RO Concentrate Softening through Induced Crystallization in a Fluidized Bed Reactor

Induced crystallization technique was adopted in a fluidized bed reactor to remove scale-forming ions from RO concentrate in this paper, the result of which is as follows: the effluent of the fluidized bed reactor reached stability in two minutes after the operation started; the optimized consumption ratio of softener was 1, under the circumstances of which the removal rate of Ca was 89.22%; the removal of Ca was mainly completed at the position 20 centimeters above the bottom of the bed; the reflux rate had no influence on the removal of Ca , but it had influence on the turbidity of effluent.


Introduction
Treating water with RO (Reverse Osmosis) process will produce RO concentrate. Especially when using RO to produce fresh water from brackish water, the concentrate yield may be higher than the fresh water yield [1]. At present, the total output of seawater desalination in China is 5×105m3/d [2], while the recovery rate of RO when treating seawater is 30% to 40% [3]. That is to say, the daily fresh water output is 7.5×105m3 1.17×106m3. Therefore, if RO concentrate is directly discharged without any recycling method, it would not only be a waste of water resources, but also affect the environment.
Induced crystallization means making oversaturated CaCO 3 and CaSO 4 grow on the added seed crystal to achieve the goal of reducing the concentration of scaleforming ions in the solution. Induced crystallization coupling with a fluidized bed reactor means using fluidized bed packing as seed crystal and remove scaleforming ions during the process of oversaturated RO concentrate flowing from bottom to top in the fluidized bed reactor.
The crystallization process depends on the equilibrium relation between the solid and the solution. If the solution is already supersaturated when solid substance comes into contact with it, then the supersaturated portion of the substance in the solution will crystallize [13]. The crystallization process of RO concentrate can be divided into two parts, one is the formation of crystal nucleus and the other is the growth of crystals, which is the growing process of crystal nucleus [14]. And the mechanism of induced crystallization lies in adding crystal nucleus. Theoretically, induced crystallization can remove every substance that can precipitate as crystals from water solution [15].  2 Material and method

Experimental water
Prepare the experimental water based on the average water quality (as shown in Table 1) of the concentrate of two single-pass two-stage RO systems in Unit 3 and 4 of a power plant. The scale inhibitor (PTP-0100) content of the experimental water is 16mg/L.

Experimental facility
The experimental facility and process consists of a drug tank, a raw water tank, a dosing pump, a feed water pump, a fluidized bed, four sample connections, three flowmeters and other devices, as shown in Figure 1. The fluidized bed reactor is made of PMMA with the height of 100 centimeters and the inner diameter of 4 centimeters. A screen is installed inside the fluidized bed reactor to prevent the inducer from flowing out. The drug injection tube is introduced into the reactor 2 centimeters above the screen. Sample connections are installed 20 centimeters, 40 centimeters, 60 centimeters and 80 centimeters above the bottom of the reactor respectively. Limestone is used as fluidized bed packing, and the static bed height is 40 centimeters. Feed water flows into the reactor at the bottom of the bed and flows out from the outlet at the top of the reactor. The packing acts not only as the inducer but also as the sludge blanket in a clarification tank, which bear the functions of induced crystallization, clarification and filtration.

Experimental method
The operation process of the experimental facility is as follows: The experiment is conducted under room temperature, and each experiment uses brand-new inducers. Turn on the feed water pump to start the experiment. Open the inlet and outlet valves and adjust the inlet flow to the set value. Test the turbidity of outlet water every 30 seconds. After the turbidity of outlet water is stabilized, open the dosing valve to add softener into the reactor. Promptly adjust the dosing flow to the set value, and the induced softening clarification process will be started. Then, test the Ca 2+ of outlet water every 2 minutes, after the value of which is stabilized, test the Ca 2+ , pH, hardness and turbidity of the outlet water from each sample connection. After that, turn on the reflux valve, and adjust the outlet valve until the reflux flow reaches the set value. Reduce the softener dosage according to the "feed water/concentrate" ratio. Test the Ca 2+ and hardness of the outlet water every 2 minutes. In order to ensure that the bed is fluidized, u must be higher than min u . min u is calculated by Formula 1 to 3 as follows.

Fluidizing velocity
Galileo Number ( a G ): where, d represents packing particle diameter (m); U represents water density, (1000kg/m 3 ); p U represents packing density (kg/m 3 ); g represents gravitational acceleration, (9.81m/s 2 ); and P represents water Because it has to be ensured that inducers with the maximum particle diameter be in a state of suspension, d in Formula 1 to 3 must take the value of the maximum packing particle diameter, which is 0.500×10 -3 m. Based on computation result, min

ICCCP 2016
To ensure that the packing will not flow out of the reactor, u must not be higher than max u . max u is related with the height of the bed and the reactor, and needs to be determined by experiment. The experiment process is as follows: Turn on the feed water pump after evacuation of all the air inside the fluidized bed reactor. Open the inlet and outlet valves and adjust the inlet flow while observing the expansion height of the bed. When the bed is expanded to a place approximately 5 centimeters below the outlet, record the current inlet flow which is the maximum flow ( max Q ) allowed in the reactor. Based on the experiment result, max Q is 90L/h. And by calculation, max u should be 71.66m/h.

Retention time
The retention time of concentrate in the bed ( t ) is calculated by Formula 4: where, t represents the retention time of concentrate in the bed (s); ) represents the inner diameter of the reactor (cm); H represents the bed height during steady operation (cm); m represents the mass of inducer inside the reactor (g); U represents the inducer density (2.50g/cm 3 ) and Q represents the inlet flow (L/h).
Based on the experiment result, H is 91cm under the condition when Q is 90L/h and m is 660g. Therefore, t is calculated to be 35.12s.

Softener
Based on the experimental water quality in Table 1, NaOH is selected as the softener. The amount of softener is expressed through the consumption ratio of softener ( N ), where N is defined as the molar ratio of NaOH to the experimental water alkalinity (JD m ). In order to ensure complete neutralization of NaOH and HCO 3 -, the theory value of N should be 1.0.

Induction time
In During the crystallization process, inducer acts as the crystal seed, lowering down the energy barrier of the reaction. Therefore, under the same circumstance, adding inducer will accelerate the crystallization process and reduce the residual Ca 2+ . As reflected in Figure 2, curve c is below curve a while curve d is below curve b.

Operation time
Reflux rate is defined as the percentage of reflux flow in the total inlet flow.
Using the experimental facility shown in Figure 1, the relation between the removal rate of Ca 2+ and the operation time is shown in Figure 3, under the condition where the reflux rate is 50%, the water temperature is 19.3 , u is 63.59m/h, t is 46.38s and N is 1.02. Here, the beginning of the operation is the time when softener starts to be added.
After the softener is added, the CO 3 2level of the inlet flow will keep rising until a supersaturated solution of Ca 2+ and CO 3 2is formed. The supersaturated solution flows through the fluidized bed reactor and induces the crystallization process. The generated crystals will adhere to the surface of inducer particles, and the Ca 2+ level will drop. This reaction process takes approximately 2 minutes according to Figure 3.

Softener dosage
Under the same condition as Section 2.2, the relation between the removal rate of Ca 2+ and N is shown in

Height of the fluidized bed reactor
Using the experimental facility shown in Figure 1, the relation between the removal rate of Ca 2+ and the height of the reactor is shown in Figure 5, under the condition where the reflux rate is 50%, the water temperature is 19.6 , u is 65.97m/h, t is 46.19s and N is 1.04.
The taller the reactor is, the longer the inlet flow will be in contact with the inducer, and the higher the removal rate of Ca 2+ will be. However, when the height of the reactor is enough to ensure adequate crystallization time, raising the height will have little effect on the removal rate of Ca 2+ .

Reflux rate
Using the experimental facility shown in Figure 1, the relation between the removal rate of Ca 2+ , the effluent turbidity and the reflux rate is shown in Figure 6, under the condition where the water temperature is 19.6 , u is 66.35m/h, t is 46.17s and N is 1.02.
In the fluidized bed reactor, the inducer layer is acting as a filter bed. Therefore, higher reflux rate means longer filtering time of the inlet flow and lower effluent turbidity. Nevertheless, if the reflux rate is too high, the pore size of the filter bed will become larger and the filter bed will be less capable of intercepting suspended solids, resulting in higher effluent turbidity.

Conclusion
RO concentrate softening through induced crystallization in a fluidized bed reactor has the following characteristics: (1) The start-up speed of the reactor is relatively fast. The effluent quality can reach stability in approximately 2 minutes after the operation starts, as shown in Figure 2 and 3.
(2) The optimized consumption ratio of softener is 1.
When softener is inadequate ( N =0.8), the removal rate of Ca 2+ merely reaches 76.18%. When softener is just enough to completely neutralize HCO 3 -( N =1.0), the removal rate of Ca 2+ can reach 89.22%. Keep adding softener would not improve the removal rate of Ca 2+ obviously. As shown in Figure 2 and 4.
(3) The removal of Ca 2+ was mainly completed at 20 centimeters above the bottom of the bed. For comparison, the removal rate of Ca 2+ is 90.05% and 91.60% respectively at 20 centimeters and 100 centimeters above the bottom of the bed. The removal rate at the latter position is merely 1.72% higher than that at the former one, as shown in Figure 5.
(4) The reflux rate has no influence on the removal of Ca 2+ , but it had great influence on effluent turbidity. When the reflux rate is higher than 20%, the effluent turbidity will be lower than 10.5 NTU, as shown in Figure 6.