Reaction kinetics of lactic acid fermentation from bitter cassava (Manihot glaziovii) starch by Lactobacillus casei

One of the utilizations of bitter cassava is modified cassava flour (Mocaf) production using the fermentation process by Lactobacillus casei. TheMocaf has potential as the future of food security products. It has a characteristic property similar to wheat flour. Lactic acid was also produced as a by‐product during fermentation. After 40 h of fermentation, the proximate composition content of Mocaf was lactic acid content of 0.000928 g/L, hydrogen cyanide levels of 0.02 ppm, starch content of 59.13%, amylose content of 12.98% and amylopectin content of 46.15%. In the scaling‐up process from a laboratory scale to a pilot and industrial scale, modeling is needed. There are five equation models used to describe the kinetic reactions of lactic acid from bitter cassava starch: Monod, Moser, Powell, Blackman, and Product Inhibitor. Each parameter was being searched by a fitting curve using sigmaplot 12.0. The best result in terms of the highest R2 (0.65913) was obtained in the Powell equation with the value of μmax of 1.668/h, Ks of 123.4 g/L, and maintenance rate (m) of 4.672. The kinetic data obtained can be used to design biochemical reactors for industrial scaleMocaf flour production.


Introduction
Modified cassava flour (Mocaf) is a cassava flour prod uct, in which cassava is processed using the principle of lactic acid fermentation to improve the nutritional content (Gunawan et al. 2019). Abundant raw materials, inexpen sive, easily obtained, and the processing that does not re quire high technology makes Mocaf the best alternative for wheat flour substitutes . Be sides, the production of Mocaf with the principle of lac tic acid fermentation using lactic acid bacteria, produces a byproduct in the form of lactic acid (Istianah et al. 2018).
In previous studies, the production of modified cas sava flour without additional nutrients at an appropriate microorganism (Lactobacillus plantarum, Saccharomyces cereviseae, and Rhizopus oryzae) obtained the best results of Mocaf flour in fermentation using L. plantarum (Gu nawan et al. 2015). The use of L. plantarum bacterial culture in cassava (Manihot esculenta; "singkong" in In donesia) (Gunawan et al. 2015), sorghum (Sorghum bi color L. Moench; "sorgum" in Indonesia) (Istianah et al. 2018), sago (Metroxylon sago; "sagu" in Indonesia) (Gu nawan et al. 2018), and yam (Dioscorea hispida Dennst; "gadung" in Indonesia) (Gunawan et al. 2019) fermen tations has been widely used in previous studies, so re searchers are interested in using other bacterial starters (such as Lactobacillus casei) that are easily obtained and easily adaptable. L. casei is a lactic acidproducing bacte ria, obtained by glucose fermentation and the production of homofermentative lactate produces pure lactate nearly 85%, and also able to ferment ribose into acetic and lac tic acid (Farinde et al. 2010). Moreover, L. plantarum and L. casei have differences in their growth rates (growth ve locity constants) at the basal media. The growth rate con stant of L. plantarum and L. casei was 0.13 cell/h and 0.16 cell/h, respectively (Zacharof et al. 2009).
To understand the process of lactic acid fermentation by L. casei, a fermentation kinetics study is needed which describes cell growth and product formation by microbes. The kinetic model is a useful parameter in the design and control of biotechnology processes to increase knowledge about microbial growth behavior using accurate mathe matical models in detailed repeated experiments. Mahanta et al. (2014) have researched the microbial growth kinet ics of Escherichia coli on glucose growth media. Previ ously, Rezvani et al. (2017) were also using the Contois and Exponential kinetic models. They reported that Con tois kinetic model is a suitable model to describe the kinet ics of five species of Lactobacillus. Therefore, the objec tive of this study was to study the reaction kinetics of lactic acid fermentation from bitter cassava (Manihot glaziovii)

Material
Bitter cassava (M. glaziovii) was obtained from Sragen Re gency, Central Java, Indonesia. It has an average planting age of more than one year with an average tuber diame ter reaching 10 cm. L. casei bacteria obtained from the Microbiology Laboratory of the Department of Biology, Airlangga University, Surabaya. All solvents and reagents were purchased from commercial sources.

Starter Preparation
In this study, the starter used was 5.2 g of MRS dissolved in 100 mL distilled water. A full loop of L. casei bacteria was added to the Erlenmeyer flask containing the media. Furthermore, the starter was incubated in a shaker incuba tor at a speed of 150 rpm at 37°C for 24 h.

Pre-Treatment
Pretreatment was done by cutting 130 g of cassava then cut into slices chips with a thickness of 0.10.5 cm. This cutting process aims to expand the contact surface between bitter cassava and water. Cassava was washed 3 times then was soaked with the weight ratio of cassava to soaking water volume of 1:10 (cassava: water) for 90 min to re duce HCN levels at room temperature. This method was adopted from Nebiyu and Getachew (2011) by modifying the fermentation time.

Fermentation
In this study, the fermentation method used was the sub merged fermentation. One hundred and thirthy grams of cassava that had been pretreated was put into an Erlen meyer flask containing 260 mL of distilled water. Then, a starter was added as much as 28 mL (10% of the total work volume) (Panesar et al. 2010). To obtain anaerobic conditions, the surface of the Erlenmeyer was closed using cotton and aluminum foil. Fermentation was done at 37°C for 40 h in a shaker incubator. After 40 h, the fermentation was stopped then the cassava was separated from the fer mentation liquid for further drying and mashing according to flour standards (80 mesh).

Proximate Analysis
The proximate analysis consists of protein, starch, amy lose, and amylopectin contents. Protein content was ana lyzed by using the AOAC (2005) procedure, while starch, amylose, and amylopectin contents were determined by using the Indonesian National Standard (SNI) procedure (SNI 2011).

Hydrogen Cyanide (HCN) Analysis
HCN content was analyzed based on SNI (2011). The step was began with distillation process of 20 g samples during 24 h. Then, distillate was titrated by solution of AgNO 3 0.02 M. Calculation of HCN levels was using the follow ing equation: where V1 , V2 , V3 , N, and W is the blank titra tion reading, sample titration reading, distillate volume, AgNO 3 normality, and sample weight, respectively. The value of 27 is the molecular weight of cyanide acid.

Lactic Acid Analysis
Product measurements of lactic acid concentration were determined using the Total Titratable Acidity (TTA) method in GEA (2006). A total of 2 mL of sample was added with three drops of PP indicator then titrated using 0.2 N NaOH solution until it turned pink. Lactic acid con tent was calculated by the following formula: where N, V1, and V2 is NaOH normality, NaOH vol ume, and sample volume, respectively. Whereas 0.090 is the milliequivalent of lactic acid.

pH Analysis
Two mililiter samples were taken from an Erlenmeyer flask using a pipette, then the pH was analyzed using a dig ital pH meter electrode (Eutech, Singapore). Before using a digital pH meter, it was calibrated first.

Analysis of Microbial Number
The number of microbes was analyzed by using the method of counting chamber hemocytometer. Briefly, 1 mL of the fermentation solution was diluted with a 1,000 dilution factor. Then, it was dropped on a hemocytometer and was covered with deck glass. The number of bacteria calculation was done under a microscope (Novel, China).

Cell growth
Growth is one of the important characteristics of microor ganisms in fermentation. The increased number of L. ca sei bacteria can be expressed as cell growth. Growth dy namics are displayed in a curve of increasing cell number with incubation time which can describe the phases of the bacterial growth cycle. The first bacterial growth cycle is the adaptation phase (lag), the phase of bacteria adjusting to the new environment, in this phase the cell increase in size but not in the number. The second is the exponen tial phase (log), the phase that bacterial propagation takes place rapidly, the cell is divide and the number increases logarithmically according to time increase. Then, the sta tionary phase is a balanced state between the rate of growth with the rate of death and finally the phase of death where the rate of bacterial death exceeds the rate of propagation (Gunawan et al. 2015).
The bacterial growth phase was used to determine the incubation time during the making of a starter in the lactic acid fermentation process. Bacterial growth was observed by analyzing the calculation of the cell number per hour during the 48 h incubation time by the counting chamber method. This tool can be used to count the cell number per unit volume.
The growth of L. casei for 48 h incubation at 37°C in MRS media is shown in Figure 1. From the results of the observations of the number of bacteria, the adaptation phase (lag) has occurred at the 0 h to the 4 th h, the exponen tial phase (log) has occurred at the 5 th to the 24 th h, and the stationary phase has occurred at the 25 th to the 48 th h. The growth of lactic acid bacteria was strongly influenced by fermentation conditions, such as temperature, pH, medium components, and oxygen. Among all factors, the type of growth media played an important role in bacteria viabil ity. This media difference was also the cause of the dif ference in the growth phase curve in the starter with MRS media and the growth curve resulting from the fermenta tion process using bitter cassava media. From Figure 1, it can be seen that L. casei had a shorter lag phase, from 0 to 2 h, log phase from 2 to 24 h, and stationary phase from 24 to 40 h.
In this study, the growth phase of L. casei was differ ent from previous studies by Suharyono et al. (2012) who cultured L. casei on skim milk media the log phase oc curred on the 8 th h to the 16 th h. Whereas Rezvani et al. (2017) reported that the log phase occurred on the 6 th h to the 24th h with whey milk fermentation media. The differ ence growth phase can be caused by differences in the me dia used in bacterial culture. However, research by Pane sar et al. (2010) that also using MRS media the log phase occurred on the 8 th h to the 24 th h. This difference was due to the greater concentration of media used which was 0.2 g/mL while in this study the concentration of MRS media used was 0.052 g/mL. This is consistent with the theory of microorganisms that are inoculated from a medium with low concentration to a medium with a higher concentra tion will require a longer lag phase since cells must pro duce enzymes for use. From the known growth phase, then it was used to de termine the incubation time for the fermentation starter. The time chosen was at the peak of the log phase, on the 24 th h where the bacteria have the biggest log which is 9.4 with the number of bacteria reaching 2.48×10 9 cells/mL so that when regenerated on the new fermentation media, the number of cells produced was increased.

Effect of pretreatment and fermentation time on HCN concentrations
The purpose of pretreatment is to reduce the concentration of HCN from bitter cassava so it can help the work of the  bacterium L. casei in the next process (fermentation). Data analysis of HCN concentration in bitter cassava after pre treatment can be seen in Figure 2. It can be seen that there was a decrease in the concentration of HCN in bitter cas sava by 63.52%, from 202.12 ppm to 73.74 ppm. This was consistent with research by Nebiyu and Getachew (2011) which showed that there was a decrease in HCN concen tration after soaking using water for 24 h by 90.01%, from 108.37 ppm to 10.83 ppm. The decrease in HCN concen tration is due to the nature of HCN which is easily soluble in water. This lower result is due to the shorter soaking time, which is 1.5 h compared to the study conducted by Nebiyu and Getachew (2011) where the soaking was car ried out for 24 h. In the soaking process, the linamarin compound is hy drolyzed to form cyanide acid which is easily soluble in water. The difference in the concentration of the solution inside the bitter cassava cell with the solution outside the cell allows osmosis during the soaking process. In this case, the concentration of the solution outside the cell was smaller than inside the cell (hypotonic) so that water en tered the cell and caused the cell to expand where the hy drolyzed linamarin formed cyanide acid which is easily soluble in water and volatile so the linamarin levels can be lowered through the soaking process. Therefore, it is necessary to soak the bitter cassava first to reduce HCN concentrations to ease the next process.
In Figure 3, it can be seen that HCN concentration    fermentation process by microorganisms can convert glu cose into organic acids, causing pH to decrease to ± 4.2.
On the other hand, optimum laminarase enzyme activity at pH 6.0. This low pH condition can reduce the activity of linamarase enzyme to decrease linamarin which will turn into cyanide acid.

Effect of fermentation time on starch content
Starch is microscopic granules found in roots, tubers, and seeds of plants. Starch consists of two separable fractions, the dissolved fraction is called amylose and the insolu ble fraction is called amylopectin. Starch usually contains about 2030% amylose and 7080% amylopectin, but the amylose content can range from <1% in waxy starch and> 70% in certain high amylose starch (Martens et al. 2018).
The effect of fermentation time on starch content is shown in Figure 4. It can be seen that the starch content of modified cassava flour has decreased. Starting with an ini tial starch content of 88.50%, the starch content decreased by 5.57%, 15.08%, 29.74%, 30.56%, and 33.19% after fer mentation time of 8, 16, 24, 32, and 40 h, respectively. It was found that the longer the fermentation time, the lower the starch content was obtained. This is consistent with the research by Gunawan et al. (2015), after 120 h of fermen tation, the starch content in cassava decreased. The lowest starch content was obtained after fermenting cassava us ing L. plantarum (55.40%), S. cerevisiae (71.03%), and R. oryzae (48.20%). Decreasing starch content is caused by the use of organic materials to fulfill the energy needed for the growth of microorganisms. It is known that during the fermentation process, starch is hydrolyzed into sim pler sugars from oligosaccharides and maltose to glucose. Furthermore, glucose is converted to lactic acid (Gunawan et al. 2015).

Effects of fermentation time on amylose and amylopectin concentrations
Starch usually contains about 2030% amylose and amy lopectin by 7080%. Amylose is a linear polymer with α(1,4) Dglucopyranose bonds. Amylose molecules have more than 1000 glucose units. The chain of the Dglucose unit with αglycosidic tends to form a helix structure (Solomons and Craig 2011). While amylopectin has a larger structure than amylose, it consists of about 106 glu cose units per molecule and forms a complex structure. Amylopectin consists of a linear chain of glucose units connected by α1,4 glycosidic bonds and is very branched in the α1 position, 6 (AlcázarAlay and Meireles 2015). Based on Figure 5, it can be seen that when the variable time increased, the amylose concentration decreased. The results obtained with bacterial concentrations of 5.5×10 7 cells/mL L. casei. Starting with an initial amylose concen tration of 28.82%, the amylose concentration decreased by 10.79%, 26.75%, 50.28%, 50.87%, and 54.96% after fer mentation time of 8, 16, 24, 32, and 40 h, respectively. It also can be seen that when the time variable increases, the amylopectin concentration decreases. The results ob tained with a bacterial concentration of 5.5×10 7 cells/mL L. casei. Starting with an initial amylopectin concentra tion of 59.68%, the amylopectin concentration decreased by 3. 05%, 9.45%, 19.82%, 20.76%, and 22.67% after fer mentation time of 8, 16, 24, 32, and 40 h, respectively. From these data, it shows that the longer the fermen tation time, both the amylose concentration and the amy lopectin concentration decreased. In a study by Setiarto and Nunuk (2017) fermentation using L. plantarum with a concentration of 10 8 CFU/mL for 24 h showed amylose concentration decreased from 21.57% to 16.29% (a de crease of 24.48%) whereas for amylopectin concentration decreased from 83.19% to 73.80% (a decrease of 11.29%).

Analysis of changes in lactic acid levels with Total Titratable Acidity (TTA)
Lactic acid is a carboxylic acid that has the isomeric form of L(+) or D() lactic acid. Lactic acid has a molecular for mula CH3CHOHCOOH. There are two ways to produce lactic acid, either by chemical synthesis or by fermenta tion. The fermentation process produces specific lactic acid: L(+) lactic acid or D() lactic acid while the chemical synthesis process produces lactic acid which is a mixture of two isomers (Narayanan et al. 2004). L(+) lactic acid is the isomer chosen for the food and pharmaceutical indus tries because the human body only produces the enzyme Llactate dehydrogenase. For the food and beverage industries, lactic acid lev els of 5080% are usually required, whereas for the phar maceutical industry required higher levels of 8590%. L. casei is a lactic acidproducing bacteria, obtained by glu cose fermentation and homofermentative lactate forma tion forming pure lactate nearly 85%, this bacterium is also able to ferment ribose into acetic acid and lactic acid (Farinde et al. 2010). Based on research by Mirdamadi et al. (2002) among the strains of lactobacilli, L. casei (ca sei PTCC 1608) produced high concentrations of L (+) lac tic acid with a purity of 98%. To analyze the amount of lac tic acid produced from the fermentation process the Total Titratable Acidity (TTA) method is used. This method has standard procedures from GEA (2006).

Effect of microbial growth on lactic acid prod ucts
Based on Figure 4, the production of lactic acid from bitter cassava fermentation (M. glaziovii) was very volatile but tends to increase with increasing numbers of bacteria. This was consistent with the research by Alvarez et al. (2010), that the concentration of lactic acid increases with the in crease in the number of microbes and a decrease in glucose levels as a substrate. At the end of the fermentation, the lactic acid concentration reached >50 g/L with a yield of 1.8 g of lactic acid/g of biomass. While the peak produc tion of lactic acid in this study occurred at the 38th h with a concentration of 0.000984 g/L on the number of bacteria of 5.675×10 8 . The type and condition of the media determine the type and ability of lactic acid bacteria to ferment starch as a sub strate for growth. For the batch fermentation lactic acid bacteria, a previous study by Bai et al. (2003) compar ing the yield of lactic acid produced from different car bon sources such as glucose, lactose, and xylose. The best results were obtained glucose as a substrate with a concen tration of 150.2 g/L with lactic acid produced at 1.34 g/L. From five types of microbes that were compared, Lacto bacillus bulgaricus was the microbe that produced the best lactic acid yield of 0.602 (Rezvani et al. 2017). So it can be concluded that the differences in substrate and type of microbes greatly affect the lactic acid produced in the fer mentation process.
Lactic acid bacteria require substrates with high nitro gen content. Sources of nitrogen needed in the fermen tation media can be supplied with additional yeast extract, soy flour, tryptone, and peptone. Whereas in the process of bitter cassava fermentation by L. casei there was no treat ment to add a source of nitrogen during the fermentation, as a result, the bacteria lacked the element nitrogen so that the production of lactic acid was very small. The results of research by Taleghani et al. (2016) stated that the produc tion of lactic acid increased with increasing concentrations of added yeast extract. Optimum results are obtained by adding 1% yeast extract with a yield of 29.5 g/L lactic acid and a yield of 79.5%.

Effect of lactic acid products on changes in pH
Changes in lactic acid production are associated with changes in the level of acidity of the media (pH). The pH measurement was carried out simultaneously with the measurement of the amount of lactic acid using a pH me ter. Based on Figure 5, the fermentation of bitter cassava (M. glaziovii) by L. casei caused a decrease in pH and re sults in a pattern of lactic acid production which fluctuated but tended to increase. The measured pH value showed a decrease in following the rate of increase in the amount of lactic acid. An increase in lactic acid in low concentra tions can affect the rate of dissociation of H + ions so that it results in changes in media pH.
According to Krischke et al. (1991), the optimum pH for producing lactic acid using the L. casei strain was in the pH range of 6.06.5, whereas in this study the measured pH was in the range of 5.694.73. This is thought to be one of the factors affecting the low lactic acid formed. Initially, an increase in lactic acid is followed by a decrease in pH, but in the log phase, an increase in total lactic acid is not always followed by a decrease in pH, during this phase the total measured lactic acid is very volatile. The rate of decrease in pH continues until it reaches a pH of 4.73 at the end of the incubation period (40 th h), as well as the rate of increase in lactic acid, continues to occur until an insignificant decrease at the 40 th h.

Analysis of the growth kinetics of L. casei
The nonlinear kinetic models of Powell, Moser, Black man, Monod, and Product Inhibitor were fitting with the correlation between the growth rate of L. casei bacteria and the concentration of the substrate as shown in equa tion 3, 4, 5, 6, and 7, respectively. The parameters of each model were estimated by the curve fitting method using the Sigmaplot 12.0 software. They are specific growth rate (h 1 ), µ m specific growth rate maximum (h 1 ), Ks sat urated substrate constant (g/L), S substrate concentration (g/L), m maintenance rate, Cp product concentration (g/L), Cpm Maximum product concentration (g/L), and n num ber of cell.
The model commonly used to describe the kinetics of mi crobial growth is the Monod equation, but in this study, the Monod model provided a result that was less appropri ate as seen from R 2 which was only 0.1638. A study by Istianah and Gunawan (2017), the Monod equation model was also less suitable for describing the kinetics of lactic acid fermentation from sorghum flour using L. plantarum, Baker's yeast and a mixture of both with R 2 of 0.6073, 0.5638 and 0.0804, respectively. In this study, the best results were obtained in the Powell equation model where R 2 had the highest value of 0.65913 with a maximum specific growth rate (µmax) of 1.668/h, saturated substrate constant (Ks) of 123.4 g/L and maintenance rate (m) of 4.672/h as shown in Table  1. The growth of L. casei produced metabolic products in the form of lactic acid which can be accumulated in the media. This lactic acid product can be an inhibitor in bac terial growth. Therefore, the effect of lactic acid concen tration on bacterial growth was also studied in this study. In the same way as the various models above, the highest R 2 results in the number of cell (n) 1.5, which was equal to 0.2204. This means that formed lactic acid products can affect the rate of bacterial growth. This was also con veyed in the study of Alvarez et al. (2010), the strong in hibitory effect of lactic acid on the growth rate of biomass was characterized and explained by simple kinetic models, where specific growth rates exponentially decrease when lactic acid accumulates. The product inhibition (λ) con stant in fermentation using L. casei was 0.34 L/g of lactic acid. The low R 2 results because the substrate concentra tion (S) data was taken from the substrate reaction calcu lation data from time to time, so S has a toosmall interval. This results in the calculation of kinetics having a large er ror. To produce high R 2 in kinetics calculations, the vari ability substrate concentration is needed.

Conclusions
Fermentation of bitter cassava (M. glaziovii) using L. ca sei can reduce cyanide acid content, starch levels and also produce lactic acid. Therefore, modified bitter cas sava flour can be used as an alternative flour to substitute wheat flour. After 40 h of fermentation, HCN decreased to 0.02 ppm, lactic acid increased to 0.000928 g/L, while starch decreased to 59.13%, as well as amylose and amy lopectin, which decreased respectively, reaching 12.98% and 46.15%. The best kinetic model that can describe the growth of L. casei in terms of the highest R 2 (0.65913) ob tained from the Powell equation. In this study, the effect of the product produced on the growth kinetics of L. casei by modifying the Monod equation to produce the best R 2 (0.2204) at an n value of 1.5 was also studied. The high error in this study was because the S interval is too small, therefore it is necessary to do a variable on the initial sub strate concentration..