Factors affecting chemical P removal

  • Mixing intensity – Vigorous mixing near the injection point “may” increase the chance of metal ions contact with phosphate ions before the metal ion forms hydroxides in theory. Moreover, the freshly formed metal hydroxides with reactive surface can contact with more phosphate ions before their surface ages. As shown in Fig. 1, initially identical soluble phosphorus level (1 mg P/L) decreases fastest as G value increases. However, G valve effect decreases as it increases. The phosphorus removal rate marginally increases when G value increases from 182 sec-1 to 425 sec-1 . According to Mueller (2002), the aeration basin of CAS with a depth of 4.6m runs at in the order of 80-125 sec-1. Since MBR requires more aeration due to the higher OUR than CAS at lower oxygen transfer efficiency as discussed here, G value also should be higher than that of CAS. Therefore, the reaction rate in a typical MBR might be best represented by the curve for 182 sec-1.

Graphics1c2aFig. 1. Phosphorus removal kinetics by ferric chloride at different G values.  Initial phosphorus concentration = 1 mg P/L. Initial Fe/P = 3.0 mole/mole.

  • Contact time –   Although metal hydroxide formation is kinetically favored, metal phosphates are favored thermodynamically. Therefore, the most of the metal ions in the initially formed metal hydroxides are able to combine with phosphate ions, if enough contact time and phosphate ions are given. As can be seen in Fig. 1 here, initially formed metal hydroxides continue to absorb soluble phosphate ions for more than 24 hours under the most vigorous mixing condition. Therefore, enough contact time must be secured to obtain low Metal/P ratio especially at slow mixing conditions, which means the long SRT of MBR is a significant advantage. One more thing noticeable in the figure is that collision frequency is not a bottleneck for phosphate ion absorption at the vigorous mixing condition because phosphorus absorption rate does not increase much at high G values. Perhaps the kinetics of hydroxyl group cleavage from metal ion or phosphate ion binding to metal ion is not fast enough to take advantage of the high collision frequencies.   Due to the contact time effect, jar tests tend to predict much higher Metal/P ratio to obtain certain effluent phosphorous level. In spite of the mixing performed near ideal condition in jar tests, mixing time typically ranges a few hours or less. Metal hydroxides do not seem to react to their maximum capacity in such short time, which results in high Metal/P ratio such as 5-10 to obtain a moderate soluble phosphorus concentration such as 0.1-0.2 mg/L (Song, 2008). Therefore, when jar tests are performed, mixing time must be taken into consideration in order to reflect the contact time in the full scale process.
  •    It is possible that all the factors affecting phosphorous removal kinetics in MBR are overwhelmed by the contact time effect. For example, injecting coagulant in one point without vigorous mixing is not favorable for initial metal hydroxide formation, but, if contact time is long enough, the way coagulants are added may not affect the final results significantly. However, there is no enough experimental data obtained from full-scale MBR plants to answer this question. Therefore, it is plausible to consider all the kinetics based factors as precautions until the contact time effect is cleared.
  • Metal/P ratio – As the metal/P ratio increase, which represent mole ratio of the two species, soluble phosphorous concentration decreases (Fig. 2). Due to the slow reaction between initially formed metal hydroxides and phosphate ions, soluble phosphate ion concentration does not drop to zero even at high Metal/P ratio.

Picture12d333Fig. 2.  Kinetics of phosphorus removal. Initial phosphorus concentration = 3.0 mg P/L. Fe/P is initial mole/mole basis. G=425 sec-1. (Smith, 2008)

  • Alkalinity – Alkalinity plays an important role by resisting against pH changes. If alkalinity is high enough, the local pH near the dosing point would not decrease much and the metal hydroxide formation can occur without being hampered by a limited hydroxyl ion supply. Therefore, High alkalinity of mixed liquor can slow down the metal-phosphate reaction by encouraging initial metal hydroxide formation in theory. In typical MBR process, alkalinity would not be a significant factor affecting chemical phosphorus removal since coagulants are added very slowly to the mixed liquor and the mixed liquor has enough alkalinity to resist against excessive pH change.
  • Target phosphorus level – As target phosphorus level decreases, more reaction time is required because it is increasingly harder for metal ions to randomly collide with phosphate ions. Alternatively, the slow kinetics can be covered by high coagulant dosage that causes high Me/P ratio.
  • Injection point – In biological nutrient removal (BNR) process, soluble phosphorus concentrations are different reactor by reactor. Phosphorus concentration is the highest in anaerobic tank, where phosphorus accumulation organisms (PAO) release orthophosphate, but it is the lowest in aeration/membrane tank, where phosphorus uptake by PAO takes place. No complete consensus exists presently, but, adding coagulant to anaerobic tank, where soluble phosphate concentration is the highest, may enhance the initial metal phosphate formation and lead to an efficient coagulant usage. According to pilot tests (Trivedi, 2004), alum concentration increased four times when dosing point was changed from anaerobic tanks to aerobic tanks to obtain 0.1 mg/L TP in permeate. In lab scale tests (Johannessen, 2006), no adverse effect of alum was observed when it was added to anaerobic tanks. One potential drawback of adding coagulant to anaerobic tank is that the weak mixing intensity can cause sludge settling in the tank.   The above argument may sound plausible, but, given the fact that extended period of time is required for metal hydroxides to reach equilibrium with phosphate ions, adding coagulant to any tank does not necessarily mean phosphate ions are depleted in the tank. In conclusion, no consensus has been reached regarding the best coagulant injection point.
  •    However, there are many contradicting literature that suggest adding coagulants to anaerobic tank can suppress the growth of PAO by not leaving enough free phosphorus in the subsequent aeration tank. According to this notion, due to the lack of enough phosphorus to grow PAO in the aeration tank, PAO population in the system eventually decreases and the coagulant dosages should be raised to compensate the lost phosphorus absorption capacity by PAO. Therefore, coagulants should be added to either in the downstream of aeration tanks or membrane tank (Crawford, 2006; Johnson, 2009; Daigger, 2010).
  • Number of injection points – Adding coagulant through multiple injection points may help expedite capturing phosphate ions from a wider range of reactor spaces before the initially formed amorphous metal hydroxides are aged.
  • Pre-neutralized coagulants – Using partially or fully pre-neutralized coagulants such as sodium aluminate and polyaluminum chloride (PAC) would not result better phosphorus removal than alum and aluminum chlorides. Phosphate ions must replace already existing hydroxyl ligands to be absorbed to the metal hydroxide, which acts as a barrier for the reaction.


© Seong Hoon Yoon