**3.2. Metabolism of corn steep liquor on AC production**

be used as nutritional supplement [32]. The addition of CSL reduced the fermentation time

On the other hand, when yeast extract was used as the nitrogen source in the fermentation media, as depicted in **Figure 2b**, after reducing 150 g/L of sugar, the results showed a rapid consumption of the reducing sugar when compared to **Figure 2a**. More than 97% of the reducing sugar was converted (during the first 84 h of fermentation), the acetoin concentration approached a maximum, and the biomass growth was constant at 8.6 and 7.7 g/L, respectively. The wide range of amino acids, peptides, vitamins, inorganic salts, and carbon in growth media of yeast extract supports the biomass growth and rapid sugar

In **Figure 2c**, the extract was used in place of CSL and yeast extract as earlier discussed (**Figure 2a** and **b**). The beef extract is a mixture of peptides and amino acids, nucleotide fractions, organic acids, minerals, and some vitamins. Its function can, therefore, be described as complementing the nutritive properties of peptone by contributing minerals, phosphates, energy sources, and those essential factors missing from peptone [34]. The glucose was almost depleted after 96 h till the end of the fermentation lapse. It can be deduced from the study that the beef extract supports the utilization of glucose consumption as more than 98% of the

To see the clear difference of the complex media earlier considered, the acetoin fermentation was done without the nitrogen sources (**Figure 2d**) but with the simple salts in the fermentation

**Figure 2.** (a) Plots of AC fermentation using corn steep liquor as nitrogen source. (b) Plots of AC fermentation using yeast extract as nitrogen source. (c) Plots of AC fermentation using beef extract as nitrogen source. (d) Plots of AC fermentation

using none of the complex media as the nitrogen source. (RS–reducing sugar; AC–acetoin; BM–biomass).

and promoted the growth and fermentation of the strain.

glucose has been consumed within 96 h of fermentation.

utilization [33].

90 Renewable Resources and Biorefineries

It was observed in the preliminary evaluation of nitrogen sources discussed earlier (**Figure 2a–d**) that corn steep liquor produces acetoin at a short time interval when compared to other nitrogen sources. Therefore, the fermentation time interval of acetoin was reduced from the initial 168 h (**Figure 2a–d**) to 48 h (**Figure 3**) and sample taken at every 2-h interval, to investigate the biomass growth, utilization of glucose, and the acetoin accumulation when corn steep liquor was used as nitrogen source. The results show that the accumulated acetoin and the biomass growth were already at the peak of 7.03 and 7.83 g/L, respectively, in the first 36 h of fermentation (**Figure 3**). The rapid utilization of reducing sugar (from 150 to 66 g/L) within 36 h of fermentation was due to high amino acids and polypeptides, which are excellent sources of nitrogen in corn steep liquor and has been reported to support the growth of most microorganism [35]. CSL comprises a mixture of reducing sugars that contribute to the nutritional growth of the bacteria with a steady increase in biomass growth (7.7–7.8 g/L) from 18 to 46 h when it starts to decline [36]. It can be affirmed from the findings that corn steep liquor hastens fermentation of acetoin at short time interval and this makes large-scale production of acetoin cost-effective.

### **3.3. Optimization of AC production using response surface method**

After the preliminary studies, RSM coupled with Box-Behnken design (BBD) was used for the optimization of the fermentation process with respect to glucose concentration, CSL

**Figure 3.** Acetoin growth profile using corn steep liquor at short fermentation time interval. RS–reducing sugar; AC–acetoin; BM–biomass.

concentration, and inoculum size with a view to maximize the AC production. **Table 2** shows the experimental conditions investigated together with the observed and predicted values. The data were fitted using the following second-order mathematical equation:

$$\begin{array}{l} \text{(Y = 7.81 - 1.37 X\_1 + 2.56 X\_2 - 0.61 X\_3 - 0.17 X\_1 X\_2 - 0.57 X\_1 X\_3 - 1.02 X\_2 X\_3 \\ \text{ - 1.25 X\_1 {}^2 - 0.50 X\_2 {}^2 - 1.09 X\_3 {}^2 \end{array} \tag{2}$$

where Y is the acetoin (AC) produced in g/L and X<sup>1</sup> is glucose concentration, X<sup>2</sup> is corn steep liquor, X3 is Inoculum size. The residuals between the observed and predicted (**Table 2**) values in this work revealed good fit of the equation as shown by the parity graph which is a measure of agreement between the observed and predicted values (**Figure 4**). These observations implied that the model developed for the fermentation process adequately described the actual relationship among the selected factors.

The three-dimensional graph and contour plot, which is depicted in **Figure 5(a)** and **(b)** shows the relationship between corn steep liquor and glucose concentration when acetoin production is at the maximum (10.70 g/L). Also, the *p*-values of the model terms were significant at *p* < 0.05 (**Table 3**). Also, the observed low *p*-value of 0.0001 together with the corresponding *F*-value of 11.09 showed that the model obtained was significant. The *F*-value and *p*-value do not differentiate between negative and positive significant effects of each term in the model [37]. **Table 3** displays the test of significance and ANOVA of the regression equation results. The coefficient of determination (*R*<sup>2</sup> ) is used to assess the goodness of fit of the regression


**Table 2.** BBD of three independent factors for AC production including the coded levels of each parameter.

equation. *R*<sup>2</sup>

corn steep liquor and glucose on acetoin production.

**Figure 4.** Parity plot of acetoin production.

of 0.930 of the model demonstrated a good correlation between the observed

Statistical Optimization of Acetoin Production Using Corn Steep Liquor as a Low-Cost Nitrogen…

http://dx.doi.org/10.5772/intechopen.79353

93

and predicted values. It showed that 93% sample variation for AC produced is attributable to the independent factors and just 0.70% of the total variations is not described by the model

**Figure 5.** Contour and response surface plots. (a) The response surface plot and (b) contour plot showing the effects of

Statistical Optimization of Acetoin Production Using Corn Steep Liquor as a Low-Cost Nitrogen… http://dx.doi.org/10.5772/intechopen.79353 93

**Figure 4.** Parity plot of acetoin production.

**Run X1**

liquor, X3

 **(g/L) X2**

 **(% w/v) X3**

The coefficient of determination (*R*<sup>2</sup>

− 1.25 *X*<sup>1</sup>

92 Renewable Resources and Biorefineries

<sup>2</sup> − 0.50 *X*<sup>2</sup>

where Y is the acetoin (AC) produced in g/L and X<sup>1</sup>

actual relationship among the selected factors.

 **(%v/v) Observed AC (g/L) Predicted AC (g/L) Residuals**

) is used to assess the goodness of fit of the regression

<sup>2</sup> (2)

is corn steep

is glucose concentration, X<sup>2</sup>

 100 (0) 10 (0) 3.5 (0) 7.68 7.81 −0.13 150 (1) 10 (0) 2 (−1) 4.32 5.28 −0.96 150 (1) 5 (−1) 3.5 (0) 2.47 2.29 0.18 100 (0) 10 (0) 3.5 (0) 7.68 7.81 −0.13 100 (0) 5 (−1) 2 (−1) 3.66 3.24 0.42 100 (0) 15 (1) 2 (−1) 10.69 10.41 0.28 50 (−1) 10 (0) 2 (−1) 7.14 6.88 0.26 100 (0) 15 (1) 5 (1) 6.49 7.15 −0.66 50 (−1) 10 (0) 5 (1) 8.00 6.8 1.20 150 (1) 10 (0) 5 (1) 2.91 2.92 −0.01 100 (0) 10 (0) 3.5 (0) 7.68 7.81 −0.13 100 (0) 5 (−1) 5 (−1) 3.53 4.06 −0.53 150 (1) 15 (1) 3.5 (0) 7.88 7.08 0.80 50 (−1) 15 (1) 3.5 (0) 9.74 10.16 −0.42 100 (0) 10 (0) 3.5 (0) 7.70 7.80 −0.11 100 (0) 5 (−1) 3.5 (0) 5.72 4.74 0.98 50 (−1) 5 (−1) 3.5 (0) 3.65 4.69 −1.04

concentration, and inoculum size with a view to maximize the AC production. **Table 2** shows the experimental conditions investigated together with the observed and predicted values.

*Y* = 7.81 − 1.37 *X*<sup>1</sup> + 2.56 *X*<sup>2</sup> − 0.61 *X*<sup>3</sup> − 0.17 *X*<sup>1</sup> *X*<sup>2</sup> − 0.57 *X*<sup>1</sup> *X*<sup>3</sup> − 1.02 *X*<sup>2</sup> *X*<sup>3</sup>

ues in this work revealed good fit of the equation as shown by the parity graph which is a measure of agreement between the observed and predicted values (**Figure 4**). These observations implied that the model developed for the fermentation process adequately described the

The three-dimensional graph and contour plot, which is depicted in **Figure 5(a)** and **(b)** shows the relationship between corn steep liquor and glucose concentration when acetoin production is at the maximum (10.70 g/L). Also, the *p*-values of the model terms were significant at *p* < 0.05 (**Table 3**). Also, the observed low *p*-value of 0.0001 together with the corresponding *F*-value of 11.09 showed that the model obtained was significant. The *F*-value and *p*-value do not differentiate between negative and positive significant effects of each term in the model [37]. **Table 3** displays the test of significance and ANOVA of the regression equation results.

is Inoculum size. The residuals between the observed and predicted (**Table 2**) val-

The data were fitted using the following second-order mathematical equation:

<sup>2</sup> − 1.09 *X*<sup>3</sup>

**Table 2.** BBD of three independent factors for AC production including the coded levels of each parameter.

**Figure 5.** Contour and response surface plots. (a) The response surface plot and (b) contour plot showing the effects of corn steep liquor and glucose on acetoin production.

equation. *R*<sup>2</sup> of 0.930 of the model demonstrated a good correlation between the observed and predicted values. It showed that 93% sample variation for AC produced is attributable to the independent factors and just 0.70% of the total variations is not described by the model


**Table 3.** Test of significance for every regression coefficient and ANOVA.

[28, 38]. The adjusted *R*<sup>2</sup> of 0.90 proved that the model was significant. It has been suggested that *R*<sup>2</sup> should be less or equal to 80% for the good fit of a model [39].
