Mixed fermentation of Aspergillus niger and Candida shehatae to produce bioethanol with ionic-liquid-pretreated bagasse

Abstract

In this study, bagasse was pretreated with ionic liquid (IL) 1-butyl-3-methylimidazolium chloride ([Bmim]Cl) and 1% NaOH solution for initial activation of bagasse. A mixed fermentation of treated bagasse by Aspergillus niger and Candida shehatae showed the optimal conditions with the addition of C. shehatae 12 h later at a 1:1 proportion to A. niger. To further improve the ethanol production and obtain optimal fermentation conditions, a Plackett–Burman design was applied to screen the significant formulation and process variables. The optimal ethanol fermentation conditions with IL pretreated bagasse were determined using response surface methodology by Box–Behnken design. Three variables “initial pH, (NH₄)₂SO₄, fermentation time” were regarded as significant factors in the optimization study. The resulting optimum fermentation conditions for bioethanol was identified as: initial pH of 5.89, (NH₄)₂SO₄ concentration of 0.40 g/50 mL, and fermentation time of 3.60 days. The verification experimental ethanol concentration was 8.14 g/L, which agreed with the predicted value. An enhancement of approximately 153.58% compared with initial fermentation conditions in ethanol production was found using optimized conditions. It demonstrated that optimization methodology had a positive effect on the improvement of ethanol production. Under the optimal fermentation medium and conditions, the ethanol production with IL-pretreated bagasse and untreated bagasse was 8.14 g/L and 5.03 g/L, respectively, which exhibited 62% increase, compared to initial conditions with production of 3.21 g/L and 2.67 g/L, respectively, which displayed 20% increase. Both under optimal and original fermentation conditions, compared to the fermentation medium with untreated bagasse, all the results indicated that IL-pretreated bagasse resulted in higher ethanol production than untreated bagasse, demonstrating that IL-pretreated bagasse successfully increased the ethanol production in the mixed fermentation by A. niger and C. shehatae.

Electronic supplementary material

The online version of this article (10.1007/s13205-019-1570-6) contains supplementary material, which is available to authorized users.

Introduction

Currently the excessive consumption of fossil energy sources such as natural gas, coal, and petroleum, particularly in large industrial areas (Sarkar et al. 2012), has already resulted in a series of problems during thepast few decades such as air pollution (Rao et al. 2017), greenhouse gas emission (Trinh et al. 2013; Hertwich et al. 2015), climate change (Riahi et al. 2017), and global warming (Kim et al. 2014). Depletion of the restricted fossil fuel resources supply, as well as the expansion of the human population and continuously increasing fuel demands, leads to increasing concerns about environmental pollution and climate change (Niemistö et al. 2013; Sarkar et al. 2012), as well as political concerns with tightening of legislation. Hence, the use and production of biomass-based renewable fuels are promoted (Niemistö et al. 2013). Production of lignocellulosic biofuels is generally considered a potential possible route to aid with resolving these problems and in particularly bioethanol is among the most intensively investigated biofuels (Kim et al. 2014; Oleskowicz-Popiel et al. 2014).

Lignocellulosic biomass is abundant, and most importantly, renewable natural source (Wang et al. 2017): CO₂ is consumed by green plants during photosynthesis, while the gas is returned to the atmosphere as the biofuel is combusted (Oleskowicz-Popiel et al. 2014). Therefore, there is increasing interest in the use of lignocellulosic bioresources, such as sugarcane bagasse, in different processes as the production of bioethanol, bio-hydrogen, and bio-diesel (Santos et al. 2012). Bagasse and sugarcane have also been identified as potential lignocellulosic materials for lignocellulosic ethanol production, which have attracted interest from scientists worldwide recently (da Silva et al. 2010).

The basic structure of lignocellulosic biomass has been modeled (da Silva et al. 2010). Bagasse consists of a network of cellulose and hemicellulose bound by lignin in an intricate structure, which is recalcitrant to decomposition (Suhardi et al. 2013). Cellulose is considered the contributor to the crystalline part, whereas hemicellulose and lignin are the amorphous parts (Kelley et al. 1987). The crystalline regions contain highly ordered cellulose molecules derived from the organization of cellulose chains. These are linked by hydroxyl groups to form intra-and inter-molecular hydrogen bonds in different arrangements, unlike in the amorphous regions where the molecules are less ordered (Teixeira et al. 2013). The crystallinity of cellulose is considered as having a significant influence on the enzymatic hydrolysis of cellulosic biomass (da Silva et al. 2010) as the crystalline regions are more recalcitrant to enzymatic attack, while the amorphous regions are readily hydrolyzed (Cao and Tan 2005). Therefore, pretreatment is a crucial process in converting lignocellulosic material into biofuels (Mosier et al. 2005). Structural changes in lignocellulosic biomass vary depending on pretreatment methods and hydrolysis conditions (Saha et al. 2005; Xu et al. 2011). An effective and suitable pretreatment procedure involves (1) breaking down the cross-linked matrix of lignin and hemicellulose, (2) disrupting hydrogen bonds in crystalline cellulose, and finally, (3) raising the porosity and surface area of cellulose for subsequent enzymatic hydrolysis (Li et al. 2010). There are several pretreatment methods including chemical pretreatment (alkali, acid, ozonolysis, organosolv and ionic liquid) in particular, alkaline and the weak acid solution can effectively remove lignin and reduce cellulose crystallinity (Dawson and Boopathy 2007, 2008). Additionally, physical pretreatment (microwave, extrusion, grinding and milling) and physicochemical pretreatment (liquid hot water, CO₂ explosion, steam explosion, wet oxidation) (Yan et al. 2013, 2016; Yan and Liu 2015) are used as the pretreatment.

Recently, ionic liquids (ILs)-based pretreatment of lignocellulosic biomass has gained much attention because of their potential application as “green solvents” (Earle and Seddon 2000). ILs are a new class of purely ionic, salt-like materials that remain in the liquid state at an unusually low temperature, which have high thermal stability and nearly complete non-volatility (Yoo et al. 2017). Additionally, ILs can be easily recycled. ILs have been reported to be capable of dissolving cellulose and lignocellulose materials such as woody biomass, rice straw, wheat straw (Nguyen et al. 2010; Sun et al. 2009; Zhu et al. 2006; Mäki-Arvela et al. 2010), and significant changes in biomass structure after ILs pretreatment have been described (Tan et al. 2011). As reported in the literature (Yoo et al. 2017), ILs pretreatments significantly altered the crystallinity structure and reduced the crystallinity of lignocellulosic biomass (Gregorio et al. 2016; Moyer et al. 2018). The ILs pretreated lignocellulosic materials became amorphous and porous and more susceptible to degradation by cellulases and hemicellulases (Moniruzzaman and Goto 2018), thereby increasing ethanol production from fermentation (Dadi et al. 2006; Chapple et al. 2007). However, the dissolved cellulose and hemicellulose in the ionic liquid can be easily regenerated with the addition of an anti-solvent, such as water and acetone (Zhu et al. 2006). The regenerated cellulose and hemicellulose exhibited significantly reduced crystallinity and increased porosity, which enhanced the digestibility of the material and thus subsequently resulted in a higher yield for the ethanol fermentation (Li et al. 2009; Kuo and Lee 2009; Tan et al. 2011).

Aspergillus niger is one of the saprobic fungus (van den Brink and de Vries 2011) and produces many extracellular enzymes to degrade the plant biomass efficiently into hexose and pentose sugars in the cellulosic hydrolysates (Martens-Uzunova and Schaap 2009; Liu et al. 2018; Fiedler et al. 2018; Devi and Kumar 2017). Candida shehatae is one of the genera that has an ability to use glucose or xylose as substrate for ethanol production (Yuvadetkun et al. 2017, 2018). Currently, most of the biomass is utilized through enzyme hydrolysis, and then bioethanol is produced by Saccharomyces cerevisiae (Kuhad et al. 2010). So far, bioethanol fermentation by A. niger and C. shehatae with IL-pretreated bagasse has not been reported. When the bagasse had been pretreated with ionic liquids, then the strains of A. niger and C. shehatae were inoculated to the fermentation medium using the IL-pretreated bagasse as the main carbon source for production of bioethanol.

In this work, production of ethanol via biological mixed fermentation of the IL-pretreated bagasse by strains A. niger and C. shehatae was performed. In general, the steps were: (1) 1% NaOH solution used for initial activation of bagasse, followed by pretreatment by ionic liquid ([Bmim]Cl); (2) The secretion of cellulase and hemicellulase by the strain A. niger, which resulted in enzymatic lysis of the hemicellulose and cellulose polysaccharides into monomers such as glucose and xylose, and finally (3) the fermentation of the xylose into bioethanol by the strain C. shehatae.

This study was undertaken in order to evaluate: (a) determination of the positive effect of the IL pretreated bagasse on improving the bioethanol production compared to the untreated bagasse; (b) investigation of the feasibility of ethanol production by A. niger and C. shehatae. A combination of single factor design, Plackett–Burman design and response surface methodology was used to optimize fermentation medium and conditions to maximize ethanol production for future industrial use.

Materials and methods

Strains, materials, and culture media

The strain of A. niger that produces cellulase and hemicellulose was screened by our laboratory, and the ethanol-producing strain of C. shehatae was obtained from China Center of Industrial Culture Collection. NaOH reagent was purchased from Sinopharm Chemical Reagent Co., Ltd. Ionic liquid ([Bmim]Cl) was purchased from Tianjin Jin Zheng Da Energy saving and environmental protection Technology Co. Ltd. The medium used for C. shehatae seed cultivation consisted of 2 g xylose, 0.3 g yeast extract and 0.5 g peptone (per 50 mL). The medium used for A. niger seed cultivation was composed of 1 g glucose, 0.1 g peptone, and 0.15 g yeast extract (per 50 mL). The initial mixed fermentation medium for ethanol production consisted of 1.5 g glucose, 0.1 g peptone, 0.3 g (NH₄)₂SO₄, 0.1 g KH₂PO₄, 0.05 g NaNO₃, 0.015 g CaCl₂, 0.015 g MgSO₄·7H₂O, 0.15 g yeast extract, 0.7 g bagasse, and 0.5 mL trace elements (per 50 mL).

Feedstock materials

Sugarcane was collected from the sugar-cane field near the northwest of Zhenjiang City, Jiangsu Province, China. The air-dried sugarcane was milled into small particles by a grinding machine, the shredded fronds were sieved to obtain fractions with a particle size of less than 1 mm, and the bagasse was dried in a drying oven prior to use (Kim et al. 2014).

Seed preparation

The strain A. niger was stored at − 10 °C and streaked in the agar plate containing A. niger cultivation medium. One loop was used to inoculate 100 mL A. niger seed media in 250-mL conical flasks. The seed culture was incubated at 30 °C, 150 rpm for 22 h. The preparation of the strain C. shehatae inoculums started with the stock culture into a liquid medium, after the growth culture, inoculated into 250 mL Erlenmeyer flasks with 50 mL of seed culture medium, and then was incubated at 30 °C, 150 rpm for 24 h.

NaOH solution and ionic liquid pretreatment

The preliminary activation of the bagasse was conducted at room temperature and atmospheric pressure with 1% NaOH solution (w/v) (Belal 2013). After 1.5 h activation, the 1% NaOH pretreated bagasse was washed repeatedly with distilled water until the pH value was about 7.0. It was then dried at 60 °C for 12 h. Ionic liquid was mixed with biomass at a ratio 20:1 (w/w) (Qiu et al. 2012) as follows: 20 g 1-butyl-3-methylimidazolium chloride ([Bmim]Cl) was mixed with 1 g NaOH-pretreated bagasse in a 150-mL flat-bottomed flask and heated to 100 °C for 1.5 h. Post pretreatment, 100 mL of deionized water was slowly added into the biomass/[Bmim]Cl slurry to recover the pretreated bagasse. The ionic liquid /water mixture and biomass were separated by vacuum filtration. The solid precipitate was washed with ethanol to remove any remaining IL, followed by repeated washes with deionized water until the washing solution appeared colorless. The solids were collected (Qiu et al. 2012; Perez-Pimienta et al. 2013). Untreated bagasse was used as the control for ethanol fermentation. Solid samples were oven dried at 60 °C for 24 h and stored in sealed bags at room temperature.

Ethanol fermentation with microorganisms A. niger and C. shehatae

Bioethanol fermentation was conducted in liquid state fermentation. The strain A. niger was incubated at 30 °C, 100 rpm for around 22 h on a rotary shaker, and the strain C. shehatae cells were incubated at 30 °C with shaking at 150 rpm for 24 h. To determine the suitable inoculation proportion and interval time of strain inoculation, the A. niger seed culture was inoculated into 250 mL Erlenmeyer flask with ethanol fermentation medium at 10% (v/v) at 30 °C on rotary shaker with 150 rpm, and the C. shehatae seed culture was transferred into the fermentation culture at the inoculation ratios 2:1, 1:2, and 1:1 (v/v) compared with the inoculum size of A. niger. The volume of all fermentation medium was 50 mL. The seed medium of strain A. niger was inoculated into the mixed ethanol fermentation medium first, followed by the C. shehatae cells. The interval times of strain inoculation were estimated at 0, 12, 24, 36, and 48 h, respectively, and all the inoculum sizes were 10% (v/v).

Determination of optimal temperature, rotation speed, initial pH, inoculum size, mixed fermentation medium volume, fermentation time, and the bagasse, as well as addition of KH₂PO₄, (NH₄)₂SO₄ were investigated in the ethanol fermentation. The parameters tested were: temperatures: 28, 30, and 32 °C, rotation speed: 100, 150, and 200 rpm, initial pH values: 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, and 7.0, fermentation medium volumes: 30, 50, 70, 90, and 110 mL, inoculum sizes: 5.0, 7.5, 10, 12.5, and 15 mL, and fermentation time: 1, 2, 3, 4, 5, 6, and 7 days. For determination of optimum bagasse, KH₂PO₄ and (NH₄)₂SO₄ contents for ethanol production following conditions were tested: bagasse content: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, and 1.0 g, KH₂PO₄ content: 0.05, 0.075, 0.10, 0.125, and 0.15 g, and (NH₄)₂SO₄ content: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9 g.

Plackett–Burman design and method of steepest ascent

Plackett–Burman (PB) design was introduced in this study to evaluate the effect of the processing factors and identify which have a significant effect (Wang et al. 2012) on ethanol production. Before designing the experiment, suitable values for the seven controllable factors were selected according to previous single-factor experiments and our preliminary tests. A 12-experiment with 7-factor PB design (Haghighi Mood et al. 2013; Subara et al. 2018) was performed. The dummy factors could be used for a statistical evaluation of the factor effects (Haghighi Mood et al. 2013). The two-level independent factors were investigated for the screening design: − 1 for low level and + 1 for high level (Wang et al. 2012). Table 1 shows the factors under investigation as well as the levels of each factor used in the experimental design, whereas Table 2 lists the design matrix. Each line of Table 2 represents a run, which is a specific set of factor levels to be applied. The response value used in this study was ethanol production. PB design was conducted using the Minitab software (Version 16, Minitab Inc., USA).

Table 1.

Variables and levels of Plackett–Burman design

Code	Variables	Unit	Low level (− 1)	High level (+ 1)
A	Initial pH	–	5.5	6.5
B	(NH4)2SO4	g	0.3	0.5
C	KH2PO4	g	0.075	0.1
D	Inoculation amount	%	5	7.5
E	Fermentation time	days	3	4
F	Temperature	°C	26	28
G	Rotation speed	rpm	80	100

Level “− 1” represents the low factor level, level “+ 1” represents the high factor

Table 2.

Results of Plackett–Burman design matrix for evaluating factors

Run	A	B	C	D	E	F	G	aYs (g/L)
1	− 1	1	1	1	− 1	1	1	6.75 ± 0.21
2	1	1	1	− 1	1	1	− 1	5.20 ± 0.13
3	− 1	− 1	− 1	1	1	1	− 1	5.46 ± 0.21
4	1	1	− 1	1	− 1	− 1	− 1	3.43 ± 0.23
5	− 1	1	− 1	− 1	− 1	1	1	6.66 ± 0.30
6	1	− 1	1	1	− 1	1	− 1	2.00 ± 0.14
7	− 1	− 1	1	1	1	− 1	1	5.53 ± 0.17
8	1	1	− 1	1	1	− 1	1	5.35 ± 0.08
9	− 1	− 1	− 1	− 1	− 1	− 1	− 1	5.14 ± 0.20
10	1	− 1	− 1	− 1	1	1	1	3.60 ± 0.18
11	1	− 1	1	− 1	− 1	− 1	1	1.03 ± 0.04
12	− 1	1	1	− 1	1	− 1	− 1	6.98 ± 0.29

^aYs represents average ethanol production of triplicate experiments

Based on the PB experiment results, the factors that influenced ethanol production were identified by initial screening. The resulting important factors were initial pH value, (NH₄)₂SO₄, and fermentation time, respectively. These three factors were used to the next path of steepest ascent experiment. The path of steepest ascent was used to approach the optimal region of the three significant factors (Pan et al. 2008), and the optimal results were used for the following response surface optimization. The path of steepest design is shown in Table S1.

Box–Behnken experimental design and response surface methodology

As the region of optical response is identified by the method of steepest ascent, it is often necessary to characterize the response in that region (Wang and Wan 2009). In Table S2, three significant variables from the results of Box–Behnken design were selected for further optimization at three levels (− 1, 0, + 1): initial pH value (5.0, 5.5, 6.0), (NH₄)₂SO₄ concentration (0.3, 0.4, 0.5 g/50 mL), and fermentation time (3, 4, 5 days). The BBD matrix (Table 4) consisted of 12 different level combinations and 3 center point runs. Response surface methodology (Chen et al. 2018) (RSM) is a collection of mathematical and statistical techniques used for modeling and analyzing data where the response of interest is influenced by several variables, and the objective is to optimize this response (Han et al. 2010). In this study, RSM based on Box–Behnken design experiments (Vyavahare et al. 2018) was applied to optimize the conditions of ethanol fermentation and to evaluate the effect of variables, which included initial pH (X₁), ammonia sulfate (X₂), and fermentation time (X₃). Experiments were randomized in order to minimize the effects of unexplained variability in the observed response due to extraneous factors, and the experimental data were fitted to the following quadratic polynomial model (Xie et al. 2013):.

$$Y={\mathit{\alpha }}_o+\sum _{i=1}^3{\mathit{\alpha }}_i{X}_i+\sum _{i=1}^3{\mathit{\alpha }}_{ii}{X}_i^2+\sum _{i,j=1}^3{\mathit{\alpha }}_{ij}{X}_i{X}_j+\mathit{\epsilon },$$ 1

where Y represents the predicted response; ${\mathit{\alpha }}_o$ is the model constant, ${\mathit{\alpha }}_i$, ${\mathit{\alpha }}_{ii}$, ${\mathit{\alpha }}_{ij}$ are the linear, quadratic, and interaction coefficients, respectively, $X_i$ and $X_j$ are the coded variables, and ε is the statistical error (Xie et al. 2013). Design Expert^® Software (Version 8.0.6, Stat-Ease Inc, Minneapolis, MN) (Baroutaji et al. 2015) was used to design the optimization experiments and conduct data processing. The significance of these variables was evaluated by variance analysis (ANOVA). Differences with p < 0.05 was regarded as obvious, p < 0.01 as significant, and p < 0.001 as very significant.

Table 4.

The Box–Behnken experimental design with three independent variables

Run #	Initial pH value		(NH4)2SO4 (g)		Fermentation time (days)		aYs (g/L)
	X 1	Code	X 2	Code	X 3	Code
1	5.0	− 1	0.3	− 1	4.0	0	2.67 ± 0.06
2	5.0	− 1	0.5	1	4.0	0	2.67 ± 0.11
3	6.0	1	0.3	− 1	4.0	0	4.66 ± 0.12
4	6.0	1	0.5	1	4.0	0	6.17 ± 0.10
5	5.5	0	0.5	− 1	5.0	− 1	5.33 ± 0.12
6	5.5	0	0.3	− 1	5.0	1	5.00 ± 0.16
7	5.5	0	0.5	1	3.0	− 1	2.67 ± 0.11
8	5.5	0	0.5	1	5.0	1	4.00 ± 0.08
9	5.0	− 1	0.4	0	3.0	− 1	2.75 ± 0.03
10	6.0	1	0.4	0	3.0	− 1	8.33 ± 0.11
11	5.0	− 1	0.4	0	5.0	1	4.95 ± 0.14
12	6.0	1	0.4	0	5.0	1	5.88 ± 0.05
13	5.5	0	0.4	0	4.0	0	8.12 ± 0.12
14	5.5	0	0.4	0	4.0	0	6.70 ± 0.10
15	5.5	0	0.4	0	4.0	0	8.44 ± 0.15

The center point was replicated three times to estimate the experimental errors

^aYs represented average ethanol production of triplicate experiments

Analytical and statistical analysis

The ethanol concentration was determined by an SBA-40D biosensor analyzer (Institute of Biology, Shandong Academy of Sciences, Jinan, China). The experiments were done in triplicate. All values are expressed as means ± standard deviation. Mean value treatments were compared by the analysis of variance. Analysis of variance (ANOVA), followed by Tukey test, was used to analyze the effect of investigated parameters. Differences at P < 0.05 were considered as significant, and P < 0.01 was considered as very significant.

Results and discussion

Determination of inoculation proportion and inoculation interval time

For the preliminary investigation of the effects of inoculation proportion and inoculation interval time on ethanol production by A. niger and C. shehatae, different inoculation proportions and inoculation interval times were tested. The results are presented in Fig. 1. As shown in Fig. 1a, the inoculation proportions of A. niger and C. shehatae were 1:2, 2:1, and 1:1, respectively. When the fermentation time was 1 day and 2 days, compared to the inoculation proportions of (1:2) and (1:1), the inoculation proportion of (2:1) showed higher ethanol production, while ethanol production with inoculation proportion (2:1) almost had no increase during the fermentation time of 3–5 days. It might be that excessive A. niger inoculum had negative effects on the growth of C. shehatae (Pal and Khanum 2010). Perhaps the generation of metabolites by A. niger could have inhibitory effects on the growth of C. shehatae (Pal and Khanum 2010; Lakshmi et al. 2009). However, when the fermentation time was in the range of 3–5 days, the inoculation proportion of (1:1) exhibited an apparent higher ethanol production than the inoculation proportion of (1:2). It possibly indicates that nutrient substances are consumed by excessive C. shehatae (Montheard et al. 2012).

Fig. 1.

Comparison of the effects of inoculation proportion and inoculation interval time on the ethanol production by strains A. niger and C. shehatae in flask fermentation is shown as a, b, respectively. The data shown are means of three repeated fermentations and the error bars represent their standard deviation

From these results, we might demonstrate that excessive A. niger inocula may have inhibitive effects on growth of C. shehatae (Pirota et al. 2013). Meanwhile, excessive C. shehatae cells consume nutrients sooner, leading to an increase in the rate of microbial death (Zhang et al. 2012). As shown in Fig. 1b, maximum ethanol production was obtained when the interval time was 12 h, whereas lower ethanol production was produced when A. niger and C. shehatae were inoculated simultaneously at the beginning of the mixed culture (interval time 0 h). This may be explained by the necessity of a period of time for the xylanases (Do et al. 2013) and cellulase produced by A. nigher to decompose the hemicellulose and cellulose (Pirota et al. 2013; Sato et al. 2010). At 0 h, almost no carbon source such as glucose or xylose would have been available for the growth of C. shehatae. When the inoculation delay time of C. shehatae was 36 and 48 h, respectively, the resulting ethanol production was significantly lower than that of the 12-h delay time. This might be due to longer fermentation time of A. niger generating metabolites with a negative effect on the growth of C. shehatae (Hickert et al. 2013, 2014). There was no significant effect on ethanol production with the delay time of C. shehatae between 12 and 24 h, throughout the fermentation period of 12–24 h. The results of ethanol production indicated that mutualism between A. niger and C. shehatae appeared useful and beneficially complimented each other (Hickert et al. 2014; Lebeau et al. 2007). These results suggest that the strategy of co-culture not only reduces the xylose concentration through consumption, but also stimulates cell growth and ethanol production by C. shehatae (Kuhad et al. 2011).

To summarize, the study identified the optimal inoculation interval time between C. shehatae and A. niger as 12 h, with an inoculation proportion of 1:1.

The optimal single factor experiment results

In order to obtain optimal fermentation conditions and achieve higher ethanol production, a series of single-factor experiments were performed in this study. These included the selections of suitable medium component supplementation, such as the addition of bagasse, (NH₄)₂SO₄, and KH₂PO₄, as well as the optimization of fermentation process parameters; temperature, rotation speed, initial pH, inoculum size, medium volume, and fermentation time. The results are presented in Figs. 2 and 3. Figure 2b–d, summarizes the results from the variation in addition of (NH₄)₂SO₄, KH₂PO₄, and bagasse. Maximum ethanol production was achieved when (NH₄)₂SO₄, KH₂PO₄, and bagasse additions were 0.5, 0.1, 0.9 g/50 mL, respectively. Further increase in (NH₄)₂SO₄ and KH₂PO₄ addition resulted in a decrease in ethanol production. Low levels of (NH₄)₂SO₄ and KH₂PO₄ supplementation benefited ethanol production, whereas high levels of (NH₄)₂SO₄ and KH₂PO₄ demonstrated a negative effect. Likely excessive (NH₄)₂SO₄ and KH₂PO₄ addition would change the pH of fermentation liquor, negatively affecting the osmotic stress for microorganisms (De Bari et al. 2014; Sittijunda and Reungsang 2012), while lower levels of salts were beneficial for growth. Ethanol production increased with increase in bagasse supplementation as could be expected due to the increased availability of carbon source. However, this appeared to be limited to 0.9 g/50 mL (Fig. 2d). Higher addition of bagasse resulted in lower ethanol production, likely due to by-product production preventing further substance degradation (Kongjan et al. 2009; Belmessikh et al. 2013; Han et al. 2011).

Fig. 2.

Effect of temperature and rotation speed on ethanol production (a), the effect of (NH₄)₂SO₄ addition on ethanol production (b), the effect of KH₂PO₄ addition on ethanol production (c), the effect of bagasse addition on ethanol production (d). Results expressed as mean values ± standard deviation (n = 3)

Fig. 3.

Effect of media volume on ethanol production (a), the effect of initial pH on ethanol production and final pH (b), the effect of fermentation time on ethanol production (c), the effect of inoculum volume on ethanol production (d). The data were the averages ± standard deviations of three independent experiments

The temperature and rotation speed for the fermentation was optimized as shown in Fig. 2a. The optimum temperature and rotation speed for maximal ethanol production was found to be 28 °C at 100 rpm. Table S3 shows that with the increase of temperature, the final pH increased correspondingly, while the ethanol production decreased. Potentially, the exorbitant temperature inhibited cell growth and ethanol production in the mixed fermentation (Pirota et al. 2013). As the rotation speed increased, the ethanol production decreased (Fig. 2a), demonstrating that an increase in dissolved oxygen from the increased stirring inhibited the metabolism for ethanol production by C. shehatae (Hickert et al. 2013; Fontana et al. 2009).

The final pH also decreased with the increase of rotation speed as shown in Table S3. This indicated that acids were produced during the fermentation process, and excessive acidic substances could have an inhibitory effect on ethanol production (Liu et al. 2010). In addition, it showed that with the increase of rotating speed, microorganisms’ metabolism increased, which led to an increase of acidic substances (Fontana et al. 2009; Porfiri et al. 2010).

The effects of media volume, initial pH, fermentation time, and inoculum on ethanol production in the flask fermentation by A. niger and C. shehatae were also studied (Fig. 3). When the volume was below 50 mL, lower ethanol concentration was produced. Two different effects may have caused this. As a results of an increase in the dissolved oxygen, the ethanol was oxidized into acetic acid (Agler et al. 2011; Bengtsson et al. 2010), which had a negative effect on the cell growth of C. shehatae and an inhibitory effect on ethanol production (Bellido et al. 2011; Steinbusch et al. 2009). Conversely, more dissolved oxygen had a positive effect on the growth of C. shehatae (Hickert et al. 2013), but most of the carbon source was used to the growth of C. shehatae, but not to produce ethanol by xylose fermentation (Hickert et al. 2013; Fontana et al. 2009). When the volume was greater than 90 mL, lower ethanol concentration was also obtained. The reason may be that less dissolved oxygen would adversely influence the growth of C. shehatae. It has been reported that C. shehatae is microaerobic, or dissolved oxygen concentration should ensure that within a reasonable range in the shake flask, which would be beneficial to the growth of microorganisms (Hickert et al. 2013; Fontana et al. 2009; Bengtsson et al. 2010) (Fig. 3a).

pH value was one of the most significant factors which had influences on the ethanol fermentation (Wang and Wan 2009; Steinbusch et al. 2009). Appropriate pH value range is key to the growth and metabolism of microbes (Sun et al. 2012; Chaganti et al. 2012). As shown in Fig. 3b, when the initial pH value was lower than 5.5 or higher than 6.5, significant lower ethanol production was gained. Maximum ethanol production was observed when the pH value was 5.5, and in general higher ethanol production was obtained when the initial pH value was between 5.5 and 6.5. It may be that as a result of lower or higher pH value, microbes lost the suitable fermentation conditions to synthesize certain substances required for growth and metabolism (Liu et al. 2010; Sun et al. 2012).

As shown in Fig. 3c, a fermentation time of 4 days appeared to be the optimum for ethanol production by A. niger and C. shehatae. When the fermentation was continued for longer than 4 days, the ethanol production dropped. Likely the death rate increased because of too long fermentation time, making metabolism of microbes weak and general bacteria were suppressed by unfavorable conditions (Ma et al. 2014; Inoue et al. 2013).

As the inoculum size increased, the ethanol production increased correspondingly reaching a maximum with 10 mL inoculation. Further increase in the inoculum, however, resulted in a decrease of ethanol production (Fig. 3d).

Screening of important factors using Plackett–Burman design

From the above one-factor experiments, it was not obvious which factor had the most significant influence on ethanol production. Plackett–Burman design is a powerful technique for screening significant variables (Abd El Aty et al. 2014). Therefore, in this study, based on single-factor experiments, it was used to analyze the effect of seven variables [initial pH, (NH₄)₂SO₄ and KH₂PO₄ concentration, inoculation amount, fermentation time, temperature, and rotation speed] using a two-level 12-run Plackett–Burman design (Oita et al. 2009). Each variable was examined at two levels: − 1 for the low level and + 1 for the high level. The values of two levels were set based on the previous single-factor experiments. Table 1 illustrates the seven variables and their corresponding levels used in the experimental design. The Plackett–Burman design and the response values of ethanol production were shown in Table 2. The data listed in Table 2 indicated that there was a wide variation in ethanol production, from 1.03 to 6.98, during the 12 trials (Chan Cupul et al. 2014).

As shown in Table 3, the effects of each variables towards the ethanol production followed the order of initial pH value > (NH₄)₂SO₄ > fermentation time > temperature > KH₂PO₄ > rotation speed > inoculation amount. The P values less than 0.05 demonstrated that the variables were significant. Analysis of P values depicted in Table 3 shows that among variables tested, initial pH value (P = 0.002), (NH₄)₂SO₄ concentration (P = 0.006), and fermentation time (P = 0.030) had significant influences on ethanol production (Guzun et al. 2014). Variables KH₂PO₄ concentration, inoculation amount, temperature, and rotation speed had P values higher than 0.05 and were regarded as insignificant. Hence they were not included in the next path of steepest ascent and Box–Behnken design experiments.

Table 3.

Main effect factors analysis results of Plackett–Burman experimental design

Variable	Effects	Coefficient	Standard error	T value	P value	Significance
Constant		4.761	0.1790	26.59	0.000
Initial pH value	− 2.652	− 1.326	0.1790	− 7.41	0.002a	1
(NH4)2SO4	1.935	0.968	0.1790	5.40	0.006a	2
KH2PO4	− 0.358	− 0.179	0.1790	− 1.00	0.374	5
Inoculation amount	− 0.015	− 0.008	0.1790	− 0.04	0.969	7
Fermentation time	1.185	0.592	0.1790	3.31	0.030b	3
Temperature	0.368	0.184	0.1790	1.03	0.362	4
Rotation speed	0.118	0.059	0.1790	0.33	0.758	6

P < 0.05 was considered as significant and P < 0.01 was considered as very significant

^aSignificant at 1% (P < 0.05)

^bSignificant at 5% (P < 0.05)

Path of steepest ascent

Based on the Plackett–Burman results, the path of steepest ascent (Kamarudin et al. 2017) moved along the direction of (NH₄)₂SO₄ concentration and fermentation time increasing, and initial pH value decreasing (Zhou et al. 2011). The experimental design and results are shown in Table S1. It was observed that the highest response value for ethanol production was 5.03 g/L with the initial pH value 5.5, (NH₄)₂SO₄ at 0.4 g/50 mL, and fermentation time for 4 days. This point was concluded to be near optimal and was chosen for the next optimization of Box–Behnken design experiment (Zhou et al. 2011; Long et al. 2010).

Results of optimization by response surface methodology

Based on the previous results of Plackett–Burman design and path of steepest ascent experiments, optimization of three process parameters (X₁ = initial pH value, X₂ = (NH₄)₂SO₄, X₃ = fermentation time) was attempted using a Box–Behnken design for the optimal production of ethanol by a mixed culture of A. niger and C. shehatae (Table S2).

The design matrix of tested variables and experimental results was shown in Table 4. The quadratic model (fitted with coded values) (Martínez-Toledo and Rodríguez-Vázquez 2010) was given by the following equation:

$${}^{\text{a}}\text{Ys}=7.75+1.50X_1-0.27X_2+0.094X_3+0.38X_1{X}_2-1.16X_1{X}_3+0.42X_2{X}_3-1.24{X_1}^2-2.47{X_2}^2-1.03{X_3}^2,$$ 2

X ₁: initial pH value; X₂: (NH₄)₂SO₄; X₃: fermentation time.

The adequacy of the model was checked using ANOVA (Zhou et al. 2011) shown in Table 5. The “F value” of the model was 5.59, and the value of “Prob > F” was 0.0363, which is less than 0.05. These results implied that the model was significant. The “F value” of lack of fit was 1.42, indicating that the lack of fit was not significant relative to the pure error. In addition, the coefficient of determination (R²) for the model was calculated as 0.9096 for ethanol production, which indicated that there was a good agreement between the experimental and the predicted values, suggesting that the model was suitable for ethanol production prediction within the range of the variables employed.

Table 5.

ANOVA for response surface quadratic model for ethanol production

Source	Sum of squares	df	Mean square	F value	P value Prob > F	Significant
Model	54.14	9	6.02	5.59	0.0363	Significant
X 1	18	1	18	16.72	0.0095
X 2	0.58	1	0.58	0.54	0.4966
X 3	0.07	1	0.07	0.065	0.8085
X 1X 2	0.57	1	0.57	0.53	0.4994
X 1X 3	5.41	1	5.41	5.02	0.0751c
X 2X 3	0.69	1	0.69	0.64	0.4600
X 1 2	5.69	1	5.69	5.29	0.0698c
X 2 2	22.51	1	22.51	20.91	0.0060b
X 3 2	3.95	1	3.95	3.67	0.1136
Residual error	5.38	5	1.08
Lack of fit	3.67	3	1.22	1.42	0.4377	Not significant
Pure error	1.72	2	0.86
Correlation total	59.52	14

P value is lower than 0.05 being regarded as significant; otherwise, it is considered as not significant. P value is close to 0.05, model P < 0.05, lack of fit P > 0.05, R² = 0.9096, adj-R² = 0.7468, CV% = 19.87, SD = 1.04, mean = 5.22, Pred R² = − 0.054, Adeq precisior = 8.421, Press = 62.52

To further study mutual interactions between the significant factors, the graphical representations of the regression model (the response surface plots) were obtained using Design-Expert software. The results are shown in Fig. 4; each response surface plot represents the effect of the interaction of two independent variables, keeping the other variables at zero level (Zhou et al. 2011) to identify the optimal level of each factor for maximum response (Hwang et al. 2012). The results showed that the effect of initial pH value and (NH₄)₂SO₄ on ethanol production was not significant (Fig. 4a, “Prob > F” = 0.4994, higher than 0.05), while the interaction between initial pH value and fermentation time was closer to significant (Fig. 4b, “Prob > F” = 0.0751, close to 0.05). The surface plot (Fig. 4c) showed the effect of interaction (NH₄)₂SO₄ and fermentation time on ethanol production was also not significant (“Prob > F” = 0.4600, more than 0.05). Thus, according to the result of the interactions between initial pH value and (NH₄)₂SO₄, (NH₄)₂SO₄, and fermentation time, it means that the effect of one-factor response value does not depend on the level of the other factor (Abdel-Fattah 2002). Response value (^aYs) increased as factors of (NH₄)₂SO₄ and fermentation time increased; however, when the response value reached the maximum value, response value decreased with the increase of independent variables. On the basis of the results of statistically designed experiments, the optimal three independent variable values were identified as follows: initial pH value 5.89, (NH₄)₂SO₄ at 0.4 g/50 mL, fermentation time for 3.6 days, with the predicted ethanol production of 8.32 g/L.

Fig. 4.

a Three-dimensional response surface plots (coded values) showing the effect of variable interaction: initial pH value (X₁) and (NH₄)₂SO₄ (X₂). b Three-dimensional response surface plots (coded values) showing the effect of variable interaction: initial pH value (X₁), and fermentation time (X₃). c Three-dimensional response surface plots (coded values) showing the effect of variable interaction: (NH₄)₂SO₄ (X₂) and fermentation time (X₃)

Validation of the model

The adequacy of the statistical model was examined by running test experiments using the derived optimal conditions with initial pH value 5.89, (NH₄)₂SO₄ at 0.40 g/50 mL, and fermentation time for 3.60 days. Under the optimized conditions, the ethanol production of three repeated fermentation values was 7.95 g/L, 8.21 g/L, 8.25 g/L, respectively. In the experiment, the average of observed experimental value was 8.14 g/L, which was 2.54 times of original ethanol concentration 3.21 g/L. The experimental value was close to the predicted response value 8.32 g/L, which demonstrated the generated model was an adequate prediction of the ethanol production (Mnif et al. 2013).

The results of the comparison of effects of untreated bagasse and IL pretreated bagasse on the ethanol production in the initial fermentation condition and optimal fermentation condition by A. niger and C. shehatae are shown in Fig. 5. From the initial to the optimal fermentation conditions, the ethanol production increased from 2.67 to 5.03 g/L with untreated bagasse, an 88.39% increase. Using IL-pretreated bagasse ethanol production increased from 3.21 to 8.14 g/L, an increase of 153.58%. These results demonstrated that the optimization technology had a significant positive influence to improve ethanol production. More importantly, using the optimal fermentation medium and conditions, the ethanol production with IL-pretreated bagasse and untreated bagasse showed a 62% increase. Under the initial fermentation medium and condition, ethanol production with IL-pretreated bagasse and untreated bagasse, however, the increase was only 20%. Both under optimal and original fermentation conditions, the results indicated that the fermentation medium with IL-pretreated bagasse had higher ethanol production than the fermentation medium with untreated bagasse in the ethanol fermentation by A. niger and C. shehatae. Therefore, the bagasse with IL pretreatment was indeed beneficial for improving ethanol production.

Fig. 5.

Comparison of the effects of untreated bagasse and ionic liquid (IL) pretreated bagasse on the ethanol production in the initial fermentation conditions and optimal fermentation conditions by A. niger and C. shehatae. The results were expressed as the mean value ± standard deviation (n = 3)

Conclusions

The present study demonstrated that ionic liquid-pretreated bagasse used to produce bioethanol is a feasible method as a high-efficient technique for biomass transformation into polysaccharide to be decomposed by A. niger. Additionally, it will not increase process costs of enzymatic hydrolysis, such as other costs of the saccharification process using cellulase, xylanase and hemicellulose etc. (Yang et al. 2010). Moreover, this process procedure will result in higher production for bioethanol fermentation.

Through the methods of single-factor experimental design, Plackett–Burman design, Box–Behnken design, and response surface plotting, the fermentation medium and conditions were improved resulting in higher bioethanol production. Therefore, application of such optimization methodology is necessary and great importance for industrial bioprocessing.

The precise mechanism of the synergism between A. niger and C. shehatae is still unclear, and further research is essential for industrial utilization of the mixed culture for bioethanol production.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Compliance with ethical standards

Conflict of interest

We confirm that this manuscript has neither been published elsewhere nor was under consideration by another journal. All authors approved this manuscript and agree with its submission to your journal. The authors have no conflicts of interest.