Phosphorus recovery from anaerobic digester supernatant by struvite crystallization: model-based evaluation of a fluidized bed reactor

Elucidating the impacts of initial supersaturation and seed crystal loading on struvite precipitation kinetics, fines production, and crystal growth

To reduce intra-plant nutrient cycling, and recover phosphorus (P) fertilizers from nutrient-rich sidestreams, wastewater utilities increasingly elect to employ struvite precipitation processes without a clear understanding of the inherent tradeoffs associated with specific design and operating decisions. Specifically, the impact of reactor conditions on struvite crystallization rate, and distribution between formation of fines particles and secondary growth onto large diameter seed crystals represent critical knowledge gaps limiting the predictive capabilities of existing process models. In this work, the relative impacts of initial supersaturation (S_i), and seed loading, on P removal kinetics, and struvite solids distribution were investigated. In experiments conducted at different levels of initial supersaturation (1.7-2.4) and seed loading (0-25 g L^-1), struvite fines represented the majority of phosphate solids formed in 10 of 12 conditions. While total P removal was dependent on S_i, and primarily attributed to formation of fines, the concentration of struvite seed granules had a significant impact on the rate of P removal. Struvite seed granules increased the rate of precipitation by reducing induction time of primary nucleation of struvite fines. Secondary crystal growth represented the majority of struvite solids formed at high seed loading and low S_i, but presented the tradeoff of low total removal and low rate of removal. To convey the significance of these findings on process modeling, we show how a prominent kinetic model with a first-order dependency on solid struvite concentration over-predicts P removal rate when total mass is dominated by large diameter seeds (0.9 mm). This works reveals the critical role of struvite fines in P removal, and highlights the need to account for their production and kinetic importance in struvite process design and operation.

Modeling phosphorus removal and recovery from anaerobic digester supernatant through struvite crystallization in a fluidized bed reactor

The cost associated with the disposal of phosphate-rich sludge, the stringent regulations to limit phosphate discharge into aquatic environments, and resource shortages resulting from limited phosphorus rock reserves, have diverted attention to phosphorus recovery in the form of struvite (MAP: MgNH4PO4·6H2O) crystals, which can essentially be used as a slow release fertilizer. Fluidized-bed crystallization is one of the most efficient unit processes used in struvite crystallization from wastewater. In this study, a comprehensive mathematical model, incorporating solution thermodynamics, struvite precipitation kinetics and reactor hydrodynamics, was developed to illustrate phosphorus depletion through struvite crystal growth in a continuous, fluidized-bed crystallizer. A thermodynamic equilibrium model for struvite precipitation was linked to the fluidized-bed reactor model. While the equilibrium model provided information on supersaturation generation, the reactor model captured the dynamic behavior of the crystal growth processes, as well as the effect of the reactor hydrodynamics on the overall process performance. The model was then used for performance evaluation of the reactor, in terms of removal efficiencies of struvite constituent species (Mg, NH4 and PO4), and the average product crystal sizes. The model also determined the variation of species concentration of struvite within the crystal bed height. The species concentrations at two extreme ends (inlet and outlet) were used to evaluate the reactor performance. The model predictions provided a reasonably good fit with the experimental results for PO4-P, NH4-N and Mg removals. Predicated average crystal sizes also matched fairly well with the experimental observations. Therefore, this model can be used as a tool for performance evaluation and process optimization of struvite crystallization in a fluidized-bed reactor.