reverse membrane bioreactor: introduction to a new ...research review paper reverse membrane...

Biotechnology Advances 34 (2016) 954–975

Contents lists available at ScienceDirect

Biotechnology Advances

j ourna l homepage: www.e lsev ie r .com/ locate /b iotechadv

Research review paper

Reverse membrane bioreactor: Introduction to a new technology forbiofuel production

Amir Mahboubi a,b, Päivi Ylitervo a, Wim Doyen b, Heleen De Wever b, Mohammad J. Taherzadeh a,⁎a Swedish Centre for Resource Recovery, University of Borås, 501 90 Borås, Swedenb The Flemish Institute for Technological Research, Vito NV, Boeretang 200, B-2400 Mol, Belgium

⁎ Corresponding author.E-mail address: [email protected] (M.J. T

http://dx.doi.org/10.1016/j.biotechadv.2016.05.0090734-9750/© 2016 The Authors. Published by Elsevier Inc

a b s t r a c t
a r t i c l e i n f o
Article history:Received 11 January 2016Received in revised form 8 April 2016Accepted 25 May 2016Available online 27 May 2016

The novel concept of reverse membrane bioreactors (rMBR) introduced in this review is a new membrane-assisted cell retention technique benefiting from the advantageous properties of both conventional MBRs andcell encapsulation techniques to tackle issues in bioconversion and fermentation of complex feeds. The rMBR ap-plies high local cell density andmembrane separation of cell/feed to the conventional immersed membrane bio-reactor (iMBR) set up. Moreover, this new membrane configuration functions on basis of concentration-drivendiffusion rather than pressure-driven convection previously used in conventional MBRs. These new featuresbring along the exceptional ability of rMBRs in aiding complex bioconversion and fermentation feeds containinghigh concentrations of inhibitory compounds, a variety of sugar sources and high suspended solid content. In thecurrent review, the similarities and differences between the rMBR and conventionalMBRs and cell encapsulationregarding advantages, disadvantages, principles and applications for biofuel production are presented and com-pared. Moreover, the potential of rMBRs in bioconversion of specific complex substrates of interest such as ligno-cellulosic hydrolysate is thoroughly studied.

© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords:Membrane bioreactorReverse membrane bioreactorDiffusionBioconversionInhibitory compoundsSuspended solidFoulingBiofilm

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9541.1. Conventional MBRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9551.2. Cell encapsulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9561.3. Challenges with MBRs and cell encapsulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 959

2. Reverse membrane bioreactors (rMBR): principles and applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9602.1. Inhibitor tolerance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9622.2. Simultaneous sugar consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9642.3. Viscosity and suspended solid content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 964

3. Diffusion in MBRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9663.1. Diffusion in rMBRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 966

3.1.1. Diffusion of compounds on the feed-side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9663.1.2. Diffusion of compounds through the membrane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9693.1.3. Diffusion of compounds on the cell-side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 970

4. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 972Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 972References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 972

aherzadeh).

. This is an open access article under

1. Introduction

Membranes and membrane related technologies have now beenaround for long, attracting the most attention in wastewater treatmentand water quality improvement technologies (Lin et al., 2012; Peters,

the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

http://crossmark.crossref.org/dialog/?doi=10.1016/j.biotechadv.2016.05.009&domain=pdf

http://creativecommons.org/licenses/by-nc-nd/4.0/

http://dx.doi.org/10.1016/j.biotechadv.2016.05.009

mailto:[email protected]

http://dx.doi.org/10.1016/j.biotechadv.2016.05.009

http://creativecommons.org/licenses/by-nc-nd/4.0/

http://www.sciencedirect.com/science/journal/07349750

www.elsevier.com/locate/biotechadv

955A. Mahboubi et al. / Biotechnology Advances 34 (2016) 954–975

2010; Radjenović et al., 2008). The records on industrial scale applica-tion of membranes in water treatment goes back to about 1970 andsince then membranes have found worldwide acceptance in differentengineering processes (Strathmann et al., 2006). Footsteps of mem-brane technology has been tracked in a wide range of applicationsfrom filtration processes to membrane bioreactors (MBR) (Judd andJudd, 2011; Mutamim et al., 2013; Wang et al., 2010; Ylitervo et al.,2013a, 2014). This vast range of membrane applications covers in situproduct recovery in MBRs (Carstensen et al., 2012; Fernandes et al.,2003), agricultural and industrial wastewater treatment, and desalina-tion processes (Alzahrani and Mohammad, 2014; Mutamim et al.,2013; Petrinić and Hélix-Nielsen, 2014; Quist-Jensen et al., 2015;Subramani and Jacangelo, 2015), metal recovery (Mack et al., 2004),oil-water separation (Padaki et al., 2015), etc.

Due to the increasing demand for alternative renewable fuel sources(Nigam and Singh, 2011) to replace depleting fossil fuels (Brown andBrown, 2013) and also to reduce greenhouse gas emissions, there hasbeen a surge of interest to find applicable biofuel production techniqueswith a high productivity (Börjesson et al., 2012; Gnansounou, 2010;Naik et al., 2010). In this regard, in a number of biotechnological appli-cationsmembranes are used to retain cells and/or enzymes inside a bio-reactor (Section 1.1). This may occur through immobilization in amembrane matrix or compartmentalization (Carstensen et al., 2012).There are different benefits sought by utilization of membrane bioreac-tors for biofuel production mainly focused on; the ease of product re-covery as a result of high separation efficiency, high product yield andbiological conversion rate due to high cell concentration, low energy de-mand and ease of operation in continuous mode and others. However,there are limitations in the application of conventional MBRs for biolog-ical treatment of different feed streams. In brief, handling feed sourcescontaining a high concentration of cell inhibitory compounds or severaldifferent prioritized substrate sources by conventional MBR technolo-gies is inefficient. Moreover, feeds with high suspended solid (SS) con-tent are problematic in MBR assisted bioconversions as theydeteriorate membrane functionality through exacerbating cake layerformation and membrane fouling. High SS loading also negatively af-fects cell/medium separation and hinders cell reuse for several batchexperiments.

On the other hand, there are cell retention and immobilization tech-niques such as cell encapsulation that can effectively deal with the is-sues confronted by conventional MBRs. Through cell encapsulation, ahigh local cell concentration is provided in a jelly capsule which sepa-rates the cells from the main bioreactor medium by a permeable mem-brane (Westman et al., 2012b). Thismicroenvironment and cell housingconfiguration gives the cells the ability to tolerate high inhibitor contentand also co-utilize different substrates in the feed (Pourbafrani et al.,2007b; Westman et al., 2012a, 2014a). However, this technique alsocomes with inherent shortcomings. Encapsulating cells is time consum-ing and laborious and simple flaw in capsule preparation and agitationduring application can cause capsule disintegration, rupture and cell es-cape (Sections 1.2, 2.1 and 2.2) (Ishola et al., 2015a, 2015b; Ylitervoet al., 2011).

Themain goal pursued in this review is to introduce the novel prom-ising technology of a reverse membrane bioreactor (rMBR) and its po-tential application in biotechnological processes. The rMBR is acombinational technique merging conventional MBR and cell encapsu-lation techniques. In this regard, rMBR provides the opportunity tohave a membrane bioreactor system functioning on basis of cell encap-sulation principles, benefiting from the advantages of both technologieswhile simultaneously covering their individual shortcomings and oper-ational limitations. Through this technology, bioconversion of complexfeed streams containing inhibitors, multi-substrates and high SS canbe efficiently handled in large scale applying a diffusion driven rMBR.The rMBRs are submerged membrane modules housing microorgan-isms in between membrane layers to provide high local cell density, in-stead of having them freely suspended in the medium as for the

conventionalMBRs. rMBRs function on the basis of diffusivemass trans-fer as for cell encapsulation. High local cell density and the diffusive na-ture of mass transfer in rMBRs, opens new horizons to biologicaltreatment of complex substrates for biofuel production.

1.1. Conventional MBRs

The centre of focus throughout this review is membrane-assistedcell retention. This technique uses a selective syntheticmembrane to re-tain cells and specific chemical compounds in the bioreactor whileallowing some lowmolecular weight solutes (depending onmembraneproperties) to diffuse freely through the membrane (Tampion andTampion, 1987). Membrane applications are generally based on theability of the membrane to efficiently separate different compoundsand/or cells/particles, being selectively permeable to some substanceswhile retaining others. In this context compounds are divided in twogroups, i.e. the compounds that pass through the membrane end up inpermeate (also called filtrate), the ones that are retained in theretentate. The selective behaviour of different membranes originatesfrom membrane pore size and morphology, and other characteristicssuch as membrane charge, affinity or hydrophobicity (Judd and Judd,2011). Membrane separation mainly occurs through application ofpressure and/or concentration gradient as the separation driving forceover the membrane (Judd and Judd, 2011) (Fig. 1). This is a criterionfor categorizing membrane systems on basis of the separation drivingforce into pressure or diffusion (concentration gradient) driven.

In biological processes where membranes are integrated with themain bioreactor either for filtration, product recovery, or cell separationor retention, the MBR configuration plays a determining role. As men-tioned by Judd and Judd (2011), MBR configuration covers both the in-tegration of the membrane with the bioreactor and also the set-up ofthe membrane module in relation with the bioreactor. In general theconfiguration of various conventional MBRs sits under one of the twocategories of immersed (iMBR), also known as submerged MBR, andside-stream (external loop) sMBR (Fig. 2). The submerged membranemodule in iMBRs can be submerged either in the bioreactor or in a sep-arate compartment connected to the main reactor through an externalloop (Carstensen et al., 2012; Judd and Judd, 2011). Considering systemenergy balance, in comparison to sMBRs, iMBRs aremore energy-savingas the module is placed in the bioreactor. In contrast, the sMBR set-uprequires pumping of great medium volumes through an external mem-branemodule housing in a cross-flow filtration system (Hai et al., 2013;Radjenović et al., 2008). Profound reviews of MBR principles and appli-cations and also the differences between iMBR and sMBR in perfor-mance, operation and application are well covered in reviews byCarstensen et al. (2012), Ylitervo et al. (2013a), Judd and Judd (2011)and (Judd, 2008).

Regarding cell positioning in conventional MBRs, in iMBRs cells arekept inside the main bioreactor in a mixture with the feed medium,while in sMBRs cells are pumped through the external membranemod-ule and then recirculated back to the main bioreactor. The ability ofMBRs in retaining high cell concentrations in the bioreactor facilitatesthe in situ product recovery in biofuel production (Carstensen et al.,2012; Ylitervo et al., 2013a). Several examples of final cell concentra-tions (cell biomass) achieved by applying differentMBRs for bioethanolproduction are presented in Table 1.

Conventionally, both sMBR and iMBR processes work based on theapplication of pressure difference (over-pressure or under-pressure).These pressure driven MBRs have long been in use for a wide range ofapplications from wastewater treatment to ethanol fermentation(Carstensen et al., 2012; Judd and Judd, 2011; Ylitervo et al., 2013a;Yoon, 2015). In the sMBRs filtration or product recovery happensthrough pumping the cultivationmediumover and parallel to themem-brane surface through amembrane compartment/unit,where permeateis withdrawn, set in an external loop to themain bioreactor (Carstensenet al., 2012). On the other hand, iMBRs have the membrane module

Fig. 1. Membrane separation due to a) pressure and b) concentration gradient.

956 A. Mahboubi et al. / Biotechnology Advances 34 (2016) 954–975

immersed in the main bioreactor or an external compartment havingthe filtration or metabolite/product recovery in place by application ofunder-pressure. The immersed and side-stream configurations havebeen long taken into practice in various continuous, fed batch andbatch fermentation processes (Carstensen et al., 2012; Judd and Judd,2011). In order to have an overview of MBR assisted bioprocesses, ex-tended examples of the applications of iMBR and sMBR in ethanol fer-mentation processes taking into account the type of microorganismand substrate used and final retained biomass content are presentedin Table 1. However, biofuel production through bioconversion of com-plex substrates containing inhibitory compounds, high suspended solid(SS) content and several different sugar sources by means of the newrMBRs technology is yet to be explored.

Fig. 2. Different MBR configurations: a) Submerged MBR, b

1.2. Cell encapsulation

In order to benefit from a high bioconversion rate and productivity,in addition to optimisation of the process conditions such as pH andtemperature, maintaining a high cell density in the bioreactor is ofgreat importance (Westman and Franzén, 2015). In this regard, variousapproaches of natural (e.g. flocculation) and artificial cell immobiliza-tion (e.g. cell encapsulation and the application of MBRs) have beentaken into consideration in order to have enhanced productivity andmaintain high cell concentration in the bioreactor while increasing thesubstrate feeding rate. Cell immobilization can happen by natural cellimmobilization throughwhich cells tend to form flocs and start to settleor float in the bioreactor. In addition, cells can also be artificially

) external submerged MBR and c) external-loop MBR.

Table 1An overview of the ethanol fermentation processes involving submerged or external-loop MBRs.

Membrane bioreactordesign

Membrane configuration/quality Pore size/MWC Microorganism Biomass(g/l)

Productivity/productconcentration

Working volumeor reactor size(l)

Medium(g/l)

Reference

External cross-flow Polyamide 0.02 μm Zymomonas mobilisATCC4126

40 120–200 g/l.h – Glucose 120–150 Lee et al. (1980)

External cross-flow Short hollow fibre cartridge(Romicon)

50,000 MWC Kluyveromyces fragilisNRRL 2415

90 240 g/l.h 0.5–3 l Lactose 150 Cheryan and Mehaia (1983)

External cross-flow Flat cellulose acetate sheet Sartocon(Sartorius)

20,000 MWC Saccharomyces cerevisiaeH1022

90 44 g/l.h 2.5 l Glucose 182 Hoffmann et al. (1985)

External cross-flow Carbon coated zirconium oxide 0.14 μm Saccharomyces cerevisiae7013

300 33 g/l.h 2.4 l Glucose 150 Lafforgue et al. (1987)

External cross-flow Flat (Pellicon-cassetts Millipore) 0.2 μm Zymomonas mobilis ZM6 18.5 49.5 g/l.h 3.5 l Glucose 100 Hoffmann et al. (1988)Capillary (Enka HB 3355E) 0.3 μm

External cross-flow Ceramic 0.2 μm Zymomonas mobilis ZM4(ATCC 31821)

– 75 g/l 5 l Glucose 100 Chun and Rogers (1988)Fructose 100

External cross-flow Hollow fibre polysulfone(A/G Technology)

30,000 MWC Saccharomyces cerevisiae 160 31 g/l.h 0.25–0.5 Whey permeate(85% lactose)

Mehaia and Cheryan (1990)

External cross-flow Tubular polyphenylenephthalamide 25,000 MWC Saccharomyces cerevisiaeSL100

10–60 15 g/l.h 0.15–0.33 Sucrose 50–300 Melzoch et al. (1991)

External cross-flow Ceramic 0.45 μm Kluyveromyces marxianusY-113

109 6.2 g/l.h 0.75 Lactose 64 Tin and Mawson (1993)

External cross-flow Multi-tubular ceramic 0.05 μm Saccharomyces cerevisiae – 33.2 g/l 3 Molasses Kaseno and Kokugan (1997)External cross-flow Tubular ceramic (Membralox 19P19-40) – Saccharomyces cerevisiae 60–100 11.5 l/l 7000 Corn starch hydrolysate

Corn steep liquorEscobar et al. (2001)

External cross-flow Durapore filter (Millipore) 0.45 μm Saccharomyces cerevisiaeATCC 9658

12 – – Lignocellulosic hydrolysate Brandberg et al. (2005)

External cross-flow Tubular ceramic (Orelis Rhodia TAMI) 150,000 MWC Saccharomyces cerevisiaeCBS 8066

59–156.89 41–65 g/l 1.5–4.5 Glucose 500 Ben Chaabane et al. (2006)

External cross-flow Tubular polyethylene (PE) 0.2 μm Saccharomyces cerevisiaeCCUG 53310

6.4 31.1 g/l 0.3 Softwood hydrolysate Ishola et al. (2013b)

External cross-flow Sartorius 0.45 μm Saccharomyces cerevisiaeNCIM 3090

~20.1 55 g/l 5 Rice straw hydrolysate Zahed et al. (2015)

External cross-flow Polyethylene (PE) (Microdyn-Nadir) 0.2 μm Saccharomyces cerevisiaeT 0936

6 35 g/l 0.6 Wheat strawhydrolysate slurry

Ishola et al. (2015b)

External (PF) Composite plate polydimethylsilicon – Saccharomyces cerevisiae 4.49–5.37 0.692–1.90 g/l.h 4 Glucose 20–70 Ding et al. (2012)

(continued on next page)

957A.M

ahboubietal./BiotechnologyAdvances

34(2016)

954–975

Table 1 (continued)

Membrane bioreactordesign

Membrane configuration/quality Pore size/MWC Microorganism Biomass(g/l)

Productivity/productconcentration

Working volumeor reactor size(l)

Medium(g/l)

Reference

(PDMS)/poly amide (PA)External (PF) Poly dimethylsiloxane 20,000 MWC Alcohol active dry yeast

(ADY)19.8 1.51 g/l.h 5 Glucose 43 Chen et al. (2012)

Submerged Ceramic Al2O3 0.3 μm Saccharomyces cerevisiae 58 13 g/l 1.5 Glucose 100 Park and Kim (1985)Submerged Stainless steel tubes 2, 10 μm Saccharomyces cerevisiae

ATCC 2485850–150 20 g/l.h 1 Glucose 100 Chang et al. (1993)

Submerged Ceramic Al2O3 5 μm Saccharomyces cerevisiae 207 1–4 g/l.h 2 – Suzuki et al. (1994)Submerged Stainless steel tubes 2 μm Saccharomyces cerevisiae 42 17 g/l.h 1.8 Tapioca hydrolysate Lee et al. (1994)

208 14.7 g/l.h Glucose 100Submerged Asymmetric ceramic α-Al2O3/β-Al2O3 0.005/0.3 μm Saccharomyces cerevisiae

ATCC 248580–60 40–61 g/l 1.8 Glucose 100 Zhang et al. (1998b)

Submerged Ceramic Al2O3 0.2 μm Saccharomyces cerevisiaeK 901

236 13.1 g/l.h 1.5 Glucose 100 Ohashi et al. (1998)

Submerged Ceramic tube 0.3 μm Saccharomyces cerevisiae 1.5 × 109 cell/ml 16.9 g/l.h 1.5 Lignocellulosic hydrolysate Lee et al. (2000)Submerged Fluoro polymer 2 μm Saccharomyces cerevisiae 15 – 0.14 Molasses Thuvander (2012)Submerged

(External chamber)Integrated Permeate Channel(IPC) (Vito NV)

0.3 μm Saccharomyces cerevisiaeCBS 8066

N50 7.94 g/l.h 0.6 Lignocellulosic hydrolysate Ylitervo et al. (2014)

Submerged Hollow fibre cartridge (ZeeWeed) 0.08–0.1 μm Saccharomyces cerevisiae(Baker's yeast)

30 62 g/l 1 Glucose 100–175 Radoaj and Diosady (2014)

Submerged (PF) Hollow fibre Rubber Silicon Silicalite – Saccharomyces cerevisiaeNRRL-Y-2034

8.10 7.3 g/l.h 0.42 Glucose 100 Cho and Hwang (1991)

Submerged (PF) Hollow fibre – Saccharomyces cerevisiae ~14.6 11.4 g/l.h – Glucose Kargupta et al. (1998)Submerged (PF) Silicalite-Silicon Rubber – Saccharomyces cerevisiae

(baker's yeast)10 – 0.15 Glucose 200 Ikegami et al. (2002)

Hollow fibrefermenter

Poly sulfone 10,000 MWC Saccharomyces cerevisiaeATCC 4126

3.5 × 109–1010 cell/ml 17–26 g/l.h – Glucose 120–150 Inloes et al. (1983)


Hollow fibre (Romicon) 50,000 MWC Saccharomyces cerevisiaeNRRL-Y-132

260 17 g/l.h 0.625 Glucose 100 Mehaia and Cheryan (1984)


Poly sulfone 10,000 MWC Saccharomyces cerevisiaeATCC 4126

– 133 g/l.h – Glucose 100 Inloes et al. (1985)


Polyamide 10,000 MWC Saccharomyces cerevisiaeATCC 24858

6 10 g/l.h 0.04 Glucose 100 Park and Kim (1985)


Polypropylene Celgard X-20 – Saccharomyces cerevisiaeNRRL-Y-132

100–300 31.6 g/l 0.075 Glucose 300 Kang et al. (1990)

PF: pervaporation fermentation, MWC: molecular weight cut-off.

958A.M

ahboubietal./BiotechnologyAdvances

34(2016)

954–975


immobilized either through entrapment in gel matrices (gel capsules orbeads) or retention in MBRs, generally referred as cell encapsulation(Fig. 3) (Tampion and Tampion, 1987). The main compounds involvedin gel formation are agar, alginates, collagen, kappa-carrageenan, aga-rose, chitosan and polyacrylamide (Tampion and Tampion, 1987). Oneof the highlighted cell immobilization techniques that provides biocon-version and fermentation processes with exceptional functional fea-tures is cell encapsulation (Westman et al., 2012a, 2012b, 2014a). Theprincipal aim of cell encapsulation is to provide very high local cell con-centration within a capsule. This high local cell concentration and thesubstrate concentration gradient present in capsules, from the capsulesurface to the core, give the entrapped cells the ability to co-consumedifferent substrates. This is also followed by extraordinary performanceof encapsulated cells in medium inhibitor tolerance and detoxification(Westman et al., 2012a, 2012b). The dominant mass transfer methodin cell capsules is diffusion as direct convection does not apply insidethe capsules. The diffusion behaviour of different compounds also con-tributes to the superior characteristics observed through fermentationusing encapsulated cells (Westman et al., 2012a, 2012b). The principleof cell encapsulation, that is providing high local cell density throughcell confinement by an external membrane while having diffusion asthe dominant mass transfer mode, is used as the backbone of therMBR system.

1.3. Challenges with MBRs and cell encapsulation

Diverse techniques of cell retention and immobilization such asmembrane cell recycling and retention, and cell immobilization throughencapsulation and flocculation have been applied in bioreactors aimingmainly at obtaining higher productivity and bioconversion rates, andbenefiting from the ease of product separation from cells and cellreuse (Carstensen et al., 2012;Westman et al., 2012b). As high cell con-centration is achieved and cell washout prevention is assured throughthe application of MBRs, the bioreactor can run in continuous mode athigh dilution rate and low hydraulic retention time. However, success-ful membrane-assisted cell retention and/or cell immobilization bycell encapsulation does not always guarantee a successful fermentation.

It is to be considered that the condition of the utilized feed (sub-strate) also determines the outcome of the MBR bioconversion process.

Fig. 3. Bioreactor with e

For example fermentation of media containing inhibitory compounds(furan aldehydes, carboxylic acids, etc.) can cause problems since toxiccompounds may affect themicroorganism's physiological andmetabol-ic activity in a negative way. Utilization of high suspended solids (SS)viscous substrates and feed streams containing different types of sugars(pentoses, hexoses, etc.) is still a great hurdle (Klinke et al., 2004). Con-ventional MBRs lack the potential to positively enhance the cell inhibi-tor detoxification ability and simultaneous sugar utilization potentialof cells. Although high cell concentration is provided in the MBR, cellsare suspended in themain reactor and exposed to uncontrolled concen-trations of toxic compounds and various sugar sources. The abovemen-tioned issues are unfavourably confronted when the purpose is toproduce second generation ethanol from lignocellulosic materials(Bertilsson et al., 2008; Klinke et al., 2004; Laluce et al., 2012). Due totheir recalcitrant structure comprised of lignin, hemicellulose and cellu-lose, lignocellulosic materials show great resistance to enzymatic hy-drolysis (Taherzadeh and Karimi, 2008). Therefore, to have the sugarsreleased and prepared for bioconversion by the microorganism the lig-nocellulosic materials should be pre-treated (Taherzadeh and Karimi,2008). The pre-treatment stage iswhere inhibitory and toxic compoundsuch as furan aldehydes, phenolic compounds and carboxylic acids areproduced (Almeida et al., 2007; Klinke et al., 2004; Taherzadeh et al.,1997; Taherzadeh and Karimi, 2008; Zaldivar et al., 2001) hindering fer-mentation. These inhibitory compounds suppress ethanol fermentationby increasing the lag phase, decreasing cell viability, stopping biocon-version by inhibition of catabolic enzymes, decreasing intracellular pH,disturbing cell membrane integrity, etc. when directly in contact withfreely suspended cells in the medium (Almeida et al., 2007). Moreover,saccharides released by means of lignocellulosic pre-treatment consistof pentoses (xylose, arabinose, etc.) and hexoses (glucose,mannose, ga-lactose, etc.) extracted mainly from hemicellulose and cellulose respec-tively, with the extent depending on the type of lignocellulosic source(softwood, hardwood, etc.) (Taherzadeh and Karimi, 2008). The actualproblem occurs as wild ethanol fermenting microorganisms have poorperformance in co-consumption of hexose and pentose sugars. This re-sults in initial hexose utilization followed by pentose consumption oncethe hexose is depleted (Sànchez Nogué and Karhumaa, 2015; StanleyandHahn-Hägerdal, 2010b). In general, due to the abovementioned fac-tors, the bioconversion rate of lignocellulosic substrates to ethanol is

ncapsulated cells.


low when freely suspended cells showing diauxic growth are in simul-taneous contact with several sugar sources (Chandrakant and Bisaria,1998). As cells in both iMBR and sMBR systems are suspended in thefeed there is actually no accurate control over cell sugar consumptionand inhibitor in situ detoxification.

Recently, increased inhibitor tolerance and co-utilization of differentsugars have been successfully investigated by cell encapsulation andflocculation (Westman et al., 2012a, 2014b). These techniques providemicroenvironments with high local cell density inwhich the concentra-tion gradient and diffusive mass transfer of compounds assist in inhibi-tor detoxification and co-utilization of sugar sources. However,preparing these microenvironments and their maintenance in workingcondition prevents them from extensive industrial use. The cell encap-sulating process is time consuming (Ylitervo et al., 2011). An additionalissue as reported by Ishola et al. (2015b), is that occasionally incompletexylose consumption occurs while using encapsulated cells. Further-more, the capsules can easily undergo rupture and break during thepro-duction process or in the bioreactor due to agitation (Ishola et al.,2015a). Capsule breakage may occur at different stage of fermentation.The cells that have escaped from the capsule may become suspended inthe medium or attach to the exterior capsule wall and start to consumeglucose faster in non-inhibitory mediums (Westman et al., 2014a). Thisproblemhas been foreseen and covered in rMBR as the cells are encasedbetween synthetic membrane layers that were prepared separately andthen inoculated with grown Saccharomyces cerevisiae.

Another obstacle requiring engineering solutions is the highsuspended solid content of specific fermentation feeds such as lignocel-lulosic hydrolysate. Lignocellulosic material, depending on the plantspecies (ex. softwood and hardwood), contains different percentagesof lignin (10–35%) (Zhao et al., 2012). Lignocellulose pre-treatment pro-cesses that aid sugar release and also enhance enzymatic accessibility ofthe lignocellulosic raw material result in production of lignin residueswith non-fermentable polymeric compounds (Zhao et al., 2012).These residual lignin particles bring along several problems. Accumula-tion of residual lignin in batches adds to themedium viscosity of the lig-nocellulosic hydrolysate slurry due to high suspended solid content(Verardi et al., 2012; Zhao et al., 2012). It has been claimed that up to30% w/w solids loading in the pre-treated hydrolysate is required toguarantee an acceptable ethanol concentration (4–5 wt%). However,rising the solid loading in a hydrolysis and fermentation process

Fig. 4. Schemes of a) conventional immerse MBR and b) re

increases the viscosity of the medium causing mass transfer limitationsin pressure driven MBRs deteriorating membrane performance by cakelayer formation and consequently MBR failure (Section 2.3). Moreoverhigh SS concentration hinders enzymatic performance, increasing theinhibitory effects of intermediate compounds and decreasing the easeof mixing of the broth (homogeneity) (Sassner et al., 2006; Verardiet al., 2012). In addition, in case cell reuse is pursued for consecutivebatch processes, this increase in solid residues increases the numberof stages and cost of downstream processes for the separation of cellsfrom solids both in MBR and cell encapsulation systems. When itcomes to MBR fermentation the SS level of the feed is of critical impor-tance as it may contribute to the membrane fouling propensity. As pre-sented in the work by Liu et al. (2015) high SS broth with elevatedviscosity reduces the effect of the shear stress induced by air bubbleson the surface of the membrane required for fouling prevention. Fur-thermore, cake layer formation by solid particles is exacerbated as theSS content increases (Judd and Judd, 2011). Adding to that is the solidresidual accumulation in the bioreactor if the process is to be run for re-peated batches (Galbe and Zacchi, 2002). In the cross-flowsMBRs, in ad-dition to cake layer formation, a high viscosity increases the energyrequired for pumping the fermentation broth from the bioreactorthrough the membrane module in a closed loop (Ishola et al., 2013b).However, an increase in viscosity can be advantageous in specialcases. Ishola et al. (2013a) proved that increasing the SS loading from8 to 12% in simultaneous saccharification and fermentation (SSF) of anon-sterile lignocellulosic hydrolysate reduced the bacterial contamina-tion activity and increased the ethanol yield.

2. Reverse membrane bioreactors (rMBR): principles andapplications

The reverse membrane bioreactor (rMBR) is a recently introducednovel immersed membrane configuration used for production ofbiofuels such as methane and ethanol (Ishola et al., 2015a;Youngsukkasem et al., 2015). The principal difference between the con-ventional iMBR and the rMBR process is that, in the latter the cells areimmobilized in between membrane layers (Fig. 4b), separated fromthe actual feed medium, whereas for the conventional iMBR the cellsare suspended in the medium in direct contact with the feed (Fig. 4a).In the novel rMBR process cells are encased between synthetic

verse MBR (arrows present the direction of the flow).

Table 2General difference between rMBRs and conventional MBRs.

Conventional MBRs rMBRs

Configuration(s) Immersed or external cross-flow ImmersedMain separationdriving force

Pressure gradient Concentration gradient

Mass transfermechanism(s)

Convection and diffusion Diffusion

Cell/feed medium Cells mixed with the feed medium Cells separate from thefeed medium

Cell positioning Freely suspended in the mediumon reactor shell side

Encased betweenmembrane layers

Product recovery Through the permeate (filtrate) Mixed with feed onreactor shell side


membrane layers in the form of membrane sachets, compact multi-layer membrane columns, Integrated permeate channel (IPC) flatsheet membranes (Doyen et al., 2010) and other membrane encase-ment configurations (Fig. 5). General differences between rMBRs andconventional MBRs are presented in Table 2.

In contrast to pressure driven submerged and side-streamMBR pro-cesses, in rMBR, active liquid permeation has been replaced by substratediffusion through the membrane to the cell side and in the opposite di-rection for the metabolic products. In rMBRs, the synthetic membraneplays a similar role as the plant cells membrane, separating the mem-brane confinedmedium (cytoplasm in plant cells) and cell interior com-ponents (Golgi, mitochondrion, nucleus, etc. in plant cells) from thesurrounding medium, only letting specific nutrients to pass throughdue to the concentration gradient over the membrane. Moreover, inrMBRs a syntheticmembrane plays the same role as themembrane cap-sule in encapsulation. A thorough analysis of the similarities in princi-ples and functions of cell encapsulation and rMBR is provided inSections 2.1 and 2.2. Membrane compartments housing microorgan-isms in rMBRs could be suspended (floating) in the reactor as closedmembrane sachets (Youngsukkasem et al., 2013a), or be fixed in placesuch as compact multi-layer membrane columns (Youngsukkasemet al., 2013b) or flat sheet membrane panel (Ishola et al., 2015a)(Fig. 5). In order to benefit from rapid bio-methanation of syngas andco-digestion of syngas and organic substances, Youngsukkasem et al.(2015) successfully used closed sachets made of flat plain PVDF(polyvinylidene fluoride) membrane sheets to entrap methanogenicbacteria in an rMBR. Syngas is a gasmixturemainly composed of carbondioxide, carbon monoxide and hydrogen formerly made from

Fig. 5. Different rMBR setups used for production of biogas: a) membrane sachet bioreactor, b)panels.

controlled combustion of coal, biomass, etc. in presence of steam thatcan be used as chemical precursor for other chemical processes. In thisexperiment the sludge encased in the sachetswas capable of convertingthe fed syngas comprised of H2, CO2 and CO into biogas in a short reten-tion time of 1 day. It was reported in the study that thermophilic condi-tions (55 °C) and co-digestion using membrane sachets contributed tohighermethane yield (Youngsukkasemet al., 2015). In another attempt,the effect of organic loading rate (OLR) on the performance of methan-ogenic freely suspended and encased (in sachet) bacteriawas examined(Youngsukkasem et al., 2013a). It was observed that the bacteriaentrapped in sachets were still viable and active at an OLR of15 g COD/l.day−1, whereas at a loading rate of 7.5 g COD/l.day−1 thesystem including freely suspended cells totally failed. Also membrane sa-chets (Fig. 5a) used in rMBR configuration prevented specific inhibitory

fixed compact multi-layer membrane column bioreactor, and ethanol: c) IPC membrane


compounds to come in contact with the acting microorganism(Wikandari et al., 2014). This effect also applies to the hydrophilic capsulemembrane in cell encapsulation (Pourbafrani et al., 2007a). As studied byWikandari et al. (2014), in the case of D-limonene, known as a potentialinhibitor for the bio-methanation process, hydrophilic PVDF membranesachets were impermeable to the naturally hydrophobic D-limonene,while glucose and volatile fatty acids present in the medium penetratedthrough themembrane layer to the cell side. Robust methane productionin an inhibitory medium by cells encased in membrane sachets in anrMBR has also been experienced by Youngsukkasem et al. (2013a) andthe results were compared with that of freely suspended cells.

Another newly developed rMBR configuration for biogas production isthe compact multi-layer membrane column (Fig. 5b). In this concept, anumber of double layer steel fixtures including membranes on bothfaces, are packed and submersed in a column bioreactor. The liquidmedi-um, either as the substrate or as the carrier of syngas or biogas (methaneand carbon dioxide), flows upward through the column, and allows theexchange of the substrates and products along the membrane surface ofthe membrane packs housing the microorganism. This type of rMBRwas used for biogas production by Youngsukkasem et al. (2013b),where they examined the performance of different membranes of hydro-phobic polyamide 46 (PA), hydroxyethylated polyamide 46 (HPA) andPVDF regarding cell retention and diffusive transfer of compounds.

It is noteworthy that in the abovementionedbiogas production setups(Fig. 5a and b) a semi-rMBR system exists as the feed is being recirculatedin the bioreactor. In these conditions there is liquid and/or gas feed flowover the exterior surface of the membrane sachets and packed layer.Therefore, mass transfer on the feed side is assisted by convection andalso diffusion in a thin stagnant liquid layer on the surface of the mem-brane. However, on the interior surface of the membrane (cell side) it isonly diffusion that dominates the transfer of compounds from the mem-brane to cells and in the opposite direction for the products.

Regarding bioethanol production, in a recent research work Isholaet al. (2015a) used integrated permeate channel (IPC) flat sheet mem-brane panels (Doyen et al., 2010) for housing recombinant S. cerevisiaecells in an rMBR set up (Fig. 5c). In this regard, the functionality of therMBR techniquewas evaluated in bioethanol fermentation from the liq-uid fraction of wheat straw hydrolysate containing different sugarsources and inhibitory compounds. It was reported that the IPC rMBRconfiguration had a significant positive effect of simultaneous utilizationof xylose and glucose, and in situ detoxification of furfural and hydroxy-methyl furfural (HMF). The result indicated complete consumption ofglucose and 87% utilization of xylose by the yeast leading to an ethanolyield of 83% of the theoretical yield. Although biogas production bymeans of different rMBRs has recently been a focus of exploration, theapplication of rMBR in bioethanol production is relatively an undiscov-ered area of research yet to be explored.

The scientifically and technologically newly developed rMBR concept,representingmembrane bioreactors performing on basis of diffusionwithcells encased in betweenmembrane layers, has a great potential to be ex-perimentally explored for biosynthesis of biochemical (organic acids,propanol, butanol, etc.) and biofuels (bioethanol, biogas and biodiesel)form feed streams with complex mixture of sugars and inhibitors suchas lignocellulosic material, citrus waste and other carbohydrate-richwaste streams. In the following Sections 2.1–2.3, the principles of rMBRand its applications in increasing cell inhibitor tolerance, simultaneousconsumption of different sugar sources and dealing with high suspendedsolid feeds are elaborated in details. Also the advantages and shortcom-ings of rMBRs in comparisonwith the conventionalMBRs, cell encapsula-tion and flocculation are thoroughly studied.

2.1. Inhibitor tolerance

One of the potential areas where rMBR technology can be effectivelyapplied is the bioconversion of inhibitory feed streams. In this sectionthe mechanism by which the rMBR set up can improve the cells

inhibitor tolerance and in situ detoxification is discussed in details.Among themain issues confronted during the fermentation of complexsubstrates such as lignocellulosic hydrolysate and citrus waste tobiofuels, specifically bioethanol, is the presence of inhibitory com-pounds in the medium. These inhibitors are either convertible throughin situ detoxification (ex. furan aldehydes) or non-convertible under an-aerobic condition (ex. phenolic compounds) by the acting microorgan-ism(s). The main industrially used microorganism for ethanolfermentation is the yeast Saccharomyces cerevisiae. S. cerevisiae haslong been themicroorganism of choice in ethanol production industriesas it has high capacity of ethanol fermentation along with high ethanoltolerance (Casey and Ingledew, 1986; Ghareib et al., 1988; Stanley et al.,2010a; You et al., 2003). Moreover, some strains of yeast are capable ofin situ detoxification and conversion of some inhibitory compoundssuch as 5-hydroxymethyl furfural (HMF) and furfural into less toxicchemicals such as furfuryl alcohol (Li et al., 2009; Liu, 2011; Tian et al.,2011). However, high inhibitory level in the feed and low cell densityin the bioreactor decreases the success level of in situ detoxification. Awell-developed bioreactor housing high yeast cell concentration andlow toxicity level of the feed streamhas great propensity to conduct fer-mentation alongwith in situ detoxification. In order to provide themen-tioned condition during fermentation different remedies can be takeninto practice e.g.; using a MBR (Ylitervo et al., 2013b) or cell encapsula-tion (Westman et al., 2012a) for higher cell density, running fermenta-tions in fed-batch mode to control the inhibitory level (Taherzadehet al., 1999), engineering of recombinant cells with higher inhibitor tol-erance and ability to convert inhibitors (Koppram et al., 2012).

Encapsulated yeast can effectively ferment inhibitor-containing lig-nocellulosic hydrolysates that are considered extremely toxic for a free-ly suspended cell system (Talebnia and Taherzadeh, 2006). Cellencapsulation provides high local cell concentration that increases theinhibitor tolerance of the cells as claimed by Westman et al. (2012a).The rMBRby retaining cells in a confined space betweenmembrane sur-faces simulate that of encapsulation condition (Ishola et al., 2015a).There are several differentmechanisms contributing to better in situ de-toxification and higher inhibitor tolerance by encapsulated yeast or bycells retained in an rMBR. These remedial mechanisms owe their in-creased inhibitor tolerance to manipulated cell stress and mass transferpatterns in the cell proximity due to high local density.

Cells subjected to encapsulation or retained in an rMBR, are inten-tionally kept in a limited space for the sake of high local cell density.The yeast cell concentration in capsules can increase up to 309 g/l ofcapsule volume (Cheong et al., 1993). As a result of high local cell den-sity, these cells experience stress inducers such as low nutrition levelor nutrition starvation. Cells located deeper into the cell aggregate areexposed to a lower level of nutrients such as glucose that is requiredas carbon and energy source due to mass transfer and diffusion limita-tions. In addition, it has been reported that the stress implied during en-capsulation is accompanied by counter stress responses by the cellsthrough expression of stress related genes (Klinke et al., 2004;Taherzadeh and Karimi, 2008). It has been observed by Sun et al.(2007) and Talebnia and Taherzadeh (2007) that the level of trehalose,involved in cell stress tolerance, rises in encapsulated cells. This initialstress response induces a protective effect and gives the cells the abilityto withstand the upcoming stress by the inhibitory compounds. In awork by Ylitervo et al. (2011) it has been proven that this initial stressresponse could also help cells to better withstand thermal stresses.This reasoning supports the idea that the inhibitor tolerance does notnecessary occur in cell encapsulation by the capsule membrane butwith cells tightly kept together (Westman et al., 2014b). The aboveput forward discussion forms the basics of cell inhibitor tolerance inrMBR cell confinement (Fig. 6b).

Other techniques of cell immobilization that pursue the goal of keep-ing cells together such as flocculation also result in an increased inhibi-tor tolerance. However, in the case of flocculation, high sugar content ofthe medium, for example in batch fermentations, blocks flocculation.

Fig. 6. The schematic fig. on medium sugar and inhibitor profile and their effect on cell metabolism in an rMBR system: a) Non-inhibitory medium with the presence of a single hexosesugar source (glucose): rapid glucose utilization by yeast cells closer to membrane surface and sugar deprivation for cells placed further from the membrane deep into the cell colony.b) Inhibitory medium with the presence of a single hexose sugar source (glucose): in situ inhibitor (convertible inhibitors) detoxification by cells adjacent to the membrane, whilecells deeper in the cell aggregate are involved with glucose consumption in a sub-inhibitory condition. c) Non-inhibitory medium with the presence of hexose (glucose) and pentose(xylose) sugar sources: cells in regions closer to the membrane favorably are involved with glucose consumption as they experience high concentrations of glucose and xylose, wherecells positioned further from the membrane in glucose-deprived regions utilize xylose as the available sugar source.


When the lectin-like cell wall proteins (flocculins) that have themissionto attach to carbohydrate residues (e.g. mannose) on the cell wall ofneighbouring cells are blocked by the saccharides present in the feed,flocculation does not occur until the end of fermentation. For examplefor mannose there are both reports on flocculation inhibition and en-hancement (Marić and Vraneš, 2007; Westman et al., 2014b). This isalso considered as a benefit in breweries as it gives better product recov-ery efficiency by having yeast coagulated in flocs. It is noteworthy thatordinary lab yeast strains are poor or non-flocculants due to defectiveor missing genes required for flocculation. Although strong flocculationis accompanied by mass transfer resistance and slower glucose con-sumption in non-inhibitory medium compared to free suspendedcells, strong flocculation in the inhibitor containingmediumhelpswith-standing higher ethanol and toxic compound level and also faster glu-cose and mannose consumption compared to suspended cells.Westman et al. (2014b) studied ethanol fermentation through biocon-version of spruce dilute acid lignocellulosic hydrolysate by three recom-binant mutants of S. cerevisiae having different flocculation strength, allbeing originated from the strain CEN.PK 113-7D, and the performancewas compared with free suspended cells. It was reported that strongflocculation has the same effect as cell encapsulation when fermentinglignocellulosic hydrolysate containing different inhibitors (Westmanet al., 2014b).

In addition to starvation stress, an inhibitor concentration gradientin the cell aggregate from the surface cells (directly exposed to highcontent of inhibitor) to cells placed deeper into the cell cluster (exposedto sub-inhibitory concentrations), is the other mechanism contributingto enhanced inhibitor tolerance and process robustness. In capsules,flocs or rMBR systems that contain cell clusters of several millimetresor more in thickness, diffusion limitation defines the concentration ofdifferent nutrients and toxic compounds that cells are exposed to. Ithas been reported that in flocs larger than 100 μm,mass transfer limita-tions lead to low biomass growth and ethanol production (Talebnia andTaherzadeh, 2007).

In these diffusion driven systems, cells closer to the capsule or flocsurface or adjacent to the membrane surface in an rMBR experience ex-tremely harsh inhibitory conditions. In inhibitor containing mediums,the cells at the frontier get involved with the detoxification of readilyconvertible inhibitors such as furan aldehydes, and reduce the amountof inhibitors diffusing deep into the cell cluster leaving the cells in theinterior unaffected and active for fermentation. In this condition, cellsnear the cluster centre benefit from the sugar sources that have notbeen consumed by the surface cells involved in detoxification(Westman et al., 2012a, 2014b). However, as the mass transfer barrier

still exists, having diffusion as the dominant mechanism, a low rate ofcell growth and fermentation occurs near the base/centre/core of thecell cluster. This slow growth near the centre has its own benefits. Ithas been proven that slowgrowing yeast has enhanced stress resistancethat sequentially leads to increased inhibitor tolerance (Elliott andFutcher, 1993). This lower growth rate in the inner cell layers can alsobe justified as the limited amount of sugar that reaches deep into thecluster as utilized as maintenance energy rather than budding andgrowth (Westman et al., 2012a).

A schematic picture of the cell metabolic activity which depends onthe cell location in the cluster andmedium condition in an rMBR is pre-sented in Fig. 6. As it can be seen in non-inhibitory fermentation medi-um, cells close to the membrane surface are involved in sugarconsumption and cells placed further from the surface experience lownutrient concentrations. For the inhibitor containing medium, the cellsat the surface have the role of inhibitor detoxificationwhile cells deeperin the cluster have the opportunity of fermenting the present sugars.

The above mentioned remedies do not protect cells or increase theinhibitor tolerance to non-convertible inhibitors such as carboxylicacids and some phenolic compound present in lignocellulosic hydroly-sate (Westman et al., 2012a). As presented by Vilela-Moura et al.(2011), in anaerobic fermentation the presence of glucose repressesacetic acid metabolism. As predicted, the microorganism utilizes lessglucose in media containing inhibitors than in non-inhibitory ones.This applies to both high local cell density systems such as encapsula-tion and freely suspended cells (Westman et al., 2012a).

On the other hand, in MBRs with freely suspended cells, althoughyeast cells are capable of detoxifying convertible inhibitors, fermenta-tion robustness is not preserved. In this condition, inhibitor toleranceenhancing mechanisms are not active as all the cells in the mediumare exposed to the same inhibitory compound with similar concentra-tion at once (Westman et al., 2012a; Ylitervo et al., 2013b). This is dueto better convectional mass transfer through stirring, gas purging orother agitation methods. Conventional submerged and external-loopMBRs provide desirable medium mixing (good mass transfer) in addi-tion to retaining high concentrations of cells in the bioreactor. As expe-rienced by Ylitervo et al. (2013b), up to 17 g/l of furfural was detoxifiedusing high cell density of 180 g/l by means of cross-flow tubular sMBR.However, in fermentation conditions where cells are not retained (con-siderably low densities), there would be a long lag phase, low biocon-version rate or even in case of high inhibitor concentration,fermentation halts, until all convertible inhibitory compounds havebeen detoxified to less toxic ones. In addition, as suspended cells areconstantly subjected to toxic medium e.g. lignocellulosic hydrolysate,


the length of time that cells can be reused is closely dependent on thelevel of toxicity and cell robustness (Talebnia et al., 2005). As investigat-ed by Westman et al. (2012a) in an inhibitor containing defined medi-um with furan aldehydes and carboxylic acids the rate of glucoseconsumption in the initial stage of fermentation for freely suspendedcells and encapsulated yeast cells were 40 and 80% of that of a non-inhibitorymedium respectively. Taking the abovementioned reasoninginto consideration, the rMBR is the preferred choice over the conven-tional sMBR and iMBR systems when it comes to biological conversionof inhibitory containing complex feeds such as lignocellulosichydrolysate.

2.2. Simultaneous sugar consumption

The othermain issue confronted in fermentation of complex feeds isthe concomitant presence of different sugars for example pentoses(monosaccharides having five carbon atoms) and hexoses (monosac-charides with six carbon atoms). The problem arises from the pointthat the wild-type S. cerevisiae strains either consume pentoses at avery low rate or do not utilize pentoses at all (Sànchez Nogué andKarhumaa, 2015; Van Zyl et al., 1989; Zaldivar et al., 2001). The primeremedy in this regard is having genetically manipulated recombinantS. cerevisiae strains that are capable of pentose (xylose, arabinose, etc.)uptake (Sànchez Nogué and Karhumaa, 2015; Stanley andHahn-Hägerdal, 2010b). However, due to the fact that there are no in-herent pentose transporter proteins in the cell membrane, xylose and/or arabinose transportation to the intracellular space only happens ifthe concentration of hexoses (glucose, mannose, etc.) is low enough(Bertilsson et al., 2008; Hamacher et al., 2002; Meinander et al., 1999;Sedlak and Ho, 2004).

A breakthrough has been made by keeping cells in close proximityfor example in the case of cell encapsulation and application of rMBR(Ishola et al., 2015a; Westman et al., 2014a). Cell encapsulation hasproven to be a successful approach when co-utilization of sugars issought.Westman et al. (2014a) have reported a 220-fold increase in xy-lose consumption rate, 50% more xylose uptake and 15% more ethanolproduction for encapsulated recombinant (genetically modified)S. cerevisiae compared to the same condition with freely suspendedcells. Encapsulated cells also showed 7% more ethanol production in alignocellulosic hydrolysate medium (containing inhibitors) than thesuspended cells. Moreover, Ishola et al. (2015a) experienced a success-ful co-utilization and fermentation of glucose and xylose present in lig-nocellulosic hydrolysate to 83% of the theoretical ethanol yield. Thisarises from the diffusive nature of mass transport in cell encapsulationand also other fermentation processes benefiting from high localizedcell density such as rMBR. The diffusion pattern, somehow, has thesame pattern as described for the inhibitor concentration gradient(Fig. 6b). In this condition when cells are packed together either in acapsule or encased in betweenmembrane layers in a reversemembranebioreactor, depending on the cells location from the surface of the cap-sule or inside syntheticmembranes in an rMBR, are exposed to differentnutrient levels and possess diverse cell physiology (Westman et al.,2012b). The cells close to the surface experience high hexose (in specif-ic, glucose) concentrations, which is in the first priority of consumptionfor recombinant cells. In this condition pentoses (in specific, xylose) dif-fuse through the cell layers to cells placed in deeper positions. Cells closeto the core of the cell aggregate are in the state of glucose starvation,while xylose content is high in those regions. Therefore, these cells areinvolvedwith xylose fermentation as there is noor very little glucose in-hibition (Fig. 6c). In cell encapsulation, the glucose conversion rate forcells near the capsule surface is limited by the diffusion ratewhereas xy-lose utilization rate by cells in deeper layers is limited by the consump-tion and reaction rate (Westman et al., 2014a). In conventional MBRs,retained and recycled recombinant freely suspended S. cerevisiae inthe bioreactor are exposed to both hexoses and pentoses simultaneous-ly. In this condition, there is very low or no xylose consumption before

glucose is completely depleted in the medium (Chandrakant andBisaria, 1998). In case xylose is the only carbon source in the fermenta-tion medium the rate of xylose consumption in both freely suspendedcell system and encapsulated yeast cells is the same (Westman et al.,2014a; Westman and Franzén, 2015). The other issue to be dealt with,when considering co-utilization of sugars in freely suspended cell sys-tems, is slow xylose uptake after total glucose depletion (Kuyper et al.,2005; Westman et al., 2014a). The slow metabolism of xylose after fullconsumption of glucose in the medium may be due to lack of interme-diary metabolites required for pentose metabolism and the relatedphosphate pathway, and severe redox imbalance (Ha et al., 2011).After glucose depletion cells have to quickly adapt themselves with adrop in NADPH and NAD+ due to anaerobic consumption of xylose(Ha et al., 2011). Shortly after total glucose consumption, xylosemetab-olism stops in suspended cell cultures (Westman et al., 2014a). Howev-er, this adaptation happens smoothly for cells in the inner layers of cellclusters in encapsulation and rMBR as they are exposed to xylose fromthe beginning of the fermentation process.

In addition to the put forward advantages of an rMBR system in si-multaneous sugar consumption, the consequences accompanying thelow pace of mass transfer should also be considered. In case of poor orslow sugar transfer through the microbial aggregate, starving cellsnear the centre of the aggregate go through the stationary and conse-quently death phase. In order alleviate such conditions, circulation ofcells and the feed medium in- and outside the membrane package canbe considered as an option. In order to have a pure diffusion dominatedmass transfer the circulation pace should induce the same amount ofpressure drop on both sides so that pressure equality conditions areapplied.

2.3. Viscosity and suspended solid content

One of the main concerns in membrane filtration, membrane-assisted product recovery and in general membrane bioprocesses driv-en by pressure is to maintain a reasonable permeate flux through themembrane in order to benefit from high productivity of permeate or fil-trate (Carstensen et al., 2012). However, membrane flux deteriorationdue to different fouling mechanisms such as cake layer formation andconcentration polarization is a common phenomenon hindering theprocess efficiency (Park et al., 1997). Concentration polarization as de-fined by Judd and Judd (2011) is the concentration of rejected soluteand precipitation of poorly soluble inorganic polymericmacromolecularcompounds on the surface of the membrane, while cake formation isknown as accumulation of rejected solids on the membrane. This maybe exacerbated by reaching the criticalflux or having changes in theme-dium condition, as in the case of increase in medium viscosity. The con-cept “critical flux” was introduced by Field et al. (1995) as the fluxbelowwhich reduction in fluxwith time does not take place. Fluxes ex-ceeding the critical flux lead to membrane fouling. In conventionalMBRs, the flux of permeate through the membrane is negatively affect-ed by the viscosity of themedium, in case all other factors such as trans-membrane pressure (TMP) (the average of feed pressure minuspermeate pressure) and membrane resistance are kept constant(Yoon, 2015). Additionally, the viscosity of the feed medium is in directrelationship with the biomass (cell) and solid content of the liquid me-diumand determines the fouling propensity, the flux through themem-brane and the gas/air bubble size of pressure driven submerged andside-stream MBRs (Lee and Yeom, 2007; Wicaksana et al., 2006). Asnoted by Sarbatly and England (2004) for starch hydrolysis in sMBR,starch hydrolysis process the starchmilkmass has to be kept at aweightconcentration of 10% in order to control viscosity rise and prevent foul-ing. Furthermore, increase in the viscosity of the medium reduces theeffectivity of shear stress induced by gas sparging on the fouling resis-tance (van den Brink et al., 2011). Itonaga et al. (2004) have concludedthat the changes in viscosity level for feed water in MBR municipalwastewater treatment is marginal up to a certain suspended solid (SS)


concentration,while beyond that, the relationship becomes of the expo-nential type.

The diffusive behaviour of mass transfer in rMBRs is directly affectedby changes in the rheological properties of the medium. In diffusiondriven rMBRs the suspended solid content of the medium is less prob-lematic in comparison with the conventional pressure driven processas there is no need for a convectional flux of compounds through themembrane. This property of rMBRs avoids membrane function deterio-ration due to SS related problems such as cake layer formation on themembrane surface. Therefore, increase in suspended solid content ofthe medium although detrimental to the diffusion rate of compoundsin the feedmedium, does not causemembrane fouling by cake layer for-mation. As rMBR bioconversion systems are strongly dependent on thediffusion of various solutes through the medium and membrane to thecell side and in reverse for metabolic products, viscosity and the param-eters influencing it such asworking temperature andmediumsolid con-tent require profound investigation.

The suspended solid content of the medium plays an important rolein defining the viscosity of the medium (Hai et al., 2013). In pressuredriven MBR systems there seems to be a complex relation betweenthe fouling susceptibility and the SS concentration of the medium(Judd and Judd, 2011; Yoon, 2015). It has been reported in the literatureinvolved with wastewater treatment that, the increase in the mixed li-quor suspended solid (MLSS) can have negative, imperceptible oreven positive (in cases with low initial SS content) impact on permeateflux (Chang andKim, 2005; Çiçek et al., 1999;Hong et al., 2002; Le-Clechet al., 2003; Lesjean et al., 2005; Rosenberger et al., 2005; Yoon, 2015).Through practical experimentation, different methods of predictingthe relation between the SS content and permeate flux through themembrane have been proposed. Most of these equations have specificconditions applied for deriving the equations leading to their limited ap-plicability for all MBR systems as some parameters are considered inone and disregarded in other (Table 3) (Fang and Shi, 2005; Krauthand Staab, 1993; Sato and Ishii, 1991; Shimizu et al., 1996). A compre-hensive list of these approaches has been provided by Judd and Judd(2011). Specific to each MBR, there is an upper limit to the SS contentof the medium. In SS concentrations exceeding that of the limit, contri-bution of SS to fouling becomes problematic (Meng et al., 2007; Yoon,

Table 3Equations on the relationship between the flux through the membrane and suspended solid co

Equation Variables

J=J0 ⋅ek((MLSS−MLVSS)Re′/MLVSS) J = specific flux rate (m3·m−2·hJ0 = 0.00001072 PTM−95.854 (m3·mRe′ = modified Reynolds numbePTM = Transmembrane pressureω = flow rate (m·s−1)D = membrane tube diameter (mρ′ = density of the activated sludη′ = dynamic viscosity of the actMLSS − MLVSS = inorganic solidMLVSS = volatile solids (kg·m−

Rt=ΔP/μJ=Rm+Rp+Rc Rt = Filtration resistance of activΔP = the trans-membrane pressμ = the viscosity of the permeateJ = the permeation flux (m·s−1)Rm = the intrinsic membrane resRp = the pore fouling resistanceRc = the cake layer resistance

Jss ¼ VL ¼ K0 � ϕ � u0:1 �MLSS0:5

Jss ¼ ΔP=η f � ðRm þ Rp þ RcÞJss = steady-state filtration flux (K′ = filtration constant, defined bMLSS = dried sludge concentratiu = feed velocity (m·s−1)ϕ = geometric hindrance coefficηf = viscosity of filtrate (Pa.s−1)VL = lift velocity (m·s−1)

R=842.7 ΔP(MLSS)0.926(COD)1.368(η)0.326 R = Filtration resistance (m−1)ΔP = the trans-membrane pressCOD = Chemical Oxygen Deman

2015). The different behaviour in membrane fouling in high and lowSS media as concluded by Bin et al. (2004) may be due to the pace offouling.Where for the former, rapid cake layer formation occurs leadingto failure in filtration process, in the latter case there is thought to be aprogressive process of colloid formation and pore blocking (Bin et al.,2004). Even in low pressure submergedMBRs, deterioration of filtrationperformance has been observed by an increase in MLSS (Jeison and vanLier, 2006; Stuckey andHu, 2003). Although there is nounanimous viewon the extent of contribution of SS tomembrane fouling, high SS contenthas direct effect on diffusion limitation, viscosity increase and creatingdead zones in the reactor as mentioned by Yoon (2015).

Another factor contributing to changes in the viscosity is the cellconcentration/density in the bioreactor (Reuß et al., 1979; Shimmonset al., 1976). However, the increase in cell density-viscosity is not linear(Shimmons et al., 1976). As presented by Bhave and Todaro (1996), incross-flow filtration on fermentation broth, the flux across the mem-brane declines by the increase in yeast cell concentration. In addition,in sMBRs and iMBRs used either for fermentation or wastewater treat-ment purposes, cell retention is of a great importance as cell concentra-tion defines hydraulic retention time, dilution rate, the bioreactor sizeand the productivity of the process (Carstensen et al., 2012;Lafforgue-Delorme et al., 1994). High cell concentration due to cell re-circulation and retention inMBRs contributes to the increase inmediumviscosity and may consequently lead to reduction in medium flow orfouling (Lafforgue-Delorme et al., 1994; Lafforgue et al., 1987; Yoon,2015). Puzanov (1999) proved that in lactic acid fermentation in a cellrecycling MBR, at cell densities above 130 g/l the rheological behaviourof the fermentation broth changed fromNewtonian to pseudoplastic. InNewtonian liquids viscosity is independent of the flow rate, where forpseudoplastics the viscosity decreases with increased shear stress(Krieble and Whitwell, 1949).

One of the parameters affecting both conventional MBR and rMBRprocesses is the working temperature. Decrease in temperature in-creases the viscosity of Newtonian fluids (Kumaran, 2010). The rule ofthumb in this regard is that with every degree increase in temperaturein normal conditions the viscosity drops about 3%. As claimed byKumaran (2010), the more viscous a fluid is the more temperature de-pendent the viscosity becomes.

ntent of the medium.

References

−1) Krauth and Staab (1993)−2·h−l)r = ωDρ′/(10−3 η′)(bar)

)ge (kg·m−3)ivated sludge (m·Pa−1·s−1)s (kg·m−3)

3)ated sludge (m−1) Fang and Shi, 2005ure gradient (N·m−2),(N·s−1·m−2)

istance

m3.m-2.d-1 or m.s–1) Shimizu et al. (1996)y equations (kg0.5·m−1.5)on (mixed liquor suspended solids) (kg·m−3)

ient of membrane module

Sato and Ishii, 1991ure gradient (N·m−2)d (g·l−1)


In another way, changes in temperature can affect the behaviour ofparticles in the fluid near the membrane surface. Increase in tempera-ture increases the kinetic energy of themolecules constituting the liquidmedium resulting to a more significance in Brownian movement of theparticle in the medium. Considering the Brownian diffusion ofsuspended particles near the membrane surface in MBRs (Yoon,2015), by reducing the temperature, the particle back transport velocitydecreases in a linear manner which results in the extent of cake layerformation (Judd and Judd, 2011). The velocity of particles/solutes depo-sition on the membrane surface is determined considering the differ-ence between permeation velocity/drag (flux), towards themembrane, and back-transport velocity away from the membranecaused by Brownian diffusion, shear-induced diffusion, charge repul-sion, etc. (Fig. 7) (Yoon, 2015).

From the biological aspect, temperature reduction and elevation islimited to a range that does not affect the biological activities of themi-croorganisms (Folasade, 2006; Singh and Viraraghavan, 2003).

3. Diffusion in MBRs

The role of diffusion in the general mass transfer pattern in conven-tional pressure driven MBRs greatly differs from that of the rMBRs. Inpressure drivenMBR processes, diffusion through themembrane occursas a result of the pressure difference applied over themembrane leadingto permeation or retention of certain compounds on basis of differentfactors such as size and polarity. The most highlighted role that masstransfer by diffusion plays in these conventional MBRs is in a delicateliquid film on the membrane surface in external cross-flowmembranes(Yoon, 2015). The velocity of the liquid at the membrane surface dropsto zero, therefore in the concentrated liquid film formed on the mem-brane surface diffusion of compounds is defined as the main mode ofmass transport in the layer (Miranda and Campos, 2002). The thicknessof the concentration polarization layer is greatly dependent on cross-flowpatterns and turbulence intensity. However, a greater extent of sol-ute build-up on themembrane surface is accompanied by a higher con-centration gradient and consequently faster diffusion (Bhattacharjeeet al., 1999; Wijmans et al., 1985).

On the contrary, diffusion is the main mass transfer mechanism inrMBRs. The substrate and products are transported by diffusion in and

Fig. 7. Forces applied to a spherical charged particle in a viscousmediumnear a flat poroussurface in a laminar flow (Yoon et al., 1999).

out of the membrane-confined space. It is noteworthy that, in specialrMBR cases shear stress on the membrane/feed-side surface also in-duces higher diffusion rate (Youngsukkasem et al., 2013b, 2015). Thefollowing is a brief description of the diffusion process and its mecha-nism in rMBRs.

3.1. Diffusion in rMBRs

The diffusion related phenomena form the basis of the rMBR systemand play a major role in cell inhibitor tolerance and co-utilization ofsugars. In an rMBR set up there are three major phases of feed bulk,membrane and cell aggregate which directly affect the diffusion behav-iour of different compounds by acting as resistant to easy diffusion. Eachof the three regions acts similar to a resistant in an electric circuit thatslows down the electron transfer through the circuit at different de-grees. Therefore, in order to find out the stage acting as the bottleneckregarding the diffusion rate and compound flow, possible diffusion pat-terns in each region have to be studied profoundly. The diffusion rate ofdifferent compounds (substrate and metabolites) through the feed-side, membrane and cell-side defines the bioconversion rate. In this re-gard, the study of diffusionmechanisms in an rMBR system and also fac-tors hindering the diffusion of compounds to and from the cells is ofgreat importance. Diffusion of chemical compounds in feed-side, mem-brane and cell-side (Fig. 8) depends on medium viscosity, membranecharacteristics (pore size, polarity, etc.) and cell biofilm respectively.

The diffusion of feed medium chemical compounds e.g. saccharidesand inhibitory compounds and metabolic products in rMBR are consid-ered at three separate phases (Fig. 8):

I. Diffusion of compounds on the feed-side (substrate) towards themembrane surface and in reverse for products.

II. Diffusion of compounds, either substrates or metabolites,through the membrane layer.

III. Diffusion of feed components and products (through the biofilmlayer) on the cell side.

3.1.1. Diffusion of compounds on the feed-sideThe first and last stage of diffusion in an rMBR system is the transfer

of substrates and metabolites to and from the membrane surface to thebulk feed-side respectively (Fig. 8). The diffusion of chemical com-pounds (solute) in a liquid medium occurs as a result of the presenceof a concentration gradient. The further a system is from concentrationhomogeneity and equilibrium, the stronger is the concentration driving

Fig. 8. Diffusion stages of substrates (sphere) and products (star) in an rMBR set up.

Table 4Relevant equations for measuring the diffusion behaviour of compounds in different stages of an rMBR system.

Diffusion stage Equation Reference

Feed-side

Fick’s first law(steady state)

JA ¼ −DABdCA

dxð1Þ

DAB = diffusivity of compound A in Bx = diffusion distancecA = concentration of AJA = molar flux (diffusion rate) of compound A

_____

Fick’s second law(non-steady state)

dCA

dt¼ DAB

∂2CA

∂x2

!ð2Þ

t = time interval

_____

Equimolar counter-diffusion

JA ¼ − JB ¼ DABdCA

dxð3Þ

JA = molar flux (diffusion rate) of compound AJB = molar flux (diffusion rate) of compound B

(American Society ofHeating, 2005)

Maxwell-Stefan diffusion

−xART

dμA

dx¼Xn

B¼1;B≠A

xBNA−xANB

ctDABð4Þ

–dμA/dx = chemical potential gradientN = molar fluxct = total molar concentrationx = molar fraction of componentsT = temperatureR = universal gas constant

(Leonardi andAngeli, 2010)

Stokes-Einstein equation

DAB ¼ RTNA

1

6πηruð5Þ

NA = Avogadro numberT = temperatureη = viscosity of the solutionru = solute radiusVA = solute volume at the boiling point

(Einstein, 1905, Miller,1924, Sharma andYashonath, 2006)

Wilke-Chang equation

DAB ¼ 7:4� 10−8 xMBð Þ1=2TηV0:6

A

ð6Þ

x = association parameter defining the effective molecular weight of thesolvent (for non-associated solvents x = 1 and for water x = 2.6)MB = molecular weight of the liquid medium

(Wilke and Chang, 1955)

Polson equation

DAB ¼ C

VAMAð Þ1=3; C ¼ R3T3

162 π2NA2η3

!13 f 0

f

� �ð7Þ

MA = molecular weight of soluteR = gas constantf = frictional constant per mole and for a spherical unhydrated molecule thefrictional constant is f0C = constant and its value can be calculated from diffusion constantand molecular weight data of proteins with low f/f0 values reported bySvedberg et al. (1940).

(Polson, 1950)

Membrane

Diffusion of solute in porous solid(Fick’s first law)

Z h

0

Jdx ¼ −Z c2

c1Ddc; J ¼ −

1

h

Z c2

c1Ddc; J∝

1

hð8Þ

h = membrane thickness

Steady state diffusion through themembrane(no membrane hindrance)

JA ¼ ε D cA1−cA2ð Þτ x1−x2ð Þ ð9Þ

Steady state diffusion through themembrane(with membrane hindrance)

JA ¼ ε Deff cA1−cA2ð Þτ x1−x2ð Þ ð10Þ

Deff ¼ D 1−dsdp

� �m

ð11Þ

D = diffusivity of the solute in the liquid in the poresDeff = effective diffusivity hindered by membrane/solute interactionsds = solute diameterdp = pore diameterc = concentration of the compoundx = distance over which diffusion is being measuredm = constant (as presented by Beck and Schultz (1972) for an isoporousmembrane m is approximately 4)Deff = D (When the pore diameter is significantly bigger than the solutediameter)

(Beck and Schultz, 1972,Cussler, 2009)

(continued on next page)


Table 4 (continued)


Diffusion in porous membrane(Derivation of Fick’s first law of diffusion)

JA ¼ DAKh

cA ð12Þ

K = membrane-vehicle partition coefficientc = solutes concentration in solutionh = diffusion distance

(Pellett et al., 1997)

Cumulative compound release(diffusion cell)

Q ¼ cnV þXn−1

i¼1ciS

� �=A ð13Þ

A = membrane surface areacn = concentration of the compound at the nth samplingV = volume of the diffusion cell receptor∑ci = sum of the concentrations till n-1 samplingS = sampling volume

(Jung et al., 2012, Ng et al.,2010)

DAB ¼ V δ τ2ε A t

; lnC01−C0

2

C1−C2

!ð14Þ

V = cell volumeδ = membrane thicknesst = timeC = concentration of solute at different run intervals (at t=0, C0) in cells 1 and2. Here cell 1 is considered as the donor cell and cell 2 as the receptor cell.

Diffusivity in membrane(double diffusion cell) ln

C f2−C f

1

� �C02−C0

1

� �24

35 ¼ Dβt ð15Þ

β ¼ Aδ ð 1

V1þ 1

V2Þ ¼ constant

A = membrane areaβ = constantCf = final concentration of soluteC0 = initial concentration of solute- Sides 1 and 2 are donor and receptor cells respectively.

(Singh et al., 1996)

Cumulative diffusion(double diffusion cell)

Q ¼ A DC0

δt−

δ2

6D

!ð16Þ

t0 ¼ δ26D

Q = total amount of solute diffused by time tC0 = initial concentration of the solute in the donor cellδ = membrane thicknesst0 = lag phase (time required for reaching diffusive equilibrium)

(Bassi et al., 1987, Hannounand Stephanopoulos, 1986)

Cell-side

The relationship between temperature,viscosity and diffusion in water

constant ¼ Dw:μT

ð17Þ

Dw = diffusivity in waterμ = viscosityT = temperature

(Stewart, 2003)

Porosity of a biofilm

ε ¼ 1−ρdw

ρdð18Þ

ε = porosityρd = dry biomass to the volume of dry materialρdw = dry mass per unit wet volume

(Melo, 2005)

Time for the solute to attain about 90% of itsconcentration over the biofilm (t90) t90 ¼ 1:03

l2

Deffð19Þ

Time for the solute to attain about 90% of itsconcentration in the centre of the biofilm(Cell cluster) (t90c)(reaction and adsorption of compound inthe biofilm are not considered

t90c ¼ 0:37R2

Deffð20Þ

Deff = effective diffusivity in the biofilmR = cell cluster radiusl = biofilm thickness

(Stewart, 2003)

Penetration depth of a solute in a biofilm(for compounds reaction in the biofilm)

a ¼ 2Deff :SpK0

� �1=2

ð21Þ

K0 ¼ μ XYxs

μ ¼ μ ;maxS

Sþksɑ = penetration depthμ = specific growth rateks =half-saturation coefficient (value of substrate when μ is half μmax)S0 = solute concentration at the biofilm interfacek0 = reaction rate of the soluteX = density of cells in the biofilmμmax = maximum growth rateYxs = yield of biomass on the solute

(Stewart, 2003)


Table 4 (continued)


Product (metabolite) concentration inbiofilm regions that there is substrate(solute) depletion

P ¼ DesYpsS0Dep

þ P0 ð22Þ

P = product concentration in the biofilmP0 = concentration of the product in the bulk liquidYps = yield of product on substrateS0 = concentration of the substrate at the biofilm surfaceDes = effective diffusion of the substrateDep = effective diffusion of the product in the biofilm

(Stewart, 2003)


force. In order to reduce the concentration difference, molecules of thesolutemove towards the low concentration area by randommovement.In this regard, the diffusion rate (molar flux) can be defined by Fick'sfirst and second law (Table 4, Eqs. (1) and (2)). Fick's law relates the dif-fusion rate (J) to the diffusivity (diffusion coefficient) (D) of a com-pound in a medium. Diffusivity is defined as the ease with whichmolecules of a solute move in a medium. In the conditions where mol-ecule/molecule and/or molecule/membrane forces are high D acts as avariable (Poling et al., 2001). For the condition that there is one dimen-sional diffusion and D is independent of distance (x), time (t) and con-centration (c), Fick's laws are presented in Table 4.

In case a medium is comprised of two components of A and B, equi-molar counter-diffusion and Maxwell-Stefan diffusion models can alsobe applied to track diffusion patterns of compounds. Equimolarcounter-diffusion describes the condition in which components A andB are diffusing in opposite directions with the same molar flux, whileMaxwell-Stefan diffusion model can be applied when a component dif-fuses and the other remains or in the case that there ismulti-componentdiffusion (Leonardi and Angeli, 2010) (Table 4, Eqs. (3) and (4)).

A number of parameters such as solute molecular volume andweight, liquid viscosity and process temperature affect the diffusivityof a compound in a liquidmedium. Regarding the estimation of diffusiv-ity (D) of a compound in a liquid, differentmodels have been developedto correlate solute diffusivity to liquid properties such as Stokes-Einstein, Wilke-Chang and Polson equations (Table 4, Eqs. (5), (6) and(7)).

Therefore, in rMBR systems by the help of the above mentionedmodels an estimate of thediffusion rate of different compounds throughthe feed to the membrane surface and also products from the mem-brane surface to the bulk feed-side can be obtained. For further informa-tion in this regard, a profound discussion on the diffusion of compoundsin liquid medium can be found in the work by Cussler (2009).

3.1.2. Diffusion of compounds through the membraneIn rMBR processes, the porous membrane is considered as a bio-

mass/liquid (cell/feed) contactor (Fig. 8). As noted by Jung et al.(2012), the synthetic porous membrane has the role of a continuouslinking channel. Depending on the membrane quality and productionmethod, membranes possess different porosity, tortuosity, hydrophilic-ity, etc. On basis of these parameters, different diffusion patterns forchemical compounds through the membrane are expected. In rMBRs,penetration of chemical compounds through the membrane is concen-tration gradient dependent. Greater concentration gradients lead tohigher flux of the compounds through the membrane. In the case ofpressure drivenMBRs, this permeation rate is dependent on transmem-brane pressure.

The defining factors during membrane diffusion measurementsare membrane hydrophilicity, physical and chemical properties ofthe compound, interaction of the membrane and compounds, etc.(Jung et al., 2012). In membrane separation techniques such aspervaporation where non-porous or very dense selective mem-branes are used, polymer quality (glass transition temperature,main chain flexibility and side-groups, polymer crystallinity, free

volume, etc.), nature of penetrant compounds (molecular weight,size, shape, etc.) and process condition (temperature, etc.) are ofcritical importance (Berens and Hopfenberg, 1982; Chen et al.,2001; Steingiser et al., 1987; Stern et al., 1987). However, in rMBRsystems porous membranes are utilized, where pores are consid-ered as facile compound diffusion pathways/channels connectingthe two media. In such set ups, after reaching a steady state, asflux becomes stable and does not change with time, according toFick's first law, flux (J) through the membrane will be inverselyproportional to membrane thickness (h) as is presented by Eq. (8)(Table 4). More detailed diffusion models for the diffusion of a sol-ute in a porous solid such as a membrane have been presented thatillustrate the dependence of the flux of the solute to membrane po-rosity (ε) and tortuosity (τ) (Table 4, Eqs. (9) and (10)) (Beck andSchultz, 1972; Cussler, 2009). Another approach to measure thediffusion rate through a membrane has been simply put by Pellettet al. (1997), as a derivative of Fick's first law of diffusion(Table 4, Eq. (12)).

One of themain approaches applied for measuring the diffusion rateof different compounds through a porous membrane is using diffusioncells. Among diffusion cells, Franz cell set up has had extended experi-mental application (Clément et al., 2000; Jung et al., 2012; Ng et al.,2010; Shah et al., 1999). The studies manoeuvring over membrane dif-fusionmeasurements bydiffusion cells havemainly aimed atmeasuringthe diffusion rate of pharmaceutical and cosmetic chemical compounds(Bonferoni et al., 1999; Clément et al., 2000; Hadgraft and Ridout, 1987;Jung et al., 2012;Ng et al., 2010; Shah et al., 1999). Dependingon theop-erationmode and cell configuration, diffusion cells are divided into stat-ic and continuousflowcells (flow-through cell). Static cells are availablein vertical and side-by-side configurations and are used for finite dosepermeation, steady flux of compound and compound uptake into themembrane measurements, whereas flow-through cells come in are tomimic blood vessels, finite and infinite dose permeation (Anon., 2014;Ng et al., 2010). These diffusion cells can be utilized to estimate the dif-fusion rate of different compound through the membrane in diffusiondriven rMBRs. Principally, cumulative amount (Q) of compound re-leased by passing through the membrane in a diffusion cell duringtime is calculated (Table 4, Eq. (13)). The slope of the graph Q versusffiffit

prepresents the release (diffusion) rate of the solute (Jung et al.,

2012; Ng et al., 2010). However, Eq. (13) is valid if only diffusion ofone single compound is being measured, the diffusion coefficient(D) of the compound stays constant and the diffusing compound hasno interaction with the membrane.

In case a side-by-side (double cell) diffusion cell is being used for themeasurement of the diffusivity in membranes, an alternative approachis taken as in Eq. (14) (Table 4). Singh et al. (1996) used a side-by-side diffusion cell to measure the diffusivity of silver bromide (AgBr)through a glass membrane to purify lead bromide (PbBr2). They mea-

sured the diffusivity (D) from the slope of ln ½ðC f2−C f

1 ÞðC0

2−C01Þ� versus time plot

(Table 4, Eq. (15)). However, unlike Eq. (14), the effects of porosityand tortuosity have been neglected for the glass membrane used inthis study. In another work, Bassi et al. (1987) measured the effectivediffusion coefficient (D) of lactose and lactic acid through a 3% agarose


membrane (pore size 0.1–1.0 μm) in a side-by-side diffusion cell usingthe formula developed by Hannoun and Stephanopoulos (1986)(Table 4, Eq. (16)).

3.1.3. Diffusion of compounds on the cell-sideThe most important stage of diffusion that is the backbone of many

of the applications of rMBRs, is diffusion in the cell-side (Fig. 8). InrMBRs the cells packed together in between membrane layers providehigh local cell density. The diffusion rate and behaviour of compoundsin the cell-side and through the imposed cell aggregate determine cellinhibitor tolerance and in situ detoxification, and co-utilization of sugars(Westman et al., 2012a, 2014a, 2014b). As bioconversion happens in themembrane-confined cell-side, diffusion rate of substrates and productsin this region plays a crucial role in defining the bioconversion rate andprocess yield. The difference in diffusion of compounds in cell-side incomparisonwith diffusion in feed-side andmembrane is that in the for-mer condition two phenomena of mass transfer and reactions occur si-multaneously. Therefore, diffusion patterns and models presented forthe cell-side (biofilm side) take into account the effect of reaction aswell as effective diffusion. In this regard, as reported by Melo (2005)there seems to be a two way effect as the physical structure of the cellaggregate or biofilm depends on the internal mass transfer and thestructure affects the diffusion characteristics of compounds. In order tobetter understand the role of biofilm and the effect of biofilm formationon the diffusion of compounds on the cell-side in an rMBR system, abasic knowledge on biofilms nature, formation and prevention isrequired.

Biofilms in general are micro-colonies of microorganisms attachedto a surface and embedded in a gel like extracellular polymeric sub-stance (EPS)matrix (Fig. 9). Biofilm is the dominant life form of bacteriaguaranteeing their survival in comparison to free floating cells (Marićand Vraneš, 2007). These biofilms are comprised of cells and a networkof EPS with water channels in between the cell colonies responsible forwater, nutrient and metabolite delivery in and out of the biofilm. Theextracellularmatrix is a combination of 90%water and 10% extracellularpolymeric substances (EPS) (Rendueles et al., 2013).

In addition to bacteria, different strains of yeast such as Candidaalbicans, Candida glabrata, S. cerevisiae, etc. also have the capability toadhere to abiotic surfaces e.g. membranes, cells and tissues (Short,2011; Verstrepen and Klis, 2006). Special proteins on the yeast cell

Fig. 9. Biofilm formation

wall surface are responsible for the attachment to other cells and alsoabiotic surfaces are called “adhesins”.

EPS play an essential role in biofilm formation accounting for, de-pending on the type of themicroorganism, from 75 to 90% of the biofilmwith the rest being composed of cells (Marić and Vraneš, 2007). Theterm EPS (high-molecular weight mixture of polymers) is noted tocover a wide range of macromolecules such as proteins, carbohydrates,nucleic acids, lipids, etc. on and around the cell exterior wall (not allnecessarily directly anchored to the cell wall) covering colonies ofcells and maintaining the stability of cell aggregate in the biofilm(Flemming andWingender, 2001; Sheng et al., 2010). EPS mainly com-prises of proteins and carbohydrates. Carbohydrates, having more hy-drophilic tendency, are mainly responsible for cell adhesion to themembrane.

3.1.3.1. Fouling due to biofilm formation inMBRs.Different parameters in-fluence the EPS content in conventional pressure driven MBRs process,to name some: shear stress near themembrane surface by gas spargingand other means which induce turbulence, the substrate condition re-garding nutrient content and organic loading rate. For example, theyeast strain C. albicans is said to formmore EPS in an agitated cultivationmedium than in a stationary condition (Henriques, 2005). In wastewa-ter treatmentMBRs, solid retention time (SRT) is of great importance forthe EPS concentration in the medium (Brookes et al., 2003).

As mentioned, membrane fouling is a major issue in MBR processes.The soluble part of the EPS, defined as solublemicrobial products (SMP),is adsorbed and deposited on themembrane surface forming a gel layerblocking pores and the flow through the membrane and providing thebasis for the formation of a biofilm on the membrane (cell/cell attach-ment of the microorganisms and EPS matrix formation) (Rosenbergeret al., 2005; Vanysacker et al., 2013). In a study by Vanysacker et al.(2013) fouling of PVDF, polysulfone and polyethylene membranes byEscherichia coli and Pseudomonas aeruginosa biofilms were investigated.It was observed that the more hydrophilic a membrane is the less bio-film associated fouling occurs in an aqueous solution.

The problem with biofilm formation in conventional MBRs and spe-cially in rMBRs is that, the EPS gel housing the cells can be a major ob-stacle to nutrient delivery (the dominant mechanism of nutrientdelivery is diffusion through water channels in the EPS matrix) to cellsand cell assisted bioconversion and also a barrier for permeation

in an rMBR system.


through membrane in pressure driven processes. Considering the roleof EPS in cell/cell attachment, in someMBR systems, disruption and de-crease in EPS content is assumed to negatively affect MBR performance,however, this has not been proved (Jang et al., 2005).

As supported by Li et al. (2014) biofilm related fouling and internalpore blocking in MBRs only requires chemical cleaning. Removal ofthe biofilm in down time of the MBR is also a hurdle as chlorine, bacte-ricides and other cleaning-in-place methods (CIP) do not necessarilywork efficiently in EPS removal as they mainly affect the bacteria(Armstrong et al., 2011).

3.1.3.2. Biofilm prevention. As the presence of biofilm can hinder the per-formance of both the conventional MBRs and rMBRs there is need for indepth studies on preventive methods and approaches. In order to facil-itate diffusion of different compounds through themembrane layer andto the cell membrane surface, it is essential to keep the extent of biofilmformation at low levels. In an rMBR system, cells are intentionally keptin a close proximity in-betweenmembrane layers. This induces the pro-pensity of biofilm formation in rMBR systems. However, this biofilm for-mation is undesirable as it limits the mass transfer by diffusion only tothe water channels present in the EPS matrix. Different environmentalfactors such as nutrient and oxygen content, pH and temperature ofthemedium can affect biofilm structure and the extent of biofilm forma-tion. In this regard, the effect of phosphate concentration on biofilm for-mation has been investigated by Kim Kwang and Frank (1995). Aspresented by Kim Kwang and Frank (1995) the presence of trehaloseand mannose enhances biofilm formation. Furthermore, it has beenclaimed that, glucose availability has a direct impact on biofilm thick-ness (Marić and Vraneš, 2007). In addition, oxygen content, tempera-ture and pH of the medium affect cell adhesion. Although somebacteria can form biofilms at acidic conditions, considerably low medi-um pH reduces cell mobility required for initiating adhesion (KimKwang and Frank, 1995). Low medium oxygen content and high tem-perature are said to be detrimental to cell/surface adhesion and cell/cell connection, respectively (Kim Kwang and Frank, 1995). Metal ionconcentration of themedium is also a defining factor to be taken into ac-count as EPS binds with the cell wall through building ion bridgingwithmultivalent metals such as Ca+2 and Mg+2. Therefore, less metal con-tent leads to less EPS resulting in less onerous biofilm (Sheng et al.,2010).

All in all, in order to have healthy functioning rMBR system the bio-film formation should be controlled at an acceptable low level. The pur-sued reduction in biofilm formation tendencymay be achieved throughdifferent approaches:

1. Cell/cell communication blockage: To initiate a cell/cell attachmentand cell/surface adhesion, cells should communicate through signal-ling molecules, blocking this communication path may be a solutionfor preventing biofilm formation (Marić and Vraneš, 2007).

2. EPS enzymatic hydrolysis: In order to break down and disrupt a bio-film layer, enzymatic hydrolysis and degradation of extracellularpolymeric matrix is another applicable option (Marić and Vraneš,2007).

3. Cell secreted anti-biofilm polysaccharides: Another option is pre-sented by the microorganism as there are different sorts of bacteriathat secrete anti-biofilm polysaccharides blocking the proteins onthe cell wall that attach to sugars on other cell surfaces and inhibitbiofilm formation (Rendueles et al., 2013).

4. Genetic inactivation: repression of genes involved in EPS formation,cell/cell and cell/abiotic surface attachment is another approach toalleviate biofilm related issues (Marić and Vraneš, 2007; Verstrepenand Klis, 2006).

5. Environmental stress factors: Environmental stress factors such asthe change in the medium carbon and nitrogen content e.g. nutrientstarvation, pH and ethanol content can trigger the yeast cell adapta-tion technique from adhesive to non-adhesive and vice versa

(Verstrepen and Klis, 2006). However, the behaviour is still to bestudied as for example different glucose concentrations in the medi-um can play roles in both enhancing and hindering adhesion(Verstrepen and Klis, 2006).

6. Cell cultivation condition: In order to reduce the amount of EPS bycontrolling the process conditions, it has been claimed by Shenget al. (2010) that aerobic cultivation conditions lead to a large con-tent of EPS, while anaerobic conditions reduces EPS production andsometimes lead to disintegration of the biofilm (Judd and Judd,2011).

3.1.3.3. The effect of biofilm formation on diffusion in rMBR. It has been re-ported that the dry density of the biofilm is in close relationshipwith ef-fective diffusivity as increase in one negatively affects the other(Beyenal and Lewandowski, 2002; Fan et al., 1990; Stewart, 1998,2003). However, there are still uncertainties in generalization of suchmatter (Casey et al., 2000; Zhang et al., 1998a). Biofilms are composedof 90–99%water. The water channels (macro-pores) of different length,diameter and tortuosity have the responsibility of carrying compoundsin and out of the biofilm. There are macro- and micro-pores in betweencell/EPS and inside cell/EPS aggregates respectively (Melo, 2005). Asbiofilms aremostlywater, the first step in estimation of the effective dif-fusivity in biofilms is to have the diffusivity of different compounds inwater (Stewart, 2003). The data on diffusion of many compounds inwater is readily available. However, missing calculations for unknowndiffusivities in water can be made through different models presentedin the previous sections such as Wilke-Chang (Perry and Chilton,1973; Wilke and Chang, 1955). It is noteworthy that, as mentioned be-fore, the diffusivity inwater, aswell as othermedia, depends on temper-ature and viscosity (Stewart, 2003). The relationship betweentemperature, viscosity and diffusion in water is presented in Eq. (17)in Table 4.

Cells are concentrated and packed in very close proximity in be-tween membrane layers resembling an imposed biofilm. The highlycompact biofilmmatrix reduces the diffusivity of compounds due to in-creased tortuosity. This means that due to the presence of different ob-stacles, such as cells, EPS, abiotic particles and gas bubbles in the biofilm,compounds are forced to take longer diffusion paths than when havingfree unrestricted diffusion (Stewart, 2003; Zalc et al., 2004). Therefore,the effect of tortuosity (τ) and porosity (ε) of the biofilm should be con-sidered while measuring the effective diffusivity ðDeff ¼ εDw

τ Þ of com-pounds in a biofilms (Dw is the diffusivity of the compound in water).According to Melo (2005) the porosity of a biofilm can be calculatedfrom density values (Table 4, Eq. (18)).

Diffusion is limited in biofilms by the diffusion distance and decreasein fluid flow (in case of bulk flow) (Stewart, 2003). According to Stewart(2003), the time required for a diffusive equilibrium in biofilms is propor-tional to the square root of the diffusion distance. Moreover, the time forthe solute added to a liquidmedium, containing a biofilm, to attain about90% of its concentration over (t90) and in the centre of the biofilm (Cellcluster) (t90c) can be estimated (Table 4, Eqs. (19) and (20)).

These equations have been built assuming that the solute does notreact or is not absorbed in the biofilm, the transfer of compounds tothe biofilm surface occurs with no resistance and the biofilm thicknessis uniform along the substrate. These equations are not always valid inreal biofilm diffusion conditions as usually sugars are utilized by cellsand consequently there would be a concentration gradient throughthe biofilms and this is in contradiction with the assumptions for t90and t90c (Stewart, 2003). It should be noted that when a solute reactsin a biofilm, it may not pass through the biofilm as all may be utilizedby cells. In conditions that the solute undergoes reaction in a biofilmand the rate of reaction is independent of solute concentration, thedepth that a solute can penetrate in a biofilm can be estimated throughEq. (21) (Table 4). These formulae have been driven on basis of zero-order kinetics, which means that the reaction rate is independent of


solute concentration (for Monod and Michaelis-Menten models that isS ≫ km) (Stewart, 2003).

The other issue faced in rMBRs is thatwhile the substrate diffusing inthe biofilm is being taken upby cellswhile participating in reactions, themetabolites produced also diffuse in reverse from the biofilm towardsthe feed-side due to concentration gradient. A simplified equation isused by Stewart (2003) to obtain the concentration of the product in lo-cations in the biofilm where the reacting substrate has depleted(Table 4, Eq. (22)). In this equation it has been assumed that productformation is in a stoichiometric relation with substrate consumption.

4. Concluding remarks

Membrane bioreactors have long been used in different biologicalprocesses. Depending on the feed medium and condition, process pa-rameters and the final product a number of different membranes andmembrane module configurations can be taken into practice. In this re-gard, the newly introduced concept of rMBR opens new horizons forfurther research and application development of MBRs. The rMBRs fea-ture exceptional properties such as high local cell density, diffusive na-ture of compound separation and the ability of cell separation and reuse.These unique specifications bring along the potential for the bioconver-sion of complex substrates that contain high concentration of inhibitorycompounds, different sugar sources and high suspended solid levels dif-ficult to be handled by conventional pressure driven MBRs.

Acknowledgments

The authors would like to express their gratitude to the Flemish In-stitute for Technological Research (VITO NV), and Swedish ResearchCouncil for their financial support.

References

Almeida, J.R.M., Modig, T., Petersson, A., Hähn-Hägerdal, B., Lidén, G., Gorwa-Grauslund,M.F., 2007. Increased tolerance and conversion of inhibitors in lignocellulosic hydro-lysates by Saccharomyces cerevisiae. J. Chem. Technol. Biotechnol. 82, 340–349.

Alzahrani, S., Mohammad, A.W., 2014. Challenges and trends in membrane technologyimplementation for produced water treatment: a review. J. Water Process Eng. 4,107–133.

American Society of Heating R.a.A.-C.E. Mass Transfer, 2005. ASHRAE Handbook: Funda-mentals - SI Edition Atlanta, Georgia 2005.

Anon., 2014. Diffusion Testing Fundamentals. Permegear, Inc.Armstrong, M.W., Gallego, S., Chesters, S.P., 2011. Removing biofilm from membranes – a

practical approach. Qingdao International Conference on Desalination and WaterReuse, Qingdao, China.

Bassi, A.S., Rohani, S., Macdonald, D.G., 1987. Measurement of effective diffusivities of lac-tose and lactic acid in 3% agarose gel membrane. Biotechnol. Bioeng. 30, 794–797.

Beck, R.E., Schultz, J.S., 1972. Hindrance of solute diffusion within membranes as mea-sured with microporous membranes of known pore geometry. Biochim. Biophys.Acta Biomembr. 255, 273–303.

Ben Chaabane, F., Aldiguier, A.S., Alfenore, S., Cameleyre, X., Blanc, P., Bideaux, C., et al.,2006. Very high ethanol productivity in an innovative continuous two-stage bioreac-tor with cell recycle. Bioprocess Biosyst. Eng. 29, 49–57.

Berens, A.R., Hopfenberg, H.B., 1982. Diffusion of organic vapors at low concentrations inglassy PVC, polystyrene, and PMMA. J. Membr. Sci. 10, 283–303.

Bertilsson, M., Andersson, J., Lidén, G., 2008. Modeling simultaneous glucose and xyloseuptake in Saccharomyces cerevisiae from kinetics and gene expression of sugar trans-porters. Bioprocess Biosyst. Eng. 31, 369–377.

Beyenal, H., Lewandowski, Z., 2002. Internal and external mass transfer in biofilms grownat various flow velocities. Biotechnol. Prog. 18, 55–61.

Bhattacharjee, S., Kim, A.S., Elimelech, M., 1999. Concentration polarization ofinteracting solute particles in cross-flow membrane filtration. J. Colloid InterfaceSci. 212, 81–99.

Bhave, R.R., Todaro, H.C.V.L., 1996. 7 - Cross-Flow Filtration. Fermentation and Biochemi-cal Engineering Handbook, second ed. William Andrew Publishing, Westwood, NJ,pp. 271–347.

Bin, C., Xiaochang, W., Enrang, W., 2004. Effects of tmp, MLSS concentration and intermit-tent membrane permeation on a hybrid submerged MBR fouling. Water Environ-ment-Membrane Technology Conference, Seoul, Korea.

Bonferoni, M.C., Rossi, S., Ferrari, F., Caramella, C., 1999. A modified Franz diffusion cell forsimultaneous assessment of drug release and washability of mucoadhesive gels.Pharm. Dev. Technol. 4, 45–53.

Börjesson, P., Ahlgren, S., Berndes, G., 2012. The climate benefit of Swedish ethanol: pres-ent and prospective performance. Wiley Interdiscip. Rev. Energy Environ. 1, 81–97.

Brandberg, T., Sanandaji, N., Gustafsson, L., Franzén, C.J., 2005. Continuous fermentation ofundetoxified dilute acid lignocellulose hydrolysate by Saccharomyces cerevisiae ATCC96581 using cell recirculation. Biotechnol. Prog. 21, 1093–1101.

Brookes, A., Judd, S., Reid, E., Germain, E., Smith, S., Alvarez-Vazquez, H., et al., 2003. Bio-mass characterisation in membrane bioreactors. International Membrane Scienceand Technology Conference (IMSTEC), Sydney, Australia.

Brown, T.R., Brown, R.C., 2013. A review of cellulosic biofuel commercial-scale projects inthe United States. Biofuels Bioprod. Biorefin. 7, 235–245.

Carstensen, F., Apel, A., Wessling, M., 2012. In situ product recovery: submerged mem-branes vs. external loop membranes. J. Membr. Sci. 394–395, 1–36.

Casey, G.P., Ingledew, W.M., 1986. Ethanol tolerance in yeasts. Crit. Rev. Microbiol. 13,219–280.

Casey, E., Glennon, B., Hamer, G., 2000. Biofilm development in a membrane-aerated bio-film reactor: effect of flow velocity on performance. Biotechnol. Bioeng. 67, 476–486.

Chandrakant, P., Bisaria, V.S., 1998. Simultaneous bioconversion of cellulose and hemicel-lulose to ethanol. Crit. Rev. Biotechnol. 18, 295–331.

Chang, I.S., Kim, S.N., 2005. Wastewater treatment using membrane filtration - effect ofbiosolids concentration on cake resistance. Process Biochem. 40, 1307–1314.

Chang, H.N., Lee, W.G., Kim, B.S., 1993. Cell retention culture with an internal filter mod-ule. Continuous ethanol fermentation. Biotechnol. Bioeng. 41, 677–681.

Chen, C., Han, B., Li, J., Shang, T., Zou, J., Jiang, W., 2001. A new model on the diffusion ofsmall molecule penetrants in dense polymer membranes. J. Membr. Sci. 187,109–118.

Chen, C., Tang, X., Xiao, Z., Zhou, Y., Jiang, Y., Fu, S., 2012. Ethanol fermentation kinetics in acontinuous and closed-circulating fermentation system with a pervaporation mem-brane bioreactor. Bioresour. Technol. 114, 707–710.

Cheong, S.H., Park, J.K., Kim, B.S., Chang, H.N., 1993. Microencapsulation of yeast cells inthe calcium alginate membrane. Biotechnol. Tech. 7, 879–884.

Cheryan, M., Mehaia, M.A., 1983. A high-performance membrane bioreactor for continu-ous fermentation of lactose to ethanol. Biotechnol. Lett. 5, 519–524.

Cho, C.W., Hwang, S.T., 1991. Continuous membrane fermentor separator for ethanol fer-mentation. J. Membr. Sci. 57, 21–42.

Chun, U.H., Rogers, P.L., 1988. Development of an ethanol/sorbitol process using a ceramicmembrane cell recycle bioreactor. Desalination 70, 353–361.

Çiçek, N., Franco, J.P., Suidan, M.T., Urbain, V., Manem, J., 1999. Characterization and com-parison of a membrane bioreactor and a conventional activated-sludge system in thetreatment of wastewater containing high-molecular-weight compounds.Water Envi-ron. Res. 71, 64–70.

Clément, P., Laugel, C., Marty, J.P., 2000. Influence of three synthetic membranes onthe release of caffeine from concentrated W/O emulsions. J. Control. Release66, 243–254.

Cussler, E.L., 2009. Diffusion: Mass Transfer in Fluid Systems. third ed. Cambridge Univer-sity Press, Cambridge, United Kingdom.

Ding, W., Xiao, Z., Tang, X., Deng, K., Fu, S., Jiang, Y., et al., 2012. Evolutionary engineeringof yeast for closed-circulating ethanol fermentation in PDMS membrane bioreactor.Biochem. Eng. J. 60, 56–61.

Doyen, W., Mues, W., Molenberghs, B., Cobben, B., 2010. Spacer fabric supported flat-sheet membranes: a new era of flat-sheet membrane technology. Desalination 250,1078–1082.

Einstein, A., 1905. On the movement of small particles suspended in stationary liquids re-quired by the molecular-kinetic theory of heat. Ann. Phys. 17, 549–560.

Elliott, B., Futcher, B., 1993. Stress resistance of yeast cells is largely independent of cellcycle phase. Yeast 9, 33–42.

Escobar, J.M., Rane, K.D., Cheryan, M., 2001. Ethanol production in a membrane bio-reactor: pilot-scale trials in a corn wet mill. Appl. Biochem. Biotechnol. 91-93,283–296.

Fan, L.S., Leyva-Ramos, R., Wisecarver, K.D., Zehner, B.J., 1990. Diffusion of phenol througha biofilm grown on activated carbon particles in draft-tube three-phase fluidized-bedbioreactor. Biotechnol. Bioeng. 35, 279–286.

Fang, H.H.P., Shi, X., 2005. Pore fouling of microfiltration membranes by activated sludge.J. Membr. Sci. 264, 161–166.

Fernandes, P., Prazeres, D.M., Cabral, J.M., 2003. Membrane-assisted extractive bioconver-sions. Adv. Biochem. Eng. Biotechnol. 80, 115–148.

Field, R.W., Wu, D., Howell, J.A., Gupta, B.B., 1995. Critical flux concept for microfiltrationfouling. J. Membr. Sci. 100, 259–272.

Flemming, H.C., Wingender, J., 2001. Relevance of microbial extracellular polymeric sub-stances (EPSs) - part I: structural and ecological aspects. Water Sci. Technol. 1–8.

Folasade, F., 2006. Anaerobic Membrane Bioreactors for Municipal Wastewater Treat-ment. Cranfield University, Cranfield, United Kingdom.

Galbe, M., Zacchi, G., 2002. A review of the production of ethanol from softwood. Appl.Microbiol. Biotechnol. 59, 618–628.

Ghareib, M., Youssef, K.A., Khalil, A.A., 1988. Ethanol tolerance of Saccharomycescerevisiae and its relationship to lipid content and composition. Folia Microbiol.33, 447–452.

Gnansounou, E., 2010. Production and use of lignocellulosic bioethanol in Europe: currentsituation and perspectives. Bioresour. Technol. 101, 4842–4850.

Ha, S.-J., Galazka, J.M., Rin, K.S., Choi, J.-H., Yang, X., Seo, J.-H., et al., 2011. Engineered Sac-charomyces cerevisiae capable of simultaneous cellobiose and xylose fermentation.Proc. Natl. Acad. Sci. U. S. A. 108, 504–509.

Hadgraft, J., Ridout, G., 1987. Development of model membranes for percutaneous ab-sorption measurements. I. Isopropyl myristate. Int. J. Pharm. 39, 149–156.

Hai, F.I., Yamamoto, K., Lee, C.H., 2013. Membrane Biological Reactors: Theory, Modeling,Design, Management and Applications to Wastewater Reuse. IWA Publishing.

Hamacher, T., Becker, J., Gárdonyi, M., Hahn-Hägerdal, B., Boles, E., 2002. Characterizationof the xylose-transporting properties of yeast hexose transporters and their influenceon xylose utilization. Microbiology 148, 2783–2788.

http://refhub.elsevier.com/S0734-9750(16)30066-0/rf0005






























































































































Hannoun, B.J.M., Stephanopoulos, G., 1986. Diffusion coefficients of glucose and ethanol incell-free and cell-occupied calcium alginate membranes. Biotechnol. Bioeng. 28,829–835.

Henriques, M.C.R., 2005. Candida dubliniensis versus Candida albicans Adhesion and Bio-film Formation. University of Minho.

Hoffmann, H., Kuhlmann, W., Meyer, H.D., Schügerl, K., 1985. High productivity ethanolfermentations with crossflow membrane separation techniques for continuous cellrecycling. J. Membr. Sci. 22, 235–243.

Hoffmann, H., Schügerl, K., Schmidt, W., 1988. Use of membranes to improve bioreactorperformance. II. Zymomonas mobilis fermentation. Chem. Eng. J. 37, B17–B22.

Hong, S.P., Bae, T.H., Tak, T.M., Hong, S., Randall, A., 2002. Fouling control in activatedsludge submerged hollow fiber membrane bioreactors. Desalination 143, 219–228.

Ikegami, T., Yanagishita, H., Kitamoto, D., Negishi, H., Haraya, K., Sano, T., 2002. Concentra-tion of fermented ethanol by pervaporation using silicalite membranes coated withsilicone rubber. Desalination 149, 49–54.

Inloes, D.S., Taylor, D.P., Cohen, S.N., 1983. Ethanol production by Saccharomyces cerevisiaeimmobilized in hollow-fiber membrane bioreactors. Appl. Environ. Microbiol. 46,264–278.

Inloes, D.S., Michaels, A.S., Robertson, C.R., Matin, A., 1985. Ethanol production bynitrogen-deficient yeast cells immobilized in a hollow-fiber membrane bioreactor.Appl. Microbiol. Biotechnol. 23, 85–91.

Ishola, M.M., Babapour, A.B., Gavitar, M.N., Brandberg, T., Taherzadeh, M.J., 2013a. Effect ofHigh Solids Loading on Bacterial Contamination in Lignocellulosic EthanolProduction.

Ishola, M.M., Jahandideh, A., Haidarian, B., Brandberg, T., Taherzadeh, M.J., 2013b. Si-multaneous saccharification, filtration and fermentation (SSFF): a novel methodfor bioethanol production from lignocellulosic biomass. Bioresour. Technol. 133,68–73.

Ishola, M., Ylitervo, P., Taherzadeh, M., 2015a. Co-utilization of glucose and xylose for en-hanced lignocellulosic ethanol production with reverse membrane bioreactors.Membranes 5, 844.

Ishola, M.M., Brandberg, T., Taherzadeh, M.J., 2015b. Simultaneous glucose and xylose uti-lization for improved ethanol production from lignocellulosic biomass through SSFFwith encapsulated yeast. Biomass Bioenergy 77, 192–199.

Itonaga, T., Kimura, K., Watanabe, Y., 2004. Influence of suspension viscosity and colloidalparticles on permeability of membrane used in membrane bioreactor (MBR). WaterSci. Technol. 301–309.

Jang, N., Ren, X., Choi, K., Kim, I.S., 2005. Comparison of Membrane Biofouling in Nitrifica-tion and Denitrification for the Membrane Bio-reactor (MBR). IWA Aspire, Singapore.

Jeison, D., van Lier, J.B., 2006. Cake layer formation in anaerobic submerged membranebioreactors (AnSMBR) for wastewater treatment. J. Membr. Sci. 284, 227–236.

Judd, S., 2008. The status of membrane bioreactor technology. Trends Biotechnol. 26,109–116.

Judd, S., Judd, C., 2011. Chapter 2 - fundamentals. The MBR Book, second ed. Butterworth-Heinemann, Oxford, pp. 55–207.

Jung, Y.J., Yoon, J.H., Kang, N.G., Park, S.G., Jeong, S.H., 2012. Diffusion properties of differ-ent compounds across various synthetic membranes using Franz-type diffusion cells.J. Pharm. Investig. 42, 271–277.

Kang, W., Shukla, R., Sirkar, K.K., 1990. Ethanol production in a microporous hollow-fiber-based extractive fermentor with immobilized yeast. Biotechnol. Bioeng. 36, 826–833.

Kargupta, K., Datta, S., Sanyal, S.K., 1998. Analysis of the performance of a continuousmembrane bioreactor with cell recycling during ethanol fermentation. Biochem.Eng. J. 1, 31–37.

Kaseno, Kokugan, T., 1997. The effect of molasses pretreatment by ceramic microfiltrationmembrane on ethanol fermentation. J. Ferment. Bioeng. 83, 577–582.

Kim Kwang, Y., Frank, J.F., 1995. Effect of nutrients on biofilm formation by Listeriamonocytogenes on stainless steel. J. Food Prot. 58, 24–28.

Klinke, H.B., Thomsen, A.B., Ahring, B.K., 2004. Inhibition of ethanol-producing yeast andbacteria by degradation products produced during pre-treatment of biomass. Appl.Microbiol. Biotechnol. 66, 10–26.

Koppram, R., Albers, E., Olsson, L., 2012. Evolutionary engineering strategies to enhancetolerance of xylose utilizing recombinant yeast to inhibitors derived from spruce bio-mass. Biotechnol. Biofuels 5.

Krauth, K., Staab, K.F., 1993. Pressurized bioreactor with membrane filtration for waste-water treatment. Water Res. 27, 405–411.

Krieble, J.G., Whitwell, J.C., 1949. The viscosity of Newtonian, pseudoplastic, and dilatantliquids: use in themeasurement of the D.P. of high polymers. Text. Res. J. 19, 253–258.

Kumaran, V., 2010. Fundamentals of rheology. Rheology of Complex Fluids, pp. 35–65.Kuyper, M., Toirkens, M.J., Diderich, J.A., Winkler, A.A., Van Dijken, J.P., Pronk, J.T., 2005.

Evolutionary engineering of mixed-sugar utilization by a xylose-fermenting Saccha-romyces cerevisiae strain. FEMS Yeast Res. 5, 925–934.

Lafforgue, C., Malinowski, J., Goma, G., 1987. High yeast concentration in continuous fer-mentation with cell recycle obtained by tangential microfiltration. Biotechnol. Lett. 9,347–352.

Lafforgue-Delorme, C., Delorme, P., Goma, G., 1994. Continuous alcoholic fermentationwith Saccharomyces cerevisiae recycle by tangential filtration: key points for processmodelling. Biotechnol. Lett. 16, 741–746.

Laluce, C., Schenberg, A.C.G., Gallardo, J.C.M., Coradello, L.F.C., Pombeiro-Sponchiado, S.R.,2012. Advances and developments in strategies to improve strains of Saccharomycescerevisiae and processes to obtain the lignocellulosic ethanol - a review. Appl.Biochem. Biotechnol. 166, 1908–1926.

Le-Clech, P., Jefferson, B., Judd, S.J., 2003. Impact of aeration, solids concentration andmembrane characteristics on the hydraulic performance of a membrane bioreactor.J. Membr. Sci. 218, 117–129.

Lee, K.R., Yeom, I.T., 2007. Evaluation of a membrane bioreactor system coupled withsludge pretreatment for aerobic sludge digestion. Environ. Technol. 28, 723–730.

Lee, K.L., Lefebvre, M., Tribe, D.E., Rogers, P.L., 1980. High productivity ethanol fermenta-tion with Zymomonas mobilis using continuous cell recycle. Biotechnol. Lett. 2,487–492.

Lee, W.G., Lee, Y.S., Chang, H.N., Chang, Y.K., 1994. A cell retention internal filter reactorfor ethanol production using tapioca hydrolysates. Biotechnol. Tech. 8, 817–820.

Lee, W.G., Park, B.G., Chang, Y.K., Chang, H.N., Lee, J.S., Park, S.C., 2000. Continuous ethanolproduction from concentrated wood hydrolysates in an internal membrane-filtrationbioreactor. Biotechnol. Prog. 16, 302–304.

Leonardi, E., Angeli, C., 2010. On the Maxwell-Stefan approach to diffusion: a general res-olution in the transient regime for one-dimensional systems. J. Phys. Chem. B 114,151–164.

Lesjean, B., Rosenberger, S., Laabs, C., Jekel, M., Gnirss, R., Amy, G., 2005. Correlation be-tween membrane fouling and soluble/colloidal organic substances in membrane bio-reactors for municipal wastewater treatment. Water Sci. Technol. 1–8.

Li, F., Tian, S., Yu, Y., Zhou, G., Yang, X., 2009. Ethanol production from dilute-acid hydro-lysate in situ detoxification by co-culture of yeasts. Acta Energiae Solaris Sin. 30,1582–1586.

Li, J., He, S., Xia, S., Qiu, C., Zhang, Y., Xuan, J., 2014. The membrane fouling monitored byscanning electron microscopy (SEM) in a submerged membrane bioreactor. J. WaterSustain. 4, 159–166.

Lin, H., Gao, W., Meng, F., Liao, B.Q., Leung, K.T., Zhao, L., et al., 2012. Membrane bioreac-tors for industrial wastewater treatment: a critical review. Crit. Rev. Environ. Sci.Technol. 42, 677–740.

Liu, Z.L., 2011. Molecular mechanisms of yeast tolerance and in situ detoxification of lig-nocellulose hydrolysates. Appl. Microbiol. Biotechnol. 90, 809–825.

Liu, X., Wang, Y., Waite, T.D., Leslie, G., 2015. Numerical simulation of bubble inducedshear inmembrane bioreactors: effects of mixed liquor rheology andmembrane con-figuration. Water Res. 75, 131–145.

Mack, C., Burgess, J.E., Duncan, J.R., 2004. Membrane bioreactors for metal recovery fromwastewater: a review. Water SA 30, 521–532.

Marić, S., Vraneš, J., 2007. Characteristics and significance of microbial biofilm formation.Period. Biol. 109, 115–121.

Mehaia, M.A., Cheryan, M., 1984. Ethanol production in a hollow fiber bioreactor usingSaccharomyces cerevisiae. Appl. Microbiol. Biotechnol. 20, 100–104.

Mehaia, M.A., Cheryan, M., 1990. Ethanol from hydrolyzed whey permeate using Saccha-romyces cerevisiae in a membrane recycle bioreactor. Bioprocess Eng. 5, 57–61.

Meinander, N.Q., Boels, I., Hahn-Hägerdal, B., 1999. Fermentation of xylose/glucose mix-tures by metabolically engineered Saccharomyces cerevisiae strains expressing XYL1and XYL2 from Pichia stipitis with and without overexpression of TAL1. Bioresour.Technol. 68, 79–87.

Melo, L.F., 2005. Biofilm physical structure, internal diffusivity and tortuosity. Water Sci.Technol. 77–84.

Melzoch, K., Rychtera, M., Markvichov, N.S., Pospíchalová, V., Basařová, G., Manakov, M.N.,1991. Application of a membrane recycle bioreactor for continuous ethanol produc-tion. Appl. Microbiol. Biotechnol. 34, 469–472.

Meng, F., Shi, B., Yang, F., Zhang, H., 2007. New insights into membrane fouling in sub-merged membrane bioreactor based on rheology and hydrodynamics concepts.J. Membr. Sci. 302, 87–94.

Miller, C.C., 1924. The Stokes-Einstein law for diffusion in solution. Proc. R. Soc. Lond. A106, 724–749.

Miranda, J.M., Campos, J.B.L.M., 2002. Mass transfer in the vicinity of a separation mem-brane - the applicability of the stagnant film theory. J. Membr. Sci. 202, 137–150.

Mutamim, N.S.A., Noor, Z.Z., Hassan, M.A.A., Yuniarto, A., Olsson, G., 2013. Membrane bio-reactor: applications and limitations in treating high strength industrial wastewater.Chem. Eng. J. 225, 109–119.

Naik, S.N., Goud, V.V., Rout, P.K., Dalai, A.K., 2010. Production of first and second genera-tion biofuels: a comprehensive review. Renew. Sust. Energ. Rev. 14, 578–597.

Ng, S.F., Rouse, J., Sanderson, D., Eccleston, G., 2010. A comparative study of transmem-brane diffusion and permeation of ibuprofen across synthetic membranes usingfranz diffusion cells. Pharmaceutics 2, 209–223.

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

Ohashi, R., Kamoshita, Y., Kishimoto, M., Suzuki, T., 1998. Continuous production and sep-aration of ethanol without effluence of wastewater using a distiller integrated SCM-reactor system. J. Ferment. Bioeng. 86, 220–225.

Padaki, M., Surya, M.R., Abdullah, M.S., Misdan, N., Moslehyani, A., Kassim, M.A., et al.,2015. Membrane technology enhancement in oil-water separation. A review. Desali-nation 357, 197–207.

Park, T.H., Kim, I.H., 1985. Hollow-fibre fermenter using ultrafiltration. Appl. Microbiol.Biotechnol. 22, 190–194.

Park, B.G., Lee, W.G., Chang, Y.K., Chang, H.N., 1997. Effects of periodic backflushing withfiltrate on filtration performance in an internal-filtration bioreactor. Bioprocess Eng.16, 253–256.

Pellett, M.A., Castellano, S., Hadgraft, J., Davis, A.F., 1997. The penetration of supersaturat-ed solutions of piroxicam across silicone membranes and human skin in vitro.J. Control. Release 46, 205–214.

Perry, R.H., Chilton, C.H., 1973. Chemical Engineer's Handbook. fifth ed. McGraw-Hill BookCo., New York, United States of America.

Peters, T., 2010. Membrane technology for water treatment. Chem. Eng. Technol. 33,1233–1240.

Petrinić, I., Hélix-Nielsen, C., 2014. Towards newmembrane-based technologies for watertreatment and reuse in the textile industry. Tekstil 63, 251–258.

Poling, B.E., Prausnitz, J.M., O'Connell, J.P., 2001. Properties of Gases and Liquids. fifth ed.McGraw-Hill Education.

Polson, A., 1950. Some aspects of diffusion in solution and a definition of a colloidal par-ticle. J. Phys. Colloid Chem. 54, 649–652.










































































































































































Pourbafrani, M., Talebnia, F., Niklasson, C., Taherzadeh, M., 2007a. Protective effect of en-capsulation in fermentation of limonene-contained media and orange peel hydroly-zate. Int. J. Mol. Sci. 8, 777.

Pourbafrani, M., Talebnia, F., Niklasson, C., Taherzadeh, M.J., 2007b. Protective effect of en-capsulation in fermentation of limonene-contained media and orange peel hydroly-zate. Int. J. Mol. Sci. 8, 777–787.

Puzanov, T., 1999. Continuous Production of Lactic Acid in a Membrane Bioreactor. Uni-versity of Toronto.

Quist-Jensen, C.A., Macedonio, F., Drioli, E., 2015.Membrane technology for water produc-tion in agriculture: desalination and wastewater reuse. Desalination 364, 17–32.

Radjenović, J., Matošić, M., Mijatović, I., Petrović, M., Barceló, D., 2008. Membrane bioreac-tor (MBR) as an advanced wastewater treatment technology. Handbook of Environ-mental Chemistry, Volume 5: Water Pollution, pp. 37–101.

Radoaj, O., Diosady, L.L., 2014. Continuous ethanol fermentation in immersed, cross-flowmicrofiltration membrane bioreactor with cell retention. J. Basic Appl. Sci. 10,543–553.

Rendueles, O., Kaplan, J.B., Ghigo, J.M., 2013. Antibiofilm polysaccharides. Environ.Microbiol. 15, 334–346.

Reuß, M., Josić, D., Popović, M., Bronn, W.K., 1979. Viscosity of yeast suspensions. Eur.J. Appl. Microbiol. Biotechnol. 8, 167–175.

Rosenberger, S., Evenblij, H., Te Poele, S., Wintgens, T., Laabs, C., 2005. The importance ofliquid phase analyses to understand fouling in membrane assisted activated sludgeprocesses - six case studies of different European research groups. J. Membr. Sci.263, 113–126.

Sànchez Nogué, V., Karhumaa, K., 2015. Xylose fermentation as a challenge for commer-cialization of lignocellulosic fuels and chemicals. Biotechnol. Lett. 37, 761–772.

Sarbatly, R., England, R., 2004. Critical review of membrane bioreactor system used forcontinuous production of hydrolyzed starch. Chem. Biochem. Eng. Q. 18, 155–165.

Sassner, P., Galbe, M., Zacchi, G., 2006. Bioethanol production based on simultaneous sac-charification and fermentation of steam-pretreated Salix at high dry-matter content.Enzym. Microb. Technol. 39, 756–762.

Sato, T., Ishii, Y., 1991. Effects of activated sludge properties onwater flux of ultrafiltrationmembrane used for human excrement treatment. Water Sci. Technol. 23,1601–1608.

Sedlak, M., Ho, N.W.Y., 2004. Characterization of the effectiveness of hexose transportersfor transporting xylose during glucose and xylose co-fermentation by a recombinantSaccharomyces yeast. Yeast 21, 671–684.

Shah, V.P., Elkins, J.S., Williams, R.L., 1999. Evaluation of the test system used for in vitrorelease of drugs for topical dermatological drug products. Pharm. Dev. Technol. 4,377–385.

Sharma, M., Yashonath, S., 2006. Breakdown of the Stokes-Einstein relationship: role ofinteractions in the size dependence of self-diffusivity. J. Phys. Chem. B 110,17207–17211.

Sheng, G.P., Yu, H.Q., Li, X.Y., 2010. Extracellular polymeric substances (EPS) of microbialaggregates in biological wastewater treatment systems: a review. Biotechnol. Adv.28, 882–894.

Shimizu, Y., Okuno, Y.I., Uryu, K., Ohtsubo, S., Watanabe, A., 1996. Filtration characteristicsof hollow fibermicrofiltrationmembranes used inmembrane bioreactor for domesticwastewater treatment. Water Res. 30, 2385–2392.

Shimmons, B.W., Svrcek, W.Y., Zajic, J.E., 1976. Cell concentration control by viscosity.Biotechnol. Bioeng. 18, 1793–1805.

Short, B., 2011. Budding yeast star in their own biofilm. J. Cell Biol. 194, 659.Singh, K.S., Viraraghavan, T., 2003. Impact of temperature on performance, microbiologi-

cal, and hydrodynamic aspects of UASB reactors treating municipal wastewater.Water Sci. Technol. 48, 211–217.

Singh, N.B., Glicksman, M.E., Coriell, S.R., Duval, W.M.B., Santoro, G.J., DeWitt, R., 1996.Measurement of diffusion coefficient using a diaphragm cell: PbBr2⋯AgBr system.J. Cryst. Growth 167, 107–110.

Stanley, G., Hahn-Hägerdal, B., 2010b. Fuel ethanol production from lignocellulosic rawmaterials using recombinant yeasts. Biomass to Biofuels: Strategies for Global Indus-tries, pp. 261–291.

Stanley, D., Bandara, A., Fraser, S., Chambers, P.J., Stanley, G.A., 2010a. The ethanol stressresponse and ethanol tolerance of Saccharomyces cerevisiae. J. Appl. Microbiol. 109,13–24.

Steingiser, S., Nempos, S.P., Salame, M., 1987. In: Mark, H.F. (Ed.), Encyclopedia ofChemical Technology, third ed. Wiley Interscience, New York, United States ofAmerica.

Stern, S.A., Shah, V.M., Hardy, B.J., 1987. Structure-permeability relationships in siliconepolymers. J. Polym. Sci. B Polym. Phys. 25, 1263–1298.

Stewart, P.S., 1998. A review of experimental measurements of effective diffusive perme-abilities and effective diffusion coefficients in biofilms. Biotechnol. Bioeng. 59,261–272.

Stewart, P.S., 2003. Diffusion in biofilms. J. Bacteriol. 185, 1485–1491.Strathmann, H., Giorno, L., Drioli, E., 2006. An Introduction to Membrane Science and

Technology.Stuckey, D.C., Hu, A., 2003. The submerged anaerobic membrane bioreactor (SAMBR): an

intensification of anaerobic wastewater treatment. The IWA Leading Edge Conferenceon Drinking Water and Wastewater Treatment Technologies, Noordwijk,Amsterdam.

Subramani, A., Jacangelo, J.G., 2015. Emerging desalination technologies for water treat-ment: a critical review. Water Res. 75, 164–187.

Sun, Z.J., Lv, G.J., Li, S.Y., Yu, W.T., Wang, W., Xie, Y.B., et al., 2007. Differential role of mi-croenvironment in microencapsulation for improved cell tolerance to stress. Appl.Microbiol. Biotechnol. 75, 1419–1427.

Suzuki, T., Sato, T., Kominami, M., 1994. A dense cell retention culture system using astirred ceramic membrane reactor. Biotechnol. Bioeng. 44, 1186–1192.

Svedberg, T., Pedersen, K.O., Bauer, J.H., 1940. The Ultracentrifuge. The Clarendon Press,Oxford, United Kingdom.

Taherzadeh, M.J., Karimi, K., 2008. Pretreatment of lignocellulosic wastes to improveethanol and biogas production: a review. Int. J. Mol. Sci. 9, 1621–1651.

Taherzadeh, M.J., Eklund, R., Gustafsson, L., Niklasson, C., Lidén, G., 1997. Characterizationand fermentation of dilute-acid hydrolyzates from wood. Ind. Eng. Chem. Res. 36,4659–4665.

Taherzadeh, M.J., Niklasson, C., Lidén, G., 1999. Conversion of dilute-acid hydroly-zates of spruce and birch to ethanol by fed-batch fermentation. Bioresour.Technol. 69, 59–66.

Talebnia, F., Taherzadeh, M.J., 2006. In situ detoxification and continuous cultivation ofdilute-acid hydrolyzate to ethanol by encapsulated S. cerevisiae. J. Biotechnol. 125,377–384.

Talebnia, F., Taherzadeh, M.J., 2007. Physiological and morphological study of encapsulat-ed Saccharomyces cerevisiae. Enzym. Microb. Technol. 41, 683–688.

Talebnia, F., Niklasson, C., Taherzadeh, M.J., 2005. Ethanol production from glucose anddilute-acid hydrolyzates by encapsulated S. cerevisiae. Biotechnol. Bioeng. 90, 345–353.

Tampion, J., Tampion, M.D., 1987. Immobilized Cells: Principles and Applications. Cam-bridge University Press, Cambridge.

Thuvander, J., 2012. Continuous Ethanol Fermentation Using Membrane Bioreactors(MBR). Lund University.

Tian, S., Zhou, G., Zhang, L., Yang, X., 2011. Characteristics of two yeast strains for produc-ing ethanol from dilute-acid lignocellulosic hydrolysate in situ detoxification. ActaEnergiae Solaris Sin. 32, 272–275.

Tin, C.S.F., Mawson, A.J., 1993. Ethanol production from whey in a membrane recycle bio-reactor. Process Biochem. 28, 217–221.

van den Brink, P., Satpradit, O.-A., van Bentem, A., Zwijnenburg, A., Temmink, H., vanLoosdrecht, M., 2011. Effect of temperature shocks on membrane fouling in mem-brane bioreactors. Water Res. 45, 4491–4500.

Van Zyl, C., Prior, B.A., Kilian, S.G., Kock, J.L.F., 1989. D-Xylose utilization by Saccharomycescerevisiae. J. Gen. Microbiol. 135, 2791–2798.

Vanysacker, L., Denis, C., Declerck, P., Piasecka, A., Vankelecom, I.F.J., 2013. Microbial adhe-sion and biofilm formation onmicrofiltrationmembranes: a detailed characterizationusing model organisms with increasing complexity. BioMed. Res. Int. 2013.

Verardi, A., De Bari, I., Ricca, E., Calabrò, V., 2012. Hydrolysis of lignocellulosic biomass:current status of processes and technologies and future perspectives. In: Lima,M.A.P. (Ed.), Bioethanol.

Verstrepen, K.J., Klis, F.M., 2006. Flocculation, adhesion and biofilm formation in yeasts.Mol. Microbiol. 60, 5–15.

Vilela-Moura, A., Schuller, D., Mendes-Faia, A., Silva, R.D., Chaves, S.R., Sousa, M.J., et al.,2011. The impact of acetate metabolism on yeast fermentative performance andwine quality: reduction of volatile acidity of grape musts and wines. Appl. Microbiol.Biotechnol. 89, 271–280.

Wang, L.K., Ivanov, V., Tay, J.-H., Guglielmi, G., Andreottola, G., 2010. Selection and designof membrane bioreactors in environmental bioengineering. Environmental Biotech-nology. Humana Press, pp. 439–516.

Westman, J.O., Franzén, C.J., 2015. Current progress in high cell density yeast bioprocessesfor bioethanol production. Biotechnol. J. 10, 1185–1195.

Westman, J.O., Manikondu, R.B., Franzén, C.J., Taherzadeh, M.J., 2012a. Encapsulation-induced stress helps Saccharomyces cerevisiae resist convertible lignocellulose de-rived inhibitors. Int. J. Mol. Sci. 13, 11881–11894.

Westman, J.O., Ylitervo, P., Franzén, C.J., Taherzadeh, M.J., 2012b. Effects of encapsulationof microorganisms on product formation during microbial fermentations. Appl.Microbiol. Biotechnol. 96, 1441–1454.

Westman, J.O., Bonander, N., Taherzadeh, M.J., Franzén, C.J., 2014a. Improved sugar co-utilisation by encapsulation of a recombinant Saccharomyces cerevisiae strain inalginate-chitosan capsules. Biotechnol. Biofuels 7.

Westman, J.O., Mapelli, V., Taherzadeh, M.J., Franzén, C.J., 2014b. Flocculation causes in-hibitor tolerance in Saccharomyces cerevisiae for second-generation bioethanol pro-duction. Appl. Environ. Microbiol. 80, 6908–6918.

Wicaksana, F., Fane, A.G., Chen, V., 2006. Fibre movement induced by bubbling using sub-merged hollow fibre membranes. J. Membr. Sci. 271, 186–195.

Wijmans, J.G., Nakao, S., Van Den Berg, J.W.A., Troelstra, F.R., Smolders, C.A., 1985. Hydro-dynamic resistance of concentration polarization boundary layers in ultrafiltration.J. Membr. Sci. 22, 117–135.

Wikandari, R., Youngsukkasem, S., Millati, R., Taherzadeh, M.J., 2014. Performance ofsemi-continuous membrane bioreactor in biogas production from toxic feedstockcontaining D-Limonene. Bioresour. Technol. 170, 350–355.

Wilke, C.R., Chang, P., 1955. Correlation of diffusion coefficient in dilute solution. AICHE J.264-270.

Ylitervo, P., Franzén, C.J., Taherzadeh, M.J., 2011. Ethanol production at elevated temper-atures using encapsulation of yeast. J. Biotechnol. 156, 22–29.

Ylitervo, P., Akinbomi, J., Taherzadeh,M.J., 2013a.Membrane bioreactors potential for eth-anol and biogas production: a review. Environ. Technol. 34, 1711–1723.

Ylitervo, P., Franzén, C.J., Taherzadeh, M.J., 2013b. Impact of furfural on rapid ethanol pro-duction using a membrane bioreactor. Energies 6, 1604–1617.

Ylitervo, P., Doyen, W., Taherzadeh, M.J., 2014. Fermentation of lignocellulosic hydroly-zate using a submerged membrane bioreactor at high dilution rates. Bioresour.Technol. 164, 64–69.

Yoon, S.H., 2015. Membrane Bioreactor Processes: Principles and Applications. CRC Press.Yoon, S.-H., Lee, C.-H., Kim, K.-J., Fane, A.G., 1999. Three-dimensional simulation of the de-

position of multi-dispersed charged particles and prediction of resulting flux duringcross-flow microfiltration. J. Membr. Sci. 161, 7–20.

You, K.M., Rosenfield, C.L., Knipple, D.C., 2003. Ethanol tolerance in the yeast Saccharomy-ces cerevisiae is dependent on cellular oleic acid content. Appl. Environ. Microbiol. 69,1499–1503.
















































































































































































Youngsukkasem, S., Akinbomi, J., Rakshit, S.K., Taherzadeh, M.J., 2013a. Biogas productionby encased bacteria in synthetic membranes: protective effects in toxic media andhigh loading rates. Environ. Technol. 34, 2077–2084.

Youngsukkasem, S., Barghi, H., Rakshit, S.K., Taherzadeh, M.J., 2013b. Rapid biogas pro-duction by compact multi-layer membrane bioreactor: efficiency of synthetic poly-meric membranes. Energies 6, 6211–6224.

Youngsukkasem, S., Chandolias, K., Taherzadeh, M.J., 2015. Rapid bio-methanation of syn-gas in a reverse membrane bioreactor: membrane encased microorganisms.Bioresour. Technol. 178, 334–340.

Zahed, O., Jouzani, G.S., Abbasalizadeh, S., Khodaiyan, F., Tabatabaei, M., 2015. Continuousco-production of ethanol and xylitol from rice straw hydrolysate in a membrane bio-reactor. Folia Microbiol.

Zalc, J.M., Reyes, S.C., Iglesia, E., 2004. The effects of diffusion mechanism and void struc-ture on transport rates and tortuosity factors in complex porous structures. Chem.Eng. Sci. 59, 2947–2960.

Zaldivar, J., Nielsen, J., Olsson, L., 2001. Fuel ethanol production from lignocellulose: achallenge for metabolic engineering and process integration. Appl. Microbiol.Biotechnol. 56, 17–34.

Zhang, S.F., Splendiani, A., Freitas dos Santos, L.M., Livingston, A.G., 1998a. Determinationof pollutant diffusion coefficients in naturally formed biofilms using a single tube ex-tractive membrane bioreactor. Biotechnol. Bioeng. 59, 80–89.

Zhang, W., Park, B.G., Chang, Y.K., Chang, H.N., Yu, X.J., Yuan, Q., 1998b. Factors affectingmembrane fouling in filtration of Saccharomyces cerevisiae in an internal ceramic fil-ter bioreactor. Bioprocess Eng. 18, 317–322.

Zhao, X., Zhang, L., Liu, D., 2012. Biomass recalcitrance. Part I: the chemical compositionsand physical structures affecting the enzymatic hydrolysis of lignocellulose. BiofuelsBioprod. Biorefin. 6, 465–482.




























reverse membrane bioreactor: introduction to a new ...research review paper reverse membrane...

Documents