Optimization of the Alcoholic Concentration Obtained From Sugary Cassava (Manihot esculenta Crantz) by Response Surface Methodology

Sugary cassava or mandiocaba is a cassava variety of potential use for bioethanol production. In this study, laboratory-scale fermentations were carried out in a bioreactor with a working volume of 1L, using the yeast strain LNF CAT-1. A central composite design (CCD) was applied to determine the extent to which pH, temperature, and yeast concentration influence ethanol production with the aim of improving the fermentation process. The individual effects and the interaction of these factors were analyzed using a surface response method. Physicochemical properties of the material were also investigated and the analysis of root characterization showed high moisture content (~91%) and a low amount of starch (~4.0%), ash values close to 1.0%, total fibers 0.4%, proteins 0.15%, and lipids 0.1%. The results obtained from the wort presented a low acidity (~0.2%), pH close to neutrality (~6.5%), total soluble solids values of ~5.8%, glucose content ~2.3%, fructose ~1.0%, and sucrose ~1.2%. The second-order polynomial regression model determined that the maximum ethanol production of 2.8% (v/v) would be obtained when the optimum pH, temperature, and yeast concentration were ~5.0, 32-36 oC, and ~10-14 g L, respectively.


Introduction
The development of sustainable energy resources and the reduction of greenhouse gases from fossil fuels have become essential topics of interest worldwide (Pradhan, Mahajani, & Arora, 2018). It is known that new sources of cheap fossil fuels are no longer available and experts have been issuing warnings about the possible depletion of current sources in the near future (Sipra & Sarwar, 2018).
It is now clear that the replacement of current fossil energy will require the development of new strategies to reduce our global energy consumption and the development of a panel of renewable energy sources (Carneiro et al., 2017). In this scenario, sustainable biofuel production is a valuable tool to curb climate change (Creutzig et al., 2015). medium from 3.2 to 16.8. The alpha value of orthogonality used was 1.35313 and the coding equations for Temperature (X 0 ), pH (Y 0 ), and yeast concentration (Z 0 ) are shown respectively in Equations 1, 2, and 3. The runs were performed randomly to avoid bias in the results. X 1 = (X 0 -31.5) 3.5 (1) (2) X 3 = (Z 0 -10) 5 (3)

Moisture
The moisture content was determined according to the AOAC method 984.25 (1984). The sample (±2 g) was placed in a previously weighed porcelain crucible. Then placed in the oven at 105 °C, cooled in a desiccator, and then weighed. This procedure was repeated until a constant weight was obtained.

Ashes
The ash content was determined according to the AOAC method 923.03 (1997). Each dry sample (±2 g) previously charred was placed in a porcelain crucible, incinerated in a muffle at a temperature of approximately 550 °C until constant weight.

Fiber
The determination of the total fiber content was obtained through the acid detergent method (FDA) according to AOAC, (method 973.18, AOAC, 1997).

Proteins
The crude protein content was determined according to AOAC method 920.87 (1997). The total nitrogen content was determined using the Kjeldahl method and the protein content calculated by multiplying the nitrogen value by the general factor 6.25.

Lipids
The total lipid content was determined by the cold extraction method using the Bligh-Dyer method (1959). Before carrying out the analysis, the moisture content of the sample was reduced to approximately 10%, due to a large amount of water present.

Total Soluble Solids
The determination of total soluble solids (TSS) consists of the measurement of the solution refractive index. The results were expressed in ºBrix through the use of a digital refractometer HISEG RTD-45 (0-32 °Brix), according to method 932.12 of AOAC (1997).

Total Reducing Sugars
The determination of the Total Reducing Sugar (TRS) concentration expressed as glucose, present in the wort, was performed by the Eynon & Lane method, using the equipment referred to as REDUTEC® (TECNAL, Brazil, model TE-088), for the titration. This method consists of the sum of reducing sugars present in the sample and those from sucrose hydrolysis, according to AOAC method 31.034-6 (1984).

Determination of Starch, Glucose, Fructose, and Sucrose
The content of glucose, fructose, and sucrose (soluble sugars), present in the wort during fermentation, was obtained based on the enzymatic method published by Stitt, Lilley, Gerhardt and Heldt (1989), where 100 μL buffer containing 200 mM Imidazole, 10 mM MgCl 2 was placed in each well of the ELISA plate. 4 mM NAD + , 2 mM ATP and 2.4 U of G6PDH; 5 μL of extract (samples centrifuged 10 000 rpm 3min -1 and diluted 1:20) and 90 μL of distilled water. Glucose, fructose, and sucrose were quantified by the addition of 5 μL of hexokinase (1.4 U), 5 μL phosphoglucoisomerase (0.6 U), and 5 μL invertase (0.8 U), respectively, each time the curve of reaction reached a plateau. The analyzes were performed in triplicate, and the readings were done at 340 nm.
For the quantification of starch, 100 mg of fresh root were homogenized in ethanol 80% (v/v), incubated at 70 °C for 90 min, and subjected to two centrifugations (15 000 g 10 min -1 ). The resulting pellet was suspended in KOH at 95 °C for one hour, neutralized in acetic acid, and again centrifuged (15 000 g 10 min -1 ). The starch was then hydrolyzed in citrate buffer (100 mM at pH 4.6) containing amyloglucosidase and α-amylase (adapted from Trethewey et al., 1998), and the released glucose was quantified as described above.

pH
The pH determination was performed using the Mettler toledo electrode model InPro® 325x in the fermentation analyzes. For the previous analysis of the wort, the T-1000 bench model meter from Tekna was used. The pH meters were duly calibrated with the buffer solutions pH 7.0 and 4.0 at 20 °C, according to the AOAC method No. 981.12 (1997).
2.6.10 Determination of Total Titratable Acidity The total titratable acidity was determined by the AOAC method 942.15 (1997). Titration was carried out with 0.1 mol L -1 NaOH until pH±8.2 (referring to the color change pH of the phenolphthalein indicator). The results were expressed as a percentage of acid per 100 mL of wort.

Determination of Ethanol Concentration
The ethanol concentration was determined by the densimetric method. For distillation, 25 mL of sample was mixed with 25 mL of distilled water, collecting 50 mL of distillate in a micro distiller, model SL. 077 (SOLAB). The samples were collected in sterilized amber glass bottles, promptly frozen, and sent for analysis in a Rudolph digital densimeter, model DDM 2911. Equation 4 was used to calculate the ethanol concentration of the samples.
Where, Ethanol is the concentration of ethanol in ° GL (v/v at 20 °C); L is the sample reading on the DDM 2911 digital densimeter; 2 is the sample dilution factor.

Fermentative Parameters
Based on the results obtained from analytical determinations of the initial and fermented wort, calculations of the following fermentation parameters were performed, using Equations 5 and 6.
Equation 5 determined the yield in gram of ethanol per gram of total reducing sugars (TRS), in percentage, after 7 hours of fermentation.
Equation 6 was used to determine productivity, which expresses the average ethanol production speed, after 7 hours of fermentation.
Where, P ethanol is the productivity of ethanol (g L -1 h -1 ); Cethanol 7 , is the concentration of ethanol (g L -1 ) at the end of 7 hours of fermentation; and t = fermentation time (h)

Determination of Yeast Concentration
Wort samples were taken from the bioreactor at times 0, 0.16, 0.36, 0.5, 1, 2, 3, 4, 5 6, and 7h during fermentation. Then, 2 mL of previously homogenized suspension was pipetted into a microtube of known tare. The tubes were centrifuged and the soluble part was separated into another tube and frozen until sugar analysis was carried out, the precipitate was, in turn, resuspended and washed twice (10 000 rpm 3 min -1 ). It was dried to constant weight in an oven at 80 °C. The dry cell mass was then determined by mass difference, and the results were expressed in terms of grams of dry cells 100 mL -1 of suspension (adapted from Sperotto, 2014).

Physicochemical Characterization of Root and Wort
In the physicochemical composition of the sugary cassava root (Table 1), based on percentage, it was evident a high moisture content and a low amount of starch. Note. *The results were expressed as average±SD (standard deviation).
Among the main characteristics, of physicochemical composition, found in the sugary cassava wort (Table 2), a low acidity, pH close to neutrality and a large amount of free sugars stand out. Note. *Results were expressed as average±SD (standard deviation); ** (mL of NaOH 100g -1 ).

Experimental Optimization of Fermentation
In fermentative runs, the alcoholic concentration% (v/v) varied from 2.33% (run 1) to 2.92% (run 4), with no great variations between the experiment responses (Table 3). In the optimization of the ethanol production process, the significance level p ≤ 0.1 was considered, due to the great variability of the experimental data inherent to bioprocesses. The experiments were evaluated for pure error jas.ccsenet.org Journal of Agricultural Science Vol. 12, No. 11;2020 and a complete second-order polynomial model for data adjustments was proposed.
The analysis of variance (ANOVA) for ethanol concentration (Table 4), provided a coefficient of determination R 2 (correlation coefficient) of 0.94 significant (p ≤ 0.1), indicating a good fit to the data. Note. SV = source of variation; SS = sum of squares; DF = degrees of freedom; MS = mean square; F tab = tabulated F; F calc = calculated F; coefficient of determination R 2 = 0.9452 (explained variance = 94.52%).
Since the model was significant, the effects of the independent variables on the response were analyzed (Table  5).

Proposal for a Second-Degree Polynomial Model
Based on the regression data of the experiment (Table 6), it was possible to develop a mathematical model for the Ethanol variable, through which it is possible to establish the optimal conditions for design, determining the critical or stationary point of the model. Regarding the optimization of the yeast concentration, Souza et al. (2014), working with the S. cerevisiae PE-2 in a Box-Behnken Design (BBD), found the initial yeast concentration of 10 g L -1 to be optimal, reaching the maximum ethanol production of 0.78 ºGL. For Shokrkar, Ebrahimi, and Zamani (2017) the optimization of the yeast concentration is one of the best-known techniques to improve the efficiency of the fermentation process.
The temperature found in this work showed to have a strong relationship with yeast growth. Similarly, Lin et al. (2012), analyzing the factors that affect alcoholic fermentation, found that maximum rates of ethanol production were obtained at 30-45 ºC, achieving higher ethanol production at 30 ºC with different glucose concentrations. According to Mukherjee et al. (2017) and Camargo et al. (2018), S. cerevisiae CAT-1 is able to tolerate high temperatures (~40 ºC) and showed capacity to fermented sugars such as glucose, fructose, sucrose, mannose, maltose, raffinose and galactose, while other sugars such as xylose, cellobiose, mannitol and lactose cannot be fermented by this strain.
As shown in Figure 2, a rapid consumption of sucrose occurred shortly after the introduction of yeast in the fermentation medium, probably due to the action of extracellular invertase, which hydrolyzes sucrose into glucose and fructose (Margetić & Vujčić, 2017;Marques et al., 2017;Fernandes et al., 2020). Therefore, there is a rapid increase in these sugars in the same proportion as sucrose decreases.
The low sugar concentration in this study was not a limiting factor for ethanol production, since lower growth rates of S. cerevisiae occur in high concentrations of sugars (200-300 g L -1 ), as reported by Nuanpeng et al. (2016). Fast consumption of sugars probably occurred due to its low concentration in the medium, which does not require a long period of adaptation by the yeasts.

Conclusion
This study illustrates the huge potential sugary cassava has for ethanol production. The plant stores free sugars in greater quantity, as an energy source, in its roots (about 5.8% of its composition in fresh mass), which favors the direct fermentation process. We observed that the maximum concentration of ethanol produced after fermentation was 2.92% (v/v), in its original conditions of chemical properties, which means 23.1 grams of alcohol per liter of wort. In addition, we also observed that an increase in temperature and yeast concentration had a positive effect on ethanol production. The pH range studied showed that a pH slightly acidic is preferable by the yeast strain used.