropeanSymposium on Computer Aided Process Engineering – ESCAPE20, 2010.

[59] H.-J. Huang, S. Ramaswamy, U.W. Tschirner, B.V. Ramarao, A review ofseparation technologies in current and future biorefineries, Sep. Purif.Technol. 62 (1) (2008) 1–21.

[60] L.M. Vane, Separation technologies for the recovery and dehydration ofalcohols from fermentation broths, Biofuels Bioprod. Biorefin. 2 (2008) 553–588.

[61] R.C. Weast, CRC Handbook of Chemistry and Physics, vol. 61, CRC Press, Inc.,Boca Raton, Florida, 1980.

[62] N. Qureshi, H.P. Blaschek, Recovery of butanol from fermentation broth by gasstripping, Renew. Energy 22 (2001) 557–564.

[63] C. Xue, J. Zhao, F. Liu, C. Lu, S.-T. Yang, F.-W. Bai, Two-stage in situ gasstripping for enhanced butanol fermentation and energy-saving productrecovery, Bioresour. Technol. 135 (2013) 396–402.

[64] T. de Vrije, M. Budde, H. van der Wal, P.A.M. Claassen, A.M. López-Contreras,‘‘In situ’’ removal of isopropanol, butanol and ethanol from fermentationbroth by gas stripping, Bioresour. Technol. 137 (2013) 153–159.

[65] C. Xue, J. Zhao, C. Lu, S.-T. Yang, F. Bai, I.C. Tang, High-titer n-butanolproduction by Clostridium acetobutylicum JB200 in fed-batch fermentationwith intermittent gas stripping, Biotechnol. Bioeng. 109 (2012) 2746–2756.

[66] T.C. Ezeji, N. Qureshi, H.P. Blaschek, Microbial production of a biofuel(acetone–butanol–ethanol) in a continuous bioreactor: impact of bleed andsimultaneous product removal, Bioprocess Biosyst. Eng. 36 (2013) 109–116.

[67] M. Setlhaku, S. Heitmann, A. Górak, R. Wichmann, Investigation of gasstripping and pervaporation for improved feasibility of two-stage butanolproduction process, Bioresour. Technol. 136 (2013) 102–108.

[68] N. Abdehagh, A. Sharif, H. Tezel, J. Thibault, In Situ Removal of Biobutanolfrom Fermentation Broth. <http://aiquruguay.org/congreso/download/TL97.pdf>.

[69] A.P. Mariano, C.B. Borba Costa, D. Franceschi de Angelis, F. Maugeri Filho, D.I.Pires Atala, M.R. Wolf Macie, R. Maciel Filho, Dynamics of a continuous flashfermentation for butanol production, Chem. Eng. Trans. 20 (2010) 285–290.

[70] A.P. Mariano, N. Qureshi, R. Maciel Filho, T.C. Ezeji, Bioproduction of butanolin bioreactors: new insights from simultaneously in situ butanol recovery toeliminate product toxicity, Biotechnol. Bioeng. 108 (8) (2011) 1757–1765.

[71] W.A. Evanko, A.M. Eyal, D.A. Glassner, F. Miao, A.A. Aristidou, K. Evans, P.R.Gruber, A.C. Hawkins, Recovery of higher alcohols from dilute aqueoussolutions, US Patent 20130059370 A1, 2013.

[72] A.P. Mariano, N. Qureshi, R.M. Filho, T.C. Ezeji, Assessment of in situ butanolrecovery by vacuum during acetone butanol ethanol (ABE) fermentation, J.Chem. Technol. Biotechnol. 87 (2012) 334–340.

[73] C. Weilnhammer, E. Blass, Continuous fermentation with product recovery byin-situ extraction, Chem. Eng. Technol. 17 (1994) 365–373.

[74] P. Anbarasan, Z.C. Baer, S. Sreekumar, E. Gross, J.B. Binder, H.W. Blanch, D.S.Clark, F.D. Toste, Integration of chemical catalysis with extractivefermentation to produce fuels, Nature 491 (2012) 235–239.

[75] Y.J. Jeon, Y.Y. Lee, Membrane-assisted extractive butanol fermentation, Ann.NY Acad. Sci. 506 (1987) 536–542.

[76] A. Ishizaki, S. Michiwaki, E. Crabbe, G. Kobayashi, K. Sonomoto, S. Yoshino,Extractive acetone–butanol–ethanol fermentation using methylated crudepalm oil as extractant in batch culture of Clostridiumsaccharoperbutylacetonicum N1-4 (ATCC 13564), J. Biosci. Bioeng. 87 (3)(1999) 352–356.

[77] D.R. Anton, J. Cirakovic, B.A. Diner, M.C. Grady, F.J. Woerner, Extractionsolvents derived from oil for alcohol removal in extractive fermentation, USPatent 20110312044, 2011.

[78] L. Adhami, B. Griggs, P. Himebrook, K. Taconi, Liquid–liquid extraction ofbutanol from dilute aqueous solutions using soybean-derived biodiesel, J.Am. Oil Chem. Soc. 86 (2009) 1123–1128.

[79] P.B. Dhamole, Z. Wang, Y. Liu, B. Wang, Extractive fermentation with non-ionic surfactants to enhance butanol production, Biomass Bioenergy 40(2012) 112–119.

[80] A. Chapeaux, L.D. Simoni, T. Ronan, M.A. Stadtherr, J.F. Brennecke, Extractionof alcohols from water with 1-hexyl-3-methylimidazoliumbis(trifluoromethylsulfonyl)imide, Green Chem. 10 (2008) 1301–1306.

[81] L.D. Simoni, A. Chapeaux, J.F. Brennecke, M.A. Stadtherr, Extraction of biofuelsand biofeedstocks from aqueous solutions using ionic liquids, Comput. Chem.Eng. 34 (2010) 1406–1412.

[82] A.G. Fadeev, M.M. Meagher, Opportunities for ionic liquids in recovery ofbiofuels, Chem. Commun. 3 (2001) 295–296.

[83] S.H. Ha, N.L. Mai, Y.-M. Koo, Butanol recovery from aqueous solution intoionic liquids by liquid–liquid extraction, Process Biochem. 45 (12) (2010)1899–1903.

[84] W.R. Pitner, M.M. Schulte, A. Górak, F. Santangelo, A. Wentink, Use of ionicliquids with tetracyanoborate anions as a solvent for extraction of alcoholsfrom aqueous solutions, US Patent 2011/0071324 A1, 2011.

[85] S. Ishii, M. Taya, T. Kobayashi, Production of butanol by Clostridiumacetobutylicum in extractive fermentation system, J. Chem. Eng. Jpn. 18(1985) 125–130.

[86] S.B. Bankar, S.A. Survase, R.S. Singhal, T. Granstrom, Continuous two stageacetone–butanol–ethanol fermentation with integrated solvent removalusing Clostridium acetobutylicum B 5313, Bioresour. Technol. 106 (2012)110–116.

[87] W.J. Groot, H.S. Soedjak, P.B. Donck, R.G.J.M. van tier Lans, K.Ch.A.M. Luyben,J.M.K. Timmer, Butanol recovery from fermentations by liquid–liquidextraction and membrane solvent extraction, Bioprocess Eng. 5 (1990) 203–216.

[88] H.-W. Yen, Y.-C. Wang, The enhancement of butanol production by in situbutanol removal using biodiesel extraction in the fermentation of ABE(acetone–butanol–ethanol), Bioresour. Technol. 145 (2013) 224–228.

[89] F. Santangelo, A. Wentink, A. Górak, W.R. Pitner, M.M. Schulte, Purification ofbiofuels using ionic liquids, in: Proceedings of ISEC 2008 – InternationalSolvent Extraction Conference, Tucson, AZ, United States, 2008, pp. 931–936.

[90] U. Domanska, M. Królikowski, Extraction of butan-1-ol from water with ionicliquids at T = 308.15 K, J. Chem. Thermodyn. 53 (2012) 108–113.

[91] S.E. Davis, S.A. Morton III, Investigation of ionic liquids for the separation ofbutanol and water, Sep. Sci. Technol. 43 (9–10) (2008) 2460–2472, http://dx.doi.org/10.1080/01496390802122089.

[92] A. Heintz, J.K. Lehmann, C. Wertz, Thermodynamic properties of mixturescontaining ionic liquids. 3. Liquid–liquid equilibria of binary mixtures of 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide with propan-1-ol, butan-1-ol, and pentan-1-ol, J. Chem. Eng. Data 48 (3) (2003) 472–474.

[93] C.T. Wu, K.N. Marsh, A.V. Deev, J.A. Boxall, Liquid–liquid equilibria of room-temperature ionic liquids and butan-1-ol, J. Chem. Eng. Data 48 (3) (2003)486–491.

[94] U. Domanska, A. Marciniak, Solubility of ionic liquid [emim][PF6] in alcohols,J. Phys. Chem. 108 (2004) 2376–2382.

[95] A. Heintz, J.K. Lehmann, C. Wertz, J. Jacquemin, Thermodynamic properties ofmixtures containing ionic liquids. 4. LLE of binary mixtures of [C2MIM][NTf2]with propan-1-ol, butan-1-ol, and pentan-1-ol and [C4MIM][NTf2] withcyclohexanol and 1,2-hexanediol including studies of the influence of smallamounts of water, J. Chem. Eng. Data 50 (3) (2005) 956–960.

[96] K. Sahandzhieva, D. Tuma, S. Breyer, A.P.S. Kamps, G. Maurer, Liquid–liquidequilibrium in mixtures of the ionic liquid 1-n-butyl-3-methylimidazoliumhexafluorophosphate and an alkanol, J. Chem. Eng. Data 51 (5) (2006) 1516–1525.

[97] M. Bendova, Z. Wagner, Liquid–liquid equilibrium in binary system[bmim][PF6]t1-butanol, J. Chem. Eng. Data 51 (6) (2006) 2126–2131.

[98] J.G. Huddleston, A.E. Visser, W.M. Reichert, H.D. Willauer, G.A. Broker, R.D.Rogers, Characterization and comparison of hydrophilic and hydrophobicroom temperature ionic liquids incorporating the imidazolium cation, GreenChem. 3 (2001) 156–164.

[99] M. Tariq, M.G. Freire, B. Saramago, J.A.P. Coutinho, J.N.C. Lopes, L.P.N. Rebelo,Surface tension of ionic liquids and ionic liquid solutions, Chem. Soc. Rev. 41(2012) 829–868.

[100] C.M. Neves, P.J. Carvalho, M.G. Freire, J.A. Coutinho, Thermophysicalproperties of pure and water-saturated tetradecyltrihexylphosphonium-based ionic liquids, J. Chem. Thermodyn. 43 (2011) 948–957.

[101] L.Y. Garcia-Chavez, C.M. Garsia, B. Schuur, A.B. de Haan, Biobutanol recoveryusing nonfluorinated task-specific ionic liquids, Ind. Eng. Chem. Res. 51 (24)(2012) 8293–8301.

[102] H.R. Cascon, S.K. Choudhari, 1-Butanol pervaporation performance andintrinsic stability of phosphonium and ammonium ionic liquid-basedsupported liquid membranes, J. Membr. Sci. 429 (15) (2013) 214–224.

[103] D. Rabari, T. Banerjee, Biobutanol and n-propanol recovery using a lowdensity phosphonium based ionic liquid at T = 298.15 K and p = 1 atm, FluidPhase Equilib. 355 (2013) 26–33.

[104] K.R. Seddon, A. Stark, M.-J. Torres, Influence of chloride, water, and organicsolvents on the physical properties of ionic liquids, Pure Appl. Chem. 72(2000) 2275–2287.

[105] M.G. Freire, P.J. Carvalho, A.M. Fernandes, I.M. Marrucho, A.J. Queimada, J.A.P.Coutinho, Surface tensions of imidazolium based ionic liquids: anion, cation,temperature and water effect, J. Colloid Interface Sci. 314 (2007) 621–630.

[106] P. Bôonhote, A.P. Dias, N. Papageorgiou, K. Kalyanasundaram, M. Graetzel,Inorg. Chem. 35 (1996) 1168–1178.

[107] P.K. Kilaru, R.A. Condemarin, P. Scovazzo, Ind. Eng. Chem. Res. 47 (2008) 900–909.











http://dx.doi.org/10.1186/1754-6834-5-44

http://www.biotechnologyforbiofuels.com/content/5/1/44

http://www.biotechnologyforbiofuels.com/content/5/1/44































http://aiquruguay.org/congreso/download/TL97.pdf

http://aiquruguay.org/congreso/download/TL97.pdf



























































http://dx.doi.org/10.1080/01496390802122089

http://dx.doi.org/10.1080/01496390802122089

























































[108] R.L. Gardas, R. Ge, N. Ab Manan, D.W. Rooney, C. Hardacre, Interfacial tensionsof imidazolium-based ionic liquids with water and n-alkanes, Fluid PhaseEquilib. 294 (2010) 139–147.

[109] G.V.S.M. Carrera, C.A.M. Afonso, L.C. Branco, Interfacial properties, densities,and contact angles of task specific ionic liquids, J. Chem. Eng. Data 55 (2010)609–615.

[110] H.R. Cascon, S.K. Choudhari, G.M. Nisola, E.L. Vivas, D.-J. Lee, W.-J. Chung,Partitioning of butanol and other fermentation broth components inphosphonium and ammonium-based ionic liquids and their toxicity tosolventogenic clostridia, Sep. Purif. Technol. 78 (2) (2011) 164–174.

[111] P. Kilaru, G.A. Baker, P. Scovazzo, Density and surface tension measurementsof imidazolium-, quaternary phosphonium-, and ammonium-based room-temperature ionic liquids: data and correlations, J. Chem. Eng. Data 52 (2007)2306–2314.

[112] M.G. Freire, P.J. Carvalho, R.L. Gardas, L.M.N.B.F. Santos, I.M. Marrucho, J.A.P.Coutinho, Solubility of water in tetradecyltrihexylphosphonium-based ionicliquids, J. Chem. Eng. Data 53 (2008) 2378–2382.

[113] A. Wandschneider, J.K. Lehmann, A. Heintz, Surface tension and density ofpure ionic liquids and some binary mixtures with 1-propanol and 1-butanol,J. Chem. Eng. Data 53 (2008) 596–599.

[114] J. Martak, S. Schlosser, Extraction of lactic acid by phosphonium ionic liquids,Sep. Purif. Technol. 57 (2007) 483–494.

[115] V.M. Egorov, S.V. Smirnova, I.V. Pletnev, Highly efficient extraction of phenolsand aromatic amines into novel ionic liquids incorporating quaternaryammonium cation, Sep. Purif. Technol. 63 (2008) 710–715.

[116] N. Nishi, T. Kawakami, F. Shigematsu, M. Yamamoto, T. Kakiuchi, Fluorine-free and hydrophobic room-temperature ionic liquids, tetraalkylammoniumbis(2-ethylhexyl)sulfosuccinates, and their ionic liquid–water two-phaseproperties, Green Chem. 8 (2006) 349–355.

[118] C. Kolbeck, J. Lehmann, K.R.J. Lovelock, T. Cremer, N. Paape, P. Wasserscheid,A.P. Froba, F. Maier, H.P. Steinruck, Density and surface tension of ionicliquids, J. Phys. Chem. B 114 (51) (2010) 17025–17036.

[119] M. Blahušiak, J. Marták, F. Miranda, Š. Schlosser, J.A. Teixeira, Effect ofviscosity of a liquid membrane containing oleyl alcohol on the pertraction ofbutyric acid, Chem. Pap. 67 (12) (2013) 1560–1568.

[120] M. Matsumura, H. Kataoka, M. Sueki, K. Araki, Energy saving effect ofpervaporation using oleyl alcohol liquid membrane in butanol purification,Bioprocess Eng. 3 (1988) 93–100.

[121] V.R. Ferro, E. Ruiz, J. de Riva, J. Palomar, Introducing process simulation inionic liquids design/selection for separation processes based on operationaland economic criteria through the example of their regeneration, Sep. Purif.Technol. 97 (3) (2012) 195–204.

[122] R.J. Cornmell, C.L. Winder, G.J.T. Tiddy, R. Goodacre, G. Stephens,Accumulation of ionic liquids in Escherichia coli cells, Green Chem. 10(2008) 836–841.

[123] M. Stoffers, A. Górak, Continuous multi-stage extraction of n-butanol fromaqueous solutions with 1-hexyl-3-methylimidazolium tetracyanoborate, Sep.Purif. Technol. 120 (13) (2013) 415–422.

[124] S.-T. Yang, C. Lu, 15. Extractive-fermentation hybrid (extractivefermentation), in: S. Ramaswamy, H.-J. Huang, B.V. Ramarao (Eds.),Separation and Purification Technologies in Biorefineries, Wiley, 2013, pp.409–437.

[125] R.P. Swatloski, J.D. Holbrey, R.D. Rogers, Ionic liquids are not always green:hydrolysis of 1-butyl-3-methylimidazolium hexafluorophosphate, GreenChem. 5 (2003) 361–363.

[126] K. Kraemer, A. Harwardt, R. Bronneberg, W. Marquardt, Separation of butanolfrom acetone–butanol–ethanol fermentation by a hybrid extraction–distillation process, Comput. Chem. Eng. 35 (2011) 949–963.

[127] Y.J. Jeon, Y.Y. Lee, In situ product separation in butanol fermentation bymembrane-assisted extraction, Enzyme Microbial. Technol. 11 (1989) 575–582.

[128] S. Tanaka, Y. Tashiro, G. Kobayashi, T. Ikegami, H. Negishi, K. Sakaki,Membrane-assisted extractive butanol fermentation by Clostridiumsaccharoperbutylacetonicum N1-4 with 1-dodecanol as the extractant,Bioresour. Technol. 116 (2012) 448–452.

[129] L.Y. Jiang, Y. Wang, T.-S. Chung, X.Y. Qiao, J.-Y. Lai, Polyimides membranes forpervaporation and biofuels separation, Prog. Polym. Sci. 34 (2009) 1135–1160.

[130] F. Liu, L. Liu, X. Feng, Separation of acetone–butanol–ethanol (ABE) fromdilute aqueous solutions by pervaporation, Sep. Purif. Technol. 42 (3) (2005)273–282.

[131] E.A. Fouad, X. Feng, Use of pervaporation to separate butanol from diluteaqueous solutions: effects of operating conditions and concentrationpolarization, J. Membr. Sci. 323 (2) (2008) 428–435.

[132] H.W. Yen, S.F. Lin, I.K. Yang, Use of poly(ether-block-amide) in pervaporationcoupling with a fermenter to enhance butanol production in the cultivationof Clostridium acetobutylicum, J. Biosci. Bioeng. 113 (3) (2012) 372–377.

[133] S. Heitmann, V. Krüger, D. Welz, P. Lutze, Experimental investigation ofpervaporation membranes for biobutanol separation, J. Membr. Sep. Technol.2 (2013) 245–262.

[134] K. Srinivasan, K. Palanivelu, A.N. Gopalakrishnan, Recovery of 1-butanol froma model pharmaceutical aqueous waste by pervaporationm, Chem. Eng. Sci.62 (2007) 2905.

[135] D.L. Vrana, M.M. Meagher, R.W. Hutkins, B. Duffield, Pervaporation of modelacetone–butanol–ethanol fermentation product solutions usingpolytetrafluoroethylene membranes, Sep. Sci. Technol. 28 (1993) 2167.

[136] A.G. Fadeev, Y.A. Selinskaya, S.S. Kelley, M.M. Meagher, E.G. Litvinova, V.S.Khotimsky, V.V. Volkov, Extraction of butanol from aqueous solutions bypervaporation through poly(1-trimethylsilyl-1-propyne), J. Membr. Sci. 186(2) (2001) 205–217.

[137] A.G. Fadeev, M.M. Meagher, S.S. Kelley, V.V. Volkov, Fouling of poly[-1-(trimethylsilyl)-1-propyne] membranes in pervaporative recovery of butanolfrom aqueous solutions and ABE fermentation broth, J. Membr. Sci. 173(2000) 133–144.

[138] N. Qureshi, M.M. Meagher, R.W. Hutkins, Recovery of butanol from modelsolutions and fermentation broth using a silicalite/silicone membrane, J.Membr. Sci. 158 (1999) 115–125.

[139] E. El-Zanati, E. Abdel-Hakim, O. El-Ardi, M. Fahmy, Modeling and simulationof butanol separation from aqueous solutions using pervaporation, J. Membr.Sci. 280 (2006) 278–283.

[140] Y. Wang, S.H. Goh, T.S. Chung, P. Na, Polyamide-imide/polyetherimide dual-layer hollow fiber membranes for pervaporation dehydration of C1–C4alcohols, J. Membr. Sci. 326 (2009) 222.

[141] N. Qureshi, H.P. Blaschek, Butanol recovery from model solution/fermentation broth by pervaporation: evaluation of membraneperformance, Biomass Bioenergy 17 (1999) 175–184.

[142] S.L. Schmidt, M.D. Myers, S.S. Kelley, J.D. McMillan, N. Padukone, Evaluationof PTMSP membranes in achieving enhanced ethanol removal fromfermentations by pervaporation, Appl. Biochem. Biotechnol. 63–65 (1997)469–482.

[143] S.-Y. Li, R. Srivastava, R.S. Parnas, Separation of 1-butanol by pervaporationusing a novel tri-layer PDMS composite membrane, J. Membr. Sci. 363 (1–2)(2010) 287–294.

[144] J. Huang, M.M. Meagher, Pervaporative recovery of n-butanol from aqueoussolutions and ABE fermentation broth using thin-film silicalite-filled siliconecomposite membranes, J. Membr. Sci. 192 (1–2) (2001) 231–242.

[145] A. Jonquieres, A. Fane, Filled and unfilled composite GFT PDMS membranesfor the recovery of butanols from dilute aqueous solutions: influence ofalcohol polarity, J. Membr. Sci. 125 (1997) 245–255.

[146] S. Claes, P. Vandezande, S. Mullens, K. De Sitter, R. Peeters, M.K. Van Bael,Preparation and benchmarking of thin film supported PTMSP-silicapervaporation membranes, J. Membr. Sci. 389 (2011) 265–271.

[147] E. Favre, Q.T. Nguyen, S. Bruneau, Extraction of 1-butanol from aqueoussolutions by pervaporation, J. Chem. Technol. Biotechnol. 65 (3) (1996) 221–228.

[148] V. Garcia, E. Pongracz, E. Muurinen, R.L. Keiski, Recovery of n-butanol fromsalt containing solutions by pervaporation, Desalination 241 (2009) 201–211.

[149] C. Tong, Y. Bai, J. Wu, L. Zhang, L. Yang, J. Qian, Pervaporation recovery ofacetone–butanol from aqueous solution and fermentation broth using HTPB-based polyurethaneurea membranes, Sep. Sci. Technol. 45 (6) (2010) 751–761.

[150] J. Niemistö, W. Kujawski, R.L. Keiski, Pervaporation performance of compositepoly(dimethyl siloxane) membrane for butanol recovery from modelsolutions, J. Membr. Sci. 434 (2013) 55–64.

[151] W.-D. Zhang, W. Sun, J. Yang, Z.-Q. Ren, The study on pervaporation behaviorsof dilute organic solution through PDMS/PTFE composite membrane, Appl.Biochem. Biotechnol. 160 (2010) 156–167.

[152] E.A. Fouad, X. Feng, Pervaporative separation of n-butanol from diluteaqueous solutions using silicalite-filled poly(dimethyl siloxane) membranes,J. Membr. Sci. 339 (2009) 120–125.

[153] H. Zhou, Y. Su, X. Chen, S. Yi, Y. Wan, Modification of silicalite-1 byvinyltrimethoxysilane (VTMS) and preparation of silicalite-1 filledpolydimethylsiloxane (PDMS) hybrid pervaporation membranes, Sep. Purif.Technol 75 (3) (2010) 286–294.

[154] G. Liu, W. Wei, H. Wu, X. Dong, M. Jiang, W. Jin, Pervaporation performance ofPDMS/ceramic composite membrane in acetone butanol ethanol (ABE)fermentation–PV coupled process, J. Membr. Sci. 373 (2011) 121–129.

[155] G. Liu, D. Hou, W. Wei, F. Xiangli, W. Jin, Pervaporation separation of butanol–water mixtures using polydimethylsiloxane/ceramic composite membrane,Chin. J. Chem. Eng. 19 (1) (2011) 40–44.

[156] E.J. Jeon, A.S. Kim, Y.T. Lee, Pervaporation of butanol/water mixtures usingsiloxane polymer/ceramic composite membranes, Desalin. Water Treat. 48(1–3) (2012) 17–26.

[157] H. Tan, Y. Wu, T. Li, Pervaporation of n-butanol aqueous solution throughZSM-5-PEBA composite membranes, J. Appl. Polym. Sci. 129 (1) (2013) 105–112.

[158] H. Wu, X.-P. Chen, G.-P. Liu, M. Jiang, T. Guo, W.Q. Jin, P. Wei, D.W. Zhu,Acetone–butanol–ethanol (ABE) fermentation using Clostridiumacetobutylicum XY16 and in situ recovery by PDMS/ceramic compositemembrane, Bioprocess Biosyst. Eng. 35 (7) (2012) 1057–1065.

[159] I.F.J. Vankelecom, D. Depre, S. De Beukelaer, J.B. Uytterhoeven, Influence ofzeolites in PDMS membranes: pervaporation of water/alcohol mixtures, J.Phys. Chem. 99 (1997) 13193–131937.

[160] P. Shao, A. Kumar, Separation of 1-butanol/2,3-butanediol using ZSM-5zeolite-filled polydimethylsiloxane membranes, J. Membr. Sci. 339 (2009)143–150.

[161] X. Liu, Y. Li, Y. Liu, G. Zhu, J. Liu, W. Yang, Capillary supported ultrathinhomogeneous silicalite–poly(dimethylsiloxane) nanocomposite membranefor bio-butanol recovery, J. Membr. Sci. 369 (2011) 228–232.

[162] H. Zhou, Y. Su, X. Chen, Y. Wan, Separation of acetone, butanol and ethanol(ABE) from dilute aqueous solutions by silicalite-1/PDMS hybridpervaporation membranes, Sep. Purif. Technol. 79 (2011) 375–384.


















































































































































































[163] S.Y. Li, R. Srivastava, R.S. Parnas, Study of in situ 1-butanol pervaporationfrom A–B–E fermentation using a PDMS composite membrane: validity ofsolution–diffusion model for pervaporative A–B–E fermentation, Biotechnol.Prog. 27 (1) (2011) 111–120.

[164] A.B. Beltran, G.M. Nisola, S.S. Choi, Y. Kim, W.J. Chung, Surface functionalizedsilica nanoparticles as fillers in polydimethylsiloxane membrane for thepervaporative recovery of 1-butanol from aqueous solution, J. Chem. Technol.Biotechnol. 88 (12) (2013) 2216–2222.

[165] T.A. Peters, C.H.S. Poeth, N.E. Benes, H.C.W.M. Buijs, F.F. Vercauteren, J.T.F.Keurentjes, Ceramic-supported thin PVA pervaporation membranescombining high flux and high selectivity; contradicting the flux-selectivityparadigm, J. Membr. Sci. 276 (2006) 42–50.

[166] T.A. Peters, N.E. Benes, J.T.F. Keurentjes, Hybrid ceramic supported thin PVApervaporation membranes: long-term performance and thermal stability inthe dehydration of alcohols, J. Membr. Sci. 311 (2008) 7–11.

[167] J. Marszałek, W. Kaminski, Efficiency of acetone–butanol–ethanol–watersystem separation by pervaporation, Chem. Process Eng. 33 (1) (2012) 131–140.

[168] S. Liu, G. Liu, X. Zhao, W. Jin, Hydrophobic-ZIF-71 filled PEBA mixed matrixmembranes for recovery of biobutanol via pervaporation, J. Membr. Sci. 446(2013) 181–188, http://dx.doi.org/10.1016/j.memsci.2013.06.025.

[169] G.G. Paradis, D.P. Shanahan, R. Kreiter, H.M. van Veen, H.L. Castricum, A.Nijmeijer, J.F. Vente, From hydrophilic to hydrophobic HybSi� membranes: achange of affinity and applicability, J. Membr. Sci. 428 (1) (2013) 157–162.

[170] H. Negishi, K. Sakaki, T. Ikegami, Silicalite pervaporation membraneexhibiting a separation factor of over 400 for butanol, Chem. Lett. 39(2010) 1312–1314.

[171] D. Shen, W. Xiao, J. Yang, N. Chu, J. Lu, D. Yin, J. Wang, Synthesis of silicalite-1membrane with two silicon source by secondary growth method and itspervaporation performance, Sep. Purif. Technol. 76 (3) (2011) 308–315.

[172] P. Izák, K. Schwarz, W. Ruth, H. Bahl, U. Kragl, Increased productivity ofClostridium acetobutylicum fermentation of acetone, butanol, and ethanol bypervaporation through supported ionic liquid membrane, Appl. Microbiol.Biotechnol. 78 (2008) 597–602.

[173] S. Heitmann, J. Krings, P. Kreis, A. Lennert, W.R. Pitner, A. Górak, M.M. Schulte,Recovery of n-butanol using ionic liquid-based pervaporation membranes,Sep. Purif. Technol. 97 (3) (2012) 108–114.

[174] Paweł-D _zygiel, P.P. Wieczorek, Supported liquid membranes and theirmodifications: definition, classification, theory, stability, application andperspectives, in: V.S. Kislik (Ed.), Liquid Membranes: Principles andApplications in Chemical Separations and Wastewater Treatment, ElsevierScience Ltd., 2009 (Chapter 3).

[175] M. Matsumura, H. Kataoka, I.M. Sueki, K.A. Yokohama, Energy saving effect ofpervaporation using oleyl alcohol liquid membrane in butanol purification,Bioprocess Eng. 3 (1988) 93–100.

[176] M. Matsumura, S. Takehara, H. Kataoka, Continuous butanol/isopropanolfermentation in down-flow column reactor coupled with pervaporation usingsupported liquid membrane, Biotechnol. Bioeng. 39 (1992) 148–156.

[177] A. Thongsukmak, K.K. Sirkar, Pervaporation membranes highly selective forsolvents present in fermentation broths, J. Membr. Sci. 302 (2007) 45–58.

[178] P. Izák, W. Ruth, Z. Fei, P.J. Dyson, U. Kragl, Selective removal of acetone andbutan-1-ol from water with supported ionic liquid–polydimethylsiloxanemembrane by pervaporation, Chem. Eng. J. 139 (2) (2008) 318–321.

[179] M. Kohoutová, A. Sikora, S. Hovorka, A. Randová, J. Schauer, M. Tisma, K.Setnicková, R. Petrickovic, S. Guernik, N. Greenspoon, P. Izák, Influence ofionic liquid content on properties of dense polymer membranes, Eur. Polym.J. 3 (2009) 813–819.

[180] M. Matsumoto, Y. Murakami, K. Kondo, Separation of 1-butanol bypervaporation using polymer inclusion membranes containing ionic liquids,Solvent Extr. Res. Dev. 18 (2011) 75–83.

[181] A. Plaza, G. Merlet, A. Hasanoglu, M. Isaacs, J. Sanchez, J. Romero, Separationof butanol from ABE mixtures by sweep gas pervaporation using a supportedgelled ionic liquid membrane: analysis of transport phenomena andselectivity, J. Membr. Sci. 444 (2013) 201–212.

[182] S.-Y. Li, R. Srivastava, S.L. Sui, Y. Li, Performance of batch, fed-batch, andcontinuous A–B–E fermentation with pH-control, Bioresour. Technol. 102(2011) 4241–4250.

[183] N. Qureshi, B.C. Saha, M.A. Cotta, V. Singh, An economic evaluation biologicalconversion of wheat straw to butanol: a biofuel, Energy Convers. Manage. 65(2013) 456–462.

[184] L.M. Vane, A review of pervaporation for product recovery from biomassfermentation processes, J. Chem. Technol. Biotechnol. 80 (6) (2005) 603–629.

[185] Y. Wang, N. Widjojo, P. Sukitpaneenit, T.-S. Chung, 10. Membranepervaporation, in: S. Ramaswamy, H.-J. Huang, B.V. Ramarao (Eds.),Separation and Purification Technologies in Biorefineries, Wiley, 2013, pp.259–300.

[186] K.W. Lawson, D.R. Lloyd, Membrane distillation, J. Membr. Sci. 124 (1997) 1–25.

[187] F.A. Banat, M. Al-Shannag, Recovery of dilute acetone–butanol–ethanol (ABE)solvents from aqueous solutions via membrane distillation, Bioprocess Eng.23 (6) (2000) 643–649.

[188] I.B. Elkina, A.B. Gilman, V.V. Ugrozov, V.V. Volkov, Separation of mineral acidsolutions by membrane distillation and thermopervaporation throughporous and nonporous membranes, Ind. Eng. Chem. Res. 52 (26) (2013)8856–8863, http://dx.doi.org/10.1021/ie3025246.

[189] A. Alkhudhiri, N. Darwish, N. Hilal, Membrane distillation: a comprehensivereview, Desalination 287 (2012) 2–18.

[190] M.S. El-Bourawi, Z. Ding, R. Ma, M. Khayet, A framework for betterunderstanding membrane distillation separation process, J. Membr. Sci. 285(2006) 4–29.

[191] E.S. Fernandez, P. Geerdink, E.L.V. Goetheer, Thermo pervap: the next step inenergy efficient pervaporation, Desalination 250 (2010) 1053–1055.

[192] I.L. Borisov, V.V. Volkov, V.A. Kirsh, V.I. Roldugin, Simulation of thetemperature-driven pervaporation of dilute 1-butanol aqueous mixturesthrough a PTMSP membrane in a cross-flow module, Petrol. Chem. 51 (7)(2011) 542–554.

[193] R.A. Diltz, T.V. Marolla, M.V. Henley, L. Li, Reverse osmosis processing oforganic model compounds and fermentation broths, Bioresour. Technol. 98(2007) 686–695.

[194] M. Ito, I. Morita, S. Yamane, K. Yamada, Butanol Manufacturing Method, US20130041187A1, 2013.

[195] S. Venkatesan, Adsorption, in: S. Ramaswamy, H.-J. Huang, B.V. Ramarao(Eds.), Separation and Purification Technologies in Biorefineries, Wiley, 2013,pp. 103–148.

[196] X. Yang, G.T. Tsao, Enhanced acetone–butanol fermentation using repeatedfed-batch operation coupled with cell recycle by membrane andsimultaneous removal of inhibitory products by adsorption, Biotechnol.Bioeng. 47 (1995) 444–450.

[197] W.J. Groot, K.Ch.A.M. Luyben, In situ product recovery by adsorption in thebutanol/isopropanol batch fermentation, Appl. Microbiol. Biotechnol. 25 (1)(1986) 29–31.

[198] N. Abdehagh, F.H. Tezel, J. Thibault, Improvements in bio-butanol separationby adsorption: adsorbent screening, kinetics and equilibrium, in: AIChE 2012Annual Meeting, Pittsburgh, 2012.

[199] J.C.S. Remi, G. Baron, J. Denaye, Adsorptive separations for the recovery andpurification of biobutanol, Adsorption 18 (2012) 367–373.

[200] A. Oudshoorn, L.A.M. Van der Wielen, A.J.J. Straathof, Assessment of optionsfor selective 1-butanol recovery from aqueous solution, Ind. Eng. Chem. Res.48 (15) (2009) 7325–7336.

[201] A. Oudshoorn, L.A.M. van der Wielen, A.J.J. Straathof, Adsorption equilibria ofbio-based butanol solutions using zeolite, Biochem. Eng. J. 48 (1) (2009) 99–103.

[202] V. Saravanan, D.A. Waijers, M. Ziari, M.A. Noordermeer, Recovery of 1-butanolfrom aqueous solutions using zeolite ZSM-5 with a high Si/Al ratio;suitability of a column process for industrial applications, Biochem. Eng. J.49 (1) (2010) 33–39.

[203] T. Remy, J.C.S. Remi, R. Singh, P.A. Webley, G.V. Baron, J.F.M. Denayer,Adsorption and separation of C1–C8 alcohols on SAPO-34, J. Phys. Chem. C115 (16) (2011) 8117–8125.

[204] A.B. Thompson, S.J. Cope, T.D. Swift, J.M. Notestein, Adsorption of n-butanolfrom dilute aqueous solution with grafted calixarenes, Langmuir 27 (2011)11990–11998.

[205] L. Nielsen, M. Larsson, O. Hoist, B. Mattiasson, Adsorbents for extractivebioconversion applied to the acetone–butanol fermentation, Appl. Microbiol.Biotechnol. 28 (1988) 335–339.

[206] X. Yang, G.T. Tsao, Enhanced acetone–butanol fermentation using repeatedfed-batch operation coupled with cell recycle by membrane andsimultaneous removal of inhibitory products by adsorption, Biotechnol.Bioeng. 47 (4) (1995) 444–450.

[207] D.R. Nielsen, K.J. Prather, In situ product recovery of n-butanol usingpolymeric resins, Biotechnol. Bioeng. 102 (3) (2009) 811–821.

[208] X. Lin, J. Wu, X. Jin, J. Fan, R. Li, Q., W. Qian, D. Liu, X. Chen, Y. Chen, J. Xie, J.Bai, H. Ying, Selective separation of biobutanol from acetone–butanol–ethanol fermentation broth by means of sorption methodology based on anovel macroporous resin, Biotechnol. Prog. 28 (4) (2012) 962–972.

[209] X. Lin, J. Wu, J. Fan, W. Qian, X. Zhou, C. Qian, X. Jin, L. Wang, J. Bai, H. Ying,Adsorption of butanol from aqueous solution onto a new type ofmacroporous adsorption resin: studies of adsorption isotherms and kineticssimulation, J. Chem. Technol. Biotechnol. 87 (7) (2012) 924–931.

[210] X. Lin, R. Li, Q. Wen, J. Wu, J. Fan, X. Jin, W. Qian, D. Liu, X. Chen, Y. Chen, J. Xie,J. Bai, H. Ying, Experimental and modeling studies on the sorptionbreakthrough behaviors of butanol from aqueous solution in a fixed-bed ofKA-I resin, Biotechnol. Bioprocess Eng. 18 (2013) 223–233.

[211] B.M. Ennis, N. Qureshi, I.S. Maddox, In-line toxic product removal duringsolvent production by continuous fermentation using immobilizedClostridium acetobutylicum, Enzyme Microbial. Technol. 9 (1987) 672–675.

[212] D.R. Nielsen, G.S. Amarasiriwardena, K.L.J. Prather, Predicting the adsorptionof second generation biofuels by polymeric resins with applications forin situ product recovery (ISPR), Bioresour. Technol. 101 (8) (2010) 2762–2769.

[213] M.-H. Eom, W. Kim, J. Lee, J.H. Lee, S. Park, Kinetic studies on biobutanolrecovery process using adsorbent resin, in: 8th IFAC InternationalSymposium on Advanced Control of Chemical Processes, ADCHEM 2012,Singapore, 10–13 July, 2012.

[214] N. Qureshi, S. Hughes, I.S. Maddox, M.A. Cotta, Energy-efficient recovery ofbutanol from model solutions and fermentation broth by adsorption,Bioprocess Biosyst. Eng. 27 (2005) 215–222.

[215] M.-H. Eom, W. Kim, J. Lee, J.-H. Cho, D. Seung, S. Park, J.H. Lee, Modeling of abiobutanol adsorption process for designing an extractive fermentor, Ind.Eng. Chem. Res. 52 (2) (2013) 603–611.



















http://dx.doi.org/10.1016/j.memsci.2013.06.025





































































http://dx.doi.org/10.1021/ie3025246


















































































[216] N. Abdehagh, F.H. Tezel, J. Thibault, Adsorbent screening for biobutanolseparation by adsorption: kinetics, isotherms and competitive effect of othercompounds, Adsorption 19 (2013) 1263–1272.

[217] N.B. Milestone, D.M. Bibby, Concentration of alcohols by adsorption onsilicalites, J. Chem. Technol. Biotechnol. 31 (1981) 732–736.

[218] N.B. Milestone, D.M. Bibby, Adsorption of alcohols from aqueous solutions byZSM-5, J. Chem. Technol. Biotechnol. 34A (1983) 73–79.

[219] T. Ikegami, H. Negishi, K. Sakaki, Selective separation of n-butanol fromaqueous solutions by pervaporation using silicone rubber-coated silicalitemembranes, J. Chem. Technol. Biotechnol. 86 (6) (2011) 845–851.

[220] A. Oudshoorn, L.A.M. van der Wielen, A.J.J. Straathof, Desorption of butanolfrom zeolite material, Biochem. Eng. J. 67 (2012) 167–172.

[221] J. Cousin Saint Remi, T. Rémy, V. Van Hunskerken, S. van de Perre, T. Duerinck,M. Maes, D. De Vos, E. Gobechiya, C.E.A. Kirschhock, G.V. Baron, J.F.M.Denayer, Biobutanol separation with the metal-organic framework ZIF-8,ChemSusChem 4 (8) (2011) 1074–1077.

[222] V.I. Águeda, J.A. Delgado, M.A. Uguina, J.L. Sotelo, Á. García, Column dynamicsof an adsorption–drying–desorption process for butanol recovery fromaqueous solutions with silicalite pellets, Sep. Purif. Technol. 104 (2013)307–321.

[223] M.M. Meagher, N. Qureshi, R. Hutkins, Silicalite membrane and method forthe selective recovery and concentration of acetone and butanol from modelABE solutions and fermentation broth, US Patent 5,755,967, 1998.

[224] A. Faisal, A. Zarebska, P. Saremi, D. Korelskiy, L. Ohlin, U. Rova, J. Hedlund, M.Grahn, MFI zeolite as adsorbent for selective recovery of hydrocarbons fromABE fermentation broths, Adsorption 2013. http://dx.doi.org/10.1007/s10450-013-9576-6.

[225] I.S. Maddox, Use of silicalite for the adsorption of n-butanol fromfermentation liquors, Biotechnol. Lett. 4 (1982) 759–760.

[226] W.J. Groot, R.G.J.M. van der Lans, K.Ch.A.M. Luyben, Technologies for butanolrecovery integrated with fermentations, Process Biochem. 27 (1992) 61–75.

[227] M.F. Doherty, M.F. Malone, Conceptual Design of Distillation Systems,McGraw-Hill, New York, 2001. p. 383.

[228] W.L. Luyben, Control of the heterogeneous azeotropic n-butanol/waterdistillation system, Energy Fuels 22 (6) (2008) 4249–4258.

[229] X. Hu, J. Yu, H. Liu, Liquid–liquid equilibria of the system 1-(2-hydroxyethyl)-3-methylimidozolium tetrafluoroborate or 1-(2-hydroxyethyl)-2,3-dimethylimidozolium tetrafluoroborate + water + 1-butanol at 293.15 K, J.Chem. Eng. Data 51 (2006) 691–695.





















http://dx.doi.org/10.1007/s10450-013-9576-6

http://dx.doi.org/10.1007/s10450-013-9576-6



















1-s2.0-s1383586614003608-main

Documents

butanol dehydration

methylimidazolium fap

methylimidazolium ipa

methylimidazolium btese

methylimidazolium hohmim

codvb pf6

youyan liu b

university of minnesota