Chemical reagents

All of the reagents were of analytical grade. A standard recipe was used to make the synthetic urine 21, 22. The chemicals were used to prepare the synthetic urine in this study including CaCl.2HO 0.65 g/L, MgCl.6HO 0.65 g/L, NaCl 4.6 g/L, NaSO 2.3 g/L, CHNaO.2HO 0.65 g/L, Na-(COO) 0.02 g/L, KHPO 4.2 g/L, KCl 1.6 g/L, NHCl 1 g/L, NHCONH 25 g/L 22. The characteristics of synthetic urine: pH = 5 – 5.5; NH-N = 261.68 mg/L; PO-P = 957.35 mg/L; Mg = 76.85 mg/L. The important chemicals needed are calcium chloride dihydrate (CaCl.2HO, 96.0%), magnesium chloride hexahydrate (MgCl.6HO, 99.0%), and sodium chloride (NaCl, 99.0%) were from Showa (Japan). We got potassium phosphate monobasic (KHPO, 99.0%), potassium chloride (KCl, 99.0%), and ammonium chloride (NHCl, 99.8%) were obtained from Sigma-Aldrich (U.S.A), etc.

Lab-scale fluidized-bed homogeneous crystallization (FBHC) reactor

The urine sample after measuring and adding the right amount of Mg, P, and K sources will be pumped into the influent line. NaOH solution placed in the right side as in Figure 2 will be pumped directly at the same time with effluent to keep pH low around 9 at first to form precipitation of NH-struvite. Furthermore, using the recirculation flow can help to reuse the nutrients left in the waste stream. The P and N after being kept in precipitation will help the sample have more recovery efficiency than runout in the effluent line. Basic calculation according to measured pumps and parameters in Table 215:

The molar ratio of Mg:P plays an important role in the crystallization of struvite, and the ratio greater than 1 is considered necessary23. At Mg:P = 1, when the pH exceeds 9, the reaction occurs enough to completely remove the P, similar to the amount of precipitated NH-struvite. Furthermore, the NH-N concentration decreases from 41 mg/L to 22 mg/L at equilibrium. This indicates that less than 1% of the PO-P was removed as struvite, the largest percentage probably being K-struvite. With the addition of MgO (Mg:P = 1.5 and 2) the pH further increased, but did not further affect the removal efficiency of P22. The molar ratio of Mg:P added to the urine input sample is usually 1.1 – 1.2. However, according to some studies, the corresponding molar ratio of Mg:P is 2:2 at pH = 10 showed that the P and K removal efficiency of K-struvite was over 75%. That’s why in this study, the ratio of Mg:P = 1.25:1 was applied to synthesis urine sample 23.

All of the analytical parameters were followed from the APHA 2005 standard book 24 as a reference to experiment with the struvite performance. For the synthesis of urine, six parameters needed to test for influent and effluent in this topic included: pH, total hardness, Mg, PO-P, NH-N. Samples were taken once a day in the morning at influent and effluent line. Filtered samples were needed before analyzing.

Efficient treatment of nitrogen and phosphorus in the form of struvite

The pH in the reactor was maintained at around 8 to avoid the appearance of other forms such as MgNHPO, Mg(OH), MgHPO, MgCO, Ca(PO), CaHPO, CaCO precipitation in the reactor 25, 26 because when the pH is higher than 8, the sample produces Mg(OH) precipitation. Currently the reactor is operating at pH around 8. It is noted that pH higher than 11 will create tri-magnesium phosphate 27. Similar to pH, the concentration of Mg input was recorded around 150 mg/L which was much smaller than the concentration of PO-P, 4,000 mg/L, while the operation condition needs the Mg:P molar ratio to be 1.25 to create struvite 15. Typically, the molar ratio of Mg:P in the composition of the real human urine does not reach the expected ratio of 1.25. Therefore, when applying with actual urine, it is necessary to add an appropriate amount of Mg to balance the above optimal ratio suggested to obtain high precipitation efficiency. Since NH-N concentration was pre-treated then the influent was almost the same to Mg concentration. After operating for several days, the removal efficiency (RE) of Mg and PO-P could reach 99% maximum (Figure 3).

The results above showed the same trend to several previous papers when the amount of RE of Mg and PO-P was followed 90 – 99% each parameter. It’s clear that there is a significant fluctuation in the first five days. According to Liu et al. (2008), the pH and concentration of NH-N are stable after five days 28. Another fact that conductivity can be used as an effective measurement to control dosing in struvite precipitation reactors 29. In this study, urea solution was prepared and analyzed the change of concentration after seven days. The urea hydrolysis occurs as following reaction30:

NH(CO)NH + 3HO 2NH + HCO + OH

Similar to the initial preparation, the concentration of NH-N was small in the same time when testing the sample in reactor (Figure 3) and urea alone (Figure 4). It both started with only 22.4 mg/L of NH-N and fluctuated four days to reach maximum concentration it could be. While the concentration was nearly three times higher after the four-day observation, the pH fluctuated then came to be stable at 7.5. This was suited to the fact stated by Liu et al. (2009) when the pH and concentration of NH-N are stable after 72 and 84 h respectively in urea solution28. It means that after seven days of pre-treatment, the urine sample can be used to run in the FBHC reactor.

It can be seen that the results obtained from the experimental model with high PO-P and Mg RE correspond to the results of previous researches. When Saba et al. (2018) studied that while the P RE reached 94% 31 with the conventional precipitation model, then Van-Giang et al. (2019), and Leng et al. (2020) managed to treat 98 – 99% RE with the upgrade FBHC reactor 15, 32. That is one of the advantages this reactor brings to this works. Besides, the RE of Mg was also higher when adjusting the pH ≤ 8 31. It can be seen that at different pH thresholds, the RE of each criterion will also be different. Accepting the optimal pH fixation to achieve the highest granulation efficiency will lead to the fact that one of the parameters will have a lower RE than the others. It is mentioned here that the RE of NH-N in this study was not high compared to that of Liu et al. (2020) when their pH was 9, they can achieve a RE up to 98.5% 33.

From day 15 to day 24, the RE of Mg and N suddenly decreased due to a lack of balance ratio and available nutrients in the synthesis urine sample. As mentioned earlier, because the pH inside the system was unstable, it caused to the formation of many foreign precipitates at different pH leading to the formation of unnecessary salts with difficult structures to break. That is why during the operation period, periodic withdrawal of precipitate is necessary because otherwise, the amount of precipitate inside the column will form an allowable overload of the FBHC reactor used in this study to operate. Nutrient dose was added as soon as the discharge was changed suddenly. After adding nutrients, the RE started to stabilize from day 16 – 19, and then started to increase again.

Struvite formation

Figure 5 showed the growth of crystal during 24 days. The mass of struvite precipitate obtained in this study was approximately 200 g/m of urine. However, the precipitation in the FBHC reactors seems to occur much more than the working volume of the reactor; hence, weekly taking out the precipitation is needed to avoid the overloading inside the reactor. No granular struvite occurred so far but only fine struvite due to the appearance of foreign ions25.

The presence of foreign ions coming from the nutrients in urine based has a major influence on struvite precipitation while there were more different types of precipitation’s structure occurring among those samples. It is possible to see an effect 34, e.g., on morphologic changes 35, 36, crystal 35 and on precipitation rates22. With different pH conditions, different precipitates will form from components present in urine such as Mg, Ca, NH, PO, CO. Some mentioned ions when combined together under ideal conditions, they will be very difficult to break. These precipitates have a different chemical formula from struvite (Table 3), as long as they get along with each other and increase the foreign precipitation inside the reactor, they will interfere with the formation of struvite, especially granular struvite. The identification of these foreign ions is very important. However, in this study, we only focused on the efficiency of P, N recovery; therefore, in the next study, we will evaluate the influence of foreign ions.

When the K index is low, it shows that the presence of precipitate in impurities will account for the majority when meeting suitable pH and temperature conditions. Since each precipitation form has a different pH value, to avoid the appearance of too many forms during struvite operation, the suggested pH range to get the best quality of struvite is 7.5 – 8. The precipitant must be adjusted and observed carefully to make sure the sample in the reactor is always in this range. Moreover, to have better NH-N RE, pre-treatment by preparing urea solution separately and stored in room condition to let N hydrolyze and let pH increase and stabilize naturally. It can help using less chemicals.