M. Yoshino*, M. Yao*, H. Tsuno** and I.

Somiya**
* Maezawa Industries, INC., 5-11, Naka-cho, Kawaguchi-shi, Saitama 332-8556, Japan
(E-mail: masaaki_yoshino@maezawa.co.jp; makoto_yao@maezawa.co.jp)
** Graduate School of Engineering, Kyoto University, Yoshida-honmachi, Sakyo-ku, Kyoto 606-8501, Japan
(E-mail: tsuno@water.env.kyoto-u.ac.jp; isoumiya@rins.ryukoku.ac.jp)
Abstract Removal of phosphorus and nitrogen is required to prevent eutrophication problems in lakes and
enclosed coastal seas. And recovery of phosphorus from wastewater has been attracting attention because
of lack in phosphorus resources in the near future. In this study, reaction kinetics and design parameters of
struvite production are experimentally investigated by using basic reaction type and a draft-tube type
reactors. Struvite production rate, which is a very important parameter in reactor design and efficiency
estimation, is formulated in an equation consisting of a rate constant (k2), and magnesium, phosphate and
ammonium concentrations. The value of k2 is shown to be increased with struvite concentration and mixing
intensity in the reactor. The developed equation is applied to the results obtained from the draft-tube type
reactor experiments and verified for its applicability. High struvite concentration of 1025% is maintained in
the draft-tube reactor experiments. 92% removal and recovery efficiency with effluent phosphorus
concentration of 17 mg/L is achieved under the conditions of 4 minutes reaction time, pH of 8.5 and Mg/P
molar ratio of 1.1.
Keywords Draft-tube type reactor; nitrogen removal; phosphorus removal; struvite; supernatant in
anaerobic digestion

Introduction

Water Science and Technology Vol 48 No 1 pp 171178 IWA Publishing 2003

Removal and recovery of phosphate and ammonium as

struvite from supernatant in anaerobic digestion

Removal of phosphorus and nitrogen is required to prevent eutrophication problems in

lakes and enclosed coastal seas. As the reserves of phosphorus rock are limited (Steen,
1998), a recovery of phosphorus from wastewater has been also attracting attention. The
removal of phosphate for recycling from wastewater has been practiced widely and a number of effective and reliable techniques have been employed. Although these are already
put in practical use, improvement for cost reduction is also required. In this study, reaction
kinetics and design parameters of struvite production are experimentally investigated to
seek an optimum method of struvite production. The experiments were conducted by using
basic reaction type and draft-tube type reactors. The problems of struvite grown on the
walls of the pipes in anaerobic digestion systems have been reported (Borgerding, 1972).
Therefore, the wastewater subject to phosphorus recovery in this study is the supernatant in
anaerobic digestion of sludge from a sewage treatment plant.
Struvite production rate, which is a very important parameter in reactor design and efficiency estimation, is formulated in an equation consisting of a rate constant (k2), and magnesium, phosphate and ammonium concentrations. The developed equation is applied to
the results obtained from the draft-tube type reactor experiments and verified for its applicability. Design parameters, such as reaction time (hydraulic retention time), magnesium
addition ratio and pH are discussed and the relationships among these parameters required
to meet the effluent phosphorus concentration are drawn based on this equation.
The constitution of the obtained crystal are checked with Electron Spectroscopy for
Chemical Analysis (ESCA).
171

Reaction kinetics

Struvite is generated by phosphate, ammonia nitrogen, magnesium ion and hydroxide ion.
Its reaction formula is shown as follows:

Mg2+ + NH4+ + HPO42 + OH + 5H2O MgNH4PO4.6H2O

(1)

M. Yoshino et al.

The value of K, which is solubility product, is calculated by the following equation, when
reaction (1) is at equilibrium:
K = [Mg2+][NH4+][HPO42][OH]

(2)

The struvite generation rate can be obtained by using k (L3/mol3.min) as a rate constant
under the condition that temperature and agitation are constant.
dX/dt = k[Mg2+][NH4+][HPO42][OH]

(3)

where X represents the generation weight of struvite in mol/L, and t represents the time in
minutes. When the concentration of NH4-N is much higher than the concentration of PO4-P
and changes little before and after the reaction, as it is in the case of supernatant in anaerobic digestion, Eq. (3) can be simplified as follows, if pH is maintained at a fixed level:
dX/dt = k2[Mg2+][HPO42]

(4)

where k2 (L/molmin) is the reaction rate constant. Factors influencing the value of k2 are
agitation intensity and the concentration of the struvite in the reaction section.
If we suppose that, in the batch experiment, there are a mol/L of Mg2+ and b mol/L of
HPO42 when the time is zero and that both Mg2+ and HPO42 react by an amount of x mol/L
at a certain time t with struvite generation in an amount of X mol/L, Eq. (4) can be expressed
as follows:
dX/dt = k2 (axa0) (bxb0)

(5)

where a0 is the equilibrium concentration (mol/L) of Mg2+ and b0 is the equilibrium concentration (mol/L) of HPO42.
Under the initial conditions that x = 0 at t = 0, the solution of this differential equation
becomes:
[1/(aa0) (bb0) ] ln [(bb0) (axa0) ]/[ (aa0) (bxb0) ] = k2t

(6)

Methods

172

Two types of experiments were conducted in this study: a batch experiment using artificial
wastewater, which was synthesized as the supernatant in anaerobic digestion, and a continuous struvite generation experiment using the actual supernatant in anaerobic digestion. A
batch experiment was conducted to understand the mechanisms of struvite generation and
factors that influence them. Figure 1 shows the experimental apparatus used in the batch
experiment. The apparatus shown in Figure 1(a) is a paddle agitation reactor, in which a jar
tester was used for agitation and a 1 litre beaker was used as a reaction tank. In the apparatus
shown in Figure 1(b), aeration was used for agitation of the mixture. A 0.6 litre glass tank

M. Yoshino et al.

(height of 145 mm; inner diameter of 96 mm; inner diameter of 26 mm of the bottom opening for aeration) was used as the reactor. KH2PO4 solution, NH4Cl solution, and
MgSO4.7H2O solution were added to ion-exchanged water in a 1 litre measuring flask to
get a 1 L mixture of a given set of concentrations of PO4-P, NH4-N, and Mg2+, which
was promptly poured into the beaker used as a reactor in which a pH meter was installed.
The mixture was then agitated by the jar tester at a fixed rotation speed. The pH level was
adjusted by adding 1 mol l1 NaOH solution. The PO4-P concentration and the NH4-N
concentration were set to be 100 mg/L and 500 mg/L, respectively, which are the same level
as those in the supernatant in anaerobic digestion. The experimental conditions set were as
follows: concentration of Mg2+ to be added to the mixture, 100 mg/L; pH level, 9.0; rotation
speed, 200 r.p.m. Agitation intensity, pH and struvite concentration in the reaction tank
were considered to be operational parameters which influence the reaction rate constant.
Aeration agitation as well as paddle agitation were tested. The experimental conditions are
shown in Table 1.
Based on the results of the batch experiment, a continuous struvite generation experiment using supernatant in anaerobic digestion was conducted. Figure 2 shows the schematic diagram of the experimental apparatus used in the study. It consists of a center section,
which is designed as a draft-tube type reactor, and a peripheral section, where generated
products are separated. The supernatant as the influent and Mg2+ solution were continuously added into the part of the draft tube. The pH level in the reaction section was adjusted
automatically by a pH controller with 1 mol l1 NaOH solution. A paddle agitator was
installed in the inner tube of the draft tube to create a descending current in the inner tube
and an ascending current in the outer tube of the draft tube. Thus, the influent, Mg2+ solution and struvite were mixed well in the reactor. Generated struvite was separated from
effluent in the peripheral section outside the draft tube by gravity separation.
Table 2 shows the quality of the supernatant in anaerobic digestion used as the influent
after sedimentation for about 60 minutes. Table 3 shows the experimental conditions. The
struvite concentration in the reaction section was expressed as apparent volume ratio of
struvite in the mixed liquor of the reactor, which was measured by sampling of the mixed
liquor and sedimentation for 5 minutes in a graduated measuring cylinder. The effective
capacity of the reaction section was calculated by deducting the volume of sedimented
struvite at the bottom of the reaction section and the volume of struvite floating inside the
reaction section from the total volume of the reaction section.
Mixer
M

Air
(b) Aeration

(a) Paddle agitation

Figure 1 The apparatus used in the batch experiment
Table 1 Experimental parameters of the batch experiment
Method for mixing

G-value

Aeration volume

pH

Concentration of the
struvite in the reaction
section

Paddle agitation
Aeration

(1/sec.)

(m3-air/m3-reactor/min)

()

(mg/L)

4~133

1.67

8.8~9.0
9.0

55~1,380
400~3,200

173

Agitator
Influent
pH

Mg2+ solution
NaOH solution

Effluent
M. Yoshino et al.

Sedimented struvite
Reaction section
Peripheral section
Figure 2 The apparatus used in the continuous experiment
Table 2 The quality of the influent of the continuous experiment

pH ()
Water temperature (c)
SS (mg/L)
PO4-P (mg/L)
NH4-N (mg/L)
Mg2+ (mg/L)
Ca2+ (mg/L)

Run 1

Run 2

7.5~8.0
19.1~25.0
18~52
198~242
441~602
5.0~19
8.0~16

7.5~7.7
8.1~17.1
38~97
266~290
511~591
7.0~11

Table 3 The experiment conditions of the continuous experiment

Item

Center section

Peripheral
section

Volume
pH
Mg/P
Concentration of the struvite
Retention time
R.p.m.. of the paddle agitator
G-Value
Overflow rate

(L)
()
()
(%)
(mg/L)
(min)
(r.p.m.)
(1/sec.)
(cm/min)

Run 1

Run 2

2.1~2.3
8.4~8.5
0.8~1.2
5~25
37,000~140,000
3.9~12.1
250
178~219
0.4~1.2

2.5
8.2~8.3
0.7~1.1
7~12
36,000~64,000
13.2~15.8
200
132~142
0.3~0.4

Results and discussion

Batch experiment

174

Figure 3 shows an example of change in soluble concentration of PO4-P and NH4-N with
reaction time in the batch experiment. This was the result obtained at a low rotation speed of
20 r.p.m. which corresponds to the G value of 4 (1/sec). The reaction was considerably
slower as compared with the reaction seen at standard rotation speed (200 r.p.m.).
However, the reaction was fast in the initial stage, when the concentration of PO4-P was
high. Meanwhile, NH4-N concentration changed little. Based on the data obtained after a
lapse of 2 hours when chemical equilibrium had almost been achieved, the product of concentration values [Mg2+][NH4+][HPO42] was calculated. Figure 4 shows the relationship
between [OH] and the obtained product of concentration values [Mg2+][NH4+][HPO42].
A straight line, with an inclination of 1 corresponding to the experimental data, was drawn
in Figure 4, to obtain k (the product of solubility values) in Eq. (3) at the equilibrium state

10-7
50

NH4-N(mg/L)

500
400

10-8

10-9

10-10
10-7

300

10-6

10-5

10-4

[OH-] (mol/L)

0
0

60

120
180
Time (min.)

240

M. Yoshino et al.

[Mg2+][NH4+][HPO42-]

PO4-P (mg/L)

100

300

Figure 3 The change of PO4-P and NH4-N concentration

Figure 4 The relationship between [OH] and

[Mg2+][NH4+][HPO42]

[1/((a-a0)-(b-b0))]In[((b-b0)(a-x-a0))/
((a-a0)(b-x-b0))]

(dX/dt = 0). The obtained value for k was 7.8 1015 (mol/L)4. Figure 5 also shows how to
calculate the reaction rate constant k2 based on Eq. (6).
The results of experiments done under the conditions indicated in Table 1 are also shown
in Figures 6 and 7. Figure 6 shows the result of the experiment using paddle agitation. When
the agitation intensity reached 40 (1/sec), its effect seemed to have reached its upper limit.
Figure 7 shows the relationship between the concentration of struvite in the reaction section
and the reaction rate constant k2. Data obtained when the G value was 4 (1/sec) was
removed from Figure 7 because of rate limitation by agitation. As Figure 7 shows, it was
found that the reaction rate constant k2 is proportional to the concentration of struvite in the
reaction section.
20

10

0
10

20

Time (min)
Figure 5 The calculation of the rate constant k2
1.2
Struvite concentration in the reactor (mg/L)
1,000<
<500
500~1,000

Rate constant k2
(L/(mmol . min))

1.0
0.8
0.6
0.4
0.2
0
0

50

100

150

200

G-value (1/sec.)
Figure 6 The relationship between mixing speed and rate constant k2

175

1.6

Rate constant k2
(L/(mmol.min))

1.4
1.2
1.0
0.8
0.6

M. Yoshino et al.

Aeration
Paddle agitation (G-value40~41(1/sec.))
Paddle agitation (G-value133(1/sec.))

0.4
0.2
0

500

1,000

1,500

2,000

2,500

3,000

3,500

Concentration of struvite in the reactor (mg/L)

Figure 7 The relationship between the concentration of struvite in the reactor and rate constant k2
Continuous struvite generation experiment

It was found, from the result of the batch experiment, that the generation rate of struvite is
considerably affected by the concentration of struvite in the reaction section. Thus, a continuous struvite generation system using paddle agitation with variable rotation speed was
designed for this experiment. Figure 8 shows the results of the experiment conducted under
the conditions shown in Table 3 (RUN1). As shown in Figure 8, the product of concentration values became smaller, which means larger values of k, as the hydraulic retention time
became longer. The concentration of struvite in the reaction section was high, reaching 5 to
25 percent (37,000 to 140,000 mg/L), when the experiment was conducted. The reaction
was faster when the concentration of struvite was 10 to 25 percent than it was at a concentration of 5 to 10 percent. In RUN 1, the molar ratio of Mg/P was changed from 0.8 to 1.2.
The concentration of PO4-P in the effluent was within the range of 9.9 to 52 mg/L. As
shown in Figure 8, the variation range of the product of concentration values was very
small. This indicates that the product of concentration values was an appropriate parameter,
proving the generation efficiency of struvite. And it was possible to obtain a concentration
of 17 mg/L and a removal rate of 92 percent for PO4-P in the effluent, even when the
hydraulic retention time was as short as 3.9 minutes, if the pH level was set at 8.5 and
the molar ratio of Mg/P at 1.1.
A continuous experiment was also conducted for a longer period of time under the
conditions shown in Table 3 (RUN 2). Struvite crystals were allowed to grow in the reaction section. Figure 9 shows the change in the struvite crystal grain size distribution.
Struvite crystals had been growing from the beginning of the experiment up until 18
hours. Figure 9 also indicates that struvite crystals with a grain diameter of more than 0.6
mm sedimented to the bottom of the reaction section, because of the higher sedimentation

[Mg2+][NH4+][HPO42-][OH-]
(mol/L)4

10-13

Concentration of the struvite in the reactor

5~10 (%)

10~25 (%)

10-14
0

10

12

14

16

Hydraulic retention time of the reaction section (min)

176

Figure 8 The relationship between the hydraulic retention time of the reaction section and the product of
concentration values [Mg2+][NH4+][HPO42][OH]

60
Operation

25

0hr
6hr
18hr
32hr
36hr
43hr

20

M. Yoshino et al.

Weight of each size / total weight (%)

70

15
10
5
0
less
than
0.125

0.125

0.15

0.18

0.212

0.25

0.3

0.355

0.425

0.5

0.6

0.71

0.85

0.15

0.18

0.212

0.25

0.3

0.355

0.425

0.5

0.6

0.71

0.85

1.00

more
than
1.00

Struvite crystal grain size (mm)

Figure 9 The change in the struvite crystal grain size distribution

rate than the speed of the ascending current in the draft tube. The constitution of the
obtained crystal was checked with Electron Spectroscopy for Chemical Analysis (ESCA)
and proved to be struvite.
Conclusion

In this study, reaction kinetics and design parameters of struvite production are experimentally investigated by using basic reaction type and draft-tube type reactors, the main results
obtained were as follows.
1. Based on the batch-experiment data obtained after a lapse of 2 hours when chemical
equilibrium had almost been achieved, the product of solubility values K was calculated
to be 7.8 1015 (mol/L)4.
2. The struvite production rate can be obtained by using k2 (L/mmol.min) as a rate constant, if the concentration of NH4-N is much higher than the concentration of PO4-P, as
it is in the case of supernatant in anaerobic digestion.
dX/dt = k2[Mg2+][HPO42]
3. As a result of the experiment using paddle agitation, when the agitation intensity
reached 40 (1/sec), its effect seemed to have reached its upper limit. And it was found
that the reaction rate constant k2 is proportional to the concentration of struvite in the
reaction section.
4. High struvite concentration of 1025% is maintained in the draft-tube reactor experiments. 92% removal and recovery efficiency with effluent phosphorus concentration of
17 mg/L is achieved under the conditions of 4 minutes reaction time, pH of 8.5 and
Mg/P ratio of 1.1.
5. Using the draft-tube reactor, struvite crystals had been growing from the beginning of
the experiment up until 18 hours. And struvite crystals with a grain diameter of more
than 0.6 mm sedimented to the bottom of the reaction section, with a higher sedimentation speed than that of the ascending current in the draft tube.
6. The constitution of obtained crystal was checked with Electron Spectroscopy for
Chemical Analysis (ESCA) and proved to be struvite.
177

References

M. Yoshino et al.
178

Borgerding, J. (1972). Phosphate deposits in digestion systems. Journal WPCF, 44(5), 813819
Driver, J., Lijmbach, D. and Steen, I. (1999). Why recover phosphorus for recycling, and how?
Environmental Technology, 20(7), pp. 651662.
Liberti, L. (1986). The 103h1 Rim-Nut demonstration plant at westbari for removing and recovering N and
P from wastewater. Water Research, 20(6), pp. 735740.
Salutsky, M.L., Dunseth, M.G., Ries, K.M. and Shapiro, J.J. (1972). Ultimate disposal of phosphate from
waste water by recovery as fertilizer. Effluent and Water Treatment Journal, Oct., pp. 509519.
Steen, I. (1998). Phosphorus availability in the 21st century: Management of a non-renewable resource.
Phosphorus & Potassium SeptemberOctober 1998, pp.2531.