Augmentation of ethanol production through statistically designed growth and fermentation medium using novel thermotolerant yeast isolates

Renewable Energy 109 (2017) 406e421

Contents lists available at ScienceDirect

Renewable Energy journal homepage: www.elsevier.com/locate/renene

Augmentation of ethanol production through statistically designed growth and fermentation medium using novel thermotolerant yeast isolates Richa Arora a, b, 2, Shuvashish Behera a, 2, Nilesh Kumar Sharma a, b, Sachin Kumar a, *, 1 a b

Biochemical Conversion Division, Sardar Swaran Singh National Institute of Bio-Energy, Kapurthala, 144601, India I.K Gujral Punjab Technical University, Kapurthala, 144601, India

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 June 2016 Received in revised form 14 March 2017 Accepted 18 March 2017 Available online 21 March 2017

Overproduction of metabolites, high product yield and process economics are greatly influenced by the media composition used for growth and fermentation. The main purpose of this study is to enhance the ethanol production through statistical tool of response surface methodology (RSM) by optimizing media components for the growth and fermentation of thermotolerant isolates Kluyveromyces marxianus NIREK1 and NIRE-K3. Five different salts were used in the Face-centered Central Composite Design (FCCD), with the responses of biomass formation and ethanol production for growth and fermentation, respectively. Yeast extract and K2HPO4 were found to be the key media components for the growth and fermentation which is revealed from their interaction in both the yeast isolates. Further studies on batch fermentation kinetics using the optimized values of the medium composition for K. marxianus NIRE-K1 and NIRE-K3 resulted in final ethanol concentration of 17.73 (86.27% of theoretical ethanol yield) and 19.01 g l
1 (94.12% of theoretical ethanol yield), respectively. An increase in the ethanol yield and productivity by 11.36, 10.42% and 2.0, 2.7% was revealed in NIRE-K1 and NIRE-K3, respectively, as compared to our previous study. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Thermotolerant yeast Kluyveromyces marxianus Ethanologen Optimization Face-centered central composite design

1. Introduction The drastic increase in the energy crisis, green house gas emissions and exhaustion of fossil fuel reserves have led to the development of renewable energy technologies [1e3]. However, one of the major challenges in the fuel production from renewable resources lies in the development of improved strains with efficient ethanol production [4,5]. Current research primarily focuses on the utilization of thermotolerant yeasts for efficient bioconversion of biomass to ethanol [6e9]. Thermotolerant ethanologenic fermentations are reported to be superior rather than the conventional mesophilic ones with higher bioconversion rates, continuous product recovery, economically viable processes due to lesser requirement of cooling and reduced risk of contamination [4,10e12]. * Corresponding author. Biochemical Conversion Division, Sardar Swaran Singh National Institute of Bio-Energy, Jalandhar-Kapurthala Road, Wadala Kalan, Kapurthala, 144601, Punjab, India. E-mail address: [email protected] (S. Kumar). 1 Present Address: Department of Chemical & Biological Engineering, South Dakota School of Mines and Technology, Rapid City, SD, 57701, USA. 2 Equal contribution. http://dx.doi.org/10.1016/j.renene.2017.03.059 0960-1481/© 2017 Elsevier Ltd. All rights reserved.

Apart from the use of thermotolerant ethanologens, the commercialization of a bioprocess and its economics depends upon the cost for the cultivation of the culture and its subsequent ethanol yield. The biochemical and nutritional requirements of the bioprocessing strains is highly influenced by carbon, nitrogen sources along with supplements like amino acids, vitamins, antibiotics, etc., which further aids in the cost [13]. Also, mineral salts are generally used in ethanol producing industries to supplement the fermentation media and provide acceptable yields [14]. Moreover, various medium components have strong interactions which may affect the competence of the process, both positively and negatively [15]. Thus, there is a need to develop a medium formulation for convenient, cost-effective and efficient bioprocess technology for bioethanol production. There are two methods for evaluating the optimal level, empirical method and statistical method. The former has several limitations because it involves substantial amount of time and labour taking OFAT (one-factor-at-a-time) approach into account. Moreover, it does not account for the interaction among the variables which strongly influences the bioprocess [16]. On the other hand, the latter involves the statistical tools like response surface

R. Arora et al. / Renewable Energy 109 (2017) 406e421

methodology (RSM) for designing experiments, constructing models suitably fitting with the data and evaluation of positive and negative interactions among various variables and finally providing the optimal solutions with reduced number of experimental runs [17,18]. The main goal of RSM is to chase the optimum values of the parameters in such a manner that the maximum response can be obtained. Optimization of different medium composition using RSM has been done by several researchers for alcoholic fermentations [14,19e23]. Several researchers have reported the optimization of growth and fermentation conditions of thermotolerant Kluyveromyces marxianus for the production of inulinase [24]; b-galactosidase [25]; ethyl acetate [26]; hydrogen [27]. However, to the best of our knowledge, the statistical optimization of medium composition for both growth and ethanol fermentation of thermotolerant K. marxianus has not been reported yet. Therefore, the present study was carried out to optimize medium components for novel thermotolerant ethanologenic yeast K. marxianus NIRE-K1 and NIRE-K3 for biomass and ethanol production using RSM, thereby, formulating a new medium for cost-effective bioethanol production with enhanced ethanol yield. In addition to this, the comparison of optimum medium components for growth and fermentation would aid in better understanding of the physiology of this yeast. 2. Materials and methods 2.1. Microorganism and culture conditions Two thermotolerant yeast K. marxianus NIRE-K1 (MTCC 5933) and NIRE-K3 (MTCC 5934) used in this study were isolated and reported in our earlier study [28]. The cultures were maintained on yeast extract-peptone (YEP) medium [(g l
1): yeast extract (Himedia, Mumbai, India), 10; peptone (Himedia, Mumbai, India), 20; glucose (Himedia, Mumbai, India), 20; phytagel (Sigma-Aldrich, USA), 15; pH, 5.5]. Both the cultures were stored in refrigerator at 4 ± 0.5  C for future use and the stock cultures were maintained in 30% glycerol at
80  C. 2.2. Inoculum preparation Inoculum was prepared in 500 mL cotton plugged erlenmeyer flask containing 100 ml salt medium (as mentioned above but without phytagel) having 1% glucose as the carbon source. A loopful pure culture of both the isolates K. marxianus NIRE-K1 and NIRE-K3 were inoculated separately in the flasks as mentioned in Arora et al. [28] through the incubation of 24 h at 45  C in shaking condition and were used for the inoculation in the subsequent runs for optimization. 2.3. Growth and fermentation medium Growth and fermentation medium for both the isolates were prepared separately in 100 mL cotton plugged and capped erlenmeyer flask containing 25 ml of salt medium (as mentioned above but without phytagel), respectively. The initial glucose concentrations in growth and fermentation optimization were 10 g l
1 and 40 g l
1, respectively. Optimization of growth media for both the isolates was carried out in aerobic condition through the inoculation of 25 ml inoculums in each flask. However, optimization of fermentation media for both the isolates was carried out in anaerobic conditions through the inoculation of 2 g l
1 harvested cells. All the optimization experiments were run for 16 h at pH 5.5, temperature 45  C and shaking at 150 rpm in an orbital shaker incubator (New Brunswick Innova 43/43R Shaker, Germany).

407

2.4. Optimization of growth and fermentation medium components using FCCD Medium components playing an indispensible role for the growth and fermentation of K. marxianus NIRE-K1 and NIRE-K3 were optimized according to (RSM) using Design Expert software version 8.0 (STAT-EASE Inc., Minneapolis, USA). FCCD (Facecentered Central Composite Design) was employed to study the combined effect of the components like yeast extract, di-potassium hydrogen phosphate (K2HPO4), sodium di-hydrogen phosphate (NaH2PO4), magnesium sulphate (MgSO4) and ammonium sulphate [(NH4)2SO4] on the biomass (g l
1) and ethanol concentration (g l
1) as the responses. All the variables were studied at three levels viz. low (
1), middle (0) and high (þ1), with alpha value of 1. The real and coded values of these variables have been presented in Table 1. The software displayed 50 experimental runs, with 8 runs at the middle points. Similar runs at the central values ensure the accuracy of the data along with reproducibility of the model. Further, to enhance the accuracy of the model, the experiments were performed in duplicate and the values of the responses were the means of two replications. The statistical significance of the model was estimated by analysis of variance (ANOVA) with p-value < 0.05 i.e. above 95% confidence level and insignificance of lack of fit test. The responses of the dependent variables were analyzed using the polynomial Equation (1) of second order. The variance for each variable was divided into linear, quadratic and interactive parts mentioned below:

Y ¼ b0 þ b1 x1 þ b2 x2 þ b3 x3 þ b4 x4 þ b5 x5 þ b11 x21 þ b22 x22 þ b33 x23 þ b44 x24 þ b55 x25 þ b12 x1 x2 þ b13 x1 x3 þ b14 x1 x4 þ b15 x1 x5 þ b23 x2 x3 þ b24 x2 x4 þ b25 x2 x5 þ b34 x3 x4 þ b35 x3 x5 þ b45 x4 x5 (1) where, Y is predicted response, x1, x2, x3, x4 and x5 are the coded levels of independent parameters, b0 is the offset term, b1, b2, b3, b4 and b5 are the linear effects, b11, b22, b33, b44 and b55 are the quadratic effects and b12, b13, b14, b15, b23, b24, b25, b34, b35, b45 are the interaction effects. The quality of the models developed were evaluated by three types of R-squared values i.e. coefficient of determination, adjusted R2 and predicted R2. The fitted polynomial equations were then expressed in the form of contour and three dimensional plots, to illustrate the relationship between the responses and any two variables to be optimized, keeping the other variables at central positions. The interaction of any two variables under the study can be examined from the prototype of the contour plots. Further, numerical optimization method was used for obtaining the optimal solutions. 2.5. Validation through growth and fermentation Confirmatory experiments under optimized conditions for both Table 1 Coded values for each variable of FCCD for biomass and ethanol production. Variables

Unit


1

0

þ1

Yeast extract di-potassium hydrogen phosphate Sodium di-hydrogen phosphate Magnesium sulphate Ammonium sulphate

g g g g g

l
1 l
1 l
1 l
1 l
1

1 0.1 0.1 0.1 0.1

3 1.05 1.05 0.55 1.05

5 2 2 1 2

408

R. Arora et al. / Renewable Energy 109 (2017) 406e421

the isolates were carried out in triplicate for both growth and fermentation to confirm the legitimacy of the model generated by the software. The results of biomass and ethanol concentration were compared with those of the software predicted values.

2.6. Batch fermentation Inoculum of K. marxianus NIRE-K1 and NIRE-K3 were prepared in the respective shake flasks with the optimized growth medium composition keeping 10 g l
1 glucose as the carbon source. The inoculum was pumped into a bioreactor of 5 L working volume (NBS BioFlo-CelliGen 115) with optimized fermentation medium composition keeping 40 g l
1 glucose concentration. The pH and temperature for both growth and fermentation were maintained at 5.5 and 45  C. The fermentation study in both the yeast cells were carried out in triplicate basis and sample distributions were described through the mean and standard deviation (SD).

2.7. Analytical procedures Samples were withdrawn at 4 h intervals, and centrifuged at 10,000 rpm for 15 min and stored at 4  C prior to analysis. Glucose and various metabolites (ethanol, glycerol and acetic acid) were analyzed using HPLC (Agilent Technologies, US) having Hi-Plex H column with oven temperature of 57  C and 1 mM H2SO4 as the mobile phase at 0.7 ml min
1 of flow rate. Dry cell weight (DCW) was estimated by collecting the cells of 1 ml of broth in pre-dry weighted Eppendorf tube by centrifuging at 10,000 rpm for 15 min using Eppendorf centrifuge 5430 R followed by washing the cell pellet twice with distilled water and drying at 80  C in a vacuum oven till a constant weight was obtained.

2.8. Fermentation kinetics Fermentation kinetics of different parameters was evaluated by using the formulae given in the following equations [29].

YX=S ¼

Mass of biomass ðyeast cellÞ formed Mass of substrate ðglucoseÞ consumed

(2)

YP=S ¼

Mass of product ðethanolÞ formed Mass of substrate ðglucoseÞ consumed

(3)

qs ¼

Mass of substrate ðglucoseÞ consumed m Mass of biomas ðyeast cellÞ formed

(4)

qp ¼

Mass of product ðethanolÞ formed m Mass of biomass ðyeast cellÞ formed

(5)

m ¼ standardized value for specific growth rate of microorganism Qs ¼ Substrate ðglucoseÞ uptake ðgÞ per liter of hydrolysate per hour QP ¼ Product formed ðgÞper liter of hydrolysate per hour

3. Results and discussion 3.1. Optimization of medium components for the isolate K. marxianus NIRE-K1 and NIRE-K3 For a cost-effective bioprocess, optimization of physical and chemical parameters plays an indispensible role. The physical parameters including temperature and pH have been already optimized for both the isolates in our previous study [28]. However, apart from physical parameters, salts have both stimulatory and inhibitory effects on the performance of the culture, so, potential growth and fermentation rate can be obtained by optimizing the medium composition [30,31]. Hence, medium components must be studied with respect to the interactions between them and their influence on biomass and ethanol production. In the present study, five components were taken into account for their study of effects on the growth and fermentation by both the yeasts. FCCD matrix for all the components with experimental and predicted values of biomass and ethanol concentrations of K. marxianus NIRE-K1 and NIRE-K3 has been shown in Tables 2 and 3, respectively. Further, ANOVA tables were used to determine the significance of the quadratic model for both the yeast through Fisher's ‘F’-test. In case of K. marxianus NIREK1, the quadratic models were found to be significant with F-value of 16.3 and 21.05 for the responses biomass and ethanol concentrations, respectively (p-value <0.05) (Tables 4 and 6). Similar significance of quadratic models were observed in case of K. marxianus NIRE-K3 with F-value of 17.70 and 23.25 for both the responses biomass and ethanol concentration, respectively (p-value <0.05) (Tables 5 and 7). Highly significant regression model is estimated from the higher Fvalues and lower (<0.05) p-values [32]. The importance of p-value also lies in understanding the trend of mutual interactions among different parameters considered in the study [28]. During the growth of both the yeast, linear effect of yeast extract and di-potassium hydrogen phosphate were found to be significant (p-value <0.05) (Tables 4 and 5). However, linear effect of magnesium sulphate also occured for the growth of K. marxianus NIRE-K3 (p-value <0.05) (Table 4). During the fermentation, linear effect of three salts viz. sodium di-hydrogen phosphate, magnesium sulphate and ammonium sulphate were found to be significant in both the yeast (p-value <0.05) (Tables 6 and 7). Also, the effect of magnesium sulphate was clearly visible from the interaction of magnesium sulphate with ammonium sulphate and yeast extract which was significant during fermentation by both the yeasts. Magnesium ions plays a significant role as the cofactor to the metabolic enzymes which further affect the biomass production and accumulation of the lipids [33,34]. Apart from this, magnesium also plays an important role in mitigating the ethanol toxicity [35]. Moreover, the interaction between yeast extract and dipotassium hydrogen phosphate was found to be significant in both the yeasts K. marxianus NIRE-K1 and NIRE-K3 during the optimization of biomass and ethanol formation. Similarly, the interaction of yeast extract with magnesium sulphate and ammonium sulphate was found to be significant for fermentation in both the yeasts. The biomass formation increased significantly with increase in the concentration of yeast extract as it is a rich source of amino acids, small peptides, vitamins and various growth factors which function as precursors of various biomolecules essential for yeast growth [13,36]. Similar results were obtained by Anvari et al. [37], where biomass production increased significantly with the addition of yeast extract. Apart from yeast extract, another major source of nitrogen, ammonium sulphate was used which profound a significant effect on the ethanol production. In addition to the maintenance of optimum carbon to nitrogen (C/N) ratio, ammonium salts also reduced the induction period, thereby, stimulating glucose fermentation [38].

R. Arora et al. / Renewable Energy 109 (2017) 406e421

409

Table 2 FCCD using RSM in K. marxianus NIRE-K1. Std

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

Factor 1 A: yeast extract (g/l)

1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

Factor 2 B: di-potassium hydrogen phosphate (g/l)

Factor 3 C: sodium di-hydrogen phosphate (g/l)

0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 1.05 1.05 0.1 2 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05

0.1 0.1 0.1 0.1 2 2 2 2 0.1 0.1 0.1 0.1 2 2 2 2 0.1 0.1 0.1 0.1 2 2 2 2 0.1 0.1 0.1 0.1 2 2 2 2 1.05 1.05 1.05 1.05 0.1 2 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05

Factor 4 D: magnesium sulphate (g/l)

0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 1 1 1 1 1 1 1 1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 1 1 1 1 1 1 1 1 0.55 0.55 0.55 0.55 0.55 0.55 0.1 1 0.55 0.55 0.55 0.55 0.55 0.55 0.55 0.55 0.55 0.55

The interaction between di-potassium hydrogen phosphate and sodium di-hydrogen phosphate was found to be significant during growth of K. marxianus NIRE-K3 with p-value of 0.0024. This is likely
to happen as they form the HPO2
4 and H2PO4 buffer which functions to maintain the pH and aids in phospholipid accumulation during growth [34]. Similar studies have been reported by Nikerel et al. [39] where (NH4)2HPO4 and KH2PO4 have been used for medium optimization, which form effective buffer system. The effects of various medium components for ethanol fermentation have been extensively reported by several researchers [15,20,21,33,40e44]. The significance of the model developed was further confirmed by the non-significance of ‘lack of fit’ with F-value of 0.68 and 1.27 for biomass and ethanol production, respectively, by K. marxianus NIRE-K1. Similarly, the lack of fit was not significant which shows Fvalue of 2.2 and 1.4 for both biomass and ethanol production,

Factor 5 E: ammonium sulphate (g/l)

0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 0.1 2 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05

Dry cell weight (g l
1)

Ethanol concentration (g l
1)

Actual value

Predicted value

Actual value

Predicted value

1.6 3.2 2.4 2.9 1.9 2.7 2.1 3.1 1.4 2.4 1.9 2.8 1.7 2.8 2.1 2.8 1.7 2.7 2 2.74 1.4 2.5 2.2 2.8 1.6 2.7 2.5 2.9 2 2.8 2.5 2.9 2.2 2.84 2.25 2.61 2.2 2 2.6 2.8 2.6 2.4 2.6 2.43 2.2 2.7 2.8 2.5 2.4 2.6

1.75 3.00 2.24 3.08 1.75 2.90 2.21 2.96 1.42 2.55 1.97 2.69 1.66 2.70 2.19 2.81 1.57 2.64 2.15 2.81 1.53 2.50 2.09 2.65 1.73 2.67 2.36 2.90 1.92 2.78 2.54 2.98 2.12 2.97 2.28 2.63 2.11 2.14 2.72 2.73 2.53 2.52 2.50 2.50 2.50 2.50 2.50 2.50 2.50 2.50

12.97 14.99 14.6 14.82 13.74 15.83 14.85 15.41 16.18 15.9 16.24 15.53 16.51 16.43 16.45 16.18 14.6 15.99 15.17 15.05 16.2 17.12 16.37 16.05 16.93 15.79 17.2 15.35 17.2 16.97 17.22 16.26 16.4 15.99 16.39 15.92 14.63 15.25 14.13 16.06 15.75 16.4 15.37 15.87 15.1 15.82 15.62 15.34 15.45 15.72

13.29 14.84 14.22 14.76 13.98 15.81 14.69 15.50 15.98 16.00 16.62 15.62 16.26 16.56 16.68 15.95 14.84 15.67 15.32 15.13 15.93 17.03 16.18 16.26 16.80 16.09 16.99 15.26 17.48 17.04 17.44 15.98 16.12 16.16 16.13 16.07 14.53 15.24 14.44 15.64 15.62 16.42 15.59 15.59 15.59 15.59 15.59 15.59 15.59 15.59

respectively, for K. marxianus NIRE-K3. On the basis of the quadratic models developed for both the responses, the equations for biomass (Equation (6)) and ethanol production (Equation (7)) by the isolate K.marxianus NIRE-K1 are as follows:

Dry cell weight ¼ 1:53 þ 0:26A þ 0:41B þ 0:87C
1:59D
0:15E
0:05A  B
0:01A  C
0:03A  D
0:02A  E
0:005B  C þ 0:03B  D þ 0:03B  E þ 0:14C  D
0:01C  E þ 0:28D  E þ 0:01A2
0:05B2
0:42C2 þ 1:10D2 þ 0:02E2 (6)

410

R. Arora et al. / Renewable Energy 109 (2017) 406e421

Table 3 FCCD using RSM in K. marxianus NIRE-K3. Std

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

Factor 1 A: yeast extract (g/l)

1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

Factor 2 B: di-potassium hydrogen phosphate (g/l)

Factor 3 C: sodium di-hydrogen phosphate (g/l)

Factor 4 D: magnesium sulphate (g/l)

0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 0.1 0.1 2 2 1.05 1.05 0.1 2 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05

0.1 0.1 0.1 0.1 2 2 2 2 0.1 0.1 0.1 0.1 2 2 2 2 0.1 0.1 0.1 0.1 2 2 2 2 0.1 0.1 0.1 0.1 2 2 2 2 1.05 1.05 1.05 1.05 0.1 2 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05

0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 1 1 1 1 1 1 1 1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 1 1 1 1 1 1 1 1 0.55 0.55 0.55 0.55 0.55 0.55 0.1 1 0.55 0.55 0.55 0.55 0.55 0.55 0.55 0.55 0.55 0.55

Factor 5 E: ammonium sulphate (g/l)

0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 0.1 2 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05

Dry cell weight (g l
1)

Ethanol concentration (g l
1)

Actual value

Predicted value

Actual value

Predicted value

1.3 2.6 1.8 2.81 2.1 2.6 2.2 2.4 1.9 2.6 2.3 2.8 2.1 2.92 2.3 2.71 1.5 2.66 2 2.6 1.9 2.4 2.2 2.5 1.8 2.6 2.2 2.9 1.9 2.6 2.1 2.2 2.2 2.7 1.9 2.3 2.3 2.6 2.6 2.7 2.6 2.7 2.5 2.4 2.4 2.42 2.5 2.3 2.6 2.4

1.39 2.48 1.89 2.65 1.98 2.67 2.19 2.55 1.77 2.75 2.22 2.88 2.20 2.79 2.36 2.61 1.52 2.54 2.04 2.73 1.91 2.53 2.13 2.42 1.74 2.66 2.22 2.80 1.97 2.48 2.15 2.33 2.11 2.74 1.99 2.16 2.39 2.45 2.55 2.70 2.66 2.59 2.47 2.47 2.47 2.47 2.47 2.47 2.47 2.47

14.77 16.6 16.59 16.7 14.8 15.5 15.72 15.26 17.08 17.43 17.8 16.88 16.37 17.28 16.3 16.44 16.63 17.69 17.71 16.6 16.6 17.25 16.75 16.64 16 15.67 15.86 14.39 15 15 14.71 14 14.63 15 17.49 17.33 17.19 17 15.9 16 16 16 16.12 16.3 16.56 16.83 16.36 16.39 16.46 16

15.10 16.26 16.45 16.39 14.50 15.85 15.52 15.66 17.06 17.62 17.71 17.06 16.30 17.06 16.62 16.17 16.87 17.44 17.59 16.94 16.47 17.23 16.86 16.41 15.71 15.69 15.73 14.50 15.15 15.33 14.84 13.81 14.83 14.89 17.50 17.41 17.46 16.82 16.16 15.84 16.19 15.90 16.33 16.33 16.33 16.33 16.33 16.33 16.33 16.33

as follows

Ethanol concentration ¼ 12:82
0:37A
0:52*B þ 1:99*C þ 6:50*D
0:04*E
0:13*A*B

Dry cell weight ¼ 0:95 þ 0:35*A þ 1:23*B þ 0:49*C
0:38*D

þ 0:04*A*C
0:43*A*D
0:10*A*E


0:27*E
0:04*A*B
0:05*A*C
0:03*A*D


0:06*B*C
0:17*B*D
0:13*B*E
0:24*C*D þ 0:11*C*E
0:43*D*E
0:14*A2 þ 0:56*B2
0:78*C 2


0:01*A*E
0:08*B*C
0:03*B*D þ 0:005*B*E
0:1*C*D
0:06*C*E
0:09*D*E
0:01*A2
0:44*B2
0:05*C 2


2:72*D2 þ 0:47*E2 (7) Similarly, the equations for biomass (Equation (8)) and ethanol production (Equation (9)) by the isolate K.marxianus NIRE-K3 are

þ 0:77*D2 þ 0:17*E2 (8)

R. Arora et al. / Renewable Energy 109 (2017) 406e421

411

Table 4 ANOVA for the experimental results of the FCCD for biomass formation by K. marxianus NIRE-K1. Source

Sum of Squares

df

Mean Square

F Value

p-value Prob > F

Model A-Yeast extract B-Di-potassium hydrogen phosphate C-Sodium di-hydrogen phosphate D-Magnesium sulphate E-Ammonium sulphate AB AC AD AE BC BD BE CD CE DE A^2 B^2 C^2 D^2 E^2 Residual Lack of Fit Pure Error Cor Total R-Squared Adj R-Squared Pred R-Squared Adeq Precision

8.735408 6.081894 1.023824 0.012812 0.000106 0.000106 0.332113 0.017113 0.027613 0.066613 0.000612 0.006612 0.017113 0.117613 0.003612 0.465613 0.004451 0.005599 0.352612 0.122361 0.001244 0.776874 0.528587 0.248288 9.512282 0.918 0.862 0.760 15.57

20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 29 22 7 49

0.436770379 6.081894118 1.023823529 0.012811765 0.000105882 0.000105882 0.3321125 0.0171125 0.0276125 0.0666125 0.0006125 0.0066125 0.0171125 0.1176125 0.0036125 0.4656125 0.004451188 0.005598736 0.352612382 0.122360782 0.001243512 0.026788773 0.024026678 0.035469643

16.30423211 227.0314523 38.21838037 0.478251264 0.00395249 0.00395249 12.39745082 0.638793713 1.03074895 2.486582687 0.022864055 0.246838477 0.638793713 4.390365266 0.134851266 17.38088169 0.166158694 0.208995593 13.16269234 4.567614236 0.046419135

<0.0001 <0.0001 <0.0001 0.4947 0.9503 0.9503 0.0014 0.4306 0.3184 0.1257 0.8809 0.6231 0.4306 0.0450 0.7161 0.0003 0.6865 0.6510 0.0011 0.0411 0.8309

significant

0.6773871

0.7731

not significant

*df- Degrees of freedom; F-Fisher's variance ratio; P-probability value; P < 0.05- significant at 5% level.

Table 5 ANOVA for the experimental results of the FCCD for biomass formation by K. marxianus NIRE-K3. Source

Sum of Squares

df

Mean Square

F Value

p-value Prob > F

Model A-Yeast extract B-Di-potassium hydrogen phosphate C-Sodium di-hydrogen phosphate D-Magnesium sulphate E-Ammonium sulphate AB AC AD AE BC BD BE CD CE DE A^2 B^2 C^2 D^2 E^2 Residual Lack of Fit Pure Error Cor Total R-Squared Adj R-Squared Pred R-Squared Adeq Precision

5.563465 3.430588 0.254224 0.033047 0.177988 0.048188 0.221113 0.32805 0.02205 0.010513 0.17405 0.00405 0.000612 0.052813 0.0882 0.0512 0.004714 0.383281 0.004714 0.060458 0.060458 0.455847 0.398247 0.0576 6.019312 0.924 0.872 0.740 18.359

20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 29 22 7 49

0.278173244 3.430588235 0.254223529 0.033047059 0.177988235 0.048188235 0.2211125 0.32805 0.02205 0.0105125 0.17405 0.00405 0.0006125 0.0528125 0.0882 0.0512 0.004713642 0.383280806 0.004713642 0.060457778 0.060457778 0.015718866 0.018102142 0.008228571

17.69677536 218.2465448 16.17314673 2.102381839 11.32322351 3.065630478 14.06669522 20.86982585 1.402772931 0.668782333 11.07268157 0.257652171 0.038965915 3.359816118 5.611091723 3.257232383 0.299871606 24.38348931 0.299871606 3.846192044 3.846192044

<0.0001 <0.0001 0.0004 0.1578 0.0022 0.0905 0.0008 <0.0001 0.2459 0.4201 0.0024 0.6156 0.8449 0.0771 0.0247 0.0815 0.5882 <0.0001 0.5882 0.0595 0.0595

significant

2.199913066

0.1436

not significant

*df- Degrees of freedom; F-Fisher's variance ratio; P-probability value; P < 0.05- significant at 5% level.

412

R. Arora et al. / Renewable Energy 109 (2017) 406e421

Table 6 ANOVA for the experimental results of the FCCD for ethanol formation by K. marxianus NIRE-K1. Source

Sum of Squares

df

Mean Square

F Value

p-value Prob > F

Model A-Yeast extract B-Di-potassium hydrogen phosphate C-Sodium di-hyphosphatedrogen D-Magnesium sulphate E-Ammonium sulphate AB AC AD AE BC BD BE CD CE DE A^2 B^2 C^2 D^2 E^2 Residual Lack of Fit Pure Error Cor Total R-Squared Adj R-Squared Pred R-Squared Adeq Precision

36.04897 0.020262 0.033674 4.306176 12.37236 5.352356 2.070613 0.148513 4.71245 1.073113 0.103513 0.17405 0.418612 0.3362 0.308113 1.0658 0.744229 0.639647 1.234402 0.752157 0.454228 2.483411 1.986824 0.496588 38.53238 0.936 0.891 0.791 22.142

20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 29 22 7 49

1.802448 0.020262 0.033674 4.306176 12.37236 5.352356 2.070613 0.148513 4.71245 1.073113 0.103513 0.17405 0.418612 0.3362 0.308113 1.0658 0.744229 0.639647 1.234402 0.752157 0.454228 0.085635 0.09031 0.070941

21.04807 0.236606 0.393222 50.28532 144.478 62.50207 24.17955 1.734253 55.02957 12.53126 1.208766 2.032467 4.888342 3.925971 3.59798 12.44587 8.690726 7.469475 14.41471 8.783299 5.304245

<0.0001 0.6303 0.5355 <0.0001 <0.0001 <0.0001 <0.0001 0.1982 <0.0001 0.0014 0.2806 0.1646 0.0351 0.0571 0.0678 0.0014 0.0063 0.0106 0.0007 0.0060 0.0286

significant

1.273031

0.3934

not significant

*df- Degrees of freedom; F-Fisher's variance ratio; P-probability value; P < 0.05- significant at 5% level.

Table 7 ANOVA for the experimental results of the FCCD for ethanol formation by K. marxianus NIRE-K3. Source

Sum of Squares

df

Mean Square

F Value

p-value Prob > F

Model A-Yeast extract B-Di-potassium hydrogen phosphate C-Sodium di-hydrogen phosphate D-Magnesium sulphate E-Ammonium sulphate AB AC AD AE BC BD BE CD CE DE A^2 B^2 C^2 D^2 E^2 Residual Lack of Fit Pure Error Cor Total R-Squared Adj R-Squared Pred R-Squared Adeq Precision

39.85259 0.030003 0.064424 3.539438 0.889706 0.741188 2.940313 0.08 0.690313 0.68445 0.2178 0.973013 0.7938 0.04805 0.078013 19.3442 5.330249 3.141374 1.630718 0.274295 0.198112 2.485679 2.025529 0.46015 42.33827 0.941 0.901 0.806 22.522

20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 29 22 7 49

1.99263 0.030003 0.064424 3.539438 0.889706 0.741188 2.940313 0.08 0.690313 0.68445 0.2178 0.973013 0.7938 0.04805 0.078013 19.3442 5.330249 3.141374 1.630718 0.274295 0.198112 0.085713 0.09207 0.065736

23.24767 0.350039 0.751618 41.29403 10.38005 8.647318 34.30413 0.933347 8.05376 7.985363 2.541036 11.35197 9.261131 0.560591 0.910159 225.6855 62.18711 36.64989 19.02531 3.200159 2.31134

<0.0001 0.5587 0.3931 <0.0001 0.0031 0.0064 <0.0001 0.3420 0.0082 0.0085 0.1218 0.0021 0.0049 0.4601 0.3480 <0.0001 <0.0001 <0.0001 0.0001 0.0841 0.1393

significant

1.400601

0.3389

not significant

*df- Degrees of freedom; F-Fisher's variance ratio; P-probability value; P < 0.05- significant at 5% level.

R. Arora et al. / Renewable Energy 109 (2017) 406e421

Ethanol concentration ¼ 12:77 þ 2:53*A
1:69*B
2:22*C þ 4:37*D þ 1:86*E
0:16*A*B þ 0:03*A*C
0:16*A*D
0:08*A*E
0:09*B*C
0:41*B*D
0:17*B*E
0:09*C*D þ 0:05*C*E
1:82*D*E
0:37*A2 þ 1:25*B2 þ 0:90*C 2
1:64*D2
0:31*E2 (9)

413

where, A, B, C, D and E are the notations for yeast extract, dipotassium hydrogen phosphate, sodium dihydrogen phosphate, magnesium sulphate and ammonium sulphate, respectively. The quality of the developed models was analyzed by the Rsquared value i.e. coefficient of determination (R2). It measures the variation in the responses due to variation in the experimental parameters and interactions among them. Better prediction of the response for the model is indicated by the R-squared values which are close to 1 [45]. However, according to Chauhan and Gupta [46], a model with R-squared value of more than 0.75 can be accepted. In case of K. marxianus NIRE-K1, the model presented in Tables 4 and 6 exhibits high R-squared values of 0.918 and 0.936,

Fig. 1. Diagnostic plot of the distribution of observed and predicted values of (A) Biomass formation (B) Ethanol concentration by K. marxianus NIRE-K1.

414

R. Arora et al. / Renewable Energy 109 (2017) 406e421

respectively which explains 91.8% and 93.6% of the variation in the response, as well as high value of the adjusted determination coefficient [adjusted R2 ¼ 0.862 (biomass concentration), adjusted R2 ¼ 0.891 (ethanol concentration)]. The high R-squared values show a strong correlation between the observed and the predicted values, suggesting accuracy of the results and high significance of the model. Similarly, the model using K. marxianus NIRE-K3 showed higher significance rather than NIRE-K1 having Rsquared values of 0.924 and 0.941, as well as high value of the adjusted determination coefficient [adjusted R2 ¼ 0.872 (biomass

concentration), adjusted R2 ¼ 0.901 (ethanol concentration)] (Tables 5 and 7). In a similar study, correlation between the observed and predicted values i.e. R2 was close to 0.9 which showed the accuracy of the model developed by RSM [28]. However, the comparison of the results of various models developed in literature is relatively difficult due to disparity in the process conditions. The actual and predicted values were very close to each other in all the models developed in this study and hence, low values of coefficient of variation i.e. 6.8 and 1.9% (K. marxianus NIRE-K1) and 5.33%

Fig. 2. Diagnostic plot of the distribution of observed and predicted values of (A) Biomass formation (B) Ethanol concentration by K. marxianus NIRE-K3.

R. Arora et al. / Renewable Energy 109 (2017) 406e421

and 1.80 (K. marxianus NIRE-K3) for biomass and ethanol production, respectively, were obtained. The low values obtained indicate the greater precision and reproducibility of the experiments performed [13]. Moreover, further accuracy of the model developed is indicated by the ratio of signal to noise when it is more than 4 [47]. An ample amount of signal was obtained for both the models developed in K. marxianus NIRE-K1 [with a ratio of 15.57 (growth) and 22.522 (fermentation)] and NIRE-K3 [with a ratio of 18.359 (growth) and 22.142 (fermentation)] which clearly specify that both the models could be used for the prediction of the responses. Figs. 1 and 2 show the diagnostic plots between the experimental and predicted values for biomass and ethanol formation by both K. marxianus NIRE-K1 and NIRE-K3 wherein all the points lie along the diagonal line, again indicating a good fit of both the models. Contour and 3-D plots for biomass formation and ethanol production in K. marxianus NIRE-K1 and NIRE-K3 have been shown in Figs. 3e6, respectively. In the contour plots, each of the contour curves indicates the combination between any two variables while maintaining the others at their zero coded values. The type of interaction between the two variables is illustrated by the shape of the contours produced. Significantly, strong interactions are indicated by the elliptical plots. On the other hand, the circular plots indicate the weaker interactions [28,48]. The surface confined in the smallest ellipse indicates the maximum predicted value. The 3D plots illustrated the effect of individual variables and their combined effect on the response produced keeping the other variables constant at their central values (Figs. 3e6) and hence, the optimal points could be easily comprehended for both maximum biomass formation and ethanol concentration. The final optimized values of all the salts were obtained by

415

numerical optimization in the software keeping the goal of obtaining the maximum response. In case of K. marxianus NIRE-K3, the optimized medium composition for biomass formation was (in g l
1) yeast extract, 4.81; K2HPO4, 1.10; NaH2PO4, 1.05; MgSO4, 0.95; (NH4)2SO4, 1.99 with predicted biomass concentration of 2.95 g l
1, whereas the optimized values for ethanol production obtained with the software were (in g l
1) yeast extract, 2.93; K2HPO4, 1.99; NaH2PO4, 0.24; MgSO4, 0.42; (NH4)2SO4, 1.34, with predicted ethanol concentration of 18.324 g l
1. Similarly, growth medium containing (g l
1) yeast extract, 5.0; K2HPO4, 1.35; NaH2PO4, 1.05; MgSO4, 1.0; (NH4)2SO4, 1.23 was found optimum with predicted biomass concentration of 3.2 g l
1, whereas for fermentation, the optimized medium compositions were (g l
1) yeast extract, 3.71; K2HPO4, 0.11; NaH2PO4, 1.51; MgSO4, 0.84; (NH4)2SO4, 2.0 with predicted ethanol concentration of 17.22 g l
1 using isolate K. marxianus NIRE-K1. 3.2. Validation of the model Final confirmatory experiments were carried out with optimized medium composition for both growth and fermentation to compare with the predicted values for K. marxianus NIRE-K1 (biomass and ethanol concentrations of 3.20 g l
1 and 17.224 g l
1, respectively) and NIRE-K3 (biomass and ethanol concentrations of 2.95 g l
1 and 18.324 g l
1, respectively). However, the experimental values obtained for biomass and ethanol concentration in K. marxianus NIRE-K1 were found to be 2.98 g l
1 and 17.73 g l
1 which were close to the predicted ones with a small variation of 6.88% and 2.85%, respectively. Similarly, in case of K. marxianus NIRE-K3, experimental values obtained for biomass

Fig. 3. Contour and corresponding 3-D plots for biomass production by K. marxianus NIRE-K1.

416

R. Arora et al. / Renewable Energy 109 (2017) 406e421

Fig. 4. Contour and corresponding 3-D plots for ethanol production by K. marxianus NIRE-K1.

and ethanol concentration were found to be 2.87 g l
1 and 19.01 g l
1 which were close to the predicted ones with a small variation of 2.71% and 3.61%, respectively. A variation of <10% indicates the legitimacy of the model [49]. Hence, both the models were confirmed to be reliable.

3.3. Batch fermentation Batch fermentations were carried out separately in bench-scale bioreactor using the optimized medium composition for both the isolates. In the present study, ethanol production started in the log

R. Arora et al. / Renewable Energy 109 (2017) 406e421

417

Fig. 5. Contour and corresponding 3-D plots for biomass production by K. marxianus NIRE-K3.

phage of the growth producing 4.07 ± 0.05 and 3.0 ± 0.10 g l
1 of ethanol using 10.48 ± 0.03 and 6.28 ± 0.05 g l
1 of sugar in 2 h of incubation time by K. marxianus NIRE-K1 and NIRE-K3. The decrease in sugar reserve might be due to its utilization in part, for initial growth and metabolism of the yeast in addition to its conversion into ethanol [50]. For 4, 6 and 8 h of fermentation, 67.08, 88.30 and 100.00% of sugar was utilized with simultaneous increase in ethanol concentration to 8.38 ± 0.04, 15 ± 0.09 and 17.73 ± 0.03 g l
1, respectively with the cells of K. marxianus NIREK1 (Fig. 7a). Similarly, for 4 and 6 h of fermentation, 14.80 ± 0.05 and 19.01 ± 0.03 g l
1 of ethanol was produced through the

utilization of 85.00 and 100.00% of sugar using the cells of K. marxianus NIRE-K3 (Fig. 7b). The results showed that there was 100% sugar utilization at the end of 8 h and 6 h of incubation period using the cells of K. marxianus NIRE-K1 and NIRE-K3, respectively. The growth and fermentation kinetics of K. marxianus NIRE-K1 and NIRE-K3 were also studied which has been depicted in Table 8. The ethanol concentration (P) obtained with the cells of K. marxianus NIRE-K3 (19.01 ± 0.03 g l
1) was 6.73% more than that of K. marxianus NIRE-K1 (17.73 ± 0.03 g l
1), whereas the volumetric substrate uptake (Qs) was found to be 25.04% more in case of K. marxianus NIRE-K3 (6.67 ± 0.04 g l
1h
1) than that of

418

R. Arora et al. / Renewable Energy 109 (2017) 406e421

Fig. 6. Contour and corresponding 3-D plots for ethanol production by K. marxianus NIRE-K3.

Fig. 7. Fermentation profile using optimized medium concentration in bench-scale bioreactor (A) K. marxianus NIRE-K1 (B) K. marxianus NIRE-K3.

R. Arora et al. / Renewable Energy 109 (2017) 406e421

419

Table 8 Growth and fermentation kinetics of K. marxianus NIRE-K1 and NIRE-K3 at optimized conditions. Kinetic Parameters

K. marxianus NIRE-K1

K. marxianus NIRE-K3

Initial sugar concentration (S, g l
1) Final ethanol (P, g l
1) Final biomass concentration (X, g l
1) Specific growth rate (m, h
1) Cell yield (Yx/s, g g
1) Ethanol yield (Yp/s, g g
1) Volumetric substrate uptake (Qs, g l
1 h
1) Volumetric product productivity (Qp, g l
1 h
1) Specific sugar consumption rate (qs, g g
1 h
1) Specific product formation rate (qp, g g
1 h
1) Conversion rate (%) into ethanol

40.00 ± 0.02 17.73 ± 0.03 3.3 ± 0.03 0.069 ± 0.003 0.035 ± 0.003 0.44 ± 0.01 5.00 ± 0.10 2.22 ± 0.02 1.97 ± 0.03 0.87 ± 0.05 86.9 ± 0.30

40.00 ± 0.02 19.01 ± 0.03 3.1 ± 0.10 0.121 ± 0.004 0.04 ± 0.01 0.48 ± 0.004 6.67 ± 0.04 3.17 ± 0.04 3.02 ± 0.07 1.44 ± 0.03 93.2 ± 0.26

Table 9 Comparison of bioethanol production in our study with other studies reported in the literature.

*

Microorganism

Fermentation temperature  ( C)/pH

Substrate(s)

By-products

Ethanol concentration (g l
1)

Ethanol Yield (% of theoretical yield)

References

K. marxianus NIRE-K1 K. marxianus NIRE-K3 B. borstelensis Caldicellulosiruptor sp. DIB 004C

45/5.5 45/5.5 45/5.73 72/6.75

Acetic acid, Glycerol Acetic acid, Glycerol Acetic acid n.d.

This study This study [54] [55]

45/5.0

17.73 19.01 1.20 0.12 0.07 n.d.

86.27 94.12 47.00 n.d.

Clostridium strain AK1

66.75

[56]

Thermoanaerobacter strain J1

65/7.0

Glucose Glucose Avicel Avicel Washed Poplar Cellulose hydrolysates Cellulose

0.35

n.d.

[57]

T. aciditolerans

65/5.0

1.11

n.d.

[58]

Cellulose hydrolysates

Acetic acid, Hydrogen Acetic acid, Hydrogen, Methane Hydrogen, Acetic acid

n.d.: not defined.

K. marxianus NIRE-K1 (5.00 ± 0.10 g l
1h
1). The ethanol yield (Yp/
1 s ¼ 0.48 ± 0.004 g g ) and volumetric product productivity
1 (Qp ¼ 3.17 ± 0.04 g g ) obtained with K. marxianus NIRE-K3 was found to be 6.76 and 29.97%, respectively higher than that of Yp/s (0.44 ± 0.01 g g
1) and Qp (2.22 ± 0.02 g g
1) of K. marxianus NIREK1 cells. Likewise, the final sugar to ethanol conversion rate (%) with K. marxianus NIRE-K3 cells was 6.76% more than that of K. marxianus NIRE-K1. However, the final biomass concentration (X) of K. marxianus NIRE-K1 cells (3.3 ± 0.03 g l
1) was 6.06% higher than that of K. marxianus NIRE-K3 cells (3.1 ± 0.1 g l
1), which might be due to more tolerance to temperature and ethanol. However, the specific sugar consumption rate (qs) was 1.97 ± 0.03 g g
1 h
1 and 3.02 ± 0.07 g g
1 h
1 for K. marxianus NIRE-K1 and K. marxianus NIRE-K3, respectively, which is useful during product separation and purification process [28]. Compared to un-optimized medium composition, nearly 11.36% and 10.42%, increase in the ethanol yield was observed under statistically optimized conditions for K. marxianus NIRE-K1 and K. marxianus NIRE-K3, respectively. Similarly, 2.00% and 2.70%, an increase in ethanol productivity was observed under statistically optimized conditions as compared to unoptimised conditions, for K. marxianus NIRE-K1 and K. marxianus NIRE-K3, respectively [28]. Table 9 shows the comparison of our study with other studies reported in the literature. Many researchers have reported batch ethanol fermentation using various yeast isolates. Krishnan et al. [51] carried out ethanol fermentation by using Saccharomyces 1400 (pLNH33) with an ethanol yield of 0.46 g g
1. The ethanol yield obtained by both the isolates K. marxianus NIRE-K1 and NIRE-K3 in the present study is analogous to Saccharomyces 1400 (pLNH33). Likewise, Tanimura et al. [52] reported 71.6% of ethanol yield using 2% glucose or xylose by an isolate ATY839, which showed 99.5% homology with Candida shehatae. In another study, batch fermentation kinetics was

reported by Kumar et al. [53] using a thermotolerant yeast isolate Kluyveromyces sp. IIPE453 (MTCC 5314) at 50  C and pH 5.0. Various kinetic parameters viz. ethanol yield, specific ethanol productivity, sugar consumption rate and volumetric ethanol productivity were found to be 0.43 g g
1, 0.3 g g
1 h
1, 1.74 g l
1 h
1 and 0.74 g l
1 h
1, respectively, which are comparable to both the isolates in the present study. Similarly, Behera et al. [50] carried out ethanol production from mahula flowers by using immobilized cells of Saccharomyces cerevisiae where ethanol yield, volumetric substrate uptake, volumetric product productivity and conversion rate to ethanol were found to be 0.455 g g
1, 0.850 g l
1 h
1, 0.387 g l
1 h
1 and 91.1%, respectively. Therefore, the exploitation of both the isolates K. marxianus NIRE-K1 and NIRE-K3 may have an immense effect on the economics of the overall process due to their thermotolerant nature which aids in easier product recovery and reduction in the cooling costs. Besides, the utilization of cheaper media further aids in the development of cost-effective and competent bioprocess technology for bioethanol production. 4. Conclusions Both the thermotolerant isolates K. marxianus NIRE-K1 and NIRE-K3 possess a great potential for bioethanol production. Our experimental results indicate that the medium components exert significant effects on growth and bioethanol production yields. However, further studies on the physiology of the isolates using lignocellulosic hydrolysate, effect of fermentation inhibitors and its metabolic flux analysis is required to exploit the potential of these isolates at commercial scale. Acknowledgements We thank Prof. Y. K. Yadav, Director General SSS-NIBE,

420

R. Arora et al. / Renewable Energy 109 (2017) 406e421

Kapurthala, India for his valuable suggestions and encouragement to carry out this research work. We gratefully acknowledge the Ministry of New and Renewable Energy, Govt. of India for providing funds to carry out the research activities. One of the authors (Richa Arora) also acknowledges Punjab Technical University, Kapurthala, India for her Ph.D. registration (Reg. No. 1315010). References [1] R. Arora, S. Behera, S. Kumar, Bioprospecting thermostable cellulosomes for efficient biofuel production from lignocellulosic biomass, Bioresour. Bioprocess 2 (2015), 12 pages. [2] R. Arora, S. Behera, N.K. Sharma, R. Singh, Y.K. Yadav, S. Kumar, Biochemical conversion of rice straw (Oryza sativa L.) to bioethanol using thermotolerant isolate K. marxianus NIRE-K3, in: N.R. Sharma, R.C. Thakur, M. Sharma, L. Parihar, G. Kumar (Eds.), Proceedings of Exploring & Basic Sciences for Next Generation Frontiers, Elsevier, New Delhi, 2014, pp. 143e146. [3] S. Behera, R. Arora, N. Nandhagopal, S. Kumar, Importance of chemical pretreatment for bioconversion of lignocellulosic biomass, Renew. Sust. Energy Rev. 36 (2014) 91e106. [4] S. Kumar, S.P. Singh, I.M. Mishra, D.K. Adhikari, Recent advances in production of bioethanol from lignocellulosic biomass, Chem. Eng. Technol. 32 (2009) 517e526. [5] S. Behera, N.K. Sharma, R. Arora, S. Kumar, Effect of evolutionary adaption on xylosidase activity in thermotolerant yeast isolates Kluyveromyces marxianus NIRE-K1 and NIRE-K3, Appl. Biochem. Biotechnol. 179 (2016) 1143e1154. [6] S. Kumar, S.P. Singh, I.M. Mishra, D.K. Adhikari, Ethanol and xylitol production from glucose and xylose at high temperature by Kluyveromyces sp. IIPE453, J. Ind. Microbiol. Biotechnol. 36 (2009) 1483e1489. [7] R. Arora, S. Behera, S. Kumar, Comparative study of fermentation efficiency for bio-ethanol production by isolates, in: S. Kumar, A.K. Sarma, S.K. Tyagi, Y.K. Yadav (Eds.), Recent Advances in Bio-energy Research, SSS-NIRE, Kapurthala, India, 2014, pp. 149e155. [8] S. Behera, R. Arora, N.K. Sharma, S. Kumar, Fermentation of glucose and xylose sugar for the production of ethanol and xylitol by the newly isolated NIREGX1 yeast, in: S. Kumar, A.K. Sarma, S.K. Tyagi, Y.K. Yadav (Eds.), Recent Advances in Bio-energy Research, SSS-NIRE, Kapurthala, India, 2014, pp. 175e182. [9] N.K. Sharma, S. Behera, R. Arora, S. Kumar, Potential role of xylose transporters in industrial yeast for bioethanol production: a perspective review, in: S. Kumar, S. Khanal, Y. Yadav (Eds.), Proceedings of the First International Conference on Recent Advances in Bioenergy Research, Springer Proceedings in Energy, Springer, New Delhi, 2016, pp. 81e93. [10] R. Arora, S. Behera, S. Kumar, Bioprospecting thermophilic/thermotolerant microbes for production of lignocellulosic ethanol: a future perspective, Renew. Sust. Energy Rev. 51 (2015) 699e717. [11] R.C. Kuhad, R. Gupta, Y.P. Khasa, A. Singh, Y.H. Percival Zhang, Bioethanol production from pentose sugars: current status and future prospects, Renew. Sust. Energy Rev. 15 (2011) 4950e4962. [12] H.A. Ruiz, D.P. Silva, D.S. Ruzene, L.F. Lima, A.A. Vicente, J.A. Teixeira, Bioethanol production from hydrothermal pretreated wheat straw by a flocculating Saccharomyces cerevisiae strain - effect of process conditions, Fuel 95 (2012) 528e536. [13] P.N. Sampaio, C.R.C. Calado, L. Sousa, D.C. Bressler, M.S. Pais, L.P. Fonseca, Optimization of the culture medium composition using response surface methodology for new recombinant cyprosin B production in bioreactor for cheese production, Eur. Food Res. Technol. 231 (2010) 339e346. [14] R.N. Comelli, L.G. Seluy, M.A. Isla, Optimization of a low-cost defined medium for alcoholic fermentation-a case study for potential application in bioethanol production from industrial wastewaters, New Biotechnol. 33 (2015) 107e115. [15] T. Chookaew, S.O.- Thong, P. Prasertsan, Statistical optimization of medium components affecting simultaneous fermentative hydrogen and ethanol production from crude glycerol by thermotolerant Klebsiella sp. TR17, Int. J. Hydrogen Energy 39 (2014) 751e760. [16] A. Kumari, P. Mahapatra, R. Banerjee, Statistical optimization of culture conditions by response surface methodology for synthesis of lipase with Enterobacter aerogenes, Braz. Arch. Biol. Technol. 52 (2009) 1349e1356. [17] O.N. Uncu, D. Cekmecelioglu, Cost-effective approach to ethanol production and optimization by response surface methodology, Waste Manage 31 (2011) 636e643. [18] J. Antoy, Design of Experiments for Engineers and Scientists, second ed., Elsevier Science & Technology Books London, 2014. [19] M.A. Palukurty, N.K. Telgana, H.S. Bora, S.N. Mulampaka, Screening and optimization of metal ions to enhance ethanol production using statistical experimental designs, Afr. J. Microbiol. Res. 2 (2008) 87e94. [20] J. Yu, X. Zhang, T. Tan, Optimization of media conditions for the production of ethanol from sweet sorghum juice by immobilized Saccharomyces cerevisiae, Biomass Bioenergy 33 (2009) 521e526. [21] F.B. Pereira, P.M.R. Guimaraes, J.A. Teixeira, L. Domingues, Optimization of low-cost medium for very high gravity ethanol fermentations by Saccharomyces cerevisiae using experimental designs, Bioresour. Technol. 101 (2000) 7856e7863.

[22] N.A.A. Adnan, S.N. Suhaimi, S. Abd-Aziz, M.A. Hassan, L.-Y. Phang, Optimization of bioethanol production from glycerol by Escherichia coli SS1, Renew. Energy 66 (2014) 625e633. [23] E. Betiku, A.E. Taiwo, Modeling and optimization of bioethanol production from breadfruit starch hydrolyzate vis-_a-vis response surface methodology and artificial neural network, Renew. Energy 74 (2015) 87e94. [24] R.S. Singh, B.S. Sooch, M. Puri, Optimization of medium and process parameters for the production of inulinase from a newly isolated Kluyveromyces marxianus YS-1, Bioresour. Technol. 98 (2007) 2518e2525. [25] A.P. Manera, J. da Costa Ores, V.A. Ribeiro, C.A.V. Burkert, S.J. Kalil, Optimization of the culture medium for the production of b-galactosidase from Kluyveromyces marxianus CCT 7082, Food Technol. Biotechnol. 46 (2008) 66e72. [26] C. Loser, T. Urit, E. Gruber, T. Bley, Efficient growth of Kluyveromyces marxianus biomass used as a biocatalyst in the sustainable production of ethyl acetate, Energy Sustain. Soc. 5 (2015), 15 pages. [27] R. Mangayil, T. Aho, M. Karp, V. Santala, Improved bioconversion of crude glycerol to hydrogen by statistical optimization of media components, Renew. Energy 75 (2015) 583e589. [28] R. Arora, S. Behera, N.K. Sharma, S. Kumar, A new search for thermotolerant yeasts, its characterization and optimization using response surface methodology for ethanol production, Front. Microbiol, Microbiotechnol. Ecotoxicol. Bioremediation 6 (2015) 889. [29] J.E. Bailey, D.F. Ollis, Biochemical Engineering Fundamentals, McGraw-Hill, New York, 1986. [30] N.K. Sharma, S. Behera, R. Arora, S. Kumar, Enhancement in xylose utilization using Kluyveromyces marxianus NIRE-K1 through evolutionary adaptation approach, Bioprocess Biosyst. Eng. 39 (2016) 835e843. [31] M.P. Anupama, D.G. Mahesh, C. Ayyanna, Optimization of fermentation medium for the production of ethanol from jaggery using box-behnken design, Int. J. Appl. Biol. Pharm. Technol. 1 (2010) 34e45. [32] E.R. Rene, M.S. Jo, S.H. Kim, H.S. Park, Statistical analysis of main and interaction effects during the removal of BTEX mixtures in batch conditions, using wastewater treatment plant sludge microbes, Int. J. Environ. Sci. Tech. 4 (2007) 177e182. [33] D. Dasgupta, S.K. Suman, D. Pandey, D. Ghosh, R. Khan, D. Agrawal, R.K. Jain, V.T. Vadde, D. Adhikari, Design and optimization of ethanol production from bagasse pith hydrolysate by a thermotolerant yeast Kluyveromyces sp. IIPE453 using response surface methodology, Springerplus 2 (2013) 159. [34] S. Bandhu, D. Dasgupta, J. Akhter, P. Kanaujia, S.K. Suman, D. Agrawal, S. Kaul, D.K. Adhikari, D. Ghosh, Statistical design and optimization of single cell oil production from sugarcane bagasse hydrolysate by an oleaginous yeast Rhodotorula sp. IIP-33 using response surface methodology, Springerplus 3 (2014) 691. [35] K.M. Dombek, L.O. Ingram, Magnesium limitation and its role in apparent toxicity of ethanol during yeast fermentation, Appl. Environ. Microbiol. 52 (1986) 975e981. [36] P. Pradeep, O.V.S. Reddy, P.R. Mohan, S. Ko, Process optimization for ethanol production from very high gravity (VHG) finger millet medium using response surface methodology, Iran. J. Biotechnol. 10 (2012) 168e174. [37] M. Anvari, G. Khayati, S. Rostami, Optimisation of medium composition for probiotic biomass production using response surface methodology, J. Dairy Res. 81 (2014) 59e64. [38] J.A. Muntz, The role of potassium and ammonium ions in alcoholic fermentation, J. Biol. Chem. 17 (1947) 653e665. [39] E. Nikerel, E. Toksoy, B. Kirdar, R. Yildirim, Optimizing medium composition for TaqI endonuclease production by recombinant Escherichia coli cells using response surface methodology, Process Biochem. 40 (2005) 1633e1639. [40] R. Balusu, R.R. Paduru, S.K. Kuravi, G. Seenayya, G. Reddy, Optimization of critical medium components using response surface methodology for ethanol production from cellulosic biomass by Clostridium thermocellum SS19, Process Biochem. 40 (2005) 3025e3030. [41] Y. Guo, J. Xu, Y. Zhang, H. Xu, Z. Yuan, D. Li, Medium optimization for ethanol production with Clostridium autoethanogenum with carbon monoxide as sole carbon source, Bioresour. Technol. 101 (2010) 8784e8789. [42] H.U. Chunkeng, Q.I.N. Qing, G.A.O. Peipei, Medium optimization for improved ethanol production in very high gravity fermentation, Chin. J. Chem. Eng. 19 (2011) 1017e1022. [43] O. Cofre, M. Ramirez, J.M. Gomez, D. Cantero, Optimization of culture media for ethanol production from glycerol by Escherichia coli, Biomass Bioenergy 37 (2012) 275e281. [44] J.P.A. Silva, S.I. Mussatto, I.C. Roberto, J.A. Teixeira, Fermentation medium and oxygen transfer conditions that maximize the xylose conversion to ethanol by Pichia stipitis, Renew. Energy 37 (2012) 259e265. [45] K. Ohtani, Bootstrapping R2 and adjusted R2 in regression analysis, Econ. Model 17 (2000) 473e483. [46] B. Chauhan, R. Gupta, Application of statistical experimental design for optimization of alkaline protease production from Bacillus sp. RGR-14, Process Biochem. 39 (2004) 2115e2122. [47] D.C. Montgomery, Design and Analysis of Experiments, John Wiley & Sons, New York, 2001. [48] O. Prakash, M. Talat, S.H. Hasan, R.K. Pandey, Factorial design for the optimization of enzymatic detection of cadmium in aqueous solution using immobilized urease from vegetable waste, Bioresour. Technol. 99 (2008) 7565e7572.

R. Arora et al. / Renewable Energy 109 (2017) 406e421 [49] L. Levin, C. Herrmann, V.L. Papinutti, Optimization of lignocellulolytic enzyme production by white rot fungus Trametes trogii in solid state fermentation using response surface methodology, Biochem. Eng. J. 39 (2008) 207e214. [50] S. Behera, R.C. Mohanty, R.C. Ray, Ethanol production from mahula (Madhuca latifolia L.) flowers with immobilized cells of Saccharomyces cerevisiae in Luffa cylindrica L. sponge discs, Appl. Energy 88 (2011) 212e215. [51] M.S. Krishnan, N.W. Ho, G.T. Tsao, Fermentation kinetics of ethanol production from glucose and xylose by recombinant Saccharomyces 1400 (pLNH33), Appl. Biochem. Biotechnol. 77e79 (1999) 373e388. [52] A. Tanimura, T. Nakamura, I. Watanabe, J. Ogawa, S. Jun, Isolation of a novel strain of Candida shehatae for ethanol production at elevated temperature, Springerplus 1 (2012) 27. [53] S. Kumar, P. Dheeran, S.P. Singh, I.M. Mishra, D.K. Adhikari, Bioprocessing of bagasse hydrolysate for ethanol and xylitol production using thermotolerant yeast, Bioprocess. Biosyst. Eng. 38 (2015) 39e47.

421

[54] A.H. Li, C.W. Lin, D.T. Tran, Optimizing the response surface for producing ethanol from avicel by Brevibacillus strain AHPC8120, J. Taiwan Inst. Chem. Eng. 42 (2011) 787e792. [55] V.A. Svetlitchny, O. Kensch, D.A. Falkenhan, S.G. Korseska, N. Lippert, M. Prinz, J. Sassi, A. Schickor, S. Curvers, Single-step ethanol production from lignocelluloses using novel extremely thermophilic bacteria, Biotechnol. Biofuels 6 (2013) 31. [56] J. Orlygsson, Ethanol production from biomass by a moderate thermophile, Clostridium AK1, Icel. Agric. Sci. 25 (2012) 25e35. [57] J.E. Jessen, J. Orlygsson, Production of ethanol from sugars and lignocellulosic biomass by Thermoanaerobacter J1 isolated from a hot spring in Iceland, J. Biomed. Biotechnol. 2012 (2012) 1e7. [58] M.A. Sigurbjornsdottir, J. Orlygsson, Combined hydrogen and ethanol production from sugars and lignocellulosic biomass by Thermoanaerobacterium AK54, isolated from hot spring, Appl. Energy 97 (2012) 785e791.