RSC Advances

PAPER

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article OnlineView Journal | View Issue

Department of Chemical Engineering, Na

620015, Tiruchirappalli – 620015, Tamil

Fax: +91-431-2500133; Tel: +91-431-250311

Cite this: RSC Adv., 2014, 4, 22129

Received 10th February 2014Accepted 17th April 2014

DOI: 10.1039/c4ra01179a

www.rsc.org/advances

This journal is © The Royal Society of C

Application of response surface methodology foroptimization of growth and lipids in Scenedesmusabundans using batch culture system

Chelladurai Chellamboli and Muthiah Perumalsamy*

Owing to an increased demand for fuel and the depletion of fossil fuels, an alternative source such as algae

is currently being exploited for biofuel production. One such potential microalgal candidate for biofuel

production is Scenedesmus abundans. The aim of this study was to optimize the growth and lipid

content of Scenedesmus abundans in a batch culture system in order to extend its applicability to the

commercial production of biodiesel. Optimization of growth parameters such as culture time,

concentration of inoculum, and concentration of sodium bicarbonate was performed by Central

Composite Design-Response Surface Methodology (CCD-RSM) to maximize production of biomass as

well as biofuel yield. The results obtained using CCD-RSM indicate that 30 days, 10% of inoculum and

8 g L�1 of sodium bicarbonate are optimum working parameters to achieve a tentative maximum

biomass yield of 0.0391 g per L per day and a maximum lipid content of 26.28% dcw. It was observed

that the model prediction values satisfactorily matched the investigated values, within �2%, for both

growth rate and lipid content. Moreover, the individual and synergistic effects of the operational

parameters on growth rate and lipid content were analyzed. The obtained experimental data were fitted

to Lineweaver–Burk, Langmuir Hanes, Eadie and Eddie Hofstee plots. Among these models, the

Langmuir Hanes method was found to provide the best fit, with a coefficient of regression of 0.9888.

Consequently, the results of the study indicate that Scenedesmus abundans would be an apt natural

source for biodiesel production.

Introduction

It is apparent that the ever-increasing demand for fossil fuelswill deplete supplies, hence causing their price to increase andrequiring the identication of alternative sources of energy.1

Biofuel is a viable alternative energy source, which can supportan eco-friendly approach to circumvent the issues over thetragedy and environmental damages caused by fossil fuel usage,such as the emission of greenhouse gases.2 Biofuels havetherefore been garnering attention over the past decade, andlipids from biosources have been identied as potentialproducers of biodiesel.3–6 Algae are the most preferred alterna-tive biological sources for biofuel production due to theirsimple and cheap cultivation techniques, efficient yield andperennial nature.7,8

Despite these attractive features, there are several difficultiesin the commercial cultivation of algae due to the lack of basicstudies to adopt the new techniques and due to environmentalconcerns. Current research focuses widely on conventionalmethods of algal cultivation, the inuence of various factors on

tional Institute of Technology, PO Box

Nadu, India. E-mail: mpsamy@nitt.edu;

2

hemistry 2014

algal growth and improving metabolic activity to stimulate lipidproduction.

Microalgae are unicellular or simple multi-cellular photo-synthetic microorganisms. Due to their simple structure,microalgae possess high-potential abilities, such as high growthrates and photosynthetic efficiency, to provide biofuel in aneconomical and environmentally sustainable manner.9–12 Priorstudies identied that the biomass yield of microalgae could be50 times more than that of switch-grass, which is the fastestgrowing terrestrial plant. Generally, microalgae are composedof 6% to 52% of proteins, 7% to 23% of lipids, and 5% to 23% ofcarbohydrates. These compositions are in the optimum rangefor algae species. Moreover, algae proteins have a relatively highC/N ratio of 10.2.13,14 Open pond cultivation of microalgaeinvolves carbon dioxide (CO2) utilization, which enriches theyield of biomass; for example, one kilogram of dry algal biomassutilizes about 1.83 kg of CO2.15,16

Exhaustive research involves the isolation and cultivation ofa local strain as well as extraction and purication of the biofuelto develop an industrially viable and economical productiontechnique. Yet, it is recommended to optimize the productionof biofuels from existing sources with respect to various inu-encing parameters, and to induce stress in the culture system soas to enrich the secretion of fatty acids, resulting in an increased

RSC Adv., 2014, 4, 22129–22140 | 22129

Table 1 Composition of M8 medium

Components Quantity required (g L�1)

KNO3 0.75KH2PO4 0.185NaHPO4 0.065CaCl2$2H2O 0.00325FeSO4$7H2O 0.0325MgSO4$7H2O 0.1Na$EDTA$Fe 0.01

RSC Advances Paper

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

yield of the biofuel. For instance, species of the genus Scene-desmus grown in a batch column reactor, and subjected toenvironmental conditions such as light, carbon dioxide (CO2),and temperature, was shown to grow at a maximum rate of 0.47per hour.17 Mandal et al.18 analyzed the growth response ofScenedesmus obliquus, which was cultivated under a distinctiverange of culture conditions, and the operating parameters wereoptimized by RSM. A maximum yield of lipid, 61.3%, wasobserved in the eighth day of culture with 0.04, 0.03, and 1.0 gL�1 of nitrate, phosphate, and sodium thiosulphate respectivelyin the culture medium. Penglin Li et al.19 conducted experi-ments in species of the genus Chlorella, grown in the presenceof rice straw hydrolysis products as a carbon source, and it wasfound that the maximum biomass concentration of 2.83 g L�1

was attained in 48 hours, and the lipid content achieved was ashigh as 56.3%. Barghbani et al.20 carried out growth experi-ments in Chlorella vulgaris and studied the effects of severalparameters such as sodium chloride, sodium bicarbonate, iron,light, and temperature, which were optimized by Taguchi'smethod. These experiments yielded a maximum biomassproduction of 3.56 g of dry matter at 10 g L�1 sodium chloride,absence of sodium bicarbonate, and 18 mmol L�1 iron at 30 �2 �C. The purpose of the present study is to activate themetabolism of microalgae by testing parameters such as carbonsource, culturing period, and concentration of inoculum in theculture system. Incorporation of CO2 in growth culture hassome disadvantages such as penetration and improper mixingof gases in the culture system that results in failure of theexperiments. Hence, the best suited carbon source for algae isinorganic carbon, i.e., bicarbonates. Algae are able to grow wellin highly alkaline conditions.21 Therefore, bicarbonate carbonsources supplied in the solid and liquid states were employed inthis study.

Optimization of growth parameters is the most importantcriterion to achieve the maximum production of biomass as wellas biofuel. The optimization tool used in this study was Design-Expert, including several methods such as a Taguchi' method,one-level factor, response surfacemethodology (RSM), etc. RSM isideally suited for most of the optimization studies as it requiresonly a small amount of data to identify the optimum response ofthe process. The recent study by Farshid Vejahati et al.22 exploitedRSM to optimize the working parameters while investigatingentrained-ow gasication of Canadian oil sand coke. In thisstudy, improvements in biomass yield and lipid content of Sce-nedesmus abundans were achieved by optimizing the operationalparameters for a batch culture system using RSM. Hence, theexperimental data on substrate utilization and species growthwere tted using different curve-tting kinetic models; inparticular, Lineweaver–Burk, Langmuir Hanes, Eadie and EddieHofstee Plots were used to identify the kinetic parameters andalso to nd the best tting model for the batch culture system.

Materials and methodsMicroalgae collection

The process for selecting the ideal strain plays a crucial role inbiofuel production; based upon preliminary studies, the

22130 | RSC Adv., 2014, 4, 22129–22140

microalgae Scenedesmus abundans was selected for this study.Scenedesmus abundans was procured from the National Chem-ical Laboratory, Pune, India. Initially, the strain was subjectedto sub-culturing techniques and a known quantity of micro-algae culture was regularly maintained in the culture room.Subsequent sub-culturing was carried out every two weeks tomaintain the activity of the fresh algal culture, and the culturesystem was continuously illuminated by an articial uorescentlamp at 2927.07 W m�2 for 12 hours alternating with 12 hoursof dark.

Experimental procedure

Known quantities of nutrients, vitamins and buffers were dis-solved in distilled water for media preparation, followed bysterilization. The growth medium was kept in an open atmo-sphere for 24 hours to facilitate carbon dioxide incorporationinto the medium. In this study, M8 medium was used as thegrowth medium, due to the relatively few components(shown in Table 1) required for its preparation.23 A knownquantity of M8 medium was subjected to an input of 2 to 10%exponential phase inoculum culture, for 2 to 30 days culturetime, and with 0 to 10 g L�1 sodium bicarbonate based onCentral Composite Design (CCD) in an Erlenmeyer ask toimprove biomass concentration and lipid productivity. Table 2shows the optimum ranges of the aforementioned independentvariables. Preparation of medium, inoculation of algal cultureand maintenance of culture system were carried out underaseptic conditions, and these experiments were performed atroom temperature. The RSM design tool was used to optimizethe independent variables such as culture time, inoculumconcentration and sodium bicarbonate. The culture system wasconstantly illuminated by articial light at the dark–light ratioof 12 : 12 hours per day for the photosynthesis reaction andmonitored at regular intervals with simultaneous measure-ments of growth rate and lipid content. 50 ml of culture waswithdrawn and centrifuged at 5000 rpm for 10 minutes, and thebiomass was dried at 105 �C until constant weights wereobtained (approximately in two hours). Lipid content was esti-mated by the Bligh and Dyer method.24,25

Methodology

Response Surface Methodology (RSM) is primarily used fordeveloping mathematical and statistical models to determinethe relationships among the several independent variables and

This journal is © The Royal Society of Chemistry 2014

Table 2 Optimum levels of independent variables for growth study inScenedesmus abundans

Factors Unit

Range and levels

�1 0 +1

Culture time Days 2 16 30Inoculumconcentration

% 2 6 10

Sodium bicarbonate g L�1 0 5 10

Paper RSC Advances

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

one or more dependent variables. In this study, CCD wasutilized to set a designed experiment to get an optimalresponse. CCD mainly creates a two-level factorial design,augmented with center points and axial points. Regular centralcomposite designs have ve levels for each factor, although thiscan be modied by choosing an alpha value of 1.0 for a face-centered CCFD. The face-centered design has only three levelsfor each factor and has been mainly developed for estimating aquadratic model. Its replicated center point provides excellentprediction capability near the center of the design space.26

Moreover, the CCFD model maintained a sodium bicarbonateconcentration between 0 to 10 g L�1, whereas the centrecomposite rotatable design (CCRD) model yielded negativevalues for a combination of inuencing factors and is hence notapplicable for the experiment. Therefore, CCFD was chosen inthis investigation. RSM has been used to identify the individualand combination effect of operating parameters.

Table 3 Design matrix and response data for Scenedesmus abundansgrowth study as determined by RSM

Standardorder

Culturetime(days)

Inoculumconcentration(%)

Sodiumbicarbonate(g L�1)

Growth rate(g per L)per day

Lipidcontent(% dcw)

1 2 2 0 0.0377 5.8102 30 2 0 0.0382 15.7893 2 10 0 0.0378 6.2454 30 10 0 0.0388 21.5895 2 2 10 0.0379 7.6256 30 2 10 0.0391 23.6897 2 10 10 0.0378 8.2358 30 10 10 0.0391 25.6799 2 6 5 0.0378 4.35810 30 6 5 0.0389 24.54011 16 2 5 0.0385 10.25912 16 10 5 0.0382 18.64013 16 6 0 0.0381 12.59014 16 6 10 0.0381 15.86015 16 6 5 0.0381 13.56916 16 6 5 0.0381 13.56917 16 6 5 0.0381 13.56918 16 6 5 0.0381 13.56919 16 6 5 0.0381 13.56920 16 6 5 0.0381 13.569

Results and discussion

As compared to other simulation models, the RSM model is arelatively simple method to optimize and estimate the operatingparameters, and this model allows for optimization with rela-tively few data points. The RSM model derived in this study wasvalidated by the examination of how the individual responsessuch as growth rate and lipid content are inuenced by inde-pendent variables, i.e., culture time, inoculum concentrationand sodium bicarbonate. The experimental data were tted tothe second-order polynomial model shown in eqn (1).

Y ¼ b0 + b1X1 + b2X2 + b3X3 + b4X4 + b12X1X2 + b13X1X3

+ b14X1X4 + b23X2X3 + b24X2X4 + b34X3X4 + b11X12

+ b22X22 + b33X32 + b44X42 (1)

In this equation, Y is the response in coded units; b0 is aconstant; b1, b2, b3 and b4 are the regression coefficients forlinear effects; b11, b22, b33 and b44 are the quadratic coefficients;b12, b13, b14, b23, b24 and b34 are the interaction coefficients; andX1, X2, X3 and X4 are the parameters considered.27,28 The modelproposed for each response by RSM can be easily applied topredict the values for any combination of the parametersconsidered in the experimental domain.29 In this study, thegrowth rate and lipid content of Scenedesmus abundans werecontrolled by independent variables, i.e., culture time, inoc-ulum concentration, and sodium bicarbonate whereas thedependent output variables were the maximum growth rate and

This journal is © The Royal Society of Chemistry 2014

lipid content. The obtained results are shown in Table 3, whichdescribes the combination of explanatory variables and thecorresponding response variables.

Response on growth rate of Scenedesmus abundans

The nal quadratic equation obtained for the growth rate ofScenedesmus abundans is given below in eqn (2).

Growth rate of Scenedesmus abundans (g per L per day),

Y1 ¼ (0.037875 � (4.00297 � 10�6A) � (9.71294 � 10�5B)

+ (7.95287 � 10�5C) + (1.34286 � 10�6AB)

+ (2 � 10�6AC) � (5.36 � 10�6BC) + (7.07236

� 10�7A2) + (9.41364 � 10�6B2)

� (5.04727 � 10�6C2)) (2)

The individual inuences of A, B, and C and the combinedeffects of AB, BC, and AC were signicant in both actual andpredicted response data. The graphical analyses of the entireeffects and interaction plots were clear. In order to obtain a betterresponse surface result for the growth rate of Scenedesmusabundans, a slight modication was applied to the modelresponse equation of the growth rate. The growth rate was moreor less showing the same response for both the predicted andmodied models. The absolute average deviation for the pre-dicted model was 0.006%, and for the customized model it was1.2382%. The root mean square deviation for modied andpredicted model was 1.7% and 0.2366% as shown in Table 4. Themodied model for the Scenedesmus abundans considering thegrowth rate as its response variable is shown in eqn (3) (ref. 30).

Proposed response model for growth rate of Scenedesmus abun-

dans (g per L per day), Y1 ¼ (0.037875 – (4.00297 � 10�6A) –

RSC Adv., 2014, 4, 22129–22140 | 22131

Table 4 Estimated data for proposed response models as a function of code operating variablesa

Standard order

Growth rate (g per L per day) Lipid content (% dcw)

Experimental Predicted Proposed model Experimental Predicted Proposed model

1 0.0377 0.0377 0.0377 5.8100 4.3053 4.30532 0.0382 0.0383 0.0376 15.789 16.375 16.3753 0.0378 0.0379 0.0369 6.2450 6.9710 6.97104 0.0388 0.0388 0.0372 21.589 22.413 22.4145 0.0379 0.0379 0.0384 7.6250 6.9808 6.76096 0.0391 0.0391 0.0389 23.689 23.143 22.9247 0.0378 0.0377 0.0372 8.2350 7.8290 7.60918 0.0391 0.0391 0.0380 25.679 27.364 27.1449 0.0378 0.0378 0.0376 4.3580 6.1871 6.132110 0.0389 0.0388 0.0380 24.540 21.990 21.93511 0.0385 0.0383 0.0382 10.259 12.367 12.31212 0.0382 0.0383 0.0373 18.640 15.811 15.75613 0.0381 0.0379 0.0374 12.590 11.958 11.95814 0.0381 0.0382 0.0382 15.860 15.771 15.55115 0.0381 0.0381 0.0378 13.569 13.809 13.75516 0.0381 0.0381 0.0378 13.569 13.809 13.75517 0.0381 0.0381 0.0378 13.569 13.809 13.75518 0.0381 0.0381 0.0378 13.569 13.809 13.75519 0.0381 0.0381 0.0378 13.569 13.809 13.75520 0.0381 0.0381 0.0378 13.569 13.809 13.755AAD (%) — 6 � 10�3 1.2382 — 1.3067 0.6333RMSD (%) — 0.2366 1.7000 — 13.3135 13.2546

a Note: AAD – absolute average deviation; RMSD – root mean square deviation.

RSC Advances Paper

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

(9.71294 � 10�5B) + (7.95287 � 10�5C) + (1.34286 � 10�6AB) +

(2 � 10�6AC) – (5.36 � 10�6BC)) (3)

Inuence of culture time on growth rate

The effect of culture time on the growth of Scenedesmus abun-dans was investigated at three different intervals, such as 2, 16and 30 days. It was observed that the growth rate increased asculture time was prolonged with the depletion of nutrients inculture media. This is because the newly cultured algae cellsrequire a certain time period to adapt to the culture

Fig. 1 Effects of different culture times and inoculum concentrations o

22132 | RSC Adv., 2014, 4, 22129–22140

environment. The proportional change in growth rate of Sce-nedesmus abundans with culture time is shown in Fig. 1. It wasnoted that the growth rate was higher on the 30th day, than on2nd and 16th day culture. The 30th day species, taking intoaccount of all the parameters that have a cause effect onbiomass yield. Usually, algal cells are made up of multi-cellularmembranes that are highly permeable in nature to allow thepassage of water rather than of solutes or ions. Water can moveacross the membranes by osmosis. The solutes available in themedium were salts as well as organic and inorganic molecules

n growth rate of Scenedesmus abundans.

This journal is © The Royal Society of Chemistry 2014

Paper RSC Advances

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

dispersed in the distilled water (medium). The solvent iscapable of penetrating into the cellular membrane, which isconsumed by algae cells for their growth.31 The consumption ofthe cellular membrane results in the form of cell division andbiomass yield increased during growth period. The time takenfor establish cell generation and cell stability is slightly longerin a parent cell divides into two daughter cells than in twodaughter cells further multiples into four daughter cells. Theprocess extend till depletion of nutrients in the medium isknown as meiosis and mitosis process. The uptake of nutrientsfrom the medium by Scenedesmus abundans results in theirdepletion aer a certain period of time, slowing down thereproduction rate, and ending the investigation when steadystate was achieved.

Inuence of inoculum concentration on growth rate

In the uncarbonated condition, the biomass yield of Scene-desmus abundans was 0.791 g L�1 for a 2% inoculum concen-tration and 1.023 g L�1 for a 10% inoculum concentration,indicating that achieving an optimum inoculum concentrationdid not greatly inuence the biomass yield. Hence, the systemoperating in the presence of 10 g L�1 of sodium bicarbonate andin 2% and 10% of inoculum concentrations revealed the same1.908 g L�1 biomass yield, which was obtained on the 30th day.The variation in inoculum concentration doesn't have much ofan inuence on biomass yield for the carbonated growth study.

Inuence of sodium bicarbonate on growth rate

The effect of sodium bicarbonate on the growth rate of Scene-desmus abundans is shown in Fig. 2 and 3. An increase in theamount of sodium bicarbonate led to high yields of biomass, asexpected due to the natural high requirement of a carbon sourceby the algal cells. Carbon source regulates the activity of theCO2-dependent cell growth and broadens the source of muta-tions in the algal cells. Naturally occurring CO2 in the air is notsufficient for algal growth at a large-scale level; an articial

Fig. 2 Effects of inoculum and sodium bicarbonate concentrations on t

This journal is © The Royal Society of Chemistry 2014

source such as sodium bicarbonate or sodium thiosulphate isthus required.

Effects of culture time and inoculum concentration on growthrate

The growth rate of Scenedesmus abundans increased with anincrease in culture time. But, the combination of culture timeand inoculum concentration didn't signicantly inuence thegrowth rate of the study species as shown in Fig. 1. Themaximum growth rate of Scenedesmus abundans was 0.0391 gper L per day for inoculum concentrations of 2% and 10%, andthe corresponding culture time was 30 days.

Effects of inoculum concentration and sodium bicarbonateon growth rate

Fig. 2 veries that the increase in inoculum concentrationslightly improved the growth rate while sodium bicarbonate washighly inuential on the growth rate of the culture. Themaximum growth rate of Scenedesmus abundans was 0.0391 gper L per day, which occurred in the presence of a highconcentration of sodium bicarbonate (10 g L�1) with both thebase inoculum concentration (2%) and also the increasedinoculum concentration (10%). High biomass productionoccurred in both low and high concentrations of initial inoc-ulum with an increase in bicarbonate, because the growth ofalgae mainly depends on the carbon source for their metabolicactivities, mitosis and meiosis processes, and sparing of CO2

gas in culture having some disadvantage in growth, it’s due toimproper mixing and penetration in parent cells and results inproduction of less number of cells.

Effects of culture time and sodium bicarbonate on growth rate

The maximum growth rate of Scenedesmus abundans was foundto be 0.0391 g per L per day for the 30th day of growth and at asodium bicarbonate concentration of 10 g L�1, as shown in

he growth rate of Scenedesmus abundans.

RSC Adv., 2014, 4, 22129–22140 | 22133

Fig. 3 Synergistic effect of culture time and sodium bicarbonate concentration on growth rate of Scenedesmus abundans.

RSC Advances Paper

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

Fig. 3. The maximum optimized value of bicarbonate wasdetermined to be 8 g L�1. The supplied sodium bicarbonatecreated the respiration environment for algae, justifying itsincorporation in the growth medium. The high amount ofcarbon source ended with a high rate of carbon dioxideproduction, which increased the growth of algal culture.

Comparison of experimental and predicted values by RSM

The predicted values of the growth rate of Scenedesmus abun-dans using the quadratic equation generated by RSM have beencompared with the empirical values. The comparison showedthat the model prediction satisfactorily matched the investi-gated values to within �2%. An analysis of variance (ANOVA) todetermine the signicant effects of process variables was con-ducted and the results are shown in Table 5. It can be noticedfrom the table that the F-values for the regressions were rela-tively high. The large F-value indicated that most of the variancein the response could be explained by the regression modelequation. The associated p-value was used to calculate whetherthe F-statistics were robust enough to indicate the statisticalsignicance. The low p-value (<0.0001) indicated that the modelwas statistically signicant.32 The model adequacy was checkedby R2 and Radj

2. The relatively high value of R2 (0.9535) showedthat the model could predict the response successfully. The

Table 5 Analysis of variance of the 23 factorial designs for the growth r

Source Degrees of freedom Sum of squares

Regression 9 3.354 � 10�6

Residual error 10 1.634 � 10�7

Lack of t 5 1.631 � 10�7

Pure error 5 3.072 � 10�10

Total 19 3.517 � 10�6

a R2 ¼ 0. 9535, Radj2 ¼ 0. 9117.

22134 | RSC Adv., 2014, 4, 22129–22140

model adequacy has also been veried with Radj2 value. The

ANOVA indicated that the second-order polynomial model wassignicant and adequate to represent the actual relationshipbetween the response and explanatory variables, with as indi-cated above a small p-value (<0.0001) and a high value of R2

(0.9535) for the growth rate of Scenedesmus abundans.Comparison of the experimental and predicted data of thegrowth rate is provided in Fig. 4.

Lipid content response on Scenedesmus abundans

The nal quadratic eqn (4) obtained for lipid content is givenbelow:

Lipid content (% dry cell weight (dcw)),

Y2 ¼ (0.327162 + (0.35542 � A) + (0.09635B) + (0.26177C)

+ (0.015056AB) + (0.014616AC) � (0.022719BC)

+ (1.43232 � 10�3A2) + (0.017466B2)

+ (2.1981 � 10�3C2) (4)

In order to improve the response model, a modication hasbeen made in the design of the predicted model responseequation. The results were found to be similar in both predictedand modied model responses to the lipid content. A summaryof the actual, predicted, and proposed model data, as well as ofassociated deviation, is shown in Table 6. The absolute average

ate response of Scenedesmus abundansa

Mean square F value P value

3.726 � 10�7 22.81 <0.0001 (signicant)1.634 � 10�8 — —3.261 � 10�8 — —6.144 � 10�11 — —— — —

This journal is © The Royal Society of Chemistry 2014

Fig. 4 Comparison of predicted and actual (experimental) values of growth rate of Scenedesmus abundans by RSM methodology.

Paper RSC Advances

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

deviation for the predicted model was 1.3067% and for themodied model was 0.6333%. Root mean square deviations forthe predicted and as well as modied model were calculated tobe 13.3135 and 13.2546% as shown in Table 4. The customizedmodel equation for lipid content is as follows.

Proposed model of lipid content of Scenedesmus abundans (% dcw),

Y2 ¼ (3.27162 + (0.3554A) + (0.09351B) + (0.26177C)

+ (0.015056 � AB) + (0.014616 � AC) � (0.022719BC)

+ (1.42324 � 10�3A2) + (0.017466B2)) (5)

Effects of culture time and inoculum concentration on lipidcontent

In this study, the maximum lipid content of 25.679% dcw wasobtained at the optimum conditions, which were a long culturetime of 30 days and a 10% inoculum concentration. Themaximum growth rate of Scenedesmus abundans was 0.0391 gper L per day, on the 30th day, and at inoculum concentrationsof 2 and 10%, as demonstrated in Fig. 5. Every biologicalorganism requires a precise time interval to activate cell divi-sion and convert the growth medium into suitable adoptedenvironment that in turn can be taken into account for furtherprocesses for example convert sodium bicarbonate into carbon

Table 6 ANOVA results of the quadratic models from lipid content of S

Source Degrees of freedom Sum of squares

Regression 9 707.45Residual error 10 30.59Lack of t 5 30.59Pure error 5 0.000Total 19 738.04

a R2 ¼ 0.9586, Radj2 ¼ 0.9213.

This journal is © The Royal Society of Chemistry 2014

dioxide. Hence, culture time as well as inoculum concentrationplayed an important role in inuencing the lipid content.

Effects of inoculum concentration and sodium bicarbonateon lipid content

Increasing the concentrations of inoculum and sodium bicar-bonate enhanced the lipid content of Scenedesmus abundans.Fig. 6 shows the maximum lipid content of 25.679% of dcwoccurred at high concentrations of inoculum (10%) and sodiumbicarbonate (10 g L�1). The lipid content started to decline oncethe treatment with sodium bicarbonate began losing its vigor.This decrease in lipid production was largely due to the lack ofbiomass yield, thus correlating the biomass yield with thefrequency of sodium bicarbonate treatment.

Effects of culture time and sodium bicarbonate on lipidcontent

Fig. 7 shows that the maximum lipid content of 25.679% dcwwas achieved at two different optimal conditions: highculture time (30th day) and sodium bicarbonate concentration(10 g L�1). From the results, it could be understood that thelipid content decreased to 24.54% dcw upon a decrease ofsodium bicarbonate to 5 g L�1. The present study conrms

cenedesmus abundansa

Mean square F value P value

78.61 25.70 <0.0001 (signicant)3.06 — —6.12 — —0 — —

— — —

RSC Adv., 2014, 4, 22129–22140 | 22135

Fig. 5 Effects of culture time and inoculum concentration on lipid content of Scenedesmus abundans.

Fig. 6 Model for the combined effect of inoculum concentration and sodium bicarbonate concentration on lipid content of Scenedesmusabundans.

RSC Advances Paper

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

sodium bicarbonate as the most functional carbon source forScenedesmus abundans growth, as the optimum bicarbonatetreatment has been advantageous for maximum biomassproduction; the study also indicates that the carbon source hasa pronounced effect on microalgal growth and rapid productionof biomass in the tested strain and hence leads to a higher rateof lipid production.

Comparison of actual and predicted values by RSM

Fig. 8 shows the predicted values of lipid content using thequadratic equation generated by RSM, which have beencompared with the investigated values. The comparison showedthat the model prediction satisfactorily matched the experi-mental values to within �2%. An analysis of variance to deter-mine the signicant effects of process variables was conducted

22136 | RSC Adv., 2014, 4, 22129–22140

and the results are shown in Table 6. The ANOVA indicated thatthe second-order polynomial model was signicant andadequate to represent the actual relationship between theresponse variable (lipid content) and the independent variables,with a small p-value (<0.0001) and a high value of R2 (0.9586) forlipid content.

Study the growth response of Scenedesmus abundans culturedin different levels of carbon source as substrate concentrationin batch culture

In this study, a signicant difference was shown by growingalgae in the presence and absence of a carbon source. Theoptimum conditions arrived at in this study were applied to abatch culture system containing M8 medium with supporting,controlling parameters such as 10% of inoculum, and a

This journal is © The Royal Society of Chemistry 2014

Fig. 8 Comparison of predicted and actual values of lipid content of Scenedesmus abundans from Design expert® 9.

Fig. 7 Effects of culture time and sodium bicarbonate concentration on lipid content in Scenedesmus abundans.

Paper RSC Advances

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

constant photosynthetic effect from illumination by articiallight at 14 photons per unit area per unit time for a time periodof 30 days. The time prole corresponding to dry weight basisbiomass is shown in Fig. 9a; the effect was found to be steadyand the maximum biomass growth was achieved in the 30th dayof culture. The biomass concentrations in the culture growthmedium both in the presence and absence of carbon sourcewere 1.608 and 1.514 g L�1, which occurred on the 30th day. Atthe same time, the study species is subjected to several levels ofcarbon source as active substrate concentration in the batchculture system. The maximum biomass concentration occurs incarbon concentrations of zero and 10 g L�1. An increase ofcarbon concentration has signicant results in the yield ofbiomass, reected in experimental run and the data are valid by

This journal is © The Royal Society of Chemistry 2014

using kinetic models. The analyzed treatments have played asignicant role in an effective result feature in experimentaldata and this data are valid by using kinetic models. Theconcept reecting form of this study has been depicted inFig. 9(a) and (b) which describes the increases in biomassconcentration with respect to extending the carbon sources upto a certain level and equivalently the specic growth rate wasdecreased.

The effect of sodium bicarbonate was amplied with anincrease in the pH of the medium due to the prevalence of astress directed towards exponential production of the biomassvia amplication of the photosynthetic activity. This studysummarizes the effects of substrates that boost the yield oflipids as well as of biomass.

RSC Adv., 2014, 4, 22129–22140 | 22137

Fig. 9 (a) Individual growth responses of Scenedesmus abundanssubjected to different levels of substrate concentrations with respectto incubation time period and dry weight of biomass concentration. (b)The effect of different substrate concentrations on the specific growthrate of Scenedesmus abundans as a function of time.

Fig. 10 Error bars for predicted and proposed models of growth rateand recovery of lipid.

Table 7 Linear plots

Curve tting model Methods

1

m¼ Ks

mmax

1

Sþ 1

mmax

Lineweaver–Burk plot

S

m¼ Ks

mmaxþ S

mmax

Langmuir Hanes plot

m

S¼ m

Ksþ mmax

Ks

Eadie plot

RSC Advances Paper

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

Determination of statistical error bars for predicted andproposed models

The accuracy of the predicted and developed models for thegrowth rate and lipid recovery percentage has been estimatedfrom the absolute average deviation of the individual response.The average percentage errors for the predicted and proposedgrowth models were 0.0006 and 1.2382; the percentage errorsfor the predicted and modied models of lipid % response were1.3067 and 0.6333. The error bars for different models areshown in Fig. 10. The average percentage error for the growthrate was relatively low in the predicted model, and was relativelyhigh for the proposed model due to the optimum conditions ofmicroalgal culture being relatively sensitive. The estimated lipidcontent was predicted to be high and the proposed modelshowed a relatively low percentage of error. Thus, the predictedmodel for growth rate and the developedmodel for lipid% (dcw)were relatively reliable for the batch study.

22138 | RSC Adv., 2014, 4, 22129–22140

Kinetics studies of linear models

The experimental data for substrate concentrations during thegrowth phase of Scenedesmus abundans in batch culture systemswere used for the determination of kinetic parameters. Thekinetic parameters identied on the curve tting models, i.e.,Lineweaver–Burk, Langmuir Hanes, Eadie and Eddie Hofsteeplots,33–36 are shown in Table 7. These plotting methods deter-mine the kinetic parameters from standard kinetic models. Thekinetic parameters are mmax, maximum specic growth rate andthe half-saturation coefficient Ks. The Lineweaver–Burk plotmethod was linearized from the Monod equation. All of themethods used in this curve tting procedure are based on thesubstrate concentration and have been validated for the twokinetic parameters. The symbols used in the above-mentionedmodels are m for the specic growth rate (per day), mmax for themaximum specic growth rate (per day), Ks for the half-satura-tion coefficients (g L�1) and S for the substrate concentration(g L�1). The result obtained from tting the model with theexperimental data had shown a relatively satisfactory relationbetween the regressions from the linear plots, i.e., 0.8259 forLineweaver–Burk, 0.9888 for Langmuir Hanes and 0.885 in bothEadie and Eddie Hofstee plots, as shown in Fig. 11. Thisconsistency among experiments show that sodium bicarbonateis a limiting substrate in functional concentrations and is nothaving any suppressing effect on the cell growth at a certainlevel of carbon source. The linearized model of Monod isLineweaver–Burk plot, the maximum specic growth rate (mmax)

m ¼ mmax � Ksm

SEddie Hofstee plots

This journal is © The Royal Society of Chemistry 2014

Fig. 11 Fitting the explanatory data in the linear plots (a) Lineweaver–Burk plot for the inverse of substrate concentration versus specific growthrate, (b) Langmuir Hanes plot, (c) Eadie plot and (d) Eddie Hofstee plot.

Paper RSC Advances

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

is 0.0598 day�1 and Monod constant (K) is 0.6576 g L�1. ForLangmuir Hanes, Eadie and Eddie Hofstee plots, the maximumspecic growth rates and half-saturation coefficients are 0.0599,0.0951, 0.0822 per day and 0.4918, 1.5770, 1.3958 g L�1,respectively.

Conclusion

This study shows that Scenedesmus abundans has ample poten-tial for high biomass yield and lipid content, which wereobserved in the presence of a carbon source, a long culture timeand both the base and high concentrations of the motherculture. The operating parameters such as culture time, inoc-ulum concentration and sodium bicarbonate concentrationwere studied and optimized using Central Composite Design ofResponse Surface Methodology (CCD-RSM). The effects of indi-vidual and combined operating parameters on the growth rateand lipid content of Scenedesmus abundans indicated amaximum growth rate of 0.0391 g per L per day and lipid contentof 26.282% dcw on the 30th day, based on a 10% of inoculumconcentration and 8 g L�1 of sodium bicarbonate. The studiedbatch processes of experimental data are also better tted withthe Langmuir Hanes model than with the Lineweaver–Burk,Eadie and Eddie Hofstee models. Hence, commercial produc-tion of biodiesel from the perennial algal source Scenedesmusabundans will be the future focus of our work.

Notes and references

1 K. Ramluckan, G. Kandasamy Moodley and F. Bux, Fuel,2014, 116, 103–108.

This journal is © The Royal Society of Chemistry 2014

2 M. Kee Lam and K. Teong Lee, Biotechnol. Adv., 2012, 30,673–690.

3 B. Liu and Z. Zhao, J. Chem. Technol. Biotechnol., 2007, 82,775–780.

4 X. Meng, J. Yang, X. Xu, L. Zhang, Q. Nie and M. Xian,Renewable Energy, 2009, 34, 1–5.

5 B. D. Wahlen, M. R. Morgan, A. T. McCurdy, R. M. Willis,M. D. Morgan, D. J. Dye, B. Bugbee, B. D. Wood andL. C. Seefeldt, Energy Fuels, 2013, 27, 220–228.

6 H. Shen, Z. Gong, X. Yang, G. Jin, F. Bai and Z. K. Zhao,J. Biotechnol., 2013, 168, 85–89.

7 L. Brennan and P. Owende, Renewable Sustainable EnergyRev., 2010, 14, 557–577.

8 S. Amin, Energy Conv. Management., 2009, 50, 1834–1840.9 G. Dragone, B. Fernandes, A. Antonio Vicente andJ. A. Teixeira, Appl. Microbiol. Biotechnol., 2010, 1355–1366.

10 L. Rodol, G. ChiniZittelli, N. Bassi, G. Padovani, N. Biondi,G. Bonini and M. R. Tredici, Biotechnol. Bioeng., 2009,102, 1.

11 H. C. Greenwell, L. M. L. Laurens, R. J. Shields, R. W. Lovittand K. J. Flynn, J. R. Soc., Interface, 2010, 7(46), 703–726.

12 R. H. Wijffels and M. J. Barbosa, Science, 2010, 329, 796–799.13 M. R. Brown, S. W. Jeffrey, J. K. Volkman and G. A. Dunstan,

Aquaculture, 1997, 151, 315–331.14 J. Singh and S. Gu, Renewable Sustainable Energy Rev., 2010,

14, 2596–2610.15 Y. Chisti, Biotechnol. Adv., 2007, 25(3), 294–306.16 G. Pokoo-Aikins, A. Nadim, M. M. El-Halwagi and

V. Mahalec, Clean Technol. Environ. Policy, 2009, 12(3),239–254.

RSC Adv., 2014, 4, 22129–22140 | 22139

RSC Advances Paper

Publ

ishe

d on

21

May

201

4. D

ownl

oade

d by

Uni

vers

ity o

f W

ater

loo

on 0

2/06

/201

4 14

:01:

48.

View Article Online

17 M. Hanuman Reddy. Application of Algal CultureTechnology for Carbon Dioxide and Flue Gas EmissionControl, MS thesis, Arizona State University, 2002, pp. 1–96.

18 S. Mandal and N. Mallick, Appl. Microbiol. Biotechnol., 2009,84, 281–29.

19 P. Li, X. Miao, R. Li and J. Zhong, J. Biomed. Biotechnol., 2011,141207.

20 R. Barghbani, K. Rezaei and A. Javanshir, Int. J. Biotechnol.Wellness Ind., 2012, 1, 128–133.

21 Y. Chisti, J. Biotechnol., 2013, 167, 201–214.22 F. Vejahati, H. Katalambula and R. Gupta, Energy Fuels, 2012,

26, 219–232.23 K. O. Cassidy, Biosystems and Agricultural Engineering,

Thesis (M.S), 2011, University of Kentucky, pp. 1–47.24 E. G. Bligh and W. J. Dyer, Can. J. Biochem. Physiol., 1959,

37(8), 911–917.25 Z. F. Li, L. Zhang, X. J. Shen, B. S. Lai and S. SQ,Microbiology,

2001, 28, 72–75.26 G. E. P. Box and W. G. Hunter, Statistics for experimenters: an

introduction to design, data analysis, and model building, ed.N. J. Hoboken, John Wiley & Sons, 2005, vol. 2.

27 G. W. Oehlert, Design and analysis of experiments: Responsesurface design, W.H. Freeman and Company, New York,2000.

22140 | RSC Adv., 2014, 4, 22129–22140

28 D. Angela and V. Daniel, Design and Analysis of Experiments,Springer, New York, 1999.

29 Design and Analysis of Experiments, ed. D. C. Montgomery,John Wiley & Sons, Inc., New York, 2002, vol. 5.

30 N. Bradley, The response surface methodology, Master ofScience, thesis, Indiana University South Bend, 2007, pp.1–72.

31 H. Lodish, A. Berk and S. L. Zipursky, et al. Molecular CellBiology, W. H. Freeman, New York, 4th edn, 2000.

32 Design and Analysis of Industrial Experiments, ImperialChemical Industries by Longman Group, ed. O. L. Davies,New York, 1978.

33 G. Najafpour, H. Younesi and K. Syahidah Ku Ismail,Bioresour. Technol., 2004, 92, 251–260.

34 R. Dutta. Fundamentals of Biochemical Engineering, Springer,Berlin Heidelberg, New York, 2008.

35 N. S. Panikov, Kinetics, Microbial Growth, Encyclopedia ofBioprocess Technology, 2002.

36 R. Leesing and S. Kookkhunthod, International Conference onFood Engineering and Biotechnology, IACSIT Press, Singapore,2011, vol. 9.

This journal is © The Royal Society of Chemistry 2014