Indirect assessment of biomass accumulation in a wastewater-based Chlorella vulgaris photobioreactor by pH variation

Abstract

Algae bloom in coastal waters is partly supported by residual nutrients in treated wastewater (WW) released from coastally located treatment plants. In response, a Chlorella vulgaris-based photobioreactor was recently proposed for lowering nutrient levels in WW prior to release. However, the solution requires maintaining biomass accumulation to within a photobioreactor capacity for optimum operation. For high density Chlorella vulgaris suspensions, this is easily done by monitoring turbidity increase, a property directly related to biomass accumulation. For low density suspensions however, direct turbidity measurement would require a cumbersome process of concentrating large volumes of Chlorella vulgaris suspensions. Here, we demonstrate that by measuring pH of the suspensions, turbidity (T) can be estimated indirectly by the following wastewater-dependent expression: pH = aT + pH₀, hence avoiding the need to concentrate large volumes. The term pH₀ is the initial pH of the suspensions and a, a wastewater-dependent constant, can be computed independently from a = - 0.0061*pH₀ + 0.052. In the event %WW is unknown, the following wastewater-independent Gaussian expression can be used to estimate T: pH = 8.71*exp(- [(T - 250)²]/[2*1.26E05]). These three equations should offer an avenue for monitoring the turbidity of dilute Chlorella vulgaris suspensions in large, stagnant municipal Chlorella vulgaris-based wastewater treatment system via pH measurements.

Figure 1

Scatter plots of pH changes as a function of turbidity (a)–(g) and associated residual plots (h)–(n) under different WW/DH₂O ratios. pH and Residual pairs (a) and (h) are for the 0%WW/100%DH₂O ratio, (b) and (i) pair: 17%/83%, (c) and (j): 33%/67%, (d) and (k): 50%/50%, (e) and (l): 67%/33%, (f) and (m): 83%/17%, and (g) and (n): 100%WW/0%DH₂O ratio.

Figure 2

Box plots showing the distribution of (a) pH₀, the initial pH values derived from Eq. (1), and (b) the accompanying slope a. (c) The linear relationship of a to pH₀, expressed as a = − (0.0061* pH₀) + 0.052. Expected a values for the 67 and 83% WW were used in (b); the observed values for 67 and 83% WW are encircled in (c) to highlight this need.

Figure 3

HCPC results showing the coalescing of 19 × 7 multivariate pH and turbidity data into four distinct groupings. (a) Dendrogram showing hierarchical similarity of measurements as a function of the length of incubation. (b) Scatter plot of scores of the measurements in (a) along the most important HCPC dimensions: Dim and Dim 2 (explaining 92.00% and 4.63% of variance, respectively). The natural groupings of scores are labeled or color-coded as cluster 1 (black), cluster 2 (red), cluster 3 (green) and cluster 4 (blue).

Figure 4

Integrating 126 bivariate data pooled from seven WW proportions used in Fig. 1 into single planes. (a) Scatter plot showing lack of linearity between pH and turbidity overall. Data can be model by the equation pH = 8.71*exp(− [(T − 250)²]/[2*1.26E05]); 95% confidence interval limits included as blue lines. (b) Associated residual plot showing lack of randomness expected from a linear relationship between pH and turbidity at all %WW. Data corresponding to each WW proportion is color-coded.