Primary Sludge Fermentation in Chem-P Plants: Decoupling P-Release from C-Release

Introduction Many wastewater treatment plants rely on chemical phosphorus removal (CPR) to meet discharge limits. CPR commonly relies on iron (Fe) addition to remove phosphorus (P), and though it provides reliability, the chemical can be costly. Plants with high enough volatile fatty acid (VFA) concentrations are often able to use biological P removal (BPR), allowing for savings on chemical costs. Additional carbon can be supplied from primary sludge (PS) fermentation to drive BPR. There is also an opportunity to use PS fermentate to support nitrogen (N) removal in plants with low discharge limits, which are generally dependent on external carbon addition, to offset chemical costs. Despite the potential cost reduction, there are challenges with implementing PS fermentate as a carbon source. During fermentation, nutrients are released in addition to VFAs and COD. N concentrations are low compared to COD, and can be managed through PdNA or reoxidation after denitrification. P is released through hydrolysis similar to N, but in CPR plants there is an additional release of Fe-bound P due to Fe reduction under anaerobic conditions. This can be a significant issue for plants with low P limits, such as the Blue Plains AWWTP which has a 0.18 mg/L discharge limit. The objective of this study was to evaluate conditions for decoupling COD and P release by managing Fe reduction of Fe-P precipitation within the fermenter. Recent studies show that vivianite, an Fe-P mineral, may precipitate at a pH greater than 6. Therefore, it was hypothesized that pH adjustment may be a viable strategy to support optimizing CPR. In addition to elutriation, lime dosing was explored as a method for raising the pH due to its availability at Blue Plains. Materials and Methods Two 60-L fermentation pilot reactors were fed with Blue Plains PS at 5 g/L and controlled to a SRT target of 1-d. The system was operated under several test scenarios including a control: mechanical treatment, low and high microaeration, and low and high lime slurry addition (100 and 200 mg/L CaCO3). Fe overdosing in full-scale treatment also occurred during testing but was not controlled as a test scenario. Measurements were performed on influent and decant fermentate to calculate concentrations and yields. A full-scale gravity thickener (GT) at Blue Plains was converted to fermenter conditions by allowing an SRT between 0.2-2 days and was operated under full-scale conditions for sludge composition and solids loading. Influent and effluent sludges were analyzed, and data was categorized by TSS: low (<3.5 g/L), medium (3.5-5 g/L), and high (>5 g/L). Yields were calculated assuming 100% elutriation in the fermenter by measuring concentration profiles within the GT. Results and Discussion Full-scale fermenter data from spring and summer 2024 and 2025 shows an estimated yield of 0.0332 mg P/mg sCOD, while literature data from BPR plants show much lower P yields per mg of sCOD, from 0.0046 to 0.0022 (Figure 1a). It was hypothesized that all additional release at Blue Plains was a result of Fe dosing and thus could be decoupled from the sCOD release (Figure 1a). Separating the data into a low (<6) and high (>6) pH scenarios show that a high pH leads to ~20% less P yield per mg sCOD, at 0.0238 mg P released per mg sCOD, compared to 0.0298 mg P/mg sCOD at a low pH (Figure 1b). Several options were explored to raise fermentate pH above 6, into a range suitable for vivianite precipitation, including alkalinity dosing, Fe overdosing, and PS dilution. Alkalinity dosing and Fe overdosing In fermentation pilot control scenarios, PS fermentation reduced sludge pH from 6.28±0.12 to 5.51±0.26, which is below the range for vivianite formation, leaving Fe and P in their soluble forms. The average control fermentate OP-P was 13.50±2.82 mg/L (Figure 2a). Lower P levels were found at fermentate pH greater than 5.83, and when sufficient iron was available for P removal (Figure 2a, b). High fermentate pH was a result of either a high lime dose (200 mg/L CaCO2), or overdosing of FeCl3 in primary treatment, which led to high levels of Fe(OH)3 in the PS. The low alkalinity scenario (CaCO3 dose of 100 mg/L) had a pH of 5.71 with 12.62 mg/L OP-P in the fermentate, while the high alkalinity scenario (CaCO3 dose of 200 mg/L) had a pH of 6.36 and 2.72 mg/L OP-P. Overall, the 200 mg/L lime dose was effective at removing P, while the lower dose may not have achieved a high enough pH for efficient P removal. Periods of higher background Fe also saw effective P removal, with pH values of 6.06 and 6.01, and OP-P at 4.76 ± 3.7 and 2.65 ± 2.4 mg/L in the decant fermentate. PS dilution Within the GT, PS elutriation resulted in a lower incoming TSS, leading to lower blanket sCOD, higher average blanket pH, and lower OP-P (Figure 3). At a pH greater than 5.75, the average OP-P was below 17.5 mg/L, while operating at a lower pH led to an average OP-P of 47.6 mg/L (Figure 3a). Fe concentrations followed a similar trend to OP-P, showing that the chemical Fe-P dynamics were modified with the change in pH (Figure 3b). Decoupling COD and P release An estimate of 0.0046 mg P/mg sCOD (Figure 1) was used to calculate P released solely from hydrolysis within the GT operating at an SRT of 0.8 to 1.5 days. Additional P release, assumed to be a result of Fe reduction, was significantly different between scenarios with a pH below and above 6 (T-test, p = 0.1126) (Figure 4). This indicates that raising the pH above 6 is a viable solution for reducing excess P release. With elutriation, the GT achieved a fermentate OP-P release of less than 17.5 mg/L, and the projected mainstream P was less than 0.4 mg/L. The lower P release at higher pH showed decoupling of OP-P and sCOD, however the projected mainstream P was not yet low enough to meet plant discharge limits of 0.18 mg P/L. This demonstrates that dilution is a low-cost practical solution to achieve target pH conditions without losing sCOD loads, however additional research is needed to understand changes in buffer capacity over different seasons and sludge characteristics. Future work includes testing in wintertime to investigate seasonal effects. Conclusions The results of the full-scale GT and fermentation pilot studies demonstrate that pH management is a viable strategy for decoupling COD and P release at CPR plants. A higher pH can be achieved through alkalinity addition, Fe overdosing, or sludge dilution. These strategies maintain good sCOD yields while decreasing P in the mainstream.