Wastewater characterization

The wastewater was characterized for the main water quality parameters as shown in Table 1. The values were determined in the Environmental Biotechnology and Bioremediation Laboratory and the Environmental Engineering Laboratory, both at UNALM.

Microbiological analysis of B-lac

The lactic acid microbial consortium, or B-lac, was characterized for key microbiological parameters in the Marino Tabusso Laboratory, at UNALM (Table 2). The aim was to determine if there is a presence of pathogens that would contaminate the sample or others that could interfere in the process of lactic acid fermentation. The microbiological analysis was carried out according to the procedures described by the Salfinger & ^{Tortorello (2015}).

Sampling and conditioning

Wastewater sampling was done on two occasions. The first consisted of the sampling of 10 l, taken from the wastewater storage box from the sanitary facilities in the eco-touristic circuit (used for the first stage of experiments and the wastewater characterization). In the second, only 5 l of wastewater were sampled which were used to evaluate the three best treatments determined in the first stage of the experiments. Samples were transported to the Environmental Biotechnology and Bioremediation Laboratory at UNALM, where they were stored at 4 °C until used for the experiments.

Experimental design and preparation of treatments

Two factors were established: the percentage (%) by volume of molasses and the percentage (%) by volume of the B-Lac solution, each in 6 levels. These were evaluated in a Completely Randomized Design (CRD) in a 6x6 factorial arrangement with three repetitions. 36 treatments were prepared in triplicate using combinations of 6 concentrations of B-Lac and 6 of molasses in proportions of 0 (0 ml), 1 (2 ml), 3 (6 ml), 5 (10 ml), 7 (14 ml), and 10% (20 ml), as shown in Table 3. These percentages were measured in respect of a wastewater volume of 200 ml (fixed volume) used for each treatment. The labeled containers were hermetically sealed to provide anaerobic conditions and were exposed to environmental conditions in the laboratory (average temperature of 34°C).

For treatments, a lactic acid bacterial consortium (B-Lac) composed of the genus Lactobacillus, Streptococcus, and Bifidobacterium, which was prepared in the Environmental Biotechnology and Bioremediation Laboratory, was used. In addition, sugar cane molasses was the main source of soluble carbohydrates, nutrients, and growth factor (^{Peralta et al., 2016}), which was obtained from the stable at UNALM.

The variation of pH as a function of time was used as the response variable, as shown in Figure 1. The pH values on the third day of measurements were analyzed with an Analysis of Variance (ANOVA) followed by Tukey´s range test, using a level of significance of 95% (p < 0.05).

Figure 1 Variation of pH in the 36 treatments as a time function (first stage). 

Evaluation of pH variation

The experiments consisted of two stages. First, the pH values were evaluated in the 36 treatments during 9 days using the potentiometric method according to ^{Londoño et al. (2010}). Subsequently, the three best treatments that met the criteria described by ^{Peralta et al. (2016}) were selected; a good treatment has to meet the following: pH ≤ 4.5, rapid pH decrease to around 4, absence of bad odor, and no formation of mold or yeast layers.

The second stage of evaluation aimed to determine the best treatment out of the three selected during the first stage. Three repetitions were considered for each treatment and the decrease in pH was evaluated during 4 days.

Microbiologic analysis of treated wastewater

After the second stage of evaluation, the best treatment was chosen. The effectiveness of each treatment was evaluated by determining the removal of fecal coliform. The methodology used was the most probable number (MPN), according to the recommendations established in the Standard Methods (^{APHA, 2017}).

First stage of evaluation of 36 treatments

Before preparing the treatments, the initial pH of the wastewater was 8.11, while the molasses had a pH of 5.05 and the B-Lac solution registered a value of 3.86. Due to this, when preparing the 36 treatments, most of them stabilized at a low pH after 3 minutes. In addition, the average temperature at which the fermentation process was evaluated was 33°C. According to ^{Delgadillo et al. (2010}), this value is within the optimum range for the development of microorganisms, so the speed at which the pH decreased in the treatments was favored. The lactic acid microorganisms improve the organoleptic properties of the substrate on which they act (Alejo et al. cited by ^{Quiñones et al., 2016}). For this reason, most of the treatments applied presented a characteristic aroma similar to that of corn liquor as fermentation is the process that drives both the preparation of this drink as well as the experiments itself.

Figure 1 shows the behavior of the pH in each of the 36 treatments during the first evaluation stage (9 days). This graph serves as a general overview of the pH trends and helps to identify the treatments with desirable behavior. It is observed that most of the treatments start with a pH in the range of [4 - 7.8]; from day 8 the pH remained constant in most of the treatments, except for T1, T7, T13, and T19. In these, no molasses was added, which is the main carbon source for lactic acid bacteria, providing them with energy to efficiently perform their fermenting activity. A high volume of B-Lac without molasses causes a process of starvation, which leads to the consumption of the nutrients contained in the B-Lac itself. The opposite effect is observed in treatments with high concentrations of molasses, where the bacteria reach the stationary phase faster, before consuming the entire substrate, which induces a stoppage in the production of biomass, due to the saturation state of the substrate compared to the concentration of microorganisms (^{Ossa et al., 2010}). Another effect that was detected in the treatments that do not have a source of molasses (T1, T7, T13, and T19) is that the pH increases showing chaotic tendencies. This pH increase causes a decrease in the growth of bacteria in a culture medium (^{Ortiz et al., 2008}). A trend of fast pH decrease to a value close to 4 is distinguished in most of the treatments; this pH decrease was slower after the fourth day of evaluation.

Due to the large number of treatments that were shown in Figure 1, only those that met the criteria described by ^{Peralta et al. (2016}) were analyzed in detail in Figure 2. Therefore, 10 treatments that showed an accelerated pH decrease were selected, reaching values below 4.5 (in the range of 4.0 - 4.2) on the third day, absence of bad odor, and no formation of mold or yeast layers. Moreover, treatments that reached low pH values on the third day but did not show a significant variation were discarded (T31 and T32).

Figure 2 presents three bars (black, gray, and dark gray) for each treatment, which represent the pH reached on days 0, 3, and 4. Also, an orange bar is shown for each treatment, which indicates the variation of pH (delta) between day 3 (D3) and day 0 (D0) of evaluation (absolute value). The analysis of Figure 2 aims to select the best treatments for the second stage of evaluation. Thus, on day 3 it is observed that two treatments have the lowest pH values: T16 (4.08) and T33 (4.06), which were selected for the second stage of evaluation. Besides, two other treatments with low pH values were found on the third day: T22 (4.13) and T29 (4.11). However, only T22 was selected because its composition has a lower requirement of molasses (10 ml) compared to that of T29 (14 ml). This criterion is based on favoring lower-cost options for water treatment since greater quantities of molasses will translate into a higher cost of the process, especially if we aim to provide a big-scale and low-cost viable treatment option.

Figure 2 Treatments that reached pH values between 4 - 4.5 on the third day and showed significant pH variation during the first stage. 

The pH decrease is attributed to the homofermentative metabolism possessed by lactic acid bacteria, in which mainly lactic acid is produced (^{Serna & Rodriguez, 2005}). When lactic acid bacteria generate this acid, the pH of the fermented material drops to a level that inhibits the presence of putrefying bacteria (^{Garcés et al., 2004}). This is expressed in the decrease of pH, an indicator that guarantees the absence of unwanted pathogens (^{Carrasco et al., 2002}). It is relevant to mention that heterofermentative bacteria may have also been present in the mix, as they could have been living in the wastewater or the molasses; however, it is safe to assume that the medium was dominated by homofermentative bacteria as the inoculum (B-lac) was prepared to contain exclusively this type of lactic acid bacteria and there were no signs of significant gas formation, which is a clear indicator of the heterofermentative metabolism.

Second stage of evaluation of the 3 best treatments

The second stage of experiments consisted of the evaluation of pH decrease applying treatments T16, T22, and T33 (in triplicate) during 4 days. Figure 3 shows that T33 has a slower pH decrease compared to that of T22 and T16 between the first and second day, therefore it is discarded. Between T16 and T22 no significant differences were noted in their fermenting activity, which leads to an analysis of the composition of each treatment. Both had the same quantity of molasses (10 ml), but T16 had a lower amount of B-Lac solution (6 ml) compared to T22 (10 ml), which is why T16 was chosen as the best treatment since it is preferable that the food source (molasses) exists in greater proportion than lactic acid bacteria to avoid problems of scarcity.

Figure 3 Variation of pH in the three best treatments (Second stage). 

Microbiologic evaluation of the three best treatments

After 3 days of fermentation, the pathogen removal efficiency of T16 was evaluated by the most probable number method (MPN) to determine the presence of fecal coliform. This group of bacteria is used as a general indicator of water quality (^{García & Iannacone, 2014}). Table 4 shows the removal of the fecal coliform bacteria in treatment T16. The enhancement of the coliform removal in T16 occurs because the efficient microorganisms inoculum have high populations of lactic acid bacteria (Lactobacillus and Pedicoccus), which produce lactic acid and other antimicrobial products as a consequence of the metabolism of carbohydrates (^{Tannock, 2004}). One of the characteristics of this treatment is that it had enough substrate, which prevents the competition for nutrients amongst lactic acid bacteria and other present microorganisms. ^{Corpas & Herrera (2012}) observed that competition during the activation of their microbial mix, coupled with other factors, hindered its ability to deploy its arsenal for the inhibition of coliform populations. As shown in Table 2, the microbiological analysis of the B-Lac solution used indicates an absence of fecal and total coliforms (MPN.ml^-1< 3), from which one can infer that the complete inactivation of the fecal coliforms in the wastewater treated with T16 was caused by the fermenting activity of Lactobacillussp. in the lactic acid microbial consortium, which is the predominant microorganism present. Another characteristic that would have favored the antagonistic behavior against lactic acid pathogens was the production of antimicrobial peptides and low molecular weight compounds such as bacteriocin class I (^{Kelly, 1998}).

Interaction of the B-lac and the molasses

Figure 4 presents the interaction effects of the B-lac (B0 to B10) and the molasses (M0 to M10) on the pH values on the third day of evaluation. This graph suggests that an interaction effect between the two factors studied exists, as the behavior of the pH cannot be explained by each variable alone. The ANOVA confirmed that there is a significant effect of the interaction, as well as that coming from the different molasses and B-lac concentrations (Table 5); the effect of the molasses levels seems to be the strongest. It is to note that although some treatments with a level of 0% of molasses or B-lac reach a pH of around 4.0 on the third day, they are significantly different from the mean of the other levels for each factor, as shown in Table 6 and Table 7. This supports the decision to discard such treatments. As for the three selected treatments (T16, T22, and T33), the pH values on the third day do not show a significant difference between them, according to Tukey’s test (p < 0.05).

Figure 4 Interaction of molasses (M) and B-lac (B) concentrations on pH on the third day of evaluation.