Performance parameters of the fermentation process
The assessment of higher alcohols formation must be conducted under fermentation conditions that provide performances comparable to those found in industry and the scientific literature. Thus, some parameters were calculated to evaluate the performance of the fermentation process concerning the microorganism employed and the target product of interest for the industry.
Table 2 presents the analysis of variance (ANOVA) of 2³ factorial design for the responses ηf, ηp, P, and YX/S. The results indicate that the treatments were statistically significant (p-value < 0.05), suggesting the presence of significant differences in the means of fermentation conditions concerning these response variables.
YX/S
|
Source of Variation
|
Degrees of freedom
|
Sum of squares
|
Mean square
|
F
|
p-value
|
Treatment
|
7
|
0.021567
|
0.003081
|
8.85
|
0.000
|
Error
|
16
|
0.00557
|
0.000348
|
|
|
Total
|
23
|
0.027137
|
|
|
|
nf
|
Source of Variation
|
Degrees of freedom
|
Sum of squares
|
Mean square
|
F
|
p-value
|
Treatment
|
7
|
4348.01
|
621.14
|
21.26
|
0.000
|
Error
|
16
|
467.55
|
29.22
|
|
|
Total
|
23
|
4815.56
|
|
|
|
np
|
Source of Variation
|
Degrees of freedom
|
Sum of squares
|
Mean square
|
F
|
p-value
|
Treatment
|
7
|
5947.36
|
849.62
|
33.43
|
0.000
|
Error
|
16
|
406.69
|
25.42
|
|
|
Total
|
23
|
6354.04
|
|
|
|
P
|
Source of Variation
|
Degrees of freedom
|
Sum of squares
|
Mean square
|
F
|
p-value
|
Treatment
|
7
|
13.4103
|
1.91575
|
13.7
|
0.000
|
Error
|
16
|
2.2381
|
0.13988
|
|
|
Total
|
23
|
15.6483
|
|
|
|
YX/S (gyeast/gTRS) substrate-to-cell conversion factor; nf (%) fermentation efficiency; np (%) process efficiency; P (g/L.h) ethanol productivity
Table 2. ANOVA of the factorial design for the response variables ηf, ηp, P and YX/S.
The means of the results obtained under the different fermentation conditions used are presented in Table 3.
Fermentation condition
|
Independent variables
|
Results (mean ± sd)
|
Coded values
|
Real values
|
|
pH
|
Sup.
|
Ref.
|
pH
|
Sup. (g/L)
|
Ref.
|
YX/S (gyeast/gTRS)
|
nf (%)
|
np (%)
|
P (g/L.h)
|
1
|
-1
|
-1
|
-1
|
3.5
|
0.0
|
without
|
0.21 ± 0.02a
|
45.65 ± 5.19c
|
35.68 ± 3.30d
|
3.92 ± 0.14d
|
2
|
1
|
-1
|
-1
|
5.0
|
0.0
|
without
|
0.18 ± 0.01a b
|
77.24 ± 2.49a b
|
75.35 ± 3.37a b c
|
5.23 ± 0.44b c
|
3
|
-1
|
1
|
-1
|
3.5
|
1.0
|
without
|
0.22 ± 0.03a
|
88.85 ± 5.67a
|
75.86 ± 5.39a b
|
5.40 ± 0.43a b c
|
4
|
1
|
1
|
-1
|
5.0
|
1.0
|
without
|
0.11 ± 0.00c
|
87.31 ± 5.10a
|
87.17 ± 5.06a
|
6.03 ± 0.09a b
|
5
|
-1
|
-1
|
1
|
3.5
|
0.0
|
with
|
0.15 ± 0.00b c
|
90.23 ± 5.93a
|
88.77 ± 5.77a
|
6.44 ± 0.16a
|
6
|
1
|
-1
|
1
|
5.0
|
0.0
|
with
|
0.18 ± 0.03a b
|
70.97 ± 7.32b
|
61.21 ± 6.26c
|
4.66 ± 0.51c d
|
7
|
-1
|
1
|
1
|
3.5
|
1.0
|
with
|
0.18 ± 0.01a b
|
78.72 ± 5,14a b
|
69.76 ± 5.04b c
|
4.68 ± 0.54c d
|
8
|
1
|
1
|
1
|
5.0
|
1.0
|
with
|
0.17 ± 0.03a b
|
79.40 ± 5.24a b
|
75.25 ± 5.34a b c
|
5.21 ± 0.37b c
|
Means followed by the same lowercase letter in the column do not differ by Tukey's test at 5% significance level
YX/S (gyeast/gTRS) substrate-to-cell conversion factor; nf (%) fermentation efficiency; np (%) process efficiency; P (g/L.h) ethanol productivity
Table 3. Results obtained by 2³ factorial design.
The substrate-to-cell conversion factor (YX/S) met the criteria established in the literature [24,25], indicating successful yeast development under the adopted fermentation conditions (Table 3).
The fermentation conditions 4 and 5 (Table 3) resulted in lower values for YX/S (0.11 ± 0.00 gyeast/gTRS and 0.15 ± 0.00 gyeast/gTRS, respectively). It is important to highlight that fermentation condition 4 differed significantly from the other conditions (p < 0,05), except for condition 5.
In the literature, a wide range of YX/S values obtained in fermentations with ethanol-producing yeasts has been reported. According to Stroppa et al. [24], the reported values range from 0.03 to 0.28 gyeast/gTRS. The authors conducted fermentations with yeasts isolated from distilleries in sugarcane must at a concentration of 9.4 °Brix, temperature of 30°C for 24 hours, and obtained values of 0.179 and 0.185 gyeast/gTRS for the strains RM01 and CV01, respectively.
Colombi et al. [25] evaluated the influence of different compounds, such as vanillin, acetic acid, vanillic acid, and 4-hydroxybenzoic acid, on the fermentation of glucose at 40 g/L by the yeast Saccharomyces cerevisiae JP1. The authors conducted fermentation at 30°C and 150 rpm for 22 hours, the pH was adjusted to 4.9, and YX/S values ranging from 0.00 to 0.22 gyeast/gTRS were obtained.
Alves [26] conducted fermentations with molasses must at 40 g/L supplemented with 2.5 g/L yeast extract, using an industrial strain of Saccharomyces cerevisiae within a temperature range varying from 28 to 38°C. The obtained values ranged between 0.087 and 0.099 gyeast/gTRS.
The wide variation in YX/S values reported in the literature resulted from different process conditions and raw materials used, as well as yeast strains.
According to Table 3, fermentation condition 1, with a pH of 3.5, and without the addition of ammonium sulfate and without refrigeration, exhibited unsatisfactory performance compared to the other tested conditions, as fermentation efficiency, process efficiency, and ethanol productivity were low.
Fermentation conditions 3, 4, and 5 showed better performance with higher fermentation and process efficiencies, and ethanol productivity. When comparing these three experiments, it is evident that condition 5, with a pH of 3.5, no ammonium sulfate supplementation, and must refrigeration, achieved the highest process performance indicators.
In the case of the need to achieve good yeast productivity, the optimal condition was number 3 (pH 3.5, supplementation of 1 g/L ammonium sulfate, and no refrigeration), which significantly differed (p < 0.05) from conditions 4 and 5 (Table 3).
An analysis of the data presented in Table 3 suggested consistency with the results described in the scientific literature (Table 4). Furthermore, the fermentation conditions 3, 4, and 5 were appropriate, resulting in satisfactory fermentation performance.
The differences in the results reported in the literature (Table 4) are attributed to the different conditions adopted by the authors, including must composition, fermentation time, and yeast strain used.
Reference
|
Medium
|
Microorganism
|
Temperature
|
Fermentation time
|
nf
|
np
|
P
|
[27]
|
sugarcane synthetic must (160 g/L of sugars)
|
Saccharomyces cerevisiae CAT-1
|
30°C
|
72 h
|
90.20%
|
nr
|
nr
|
[28]
|
non-sterile molasses must (26°Brix)
|
Saccharomyces cerevisiae CAT-1
|
30°C
|
24 h
|
79.88%
|
nr
|
4.27 g/Lh
|
[29]
|
molasses + sugarcane juice must (270 g/L of sugars)
|
Saccharomyces cerevisiae Y-904
|
32°C
|
24 h
|
nr
|
92.80%
|
4.27 g/Lh
|
[30]
|
synthetic must (250 g/L of sugars)
|
Saccharomyces cerevisiae flocculant
|
28°C
|
12 h
|
nr
|
82.58%
|
9.6 g/Lh
|
[31]
|
sterile sugarcane juice must (25°Brix)
|
Saccharomyces cerevisiae CAT-1
|
30°C
|
24 h
|
92.73%
|
nr
|
4.69 g/Lh
|
nr not reported
Table 4. Performance parameters of the fermentation process, which are reported in the literature.
Isoamyl alcohol production
The significant effects on isoamyl alcohol production (A) can be seen in the Pareto chart in Figure 1.
There was no significant third-order interaction effect. Among the main effects, only supplementation had a significant effect (p<0.05) on A, but the second-order interaction effects, pH*Ref. and Sup.*Ref., were also significant. Therefore, factors should be evaluated together, as the response obtained by varying one factor depends on the levels of the other factors.
The regression model with coded units, considering only the significant effects and relating isoamyl alcohol production to the factors pH, refrigeration, and supplementation is given by Equation (5). The corresponding graphs can be viewed in Figures 2a and 2b.
ISOAM. (A) = 0.2548 + 0.0488 Sup. – 0.0424 pH*Ref. – 0.0310 Sup.*Ref. (5)
The ANOVA results for the regression model of Equation (5) are presented in Table 5.
Source of Variation
|
Degrees of freedom
|
Sum of squares
|
Mean square
|
F
|
p-value
|
Regression
|
3
|
0.12359
|
0.041197
|
10.04
|
0.000
|
Lack of fit
|
4
|
0.02715
|
0.006786
|
1.98
|
0.147
|
Pure error
|
16
|
0.05495
|
0.003434
|
|
|
Total
|
23
|
0.20568
|
|
|
|
R² = 60.09%
Table 5. ANOVA results for the regression model for the response variable isoamyl alcohol produced (A).
When evaluating the pH*Ref. interaction (Figures. 2a and 2b), it was observed that, in fermentations conducted without refrigeration, an increase in pH and must supplementation led to an increase in isoamyl alcohol production (A). In these same figures, it is noted that by maintaining the absence of refrigeration and decreasing both the pH and must supplementation, there is a decrease in isoamyl alcohol production (A).
In the Sup.*Ref. interaction (Figure 2b), it was observed that must supplementation increased isoamyl alcohol production (A), both in fermentations conducted without refrigeration and those with refrigeration, confirming the significant effect of supplementation (Figure 1).
Isobutanol production
The Pareto chart in Figure 3 shows that the main effects - pH, refrigeration, and supplementation - were significant, and additionally, the second-order interaction effects - pH*Ref. and Sup.*Ref. - were also significant. Therefore, these factors should be evaluated together.
Considering only significant effects, the regression model with uncoded units relates isobutanol production to the factors pH, refrigeration, and supplementation, as represented in Equation (6). The corresponding graphs can be viewed in Figures 4a and 4b.
ISOBU. (B) = – 0.0222 + 0.03120 pH + 0.0549 Sup. + 0.1269 Ref. – 0.02964 pH*Ref. – 0.0394 Sup.*Ref. (6)
The ANOVA results for the regression model of Eq. 6 are presented in Table 6.
Source of Variation
|
Degrees of freedom
|
Sum of squares
|
Mean square
|
F
|
p-value
|
Regression
|
5
|
0.06082
|
0.012163
|
11.75
|
0.000
|
Lack of fit
|
2
|
0.00359
|
0.001796
|
1.91
|
0.180
|
Pure error
|
16
|
0.01504
|
0.000940
|
|
|
Total
|
23
|
0.07945
|
|
|
|
R² = 76.55%
Table 6. ANOVA of the regression model for the response variable isobutanol produced (B).
An examination of Figures 4a and 4b shows that the production of isobutanol (B) exhibited a similar trend to that of isoamyl alcohol production (A) (Figures 2a and 2b) concerning the interaction effects of pH*Ref. and Sup.*Ref. In other words, in fermentations conducted without refrigeration, an increase in pH and must supplementation increased isobutanol production (B).
As shown in Figures 4a and 4b, must supplementation and an increase in pH increased isobutanol production (B), both in fermentations conducted without refrigeration and in those conducted with refrigeration, confirming the significant effects of supplementation and pH (Figure 3). However, must refrigeration resulted in a less significant increase in isobutanol production (B), also confirming the important effect of refrigeration (Figure 3).
A/B ratio between isoamyl alcohol (A) and isobutanol (B) produced
As shown in the Pareto chart in Figure 5, only the main effects of pH and refrigeration were significant. Additionally, the second-order interaction effects pH*Ref. and pH*Sup., as well as the third-order effect pH*Sup.*Ref. were also significant. Therefore, these factors should be evaluated together.
The regression model with coded units, considering only the significant effects and relating the A/B ratio to the factors pH, refrigeration, and supplementation, is represented in Equation (7). The corresponding graphs can be viewed in Figures 6a and 6b.
A/B RATIO = 1.8828 – 0.1482 pH + 0.0955 Ref. + 0.0274 pH*Sup. - 0.0524 pH*Ref. + 0.0778 pH*Sup.*Ref. (7)
The ANOVA results for the regression model of Eq. 7 are presented in Table 7.
Source of Variation
|
Degrees of freedom
|
Sum of squares
|
Mean square
|
F
|
p-value
|
Regression
|
5
|
0.97470
|
0.19494
|
73.30
|
0.000
|
Lack of fit
|
2
|
0.00457
|
0.002283
|
0.84
|
0.448
|
Pure error
|
16
|
0.04330
|
0.002706
|
|
|
Total
|
23
|
1.02257
|
|
|
|
R² = 95.32%
Table 7. ANOVA of the regression model for the response variable A/B ratio.
When evaluating the pH*Ref. interaction (Figure 6a), it was observed that the peak of the A/B ratio occurred at a lower pH and under refrigeration, while the lowest A/B ratio was obtained at a higher pH and without refrigeration.
In the pH*Sup. interaction (Figure 6b), the peak of A/B ratio occurred at a lower pH and without must supplementation. Conversely, the opposite trend was observed with higher pH and must supplementation.
As shown in Figures 6a and 6b, the influence of pH on the A/B ratio is clear, with a lower pH favoring an increase in this ratio.
Comprehensive analysis of the results: isoamyl alcohol produced (A), isobutanol produced (B) and the A/B ratio
The results presented in Table 8 indicate significant differences among the fermentation conditions for the evaluated response variables (production of isoamyl alcohol (A), isobutanol (B), and the A/B ratio).
Fermentation condition
|
Independent variables
|
Results (mean ± sd)
|
Coded values
|
Real values
|
|
pH
|
Sup.
|
Ref.
|
pH
|
Sup. (g/L)
|
Ref.
|
Isoamyl alcohol (A) (g/L)
|
Isobutanol (B) (g/L)
|
A/B Ratio
|
1
|
-1
|
-1
|
-1
|
3.5
|
0.0
|
without
|
0.1487 ± 0.0189b
|
0.0811 ± 0.0113b
|
1.84 ± 0.03c d
|
2
|
1
|
-1
|
-1
|
5.0
|
0.0
|
without
|
0.2406 ± 0.0040b
|
0.1379 ± 0.0046b
|
1.75 ± 0.03d e
|
3
|
-1
|
1
|
-1
|
3.5
|
1.0
|
without
|
0.2717 ± 0.0034a b
|
0.1409 ± 0.0055b
|
1.93 ± 0.06c
|
4
|
1
|
1
|
-1
|
5.0
|
1.0
|
without
|
0.4372 ± 0.0445a
|
0.2666 ± 0.0180a
|
1.64 ± 0.06e
|
5
|
-1
|
-1
|
1
|
3.5
|
0.0
|
with
|
0.2514 ± 0.0746b
|
0.1119 ± 0.0373b
|
2.26 ± 0.09a
|
6
|
1
|
-1
|
1
|
5.0
|
0.0
|
with
|
0.1831 ± 0.0873b
|
0.1109 ± 0.0529b
|
1.65 ± 0.02e
|
7
|
-1
|
1
|
1
|
3.5
|
1.0
|
with
|
0.2598 ± 0.1047b
|
0.1240 ± 0.0493b
|
2.09 ± 0.04b
|
8
|
1
|
1
|
1
|
5.0
|
1.0
|
with
|
0.2460 ± 0.0309b
|
0.1297 ± 0.0198b
|
1.90 ± 0.05c
|
Means followed by the same lowercase letter in the column do not differ by Tukey's test at 5% significance level.
Table 8. Isoamyl alcohol and isobutanol produced under the different fermentation conditions.
The fermentations conducted at pH 5.0, with must supplementation and without refrigeration (fermentation condition 4), resulted in greater formation of isoamyl alcohol and isobutanol. However, the A/B ratio was low under this condition. Despite a low substrate-to-cell conversion (YX/S = 0.11 gyeasts/gTRS), both the fermentation and process efficiencies were satisfactory, as was the high ethanol productivity (87.31%, 87.17%, and 6.03 g/L.h, respectively). These results are consistent with previous studies by Pons et al. [32] and Cachot et al. [33], which demonstrated a positive correlation between the formation of these alcohols and ethanol production. In fact, our results indicate that fermentation conditions leading to higher ethanol production also resulted in increased formation of isoamyl alcohol and isobutanol.
The evaluated factors (pH, refrigeration, and supplementation) simultaneously affect the production of isoamyl alcohol and isobutanol. However, at pH 3.5, the formation of isobutanol was lower than that at other pH values, resulting in a higher A/B ratio between these alcohols. Therefore, a lower pH is suggested to favor an increase in the A/B ratio.
Fermentation condition 5 exhibited the highest A/B ratio (2.26) and significantly differed from the other conditions. Despite a low substrate-to-cell conversion (YX/S =0.15 gyeasts/gTRS), both fermentation and process efficiencies, as well as ethanol productivity, were high (90.23%, 88.77%, and 6.44 g/L.h, respectively).
Even though it was not possible to establish a condition that reduces fusel oil formation, the obtained results are consistent with previous studies [34,35,36], which report the influence of different factors on the production of higher alcohols during fermentation. These results underscore the importance of jointly evaluating the studied factors (pH, refrigeration, and supplementation), considering the interaction effects that occur among them.
Therefore, the results provide a better understanding of these interactions and their impact on the formation of isoamyl alcohol and isobutanol, which can contribute to the development of more efficient processes.