Chemical Engineering Journal 264 (2015) 863–872

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

Heavy metals removal from aqueous solutions using Saccharomycescerevisiae in a novel continuous bioreactor–biosorption system

http://dx.doi.org/10.1016/j.cej.2014.12.0161385-8947/� 2014 Elsevier B.V. All rights reserved.

⇑ Corresponding author. Tel.: +1 (519) 661 2146.E-mail address: amarg@uwo.ca (A. Margaritis).

Shahram Amirnia, Madhumita B. Ray, Argyrios Margaritis ⇑Department of Chemical & Biochemical Engineering, University of Western Ontario, London, Ontario N6A5B9, Canada

h i g h l i g h t s

� The possibility of on-line production of yeast S. cerevisiae and biosorption of Cu2+ and Pb2+ was investigated.� Diffusion and chemical-based kinetics models were used to study the adsorption behaviour of yeast cells.� Continuous adsorption system was an effective and inexpensive method for removal of heavy metals.

a r t i c l e i n f o

Article history:Received 1 October 2014Received in revised form 2 December 2014Accepted 3 December 2014Available online 9 December 2014

Keywords:Heavy metalsContinuous bioreactor–biosorption systemS. cerevisiae

a b s t r a c t

The adsorption behaviour of unmodified yeast cells of Saccharomyces cerevisiae to remove Pb(II) andCu(II) ions from aqueous solutions in continuous mode was studied. Yeast biomass showed mediocrecapacity for Cu(II) ions compared to that of Pb(II) ions in the metals concentrations range of10–180 mg/l. Metal-binding capacity of yeast cells reached to a maximum of 29.9 mg/g and 72.5 mg/gfor Cu(II) and Pb(II) ions, respectively, under similar experimental conditions. The rate of biosorbent pro-duction in the continuous bioreactor, governed by dilution rate equal to maximum specific growth rate ofthe cells, was the limiting factor of the biosorption system. At low metal concentration, Cu(II) removal byyeast cells was higher than previously studied heat-deactivated yeast biomass suggesting involvement ofboth intracellular and surface-based sequestrations. The removal efficiencies of the test metals decreasedas the initial metal concentrations increased. Equilibrium adsorption of the metals by yeast cells was welldescribed by the Langmuir isotherm model. The adsorption kinetics data fitted to diffusion-based andchemical reaction-based models showed intracellular diffusion as the rate controlling step at low metalloads. The results obtained suggest that the use of live yeast cells in a self-contained continuousadsorption system is an effective method for removal of heavy metals from industrial effluents reducingbiomass pretreatment and preparation steps as well as achieving on-line adsorbent production, biosorp-tion, and effluent treatment.

� 2014 Elsevier B.V. All rights reserved.

1. Introduction

Use of biology based technologies, such as biosorption, is anattractive method for heavy metal removal from metal-laden efflu-ents due to low cost and high efficiency of the process. Despiteincreased understanding of biosorption phenomenon and abun-dance of research in this field, an industrially relevant methodfor biosorption technology has not been fully realized yet[44,41,19,45,7]. Most of the studies in the field deal with batchequilibrium studies relating adsorbate, adsorbent, and operatingconditions. Although biosorption is defined as the property of cer-tain non-living biomaterials to bind and concentrate selected ions

or other molecules from aqueous solutions [42], it can occur inboth living and dead microorganisms [38,7]. Metal uptake is acombination of a metabolism independent physical process, fol-lowed by a metabolic step known as bioaccumulation [47]. Tothe best of our knowledge, no studies have been reported on theuse of yeast cells in a combined continuous bioreactor–biosorptionsystem for the removal of heavy metals from aqueous solutions.

Removal of heavy metals in continuous mode was earlierreported as a preferred choice in some metal adsorption studies.For instance, Kapoor and Viraraghavan [16] used immobilized cellsof Aspergillus niger in a continuous operation for the removal ofmetal solutions containing cadmium, copper lead, and nickel.Marques et al. [26] used a fixed-bed reactor for Cd removal usingimmobilized cells of an industrial strain of Saccharomyces cerevisiae.There are certain potential limitations of continuous fixed bed

864 S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872

adsorption: low density of fungal biomass reported to be problem-atic in fixed-bed operation due to clogging and subsequent releaseof biomass in the treated wastewater. Immobilization of biomassalso causes mass transfer limitations by hindering the access ofthe metals to the biosorbent sites compared to suspended biosor-bents [37,4]. Moreover, since the regeneration capacity of immobi-lized cells is limited, biomass needs to be frequently replaced,which is a costly process. However, continuous operation is theonly viable way of treating large volume of wastewater in a reason-able time, and this is where most of the bench scale batch biosorp-tion studies are limited in their scope. Biosorbents with differentdegrees of metal adsorption capacities, availability, and selectivityhave been shown to be promising for metal removal from aqueoussolutions in batch laboratory scale, but it is their applicability incontinuous mode that makes this treatment method attractivefor industrial application. In addition, a waste stream containingnecessary food source and nutrients can be used to cultivate thebiosorbents, thereby reducing the environmental footprint of theprocess by resource recovery and waste reduction. Based on ourearlier work [2] on the removal of Cu, Pb and Zn by S. cerevisiaein a batch system, continuous metal adsorption using live S. cerevi-siae in an air-lift fluidized vessel was attempted in this work. Verylittle information is available in literature on the use of unmodifiedlive and resting cells for biosorption [8]. The main objective of thiswork was to examine the ability of yeast cells to be produced andutilized in a continuous fashion for removing metal ions in water.This is regarded as one of the desirable characteristics for a goodand useful biosorbent [22]. The yeast cells produced in the contin-uous bioreactor provided continuously fresh surface for adsorptionof the metals without the need for costly pretreatment steps, suchas harvesting, separating, heat-killing, drying, and storage of bio-mass. Copper and lead ions were chosen as the model metals forbiosorption due to their widespread use in metal industry and theirtoxicity for humans and environmental health [12,31].

2. Material and methods

A schematic of the continuous biosorption system is shown inFig. 1. A 1.2 L plexiglass double draft fluidized vessel was used asthe adsorption column whereas yeast biomass as biosorbent wasproduced in a 1.5 L stirred tank bioreactor. The sampling points,

Pump 1

Pump 3

Liquid MediumStorage

Ai

Metal Solu�onStorage

Pump 2

Bioreactor

S

Air FilterFlowmeter

Fig. 1. Schematic process flow diagram of the continuous biosor

biosorbent and metal solution flow lines, adsorption and biomassseparation sections are illustrated in Fig. 1.

2.1. Biomass preparation

Yeast inoculum was prepared by aseptically transferring theactive dry yeast (Fleischmann’s Co) to a 500 ml flask containingthe sterile medium. The flask was incubated at 25 �C in a shakerat 200 rpm and the bioreactor was inoculated with the yeast sus-pension harvested at its late exponential phase of batch growth(�16–20 h of incubation). The bioreactor was agitated at 200 rpmand aerated with an air flow of 1vvm (1.5 L/min) at room temper-ature. The liquid medium was comprised of 20 g/l dextrose, 5 g/lpeptone, 3 g/l yeast extract, and 3 g/l malt extract, similar to thenutrients used by Volesky and May-Phillips [43]. A steady expo-nential growth phase was maintained in the bioreactor by alteringthe liquid medium volumetric flow rate into the reactor and theoutflow of biomass from the reactor.

2.2. Analytical method

Copper and lead were used as model metals for biosorption.Stock solutions of Cu2+ and Pb2+ ions were prepared using cop-per(II) sulfate pentahydrate (CuSO4�5H2O) and Pb(NO3)2, respec-tively. Samples were regularly collected from the feed tank, airliftvessel’s sampling port (C), liquid outlet from the airlift column(F), and the overflow of the 2nd settling tank (details illustratedin Figs. 1 and 2). Metal concentrations of the samples wereanalyzed using inductively coupled plasma optical emissionspectrometry (ICP-OES, Varian Varian, Inc., Vista-Pro Axial) aftercentrifugation and removal of the biomass. Metal uptake capacity,qtðmg=gÞ and removal efficiency, RE (%) of biomass were obtainedbased on mass balance in the following forms:

qt ¼ðC0 � CtÞCBiomass

ð1Þ

RE ð% Removal EfficiencyÞ ¼ ðC0 � CtÞC0

� 100 ð2Þ

where C0ðmg=lÞwas the initial metal concentration at the bulk solu-tion, Ctðmg=lÞ outlet metal concentration at time t, and CBiomass theconcentration of biomass ðg=lÞ at time t. At equilibrium, Ct ¼ Ce.

r

Airli�Biosorp�on

Vessel

Se�lingTank 1

Metal freeEffluent

Metal loadedBiomass

Se�lingTank 2

MetalRecovery

P = Primary Sampling PointS = Secondary Sampling Point

P

Acid

BasePump 4

Pump 5S

SOverflow

ption system (sampling points are shown on the diagram).

Fig. 2. Schematic diagram of the 2-L double draft tube fluidized adsorption column:(A) liquid level, (B) head space, (C) sampling ports, (D) air jet orifices, (E) air inlet, (F)liquid outlet, (G) jacket (adapted from Bashar et al. [3].

S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872 865

2.3. Maximum specific growth rate of yeast cells

The growth curve of S. cerevisiae (Fig. 3) was determined by tak-ing samples from the bioreactor and measuring the optical densityand dry weight of S. cerevisiae cells. The exponential phase repre-sented the cells maximum specific growth rate (lmax). The valueof lmax was obtained from the slope of the plot of ln x vs. time.

The rate of biomass production in batch system during theexponential phase is expressed as [48];

dxdt¼ lmaxx ð3Þ

lnxx0

� �¼ lmaxt or; x ¼ x0elmaxt ð4Þ

Fig. 3. Yeast growth curves in batch and continuous culture. In the continuoussystem, the growth curve became constant at the biomass concentration of 3 g/l atthe end of exponential growth phase.

In continuous system, fresh medium for yeast growth was sup-plied in a steady rate into the bioreactor; therefore, there was nodepletion of nutrients, and yeast cells experienced prolonged per-iod of exponential growth without entering into deceleration andstationary phases [48]. The change in biomass concentration incontinuous culture system is described as:

dxdt¼ lmaxx� Dx ð5Þ

where D is the dilution rate (h�1). At steady state, the yeast growthrate remained constant (dx=dt ¼ 0) and the maximum growth rateof the cells (lmax ¼ D) was dictated by the dilution rate. Accordingto Eq. (5), for dilution rates more than lmax, the biomass concentra-tion in the bioreactor will decrease with time as the cells will bewashed out from the bioreactor.

Optimized biomass productivity required for biosorption withmaximum surface area possible necessitates that the specificgrowth rate in the bioreactor be as high as possible. As shown inFig. 3, yeast cells attained a maximum specific growth rate of0.0994 h�1, so the biomass discharge flowrate from bioreactor foran operating volume of 1.5 L was adjusted to be 2.49 ml=min.The biomass concentration in the bioreactor was examined by tak-ing periodic samples from the bioreactor and measuring the opticaldensity.

2.4. Continuous biosorption operation

As illustrated in Fig. 1, the bioreactor–biosorption system wasoperated in a continuous mode. A one-time batch incubation wasrequired before cells inoculation in the bioreactor containing liquidmedium reached steady state. When biomass concentrationreached around 3 g/l in the bioreactor, feed and liquid culturepumps (pump 1, and 3 in Fig. 1) were started simultaneously todeliver the nutrients to bioreactor and the biomass solution tothe airlift adsorption column, respectively. Metal solution wasfed into the adsorption column by pump 2 co-current to thebiomass flow. When the liquid level reached to a point that airbubbles circulation was complete in the adsorption column (LevelA in Fig. 2), the discharge valve was opened to keep the liquid levelconstant in the adsorption column. Gravity flow was used to trans-fer the metal-loaded biomass from adsorption column to settlingtanks. Biomass was separated from liquid using two settling tanksin series. Separated biomass was periodically discharged from thesettling tanks to be used in batch desorption process, which wasnot used in the present work. The stable and steady state operationof the continuous system ran for 8 weeks in a row until all therequired data were collected.

2.5. Adsorption column

A double-draft airlift column was used for biosorption experi-ments, in which, air bubbles through the draft tubes concentricallylocated inside the column mix the contents of the column. Fig. 2shows the column dimensions and air bubbles circulation patternwhich created a good mixing and fluidization of biosorbents inthe system. The air orifices were located in the annular regionbetween the outer and inner draft tubes. This type of contactorwas used to create a low shear mixing without the use of an agita-tor to allow settling of the agglomerated bioparticles in the adsorp-tion column. Scaled up airlift system would also be less energyintensive compared to a stirred tank. White and Gadd [50] showedthat an airlift bioreactor can remove thorium by fungal biomassmore effectively than a stirred tank. Pilkington et al. [29] alsofound that gas lift draft tube systems are attractive for large scaleindustrial applications as these systems provide good mixing,consume low power, and are of simple construction.

866 S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872

2.6. Adsorption surface area

Total yeast cell count in the column was measured using aNeubauer hemocytometer. The diluted cell suspension was placedin the counting chambers including 25 squares (each square has anarea 0.025 mm2 and depth of 0.1 mm) and the number of cells wascounted using an optical microscope with a 40� objective. Thenumber of cells per unit mass of biomass was calculated to be19.2 � 109 (±15%) cells/mg dry wt. The sizes of yeast cells in thebioreactor varied from 5 to 9 lm. Considering a prolate spheroidshape for yeast cells with polar diameter of 7.5 lm and equatorialdiameter of 6 lm, the surface area of cells, S, available for adsorp-tion of metals can be calculated as [32];

S ¼ 2pa2 1þ ba � e Sin�1e

� �ð6Þ

where a is the equatorial radius, b the polar radius, e2 ¼ ð1� a2=b2Þ,and arcsine value is in radians. Using Eq. (6), the external surfacearea of each cell is 132.4 lm2. Therefore, considering the above cal-culated cell population, the estimated total area available for masstransfer was �2.54 (±15%) m2/mg dry wt. cells.

Fig. 4. Mixing time comparisons for the airlift and stirred contactors of same

3. Results and discussion

3.1. Optimal physicochemical conditions

Environmental conditions such as solution pH, biomass concen-tration, and metal concentration are important factors affecting thebiosorption of metals. Among them, solution pH is of paramountimportance as it affects metal ion speciation in solution, surfacecharge of the biomass, and chemistry of biomass binding sites[2]. In the continuous biosorption system used in this study, thebiomass concentration in the bioreactor was constant (3 g/l). How-ever, biomass and metal concentrations varied in the adsorptioncolumn based on the feed concentration and the metal/biomassflowrate ratio. The mass balance results in the biosorption columnfor metals are summarized in Table 1. All experiments were con-ducted at the ambient temperature of 22 �C. Solution pH was auto-matically controlled with a pH-controller connected to acid andbase pumps with variable speeds (Fig. 1). The optimal biosorptionpH for each metal (�pH = 5.5 for Cu and pH = 5.0 for Pb) was usedbased on our earlier results obtained from batch adsorption exper-iments [2]. At steady state, the sugar content in the bioreactor’s

Table 1Metal and biomass concentration fluctuations in the inlet of biosorption tank affectedby the flowrate ratio of metal over biomass.

Metal FeedConcentration (mg/l)

Biomass (g/l)

2.5 1.5 0.75

Flowrate Ratio (metal/biomass)

0.2 1 3

C0, Initial Metal Concentration in Adsorption Column(mg/l)

10 – – 820 – 10 1530 – 15 22.560 10 30 45100 16.7 50 75120 20 60 90140 23.3 70 105180 30 90 135240 40 120 180360 60 180 –720 120 – –1080 180 – –

discharge was monitored using DNS method and was approxi-mately 0.1 g/l.

3.2. Mixing time

Mixing times of the airlift fluidized column and a same size stir-red tank equipped with one set of 45� pitched turbine 6-bladeimpeller were evaluated in different airflow rates and rpm’s,respectively. Mixing time was measured by adding 5 ml of 2 MHCl and 2 M of NaOH to each system and monitoring the rate ofchange of pH by a pH-meter connected to LabTech software, whichhad the ability of recording 60 pH values per second. Airflow of 2–8LPM (Std L min�1), and mixing rates of 150–350 rpm were usedand the mixing times were compared. This experiment confirmedthat mixing time does not improve beyond 3 LPM and 200 rpmfor the same volume airlift column and stirred tank, respectively(results not shown). Test results also showed that mixing timesfor the airlift fluidized contactor at 3LPM airflow is equivalent toa 200 rpm stirred tank (Fig. 4).

3.3. Continuous adsorption studies

Fig. 5 shows the adsorption test results for Cu(II) and Pb(II) ionswith yeast cells in the continuous system at steady state. Theamount of metal ions adsorbed per unit mass of biomass at equilib-rium increased with increase in initial metal concentrations. Yeast

volume.

Fig. 5. Metal uptake (mg/g) and percentage removal (%) plots for biosorption of testmetals at different initial concentrations and biomass dose of 1.5 g dry wt./l incontinuous biosorption system (data represents an average of three independentexperiments).

S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872 867

cells had good adsorption capacity for Pb2+ ions, exceeding 20 and74 mg of Pb per g of biomass for the 5 and 40 mg/l equilibriumconcentrations of Pb2+, respectively. For Cu2+ ions, the uptake ofmetals by cells at the same biomass and metal equilibrium concen-trations reached to about 4 and 16 mg Cu2+/g, respectively. Theincrease in adsorption capacity for metals by the increase in initialmetal concentration from 10 to 180 mg/l can be attributed toincrease in concentration difference as driving force for mass trans-fer of metals between the liquid phase and the adsorbent yeast bio-mass. Metals removal efficiencies were also influenced by theinitial concentration of the metals in the liquid phase (Fig. 5).Steady state removal of metals decreased with the initial metalconcentration due to fixed amount of biosorbent available for bio-sorption. On the other hand, equilibrium adsorption increased withthe increased concentration due to higher values of the numeratorin Eq. (1).

3.4. Continuous vs. batch biosorption

Batch studies for metal biosorption by yeast cells are widelyreported. The studies by Volesky [39] and Wang and Chen [44]on metal adsorption properties of S. cerevisiae indicated that yeastbiomass is a mediocre biosorbent due to fermentation broth resi-dues adhered to the surface of cells that affects metal uptakes.Other studies such as Bashar et al. [3] and Volesky [40] reportedhigh affinity of yeast cells for some cations such as Cd and Pb upto 10% of the mass of biomass. On the other hand, Volesky andMay-Phillips [43] observed a 3-fold increase in copper uptake bylive cells of Baker’s yeast compared to dead yeast S. cerevisiae.However, the study by [24] indicated that for Cu2+ ions, live anddead brewer’s yeast cells have similar accumulations. In this work,as depicted in Fig. 6, live yeast cells showed 25.2%, 21.3%, and 11.3%higher adsorption capacities for Cu2+ than previously reported fornon-living cells [2] at initial metal concentration of 10 mg/l,15 mg/l, and 30 mg/l, respectively, and at a biomass dose of1.5 g/l. However, metal removal percentage for living cells wasalmost the same as non-living cells at metal concentration of50 mg/l, and 14.3% lower at metal concentration of 60 mg/l.Biosorption of metals on non-living cells is known to be a passivephenomenon [44,42]. The Higher binding capacity of living yeastcells for copper implies an involvement of an active and dynamiccopper-binding mechanism. On the other hand, an averagedecrease of 9.5% was observed for the percentage removal of Pb2+

Fig. 6. Batch vs. continuous adsorption performance comparisons for Cu(II) andPb(II) ions under similar environmental conditions (e.g. biomass dose; 1.5 g/l, pH5.5 for Cu, and 5.0 for Pb adsorption). Heat-dried non-living cells used in previouslystudied batch experiments [2].

ions at the same range of metal concentrations. The lower Pb bind-ing capacity of live yeast cells than that of non-living cells could beexplained by live cells’ toxicity and metal tolerance mechanismsimpeding the excessive Pb bindings on the cells. Suh et al. [36]observed that Pb2+ ions penetrate into the inner cellular parts oflive S. cerevisiae after 24 h exposure to the metal, which may resultin a higher accumulation of Pb2+ ions by live cells than that of non-living cells. However, they showed a lower initial Pb2+ accumula-tion rates (<24 h) for live cells in comparison to dead cells, whichis in agreement with our findings in this work having a limitedcontact time in the biosorption column. The results for adsorptionof Cu(II) and Pb(II) ions shown in Fig. 6 indicated that live yeastcells may control the intracellular and surface uptake of the metalsbased on the possible level of toxicity of the aqueous phase sur-rounding the cells. The degree of toxicity in this work was relatedto the type of the essential metal Cu(II) vs. non-essential Pb(II) andtheir concentration level in the liquid.

3.5. Retention of metal-loaded biomass

For continuous adsorption, it is important to separate the spentbiomass followed by regeneration, a practice seldom is followed inbiosorption due to low cost of biosorbents. In this work, the poten-tial of regeneration of the spent yeast was evaluated. As shown inFig. 1, two overflow settling tanks in series are used for biomassseparation after biosorption. The good flocculation capability ofbrewing yeast cells is one of the major factors when brewers selectstrains for beer production. Yeast flocculation is facilitated in pres-ence of heavy metals, for example, cells usually settled in presenceof Ca2+ ions, which allows binding of specific cell wall proteins ofneighboring cells, called as ‘lectins’ [23,34]. Machado et al. [23]showed that more than 95% of the brewer’s yeast biomass was set-tled after 5 min at the presence of heavy metals acting as floccu-lants similar to Ca2+ions. In this work, settling time for copper-loaded biomass was faster than lead-loaded biomass. About 50%of biomass was settled naturally in 30 min after copper adsorption;however, Pb ions did not accelerate the settling time of the bio-mass after biosorption (Fig. 7). Similarly, Gouveia and Soares [9]showed that flocculent cells of S. cerevisiae remained dispersed inthe presence of Pb2+ ions, while they were able to flocculate inthe presence Cu2+, Ni2+, Zn2+ and Cd2+ions. Allowing higher cell-to-cell contact by decreasing the negative electrostatic charges inthe yeast cell walls, such as lowering the pH, can cause cells floc-culation [35]. Metal binding on the cell surface can also lowerthe negative cell surface charge and therefore enhance the poten-tial for cells’ flocculation. Since we were using baker’s yeast strainsof S. cerevisiae, the cells agglomeration was more likely to be the

Fig. 7. Comparison of biomass separation efficiency for metal-free biomass andmetal-loaded biomass.

Fig. 8. Comparison of sorption performance of yeast cells for metals at differentbiomass concentrations obtained at different metal inlet flowrates into thebiosorption vessel in continuous system.

868 S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872

responsible mechanism for facilitating the separation of Cu-loadedbiomass. The adsorption capacity for Pb2+ was much higher thanthe Cu2+ for S. cerevisiae; therefore, it was possible to have a chargereversal of the yeast biomass in presence of surface adsorbed Pb2+

causing lower agglomeration of the yeast cells.

3.6. Effect of metal mass flowrate

The continuous system’s response to changes in metal massflowrates is shown in Fig. 8. Increase in metal mass flowratesresulted in a decrease in the uptake of metals and percentageremoval due to decrease in biomass dose and contact time in theairlift biosorption column. As shown in Table 1, with the changeof metal/biomass flowrate ratio from 0.2 to 3, biomass concentra-tion in the adsorption column decreased from 2.5 g/l to 0.75 g/l.Fig. 8 shows the equilibrium isotherm sorption plots (Ce vs. qe)for different biomass concentrations in the continuous system. Itwas observed that, with the increase in the concentration of adsor-bent, the amount of uptake of metals also increased, which can beattributed to greater surface area and the availability of greateradsorption sites at higher amounts of biomass. Metal uptake valuesgradually achieved a plateau at higher metal equilibrium concen-trations. Increase in metal inlet concentration and also decreasein biomass concentration led to a decrease in metal percentageremoval of ions by yeast cells.

3.7. Adsorption isotherms

The effectiveness of metal-biomass interactions can be evalu-ated by sorption isotherm models, which can be utilized for opti-mization of adsorbent use [25]. The Langmuir model [21] is thewell-known monolayer sorption isotherm that relates the sorbate’sequilibrium concentration to sorbent’s capacity,

q ¼ qmaxbCe

1þ bCeð7Þ

Table 2Langmuir isotherm model parameters for biosorption of metals in continuous system.

Metal Initial metal Biomass qmax

Conc. (mg/l) Conc. (g/l) (mg/g)

Cu2+ 10–180 0.75 42.55Cu2+ 10–180 1.5 29.94Cu2+ 10–180 2.5 21.93Pb2+ 10–180 1.5 72.46

where qmax (mg/g) represents the maximum capacity of adsorbentfor metals and b (l/mg) is the Langmuir equation coefficient, andCe is the equilibrium metal concentration. The data shown inFig. 8 indicate that the Langmuir isotherm suited well for Cu2+

and Pb2+ adsorption by yeast cells. The Langmuir model values(qmax; b) calculated by fitting the model to the equilibrium adsorp-tion data and the model correlation coefficients are shown inTable 2. The maximum sorption capacity of yeast biomass forCu2+ ions decreased from 42.5 mg/g to 21.9 mg/g as the metal-bio-mass flowrate ratio decreased from 3 to 0.2 (Tables 1 and 2) in thecontinuous system. At the same biomass doses, yeast cells showed ahigher biosorption capacity for Pb2+ ions than of that for Cu2+ ionsfrom the values of qmax presented in Table 2.

To identify the favorable adsorbents and adsorption process,Hall et al. [11] introduced the following dimensionless equilibriumparameter RL based on the Langmuir isotherm coefficient b;

RL ¼1

1þ bC0ð8Þ

Eq. (8) was used by Ho et al. [15] and Chen and Wang [5] topredict whether a sorption process is favorable (0 < RL < 1), orunfavorable (RL > 1). The lower the value of RL is, the higherthe affinity of adsorbent to the adsorbed species. From Table 2,Pb2+ ions have lower RL values (0.061–0.537) than Cu2+ ions(0.155–0.768) at similar environmental conditions; thus, yeastbiomass has higher affinity for Pb2+ ions. By increasing biomassdose in the biosorption tank from 0.75 g/l to 2.5 g/l, RL values ofCu2+ ions slightly decreased suggesting more favorable adsorptionof metals at higher biomass concentrations. However, for all stud-ied test conditions in Table 2, the values of RL was less than one(0 < RL < 1) indicating favorable adsorption processes for bothmetals.

3.8. Adsorption kinetics

The time profile of Cu(II) adsorption by yeast biomass at differ-ent metal concentrations showed that metal biosorption processtook place in two steps (figure not shown): (1) rapid stage whichrepresented a surface adsorption mechanism; (2) slow stage untilthe biomass saturation was achieved, which was controlled by anintracellular diffusion process. Kinetics of adsorption of Cu(II) onyeast cells was tested by the first-order and second-order kineticmodels. First-order kinetic equation presented by Lagergren [18]based on the capacity of solid is in the following form;

dqt

dt¼ k1ðqe � qtÞ ð9Þ

where k1 is the rate constant of first order adsorption, qe is theamount of metals taken-up by the adsorbent at equilibrium(mg=g), and qt (mg=g) is the amount of metals adsorbed at anygiven time t (min).

The integrated expression of Eq. (9) by applying boundary con-dition of qt ¼ 0 at t = 0 can be written as follows;

lnðqe � qtÞ ¼ �k1t þ ln qe ð10Þ

b RL R2

(mmol/g) (l/mg)

0.670 0.0269 0.171–0.788 0.9970.471 0.0303 0.155–0.768 0.9950.345 0.0322 0.147–0.756 0.9950.350 0.0861 0.061–0.537 0.995

Table 3Copper adsorption rate constants associated with pseudo-first-order and pseudo-second-order kinetics equations.

Metal Conc. (mg/l) qe(exp) mg/g Pseudo-second-order Pseudo-first-order

k2 (g.mg�1.min�1) qe(Cal) mg/g R2 k1 (min�1) qe(Cal) mg/g R2

10 3.67 0.2625 3.70 0.9999 0.1709 2.43 0.945530 9.03 0.1934 9.07 1 0.2446 5.82 0.941160 15.90 0.1779 15.95 1 0.3110 9.56 0.9242

120 23.65 0.1712 23.70 1 0.4202 9.73 0.9668

S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872 869

The pseudo-second-order kinetic model is given as [14,13];

dqt

dt¼ k2ðqe � qtÞ

2 ð11Þ

where k2 is the rate constant of pseudo-second-order adsorptionðg=mg minÞ. By integrating Eq. (11) over the boundary conditionsof qt ¼ 0 at t ¼ 0, and qt ¼ qt at t ¼ t, and rearranging the followinglinear form of the pseudo-second-order kinetic model is obtained;

tqt¼ 1

qet þ 1

k2q2e

ð12Þ

The rate constants k1 and k2, the predicted metal uptake values(qe(cal)) are presented in Table 3. It can be seen that the secondorder kinetics equation fitted the experimental data much betterthan the first order.

For many adsorption systems, the pseudo-first order modelfound to fit the experimental data for an early stage of sorption[14], which was the case in this work. As shown in Fig. 9,pseudo-second order kinetics equation found to fit well in thewhole range of metal-biomass interaction time as the values ofqe(cal) obtained from these plots correlated well with the valuesmeasured experimentally with high correlation coefficientsR2 > 0.99 (Table 3). Similar results were reported by Chen andWang [5] for the removal of metals by biomass of S. Cerevisiae[10,25].

The rate constant values obtained from the plots variedbetween 0.1712 and 0.2625 g mg�1 min�1for 120 and 10 mg/lCu(II) concentrations, respectively. The decrease in the rate con-stant k2 indicated that the biomass saturation time is slightlylonger for the higher initial metal concentration [10]. This can bedue to active adsorption by live yeast cells that can hinder theintracellular accumulation of metals at higher metal concentra-tions for survival. This may lead to the hypothesis of surfaceadsorption at all metal concentrations, which occurs during theinitial fast adsorption stage, followed by an intracellular adsorptionat lower metal concentration (650 mg/l).

Fig. 9. A pseudo-second-order kinetic model applied to examine the effect of initialmetal concentration on rate constant (biomass dose 1.5 g/l).

3.9. Diffusion-based kinetics

The kinetic results were further analyzed by using Weber andMorris [49] intraparticle diffusion model as in Eq. (13);

qt ¼ kIt1=2 þ I ð13Þ

where kI is the inter-particle diffusion rate constant ðmg=g min0:5Þ,and I is the equation constant related to thickness of diffusion layer.Generally, as illustrated in Fig. 10, there are three steps involved inadsorption of species from liquid phase on a solid phase adsorbent[28,17] dictating the rate of sorption process: (I) bulk diffusion; (II)external diffusion through liquid film layer around the sorbent par-ticles; and (III) intraparticle diffusion. In the metal-biomass contac-tor used in this work, the mass transfer step (I) was very rapid,attributed to a well-mixed biosorption system, and thus steps (II)and (II) played important role in overall mass transfer rate of metalsfrom aqueous solution into the yeast cells. In order to evaluate therelative importance of these two steps, metal sorption kinetics wereevaluated using the intraparticle (Eq. (13)) and external mass trans-fer models. At stage (II) of sorption, metals are diffused from theliquid phase through the external diffusion layer (Fig. 10). Thechange in the adsorbate’s concentration in liquid phase at this stage,which is the initial rate of sorption, is given by the following equa-tion [27];

dCt

dt¼ �kEðCt � CsÞ ð14Þ

where kE (min�1) is the external mass transfer rate constant, and Cs

(mg/l) is the metal concentration at the surface of sorbate. Applyingboundary conditions of t ¼ 0; Cs ¼ 0 and Ct ¼ C0, Eq. (14) becomes;

dðCt=C0Þdt

� �t!0¼ �kE ð15Þ

As shown in Fig. 11, the uptake of copper by yeast biomass at allmetal concentrations followed three distinct stages. The metaluptake increased quickly at the beginning (0–5 min), which couldrepresent the negligible bulk diffusion and start of surface adsorp-tion step. Then, the change in the adsorption capacity qt sloweddown (5–20 min), which could be due to intraparticle diffusion,film diffusion, or both. Film diffusion could came into play whenmetal concentration difference (or mass transfer rate) betweenthe bulk liquid and the surface of the adsorbent decreased as the

External DiffusionLayer

Bulk Liquid

MembraneAdsorp�on

IntercellularDiffusion

Fig. 10. Schematic of the transfer of metals from the bulk liquid phase to a yeastcell.

Fig. 11. Intraparticle diffusion model for Cu biosorption on yeast cells at differentmetal concentrations. Multi-linearity of plots (three distinct regions) was observedfor metal uptake as the metal-biomass contact time increased.

Fig. 12. Slopes of second stage intraparticle diffusion plots (1.4 < t1/2 < 4) tocalculate the KI coefficient values.

Fig. 13. Time profile of Cu(II) ions dimensionless concentration Ct/C0 at differentinitial solute concentrations (biomass dose: 1.5 g/l). The external mass transferconstant (KE) for Cu adsorption on yeast cells was obtained from the initial slopes ofthe plots.

870 S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872

contact time increased, which denoted that the number of remain-ing adsorption sites on the biomass surface for binding of metalsbecame limited. Saturation of sorbent particles with metalsoccurred as the final plateau occurs. Value of I varied from2.1061 to 17.984 mg/g for 10 mg/l and 120 mg/l initial metal con-centrations, respectively (Table 4). Low value of I (I! 0) at lowmetal concentrations indicates that the qt vs. t0.5 plots should passthrough the origin (Fig. 12) and intracellular diffusion would con-trol the adsorption of metals by yeast cells. On the contrary, thelarger values of I means the greater boundary layer effect andtherefore, the greater the contribution of the surface sorption inthe rate controlling step. Furthermore, if the plots of Eq. (13) inall stages of adsorption do not go through the origin, it is an indic-ative of imprecision of the postulated intraparticle diffusion mech-anism [49,30]. The second stage of biosorption in Fig. 11 was usedto obtain the intraparticle diffusion rate constant, kI. The slopes ofthis stage are illustrated in Fig. 12, and the kI values with their cor-relation coefficients R2 (0.9328–0.9894) are tabulated in Table 4.Intra-particle diffusion rate constant of Cu(II) increased as the con-centrations of metal in the liquid phase increased. This suggestedthat the intraparticle diffusion model (Eq. (13)) predicts anincrease in internal mass transfer of Cu(II) ions; however, due toactive adsorption behaviour of live yeast cells, the natural resis-tance of the cells for internal uptake would be increased at highermetal concentrations. In this study, the plots of metal uptake qt

against t0.5 were linear in the second stage (Fig. 12), in which,the departure from the origin is larger at increasing metal concen-trations. This reveals possible involvement of some other mecha-nism along with intraparticle diffusion in controlling the rate ofsorption process. Such complex mechanism have also beenreported for biosorption processes by other researchers [1,6].

Graphical representations of Eq. (15), the external mass transfermodel in liquid phase at different initial copper concentrations, arepresented in Fig. 13. At time zero, Ct/C0 = 1, and then started todecrease with increasing the contact time. The slope of these plotswas calculated for the initial external mass transfer period(0–2 min), and the external mass transfer rate constants KE forthe tested metal concentrations (0.96 < R2 < 0.99) are listed in

Table 4External mass transfer and intraparticle diffusion parameters for biosorption of Cu ions on

Metal Conc. (mg/l) Intraparticle diffusion

KI (mg.g�1.min0.5) I (mg/g)

10 0.3406 2.106130 0.7731 5.932160 1.1321 11.657

120 1.5906 17.984

Table 4. The external diffusion rate KE of Cu(II) ions decreased withincreasing metal concentration in the solution.

Overall, the diffusion-based kinetics study revealed that thebiosorption of Cu(II) ions with live yeast cells tended to follow bothexternal and the intraparticle diffusion models, whereas intracellu-lar diffusion was the presumptive dominant mean of adsorption atlower metal concentrations.

3.10. Biosorption mechanism – further insights

Fermentation processes are relied on viable yeast cells; how-ever, biosorption of metals can be carried out by both living andnon-living cells. A simple cell viability count method was per-formed using methylene blue stain to determine the viability ofthe cells in the continuous system before and after being exposedto metals (non-viable cells do not have the metabolic capability todegrade the intruding methylene blue so they are stained darkblue). The test results showed that there was no significant

yeast cells (biomass concentration: 1.5 g/l).

R2 External diffusion

KE (min�1) R2

0.9894 0.1925 0.99060.9893 0.1742 0.99270.9472 0.1623 0.98740.9328 0.1245 0.9633

S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872 871

difference in viability for cells within first hour of exposure to themetals. The results were in agreement with findings of Liang andZhou [20]. They showed that yeast S. cerevisiae cells death underextracellular Cu stress starts at around 6 mM level, which wasquite higher than concentrations used in the present study(<2.9 mM). Thus, the biosorption process in this work was accom-plished by live yeast cells.

The metal biosorption mechanism is complicated and not fully-understood [44]. The analysis of the FTIR spectra of yeastbiosorbent showed the presence of several functional groups, suchas carboxyl, hydroxyl, amino and carbonyl groups on the cell wallsand their interaction with metal ions was responsible for metalbinding [2,46]. Results of this work confirmed that biosorption ofCu ions by yeast cells was a combination of both intracellular metalaccumulation and a metabolic-independent surface phenomenon.In contrast, comparison of passive biosorption results [2] with thiswork indicates that a non-metabolic biosorption occurred for Pb2+

(Fig. 6) due to high toxicity of lead and the active defense mecha-nism of the cells prevented the intake of toxic lead, and reducedthe amount adsorbed on the cell surface.

Although, the removal of Pb2+ in the continuous system waslower than that of in the batch system [2], maximum adsorptioncapacity of Pb2+ was still better than that of Cu2+. This is probablydue to the lower steric hindrance of Pb2+ due to smaller hydratedionic radii of Pb2+ ions as compared to hydrated ionic radii ofCu2+ [33]. Cu2+ ions with higher free enthalpy of hydration thanPb2+ ions prefer to stay in the liquid rather than adsorbing ontothe solid phase [25].

3.11. Large-scale applicability

Assessing the potential of biosorption process by live yeast cellsfor removal of heavy metals in a continuous system, similar to con-ventional technologies such as ion exchange, is of paramountimportance for large-scale design of this technology. The commer-cial development of this system can only be considered if its costand efficiency are reasonably better than commercially availableion-exchange resins. In the lab-scale system, the major cost ofthe process was related to the nutrition feed used for growth andproduction of yeast cells. Replacing the nutrients with an inexpen-sive waste stream containing sugar used for the growth of the cellswould improve the economic feasibility of the system.

The study by Machado et al. [23] on cell separation after bio-sorption of metals by brewer’s yeast strain of SaccharomycesCerevisiae showed that the yeast strain was able to sediment inthe presence of heavy metals acting as flocculants. Since the bio-sorption process is also rapid, removal of the heavy metals and cellseparation can be simultaneously achieved. The concept that heavymetal adsorption facilitates separation of biomass could be used inlarge scale continuous systems.

The challenges of industrial application of biosorption processwas addressed by Wang and Chen [45] proposing the developmentof hybrid technologies for removal of pollutant, especially usingliving cells. In this study, an application of a continuous systemfor cell growth and metal biosorption combining physical adsorp-tion and metabolic uptake of metals was demonstrated. Thisapproach has a potential for commercial application of biosorptionas it may lead to simultaneous removal of organic pollutants andother inorganic impurities including heavy metals.

If biosorption is to be used in treatment of industrial wastewa-ter effluents containing heavy metals ions, the regeneration pro-cess of biomass needs to be investigated to keep the operatingcosts down by opening the potential of metal recovery from bio-mass. Although, recovery of metal was not attempted in this work,due to low cost and abundance of biomass, recovery of metals fromthe spent biomass by acid wash is economically feasible.

4. Conclusions

In this work, the possibility of using a continuous system forsimultaneous production of yeast S. cerevisiae and removal of Cu(II)and Pb(II) ions from aqueous solutions by the produced biomasswas investigated. The yeast cells showed a higher binding capacityfor Pb2+ than Cu2+ ions. The biosorption data fitted Langmuiradsorption isotherm model and the kinetics of Cu(II) removal fol-lowed the pseudo-second-order model. The use of an air-liftadsorption tank promoted efficient mixing and effective contactbetween the metal solution and the biomass. Though live cellswere used, the biosorption on the cell surface was the major metalseparation mechanism. Intracellular metal uptake represented<10% of the overall metal binding capacity for copper ions. The dif-fusion-based kinetics models suited well to the time-course metaladsorption by live yeast cells suggesting involvement of bothextracellular and intracellular diffusion processes. The mainfinding of this work was that the continuous system is a practical,inexpensive, and self-contained method for removal of metals,where the adsorbent can be continuously produced and used formetal adsorption. Yeasts have minimal nutritional needs and theygrow easily in simple medium; therefore, the cost associated withnutrient medium can be overcome by the use of waste agriculturalmaterial such as molasses as a hydrocarbon source. The challengeof applying yeast biomass for continuous removal of heavy metalsin this work lied with the efficient separation of metal-loaded bio-mass from the solution. The biotechnological application of contin-uous metal biosorption by yeast is possible if the separation ofbiomass after the metal sorption is accelerated using a flocculentstrain of S. cerevisiae.

Acknowledgements

This work was supported by individual Natural Sciences andEngineering Research Council of Canada (NSERC) Discovery Grantsawarded to Dr. A. Margaritis and Dr. M.B. Ray.

References

[1] M.A. Ahmad, N.K. Rahman, Equilibrium, kinetics and thermodynamic ofRemazol Brilliant Orange 3R dye adsorption on coffee husk-based activatedcarbon, Chem. Eng. J. 170 (1) (2011) 154–161.

[2] S. Amirnia, A. Margaritis, M.B. Ray, Adsorption of mixtures of toxic metal ionsusing non-viable cells of Saccharomyces cerevisiae, Adsorpt. Sci. Technol. 30 (1)(2012) 43–63.

[3] H. Bashar, A. Margaritis, F. Berutti, B. Maurice, Kinetics and equilibrium ofcadmium biosorption by yeast cells S. cerevisiae and K. fragilis, Int. J. Chem.Reactor Eng. 1 (2003) 1–16.

[4] M.B. Cassidy, H. Lee, J.T. Trevors, Environmental applications of immobilizedmicrobial cells: A review, J. Ind. Microbiol. 16 (1996) 79–101.

[5] C. Chen, J. Wang, Removal of heavy metal ions by waste biomass ofSaccharomyces Cerevisiae, J. Environ. Eng. 136 (1) (2010) 95–102.

[6] F. Deniz, S. Karaman, Removal of Basic Red 46 dye from aqueous solution bypine tree leaves, Chem. Eng. J. 170 (2011) 67–74.

[7] M. Fomina, G.M. Gadd, Biosorption: current perspectives on concept, definitionand application, Bioresour. Technol. 160 (2014) 3–14.

[8] F. Gonen, Z. Aksu, Use of response surface methodology (RSM) in theevaluation of growth and copper (II) bioaccumulation properties of Candidautilis in molasses medium, J. Hazard. Mater. 154 (2008) 731–738.

[9] C. Gouveia, E.V. Soares, Pb2+ inhibits competitively flocculation ofSaccharomyces cerevisiae, J. Inst. Brew. 110 (2) (2004) 141–145.

[10] S.S. Gupta, K.G. Bhattacharyya, Kinetics of adsorption of metal ions oninorganic materials: a review, Adv. Colloid Interface Sci. 162 (1–2) (2011)39–58.

[11] K.R. Hall, L.C. Eagleton, A. Acrivos, T. Vermeulen, Pore-and solid-diffusionkinetics in fixed-bed adsorption under constant-pattern Conditions, I&ECFundam. 5 (1966) 212–223.

[12] Health Canada, Risk Management Strategy for Lead, available on Internet at:http://www.hc-sc.gc.ca/ewh-semt/pubs/contaminants/prms_lead-psgr_plomb/index-eng.ph’’, ISBN: 978-1-100-21305-7 (2013).

[13] Y.S. Ho, Review of second-order models for adsorption systems, J. Hazard.Mater. 136 (2006) 681–689.

[14] Y.S. Ho, G. McKay, Pseudo-second order model for sorption processes, ProcessBiochem. 34 (5) (1999) 451–465.

872 S. Amirnia et al. / Chemical Engineering Journal 264 (2015) 863–872

[15] Y.S. Ho, T.H. Chiang, Y.M. Hsueh, Removal of basic dye from aqueous solutionusing tree fern as a biosorbent, Process Biochem. 40 (1) (2005) 119–124.

[16] A. Kapoor, T. Viraraghavan, Removal of heavy metals from aqueous solutionsusing immobilized fungal biomass in continuous mode, Water Res. 32 (6)(1998) 1968–1977.

[17] D. Kumar, J.P. Gaur, Chemical reaction- and particle diffusion-based kineticmodeling of metal biosorption by a Phormidium sp.-dominated cyanobacterialmat, Bioresour. Technol. 102 (2011) 633–640.

[18] S. Lagergren, Zur theorie der sogenannten adsorption gelöster stoffe, K. Sven.Vetenskapsakad. Handl. 24 (4) (1898) 1–39.

[19] S.O. Lesmana, N. Febriana, F.E. Soetaredjo, J. Sunarso, S. Ismadji, Studies onpotential applications of biomass for the separation of heavy metals fromwater and wastewater, Biochem. Eng. J. 44 (2009) 19–41.

[20] Q. Liang, B. Zhou, Copper and manganese induce yeast apoptosis via differentpathways, Mol. Biol. Cell 18 (12) (2007) 4741–4749.

[21] I. Langmuir, The adsorption go gases on plane surface of glass, mica andplatinum, J. Am. Chem. Soc. 40 (1918) 1361–1403.

[22] N. Lokeshwari, K. Joshi, Biosorption of Heavy Metal (Chromium) UsingBiomass, Global J. Environ. Res. 3–1 (2009) 29–35.

[23] M.D. Machado, M.S.F. Santos, C. Gouveia, H.M.V.M. Soares, E.V. Soares,Removal of heavy metals using a brewer’s yeast strain of Saccharomycescerevisiae: the flocculation as a separation process, Bioresour. Technol. 99(2008) 2107–2115.

[24] M.D. Machado, S. Janssens, H.M. Soares, E.V. Soares, Removal of heavy metalsusing a brewer’s yeast strain of Saccharomyces cerevisiae: advantages of usingdead biomass, J. Appl. Microbiol. 106 (6) (2009) 1792–1804.

[25] S. Malamis, E. Katsou, A review on zinc and nickel adsorption on natural andmodified zeolite, bentonite and vermiculite: examination of processparameters, kinetics and isotherms, J. Hazard. Mater. 252–253 (2013) 428–461.

[26] P. Marques, H.M. Pinheiro, M.F. Rosa, Cd(II) removal from aqueous solution byimmobilised waste brewery yeast in fixed-bed and airlift reactors,Desalination 214 (2007) 343–351.

[27] G. McKay, M. El Geundi, M.M. Nassar, External mass transport processesduring the adsorption of dyes onto bagasse pith, Water Res. 22 (12) (1988)1527–1533.

[28] A.E. Ofomaja, Sorptive removal of methylene blue from aqueous solution usingpalm kernel fibre: effect of fibre dose’’, Biochem. Eng. J. 40 (2008) 8–18.

[29] P.H. Pilkington, A. Margaritis, N.A. Mensour, I. Russell, Fundamentals ofimmobilized yeast cells for continuous beer fermentation: a review, Inst. Brew.104 (1998) 19–31.

[30] V.J.P. Poots, G. McKay, J.J. Healy, The removal of acid dye from effluent usingnatural adsorbents � I. Peat, Water Res. 10 (1976) 1061.

[31] D.J. Reisman, W.B. Peirano, J.B. Lewis, D.K. Basu, D. Hohrseiter, SummaryReview of the Health Effects Associated with Copper: Health Issue Assessment,

U.S. Environmental Protection Agency, Washington, D.C., EPA/600/8-87/001,NTIS PB87137733 (1987).

[32] S.M. Selby (Ed.), CRC Standard Mathematical Tables, 14th ed., The ChemicalRubber Co., Cleveland, Ohio, 1969, p. 496.

[33] I.A. Sengil, M. Özacar, Competitive biosorption of Pb2+, Cu2+ and Zn2+ ions fromaqueous solutions onto valonia tannin resin, J. Hazard. Mater. 166 (2009)1488–1494.

[34] E.V. Soares, M. Mota, Quantification of yeast flocculation, J. Inst. Brew. 1997(103) (1997) 93–98.

[35] M. Stratford, Yeast flocculation: a new perspective, Adv. Microbiol. Physiol. 33(1992) 2–71.

[36] J.H. Suh, J.W. Yun, D.S. Kim, Comparison of Pb21 accumulation characteristicsbetween live and dead cells of Saccharomyces cerevisiae and Aureobasidiumpullulans, Biotechnol. Lett. 20 (3) (1998) 247–251.

[37] M. Tsezos, Engineering aspects of metal binding by biomass, in: H.L. Ehrlich,C.L. Brierly (Eds.), Microbial Mineral Recovery, McGraw-Hill, USA, 1990, pp.325–340 (Chapter 14).

[38] J.M. Tobin, C. White, G.M. Gadd, Metal accumulation by fungi: applications inenvironmental biotechnology, J. Ind. Microbiol. 13 (1994) 126–130.

[39] B. Volesky, Advances in biosorption of metals: selection of biomass types,FEMS Microbiol. Rev. 14 (1994) 291–302.

[40] B. Volesky, Sorption and Biosorption, BV-Sorbex Inc, St. Lambert, Quebec,2003.

[41] B. Volesky, G. Naja, Biosorption technology: starting up an enterprise, Int. J.Technol. Transf. Commercial 6 (2007) 196–211.

[42] B. Volesky, Biosorption and me, Water Res. 41 (2007) 4017–4029.[43] B. Volesky, H.A. May-Phillips, Biosorption of heavy metals by Saccharomyces

cerevisiae, Appl. Microbiol. Biotechnol. 42 (1995) 797–806.[44] J. Wang, C. Chen, Biosorption of heavy metals by Saccharomyces cerevisiae: a

review, Biotechnol. Adv. 24 (2006) 427–451.[45] J. Wang, C. Chen, Biosorbents for heavy metals removal and their future,

Biotech. Adv. 27 (2009) 195–226.[46] J. Wang, Biosorption of copper(II) by chemically modified biomass of

Saccharomyces cerevisiae, Process Biochem. 37 (8) (2002) 847–850.[47] B. Wehrheim, M. Wettern, Biosorption of cadmium, copper and lead by

isolated mother cell walls and whole cells of Chlorella fusca, Appl. Microbiol.Biotechnol. 41 (1994) 725–728.

[48] G.M. Walker, Yeast Physiology and Biotechnology, J. Wiley & Sons, 1998. pp.133–136.

[49] W.J. Weber, J.C. Morris, Kinetics of adsorption on carbon from solution, J.Sanitary Eng. Div. 89–2 (1963) 31–60.

[50] C. White, G.M. Gadd, Biosorption of radionuclides by fungal biomass, J. Chem.Tech. Biotechnol. 49 (1990) 331–343.