Growth kinetics and purifying performance of A. Platensis in stirring-free culture based on brewery effluent's supernatant

Algal cultures in outdoor environments are generally carried out in high-rate algal ponds (HRAP), which are shallow agitated basins (Borowitzka 2005). In such ponds the shallow depth (15–30 cm) maximizes algae exposure to light, critical for the photosynthesis. Moreover, the agitation prevents algae from settling. However, besides the technicality necessary to successfully agitate the system at about 30 cm·s⁻¹ (Borowitzka 2005), operating HRAP is costly given the investment costs and operating costs related to power consumption and maintenance. These aspects make the technology unaffordable to small fish farmers in developing countries. Algal species including Athrospira platensis, commonly known as spirulina, have gas vesicles which allow them to stay afloat naturally and increase their exposure to light without requiring stirring (Tomaselli 1997). In this context, this work aims at assessing productivity kinetics, and water purifying capability of A. platensis in unstirred brewery effluent-based media in outdoor environment.

Experiments were carried out at the Wetland Research Laboratory of the University of Abomey-Calavi, Benin. In this laboratory, research themes focused on fish farming diversification and its development based on unconventional resources such as biomasses produced from recovering agri-food industrial effluents. The overarching purpose of this study is to contribute to the profitability of fish farming in developing countries through low-cost productions of proteins sources and other nutrients essential in fish farming.

The nutritives potential of the effluent supernatant in our study was determined by comparing its characteristics according to Liady et al. (2020) to those of the modified spirulina medium (Schlösser 1994, cited by Andersen et al. 2005). Characteristics of the brewery effluent supernatant and those of the Schlösser medium are presented in Table 1.

Liady et al. (2020) reported that nitrate and phosphate concentrations in brewery effluent supernatants similar to that used in this study followed a Gaussian distribution. The parametric student t-test was used to compare mean values of these elements in the supernatant and those of Schlösser's medium.

Two mother cultures left outdoors for several weeks were used to inoculate experimental cultures. The first one consists of a culture on brewery effluent's supernatant amended with sodium bicarbonate (7 g·L⁻¹) and sodium nitrate (1 g·L⁻¹). It was used for seeding brewery effluent treatments. The second one consists of a culture on the modified spirulina medium (Schlösser 1994, cited by Andersen et al. 2005).

Because of their gaseous vesicles (Tomaselli 1997), spirulina cannot be harvested by centrifugation. The addition of certain salts can maximize the harvest of spirulina after flotation of its filaments (see for example Kim et al. 2005). However, this technique was not used in this study to avoid interference with the reagents used.

In this study, spirulina were harvested through three stages: before starting the experiment, each mother culture media was transferred to a new tank; it was cleared off from its flocs by filtration through a 200 μm mesh size's sieve; and finally, the filtrate was filtrated again through a 50 μm mesh size plankton net in order to concentrate the spirulina by separating it as well as possible from the initial culture medium.

In addition, observations were systematically made under a tri-ocular microscope (BIMICRO) after weighing, to assess any potential contamination of the culture.

Various physicochemical parameters including pH, temperature and conductivity were monitored every 2 days at 13:00 using a multiprobe (Hanna HI 991301).

COD, BOD, NTK and P_total measurements were carried out in accordance with Rodier et al. (2009).

All statistical analyses were carried out using Statistica software. Growth rates were determined using linear regression by adjusting exponential phase data to a linear model without intercepts. They were then compared using a nonparametric analysis of variance (ANOVA) (i.e. the Kruskal-Wallis test). The statistical significance threshold was 5%. Effects of treatment, time, and time and treatment interaction on produced biomass were studied using a repeated-measures ANOVA after satisfaction of the sphericity hypothesis (Mauchley Sphericity Test). When these assumptions were not verified, a general linear model (GLM) was used. Multiple comparisons were made using Tukey's HSD test (α = 0.05).

The effluent was poorer in nitrate than the Schlösser medium but was richer in orthophosphates, organic nitrogen and organic phosphorus (Table 2). These organic forms, whose concentrations are very important, could be made available after mineralization.

Table 2

Comparisons of nutrient potential of the effluent's supernatant to those of Schlösser medium

Parameter .	Mean .	Standard deviation .	N .	Calculated t-value .	Reference value (Schlösser medium) .	Critical t-value .	df .	p .
(mg·L−1)	4.33	2.02	27	1,046.1	411	1.706	26	0.000
(mg·L−1)	352.13	115.30	27	11.90	88	1.706	26	0.000

Parameter .	Mean .	Standard deviation .	N .	Calculated t-value .	Reference value (Schlösser medium) .	Critical t-value .	df .	p .
(mg·L−1)	4.33	2.02	27	1,046.1	411	1.706	26	0.000
(mg·L−1)	352.13	115.30	27	11.90	88	1.706	26	0.000

View Large

Temperature evolution followed the same trend in all treatments. Average temperatures recorded were 36.17, 38.00 and 37.61 °C, respectively in the T0, T7 and T8 treatments. No significant difference (p > 0.05) was noted either between the average daily temperatures recorded over the period independently of the treatments (Table 3) or between average daily temperatures recorded in the different treatments (Table 4). Unlike temperature, pH and conductivity did not follow the same trend in all treatments. Significant differences (p < 0.05) were noted between average daily pH, average daily conductivities (Table 3), between average pH of different treatments and between average conductivities of different treatments (Table 4).

The average pH and conductivity recorded for T0, T7 and T8 were 9.76 and 20.00, 9.22 and 10.53 and 7.86 and 3.07 mS·cm⁻¹, respectively.

Microscopic observations made in relation to culture contamination showed no contamination in T0, and contamination by rotifers in T7 and T8 (lower contamination in T7 compared to T8).

Table 6 presents the calculated generation times. The generation time in the three treatments was not statistically significant (p > 0.05).

During the exponential growth phase, the average spirulina productivity was noticeably higher in treatment T7 (0.55 ± 0.05 g DW·L⁻¹·d⁻¹) than in T0 (0.22 ± 0.01 g DW·L⁻¹·d⁻¹) and T8 (0.20 ± 0.04 g DW·L⁻¹·d⁻¹). For the latter two there was no statistical difference (p > 0.05) (Table 7).

Regarding the capability to purify the brewery effluent, the concentrations of COD and total nitrogen were reduced to 32.5 and 64.91%, respectively, in T7 after 6 days into the culture.

The dilution ratio used in this work was 10% compared to the ratio (20%) used in Lu et al. (2017). This difference was justified by the fact that the supernatant used in this work was doubly loaded (23,137 ± 11,457 mg COD·L⁻¹). In Lu et al. (2017) the organic load was 10,120 ± 233 mg COD·L⁻¹.

The similarity of temperature trends in the three treatments indicated that they were subjected to similar temperature conditions. The statistically significant differences for pH and conductivity found in this study can be explained by differences in chemical conditions between treatments, particularly the differences in alkalinity.

Indeed, there seems to be a good linear correlation between sodium bicarbonate intakes and observed conductivity and pH values. High values were observed in T0 which initially experienced an intake of 13.61 g NaHCO_3·L⁻¹, followed by T7 which experienced 7 g·L⁻¹, and T8 which experienced no intake and whose initial pH was slightly acidic (6.63 ± 0.18).

The relatively high density of rotifers observed in T8 compared to T7 does not explain the significant difference noted between productivities in these two treatments. Indeed, Mitchell & Richmond (1986) have shown that rotifers rather preserve spirulina crops against contamination by unicellular algae by feeding on them. The difference in productivity found between these two treatments can be explained by alkalinity (here related to bicarbonates) and concentrations in nitrates. Bicarbonates improve the availability of mineral carbon sources and create selective conditions that prevent the proliferation of other microorganisms by raising the alkalinity of the medium. This was confirmed by the absence of rotifers in the T0 medium in our study. Nitrates provide mineral nitrogen (whose brewery effluent is poor) to satisfy the needs of spirulina. The productivity observed in T7 in this study was not significantly different from that obtained in Lu et al. (2017) in the pre-treated brewery effluent, nor from those obtained respectively by Raoof et al. (2006) cited by Lu et al. (2017) who worked on the modified Zarrouk medium (Table 8). However, it was higher than those obtained in Volkmann et al. (2008), Jung et al. (2014), Salla et al. (2016), and da Rosa et al. (2016) (all cited by Lu et al. (2017)).

The purification performance observed in this study can be explained by the same mechanisms as those observed in the last basins of extensive wastewater treatment systems such as the natural lagoon that algae colonize naturally.

In these basins where organic loads are low and mineral loads are high, algae maintain a symbiotic relationship with other microorganisms such as bacteria, consuming nutrients from the mineralization of organic matter and providing oxygen favorable to these microorganisms for their respiratory activities. The reduction of the organic load is attributable to microorganisms such as bacteria that are likely to grow in this environment, despite the adversity related to the alkalinity of the environment.

Some depollution performances by algae are given in Table 9. The difference between organic matter (COD) yield found (lower than that reported in Lu et al. 2017) could be explained by differences in accounting methods. In their method, Lu et al. (2017) considered the contribution of the preliminary stage of anaerobic digestion of the effluent. This was not the case in our study.

Reductions of pollutant loads observed in our growing conditions are similar to those reported in Raposo et al. (2010) for total nitrogen on production of Chlorella vulgaris in laboratory condition, though during a longer time than in this study (Table 9).

We investigated the growth kinetic, productivity, and purifying performance of A. platensis in an unstirred brewery effluent-based media in an outdoor environment. Results indicated that given their gas vesicles which allow them to stay afloat naturally and increase their exposure to light, A. platensis can be successfully used as a low-cost alternative of electrical stirring systems for cultivating monospecific algae. Valuing agri-food industrial effluent in certain outdoor settings contribute both to lower the production cost and depollute. Such results highlight the potential for low-cost production of certain algae in transparent and relatively high barrels, thus breaking the classic barriers traditionally considered as critical to the success of algal production.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.