Chemosphere 93 (2013) 27382747

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Characterization of induced struvite formation from source-separated

urine using seawater and brine as magnesium sources
Bianxia Liu a, Apostolos Giannis a, Jiefeng Zhang b, Victor W.-C. Chang a,b,, Jing-Yuan Wang a,b
a
Residues and Resource Reclamation Centre (R3C), Nanyang Environment and Water Research Institute, Nanyang Technological University, 1 Cleantech Loop,
CleanTech One, Singapore 637141, Singapore
b
Division of Environmental and Water Resources, School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue,
Singapore 639798, Singapore

h i g h l i g h t s
 Seawater and brine enable struvite formation from source-separated hydrolyzed urine.
 Modied chemical equilibrium model agrees with experimental results.
 Struvite from synthetic and real urine have feather and cofn shape, respectively.
 Co-precipitates magnesium-calcite and calcite compromise induced struvite purity.

a r t i c l e

i n f o

Article history:
Received 7 March 2013
Received in revised form 5 September 2013
Accepted 7 September 2013
Available online 14 October 2013
Keywords:
Urine
Struvite
Phosphorus recovery
Seawater
Brine
Equilibrium model

a b s t r a c t
Struvite (MgNH4PO46H2O) precipitation is widely used for nutrient recovery from source-separated
urine in view of limited natural resources. Spontaneous struvite formation depletes the magnesium in
hydrolyzed urine so that additional magnesium source is required to produce induced struvite for Precovery. The present study investigated the morphology and purity of induced struvite crystals obtained
from hydrolyzed urine by using seawater and desalination brine as low cost magnesium sources. The
results demonstrated that both seawater and brine were effective magnesium sources to recover phosphorus from hydrolyzed urine. Crystals obtained from synthetic and real urine were revealed that the
morphology was feather and cofn shape, respectively. Structural characterization of the precipitates
conrmed that crystallized struvite was the main product. However, co-precipitates magnesium calcite
and calcite were observed when seawater was added into synthetic and real urine, respectively. It was
found that the presence of calcium in the magnesium sources could compromise struvite purity. Higher
struvite purity could be obtained with higher Mg/Ca ratio in the magnesium source. Comparative analysis
indicated that seawater and brine had similar effect on the crystallized struvite purity.
2013 Elsevier Ltd. All rights reserved.

1. Introduction
Phosphorus recovery from wastewater streams is getting more
and more interests globally because of the depletion of phosphorus
resources (Smil, 2000). It is also shown that the demand of phosphorus fertilizer is expected to peak around 2030 (Cordell et al.,
2009). Moreover, macronutrients in the wastewater are often discharged to the natural water bodies without control causing eutrophication and other issues in water cycle. It is of great interests to

Corresponding author at: Division of Environmental and Water Resources,

School of Civil and Environmental Engineering, Nanyang Technological University,
50 Nanyang Avenue, Singapore 639798, Singapore. Tel.: +65 67904773; fax: +65
67921650.
E-mail address: wcchang@ntu.edu.sg (V.W.-C. Chang).
0045-6535/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.chemosphere.2013.09.025

develop proper technologies that facilitate nutrient recovery and

improve the quality of wastewater efuent.
Human urine has been characterized as the main contributor of
nutrients in domestic wastewater with 85% of total N, 50% of the
total P, and 55% of total K, but it consists only 1% of raw wastewater volume (Larsen and Gujer, 1996). The separate collection of human urine reduces nutrients loading to domestic wastewater and
facilitates nutrient recycling as well (Jnsson et al., 2005; Larsen
et al., 2009). A number of physico-chemical and biological techniques have been developed to harvest nutrients (mainly N, P)
from the urine stream. Phosphorus recovery can be readily
achieved by crystallization as struvite (magnesium ammonium
phosphate hexahydrate), which can be further utilized as a valuable slow-release fertilizer (Johnston and Richards, 2003; Maurer
et al., 2006; Rmer, 2006).

2739

B. Liu et al. / Chemosphere 93 (2013) 27382747

After urinary excretion, urea is quickly hydrolyzed and converted into ammonia and CO2. The elevated pH-value causes spontaneous precipitation of struvite in the hydrolyzed urine,
exhausting most of the magnesium and about 30% of phosphorus
(Zhang et al., 2013). Magnesium is the major limiting factor to harvest the remaining phosphorus. Although the direct usage of commercial magnesium salt is effective, the additional cost is likely to
restrict further P-recovery from urine, which indicates the importance of exploring cost-effective magnesium sources. Kumashiro
et al. (2001) used seawater as magnesium source for struvite crystallization at pilot scale wastewater treatment. It was found that
the use of seawater was superior because it could simplify facilities
and operation without affecting subsequent processes. Bittern, a
by-product from salt manufacture, was investigated as low-cost
magnesium source to recover phosphorus and nitrogen from
wastewater (Li and Zhao, 2002; Lee et al., 2003; El Diwani et al.,
2007). Results showed that application of bittern at treated wastewater could achieve high phosphate and ammonia removal efciencies. Etter et al. (2011) investigated different magnesium
sources for phosphorus recovery from human urine in Nepal. The
result showed that bittern was a practical magnesium source, but
it had to be imported from India. Also, magnesium oxide produced
from locally available magnesite was an alternative option. In a recent study, wood ash was also tested as magnesium source (Sakthivel et al., 2012). The cost-benet analysis revealed that wood ash
was considerably cheaper than industrially-produced magnesium
sources and even cheaper than bittern. However, due to high content of heavy metals, wood ash was concluded as an unsafe magnesium source for struvite production. It is obvious that among
the magnesium sources, some of them are limited with high cost,
while some others produce low quality products.
Chemical equilibrium models have frequently been applied to
study the struvite precipitation process. These models were developed to optimize and control the struvite reaction in order to make
it practical for application. Gadekar and Pullammanappallil (2010)
developed a mathematical model to predict struvite precipitation
in the solution of ammonium, magnesium and phosphate. The
model was validated against experimental data and exhibited satisfactory prediction efciency. Harada et al. (2006) predicted the
struvite formation from urine by a chemical equilibrium model
having good agreement with experimental results. Several
researchers have employed Visual Minteq to model struvite formation. Lee et al. (2003) used Visual Minteq to model struvite formation for nitrogen and phosphorus removal from wastewater by
bittern addition. Celen et al. (2007) amended Visual Minteq to predict struvite precipitation from liquid swine manure. However,
chemical equilibrium modeling and experimental validation were
mainly focused on the precipitation process overlooking struvite
purity and co-precipitates identication. Therefore, it is of great
importance to ensure the efciency of inexhaustible magnesium
sources like seawater or brine on struvite quality in terms of reaction conditions and products obtained.
The above mentioned challenges prompt us to investigate the
feasibility of using seawater, brine and MgCl2 as magnesium
sources for phosphorus recovery from source-separated urine.

Visual Minteq was modied in order to predict the struvite formation reactions under different conditions. Effects of pH and Mg/P
molar ratio on the phosphorus recovery efciency were simulated
and outcomes were validated against experimental data. In addition, the morphology and purity of the struvite obtained were
characterized using various techniques in order to help us to evaluate the potential for using low cost magnesium alternatives for
large scale application.
2. Materials and methods
2.1. Urine samples and magnesium sources
Synthetic and real urine samples were used after complete urea
hydrolysis in bench-scale batch experiments. Synthetic urine was
prepared according to Grifth et al. (1976), and it was free from
any unspecied substances that could affect the struvite precipitation process. Fresh human urine was collected from 25 healthy
adults on a voluntary basis. Small amount (1 mg) of urease enzyme (Merck) was added to accelerate urea hydrolysis.
Three types of magnesium were investigated as additional magnesium sources, namely, seawater, brine and MgCl2. Seawater was
collected from west coast of Singapore and ltered through
0.45 lm pore size membrane lter to remove suspended particles
before usage. Brine solution was prepared by dissolving 70 g commercial sea salt (SigmaAldrich) in 1 L of deionized water (Ahmed
et al., 2003; Zhou et al., 2013). In addition, commercial MgCl2 (SigmaAldrich) was also tested for comparison purposes. The major
characteristics of synthetic and human hydrolyzed urine, seawater
and brine are presented in Table 1.
2.2. Crystallizations experiments
During hydrolysis in both synthetic and real urine, almost all
magnesium and calcium ions and 30% of initial phosphates were
naturally precipitated. The hydrolyzed urine still contained
1054 mg L1 phosphate with strong buffering capacity. Magnesium-to-phosphate (Mg/P) molar ratio and pH-value were considered the most essential factors during struvite precipitation.
Hence, batch experiments were carried out to investigate the optimum Mg/P molar ratio and pH for P-recovery from hydrolyzed urine. Ronteltap et al. (2007) have mentioned that thermodynamics
and kinetics of precipitation in synthetic urine is almost identical
to real urine.
Previously to precipitation experiments, the hydrolyzed urine
was ltered through 0.45 lm pore size membrane lter to remove
the spontaneous struvite crystals. To test the effect of Mg/P molar
ratio, standard amount of the magnesium sources was added to
50 mL urine sample in order to reach stoichiometric Mg/P molar
ratios 0.7, 1, 1.1, 1.5 and 2. To study the effect of pH, the optimum
Mg/P molar ratio was used. The pH-values were adjusted to 6.3,
7.0, 8.0, 9.3 (no pH adjustment), 10.2 and 11.6 with 1 M sulfuric
acid or 1 M sodium hydroxide. Furthermore, estimations of nutrients recovery were conducted for fresh and hydrolyzed urine at
different dilution ratios (1:2, 1:5), which represent the typical

Table 1
Chemical composition of hydrolyzed urine, seawater, and brine.
Component

Na+ (mg L1)

K+ (mg L1)

Mg2+ (mg L1)

Ca2+ (mg L1)

1
PO3
4 P (mg L )

1
NH
4 N (mg L )

1
SO2
4 (mg L )

1
CO2
3 (mg L )

Cl (mg L1)

Human urine
Synthetic urine
Seawater
Brine

2600
2600
8850
21600

2000
2000
565
916

1248
2374

410
656

344
565

4250
3515

1500
1500
900
1800

5448
4989

3800
4694
13467
32869

: Denotes the concentration below detection limits

2740

B. Liu et al. / Chemosphere 93 (2013) 27382747

Table 2
Estimated precipitates by using chemical equilibrium model (modied Visual Minteq).
Precipitate

Chemical formula

Database
consideration

Remark

Reference

Hydroxyapatite
Tricalcium
phosphate
Dolomite
Magnesite
Hydromagnesite
Hunitite

Ca5(PO4)3OH
Ca3(PO4)2

Not included
Not included

Slow precipitation kinetics

Slow precipitation kinetics

Musvoto et al., 2000

Ferguson and McCarty, 1971

CaMg(CO3)2
MgCO3
Mg5(CO3)4(OH)24H2O
CaMg3(CO3)4

Not
Not
Not
Not

Mamais et al., 1994

Case et al., 2011
Cheng and Li, 2010
Kangal and Gney, 2006

Aritinite
Struvite
Monenite
Calcite
Brucite

Mg2CO3(OH)23H2O
MgNH4PO46H2O
CaHPO4
CaCO3
Mg(OH)2

Not included
Included
Included
Included
Included

Slow precipitation kinetics

Prior formation of its metastable phase hydromagnesite
High formation temperature (>40 C)
Complex mineral with hydromagnesite, magnesite,
aragonite
and dolomite
Formation on the surface of existing magnesite
Precipitation at pH (711); Mg2+/Ca2+ ratio >0.6
Thermodynamically stable
Thermodynamically stable
Precipitation at high pH

included
included
included
included

(a-1)

Muchi and Matsumoto, 1979

Song et al., 2007
Musvoto et al., 2000
Harada et al., 2006
Musvoto et al., 2000

(b-1)

Synthetic urine + MgCl2

(a-2)

Real urine + MgCl2

(b-2)

Synthetic urine + seawater

Real urine + seawater

Fig. 1. Effect of Mg/P molar ratio on the recovery rates of phosphate, ammonium, calcium and magnesium from hydrolyzed synthetic urine (pH = 9.3), : phosphate (model),
d: phosphate (experimental), : ammonium (model), j: ammonium (experimental), - - - -: magnesium (model), N: magnesium (experimental),   : calcium (model), and
s: calcium (experimental).

dilution for low ushing toilets. In all the experiments, magnetic

stirring was applied for 15 min to enhance the reactions. The mixture was then centrifuged at 5000 rpm for 10 min to separate the
precipitates. The obtained solids were rinsed twice with 99.9% ethanol and three times with deionized water to remove any organic

impurities. The precipitates were dried overnight at 80 C for various characterization analyses. According to Bhuiyan et al.
(2008), during drying process there was 10% struvite mass loss at
80 C which was considered into the purity calculations. Furthermore, the supernatants were ltered through 0.45 lm pore size

2741

B. Liu et al. / Chemosphere 93 (2013) 27382747

membrane lter before analysis of the remaining ammonium,

phosphate, calcium and magnesium. All batch experiments were
carried out in duplicates at room temperature (23 1 C).

database because they are most probably precipitated under the

specic experimental conditions.
3. Results and discussion

2.3. Analytical methods

3.1. Optimization of struvite recovery process
After struvite precipitation, total ammonium concentration was
measured colorimetrically based on salicylate method using spectrophotometer (Hach DR2800, USA). Orthophosphate and chloride
concentration in the supernatant was measured by ion chromatography (ICS-1100, Dionex). The concentrations of magnesium and
calcium were analyzed with an Inductively Coupled Plasma Optical Emission Spectrometer (Spectro Analytical Instruments). Carbonate concentration was measured by TOC analyzer (Shimadzu,
Japan). Electrical conductivity and pH-values were measured with
a portable D-54 conductivity/pH meter (Horiba), respectively. In all
experiments, phosphate, ammonium, magnesium and calcium
recovery efciencies (Ri) were calculated by the following
equation.

Ri 1 

C i;sup
C i;ini

where Ci,sup and Ci,ini are ion concentrations in the supernatant and
initial urine sample, respectively.
Regarding the precipitates production, it can be estimated by
the following equation.

K sp;struv ite Mg2 NH4 PO3

4 

3.1.1. Effect of Mg/P molar ratio

Fig. 1a presents the effect of Mg/P molar ratio on the recovery
efciencies of phosphate, ammonium, calcium, and magnesium
with the addition of MgCl2 and seawater in synthetic urine. From
the modeling results, phosphate recovery efciency increased from
50% to 98% when Mg/P molar ratio increased from 0.5:1 to 1.1:1,
and reached almost 100% after that. Likewise at Mg/P molar ratio
1.1:1, magnesium recovery was nearly 100%, while increasing the
ratio made magnesium surpass the demand for phosphate recovery and resulted in residual magnesium in the supernatant. Under
the same ratio, the recovery efciency of ammonium was low (only
7%), because of the high initial ammonium concentration in the
solution. Furthermore, calcium was precipitated almost 100%
when seawater was added to hydrolyzed urine without any inuence of Mg/P molar ratio.
Laboratory experiments were done for further validation of Visual Minteq. Fig. 1a demonstrates that experimental results on
phosphate and ammonium recovery efciencies using MgCl2 were
in accordance with the model. However, when seawater was

where Ksp,struvite is the solubility product constant of struvite;

3
[Mg2+], [NH
4 ] and [PO4 ] are magnesium, ammonium and phosphate concentrations in the balanced solution after struvite
precipitation.
X-ray diffraction (XRD, Scintag, USA) analysis and Fourier Transform Infrared Spectroscopy (FTIR, Perkin Elmer PE 1600, USA) were
applied to characterize the precipitates. Energy Dispersive X-ray
Detector (EDX, Oxford Instrument, UK) and X-ray Fluorescence
(XRF, Bruker AXS, Germany) were used to quantify the compositions of the harvested precipitates. Furthermore, Field Emission
Scanning Electron Microscopy (FESEM, JEOL Asia Pte, Japan) images
were obtained to characterize the morphology of the precipitates.

(a)

Synthetic urine + MgCl2

2.4. Chemical equilibrium modeling

The chemical equilibrium model Visual Minteq 3.0 has been
widely used for calculating speciation, solubility and sorption equilibrium of solids in an aqueous solution (Gustafsson, 2011). Since
struvite was not included in the current solid database, the solubility constant pKsp = 13.15 was manually incorporated into the model database based on Buchanans study (Buchanan et al., 1994).
Visual Minteq involves a management tool to add or delete mineral
species. The activity coefcient computation of electrolyte solution
at relatively high concentrations was based on the Davies approximation on DebyeHuckle equation. Default values were used for
all other equilibrium constants.
It has been reported that struvite precipitation can be completed within few minutes (Lee et al., 2003). Since the precipitation
kinetics was not considered in the current model, those minerals
with low formation rates (in the scales of hours or days) were
intentionally omitted from the modeling calculation. Real data
based in Table 1 were loaded in the equilibrium model to estimate
the probable precipitates (Celen et al., 2007). Theoretically, eleven
minerals could be formed as listed in Table 2. A further literature
review was conducted on each mineral to determine the possibility
of its formation under the dened conditions. At last, the minerals
struvite, monenite, calcite, and brucite were included in the

(b)

Synthetic urine + seawater

Fig. 2. Effect of pH on the recovery rates of phosphate, ammonium and the

magnesium from hydrolyzed synthetic urine (Mg/P molar ratio of 1.1:1), :
phosphate (model), d: phosphate (experimental), : ammonium (model), j:
ammonium (experimental), - - -: magnesium (model), and N: magnesium
(experimental).

2742

B. Liu et al. / Chemosphere 93 (2013) 27382747

added to synthetic urine, the experimental data on magnesium

and calcium precipitation efciencies were different from the
model. Compared to the model, magnesium had higher precipitation efciency, while calcium had lower efciency. The reason
involving the struvite purity will be discussed in the following
sections.
Fig. 1b shows the results using real hydrolyzed urine. Visual
Minteq predicted similar results to those of synthetic urine. A good
agreement was observed between experimental and modeling data
on phosphate, ammonium, and magnesium recovery. However, the
calcium recovery efciency in the model was higher than that in
the experiment (100 vs 90% in this case). This discrepancy might
be attributed to: (1) model assumed that all reactions were in equilibrium state, whereas the actual conditions might yet not reach
the equilibrium; and (2) the reaction efciency of calcium might
be lower than the theoritical value due to low actual concentration
as indicated by Harada et al. (2006).

3.1.2. Effect of pH
Fig. 2 shows the effect of pH on the recovery efciencies of
phosphate, ammonium, and magnesium at optimum Mg/P molar
ratio 1.1:1. The experimental data showed that phosphate, ammonium and magnesium recovery were well matched with the predicted results in the modied model. It indicated that pH range
810.5 is favorable for phosphate recovery. Similar trends were reported by Buchanan et al. (1994); Ali et al. (2003). As a matter of
interest, hydrolyzed urine has strong buffer capacity with pH-value
around 9.3 based on our previous study (Zhang et al., 2013). Phos-

Fig. 3. Maximum nutrient recovery from fresh urine (FU) (pH 6.5) and hydrolyzed
urine (HU) with no dilution (pH 9.3), 1:2 dilution (pH 9.2) and 1:5 dilution (pH 9.1).
Others include carbon, hydrogen and oxygen in the solid products.

phate recovery from hydrolyzed urine could be effectively

achieved without any pH adjustment, which carries favorable
implication for the engineering application.
3.1.3. Effect of urine hydrolysis and dilution
Maximum nutrients recovery from fresh and hydrolyzed urine
with the addition of seawater was rst estimated by modeling
(Fig. 3). For fresh urine, phosphate recovery efciency was 33%
while the calcium recovery efciency was high (87%) due to its
lower concentration. There was no ammonium recovery because
urea was not hydrolyzed yet. As a result, there was no struvite formation, and the main precipitate was CaHPO4. For hydrolyzed urine, struvite formation could proceed because of the ammonium
release from the hydrolysis process. The maximum amount of precipitates that could be formed was 2.95 kg m3 of hydrolyzed
urine.
The harvested amount of precipitates at experimental conditions was relatively lower (about 22.5 kg m3), probably due to
mass loss during ltration, drying and transferring procedures.
When hydrolyzed urine was diluted at 1:2 and 1:5, the estimated
amount of precipitated mass was 1.46 and 0.55 kg m3, respectively. That was slightly less than half and fth of estimated
amount of precipitates for undiluted hydrolyzed urine.
3.1.4. Comparisons of magnesium sources
Three types of magnesium were investigated as additional magnesium sources, namely, seawater, brine and MgCl2 (Table 3).
Based on modeling and experimental results, brine was an efcient
additive for struvite formation. According to the data in Table 1, reverse osmosis brine and sea water have the same chemical compositions, except that brine has higher concentrations (almost twice).
At optimum Mg/P molar ratio 1.1:1, 99.5% of phosphate could be
recovered through struvite precipitation.
As summarized in Table 3, a cost-effective analysis was also
conducted to compare different low-cost magnesium sources,
including bittern and wood ashes. The primary cost of the magnesium sources come from transportation cost. For coastal cities like
Singapore, seawater is readily available due to short transport distance (<20 km). Desalination brine can be obtained from local
desalination plants. Compared with seawater, less volume of brine
is needed to produce the same amount of struvite because of higher Mg2+concentration in brine. This can dramatically reduce the
transportation cost, making brine a cheaper magnesium source.
Besides, the reuse of brine is eco-friendly, as brine is generally discharged back to the ocean, which might affect the oceanic eco-system. The required amount of bittern was much lower than
seawater or brine, but this source is not easily available in local
context. The transport cost will make it more expensive than

Table 3
Comparison of different magnesium sources for phosphorus recovery from urine.

a
b
c

Magnesium source

Seawater

Brine

Bittern

Wood ash

Optimal Mg/P molar ratio

Required input (kg m3 urine)
Struvite purity, %
P recovery efciency, %
N recovery efciency, %
Product price (dollar kg1)
Transport distance (km)
Transport cost (dollar kg1 km1)
Total input cost (dollar m3 urine)
Issues

1.1
247a or 0.24 m3
95.6
99.0
5.2

020
0.0002c
00.99
Further treatment

1.1
146 or 0.14 m3
96.6
99.0
7.4

020
0.0002c
00.58
Further treatment

1.1
3.2b

>99.0

502000c
0.0002c
0.031.27
Limited source

1.9c
16.2c
23c
99.0

050c
0.0002c
00.16
Heavy metal contamination, low purity

The density of seawater is 1.03 kg L1.

Data calculated based on Lee et al. (2003).
Data calculated based on Sakthivel et al. (2012). Product prices for raw materials are considered negligible.

B. Liu et al. / Chemosphere 93 (2013) 27382747

2743

Fig. 4. FT-IR spectrum of pure struvite and precipitates produced adding seawater
and MgCl2 to synthetic urine.

seawater and brine. Wood ash is also cheap and easy to transport,
but the precipitates might suffer from low purity due to the contamination of heavy metals.
3.2. Precipitate characterization
In order to conrm the presence of struvite, extensive analyses
were conducted on the precipitates that obtained by adding seawater, brine, and MgCl2 to synthetic and real urine at Mg/P molar
ratio 1.1:1 and under pH-value 9.3.
3.2.1. Struvite formation
The FT-IR spectra of pure struvite and the precipitates produced from MgCl2 and seawater using synthetic urine are presented in Fig. 4. The FT-IR spectra are assigned into four regions
depicting the main identication peaks due to the presence of
water in the crystallized struvite. According to identication
peaks, intensities between 16301750 cm1 was attributed to
NH bending vabrition in ammonium group (Chauhan and Joshi,
2013). Intensities of bands between 10171163 and 894.2 cm1
were associated with phosphate stretching vibration, while intensities between 400650 cm1 was associated to metaloxygen
bonds. It can be conrmed that the precipitates obtained in the
experiments were struvite crystals as the FT-IR spectra were
identical.
Fig. 5 shows the XRD patterns of precipitates after the addition
of MgCl2 and seawater to synthetic and real urine. For synthetic urine, the precipitate derived from MgCl2 addition was determined to
be struvite (Fig. 5a). This was in accordance to the FT-IR results.
When seawater was added to synthetic urine, Mg-calcite
(Mg,Ca)CO3 was also identied by the XRD in the precipitates
(Fig. 5b). From the chemical equilibrium model, the predicted
product was the mixture of struvite and CaCO3. The formation of
Mg-calcite in the experiment was unexpected. This result was consistant with the observation of higher magnesium recovery efciency in the experiment than that in the model.
For real urine, the precipitate produced by MgCl2 addition was
also struvite as in synthetic urine (data not shown). When seawater was added to real urine, an additional precipitate, CaCO3,
was also observed (Fig. 5c). The precipitates were a mixture of
struvite and CaCO3. In this case, the mineral species found from
experiments and predicted by Visual Minteq were in agreement.
The XRD results from brine addition were similar to those of
seawater.

Fig. 5. XRD patterns of precipitates after the addition of magnesium sources to

synthetic and real urine: (a) struvite peaks; (b) struvite,    Mg-calcite; and (c)
struvite,    calcite.

Comparing synthetic and real urine, different precipitates were

formed after the addition of magnesium sources. The formation of
Mg-calcite in synthetic urine might be related to the high Mg/Ca
ratio and supersaturation effect as proposed by Berner (1975). Under high magnesium and calcium concentrations, the magnesium
could be incorporated within the calcite lattice forming Mg-calcite.
Loste et al. (2003) have briey indicated that the increase of occluded Mg was observed along with the increasing Mg/Ca ratio
in the solution. Actually, the formation of Mg-calcite was complicated because the Mg/Ca ratio was continuously changing during
the precipitation process. This was most likely the main reason
for the incorrect precipitates predicted by the Visual Minteq.

2744

B. Liu et al. / Chemosphere 93 (2013) 27382747

(a-1) Synthetic urine + MgCl2

(b-1) Real urine + MgCl2

(a-2) Synthetic urine + seawater

(b-2) Real urine + seawater

(a-3) Synthetic urine + brine

(b-3) Real urine + brine

Fig. 6. SEM images of precipitates after the addition of magnesium sources to synthetic and real urine.

No other crystals were abundant enough to be recognize by the

XRD measurements. However, it is still possible that they were
formed in small amounts as transient compounds due to the effect
of other ions such as Na+, K+, and SO2
(Sakthivel et al., 2012).
4
Although seawater or brine contained potassium, the potassium
struvite could precipitate only in the case of low ammonium concentrations. The specic compound was not found in any crystal
analysis nor in the model simulations.
3.2.2. Crystal morphologies
Fig. 6 shows the SEM images of precipitates from the matrix of
different magnesium sources and type of urines. For synthetic urine, the morphology of the precipitates was mainly feather shape.
When MgCl2 was added, the surface of the precipitates was smooth
(Fig. 6a-1), whereas there were some ne crystals on the surface of
the precipitates with the addition of seawater and brine (Fig. 6a2,3). For real urine, the morphology of the precipitates was cofn
shape (Fig. 6b-1,2,3). The precipitates with addition of seawater

and brine had some ne crystals on their surfaces as well

(Fig. 6b-2,3). This observation has also been reported in other studies (Le Corre et al., 2005). Regarding the crystal morphology, Chauhan and Joshi (2013) carried out struvite growth studies and
reported that struvite can exist in different morphologies (prismatic type, pyramid type, cofn shaped, feather shaped, needle
type, etc.) which are strongly dependent on growth parameters.
In order to determine the composition of the ne crystals appeared on the surface of the precipitates, scanning electron microscopy coupled with EDX has been used on a specic area as labelled
on the SEM images (Fig. 7). The results are presented in Table 4. For
synthetic urine, the rst analyzed area was at the smooth part of
the feather shaped crystal (Fig. 7a-1); the main compositions were
Mg, P, O and C while calcium concentration was very low at 0.4%.
The second analyzed area was at the ne crystal (Fig. 7a-2). It was
found that the relative calcium concentration was more than 5%.
The difference at the elemental composition conrmed the
presence of two distinct entities. For further verication, the

2745

B. Liu et al. / Chemosphere 93 (2013) 27382747

(a-1) Synthetic urine + brine

(a-2) Synthetic urine +brine

(b-1) Realurine + brine

(b-2) Realurine + brine

Fig. 7. EDX images of precipitates after the addition of brine to synthetic and real urine.

Table 4
Compositions of the analyzed areas by EDX.
Element atomic (%)

Mg
P
Ca
O
C

Brine addition to synthetic urine

Brine addition to real urine

a-1 image

a-2 image

b-1 image

b-2 image

10.6
11.3
0.4
59.1
18.5

5.0
3.5
5.5
65.1
20.4

7.0
11.3
0.4
39.7
41.3

1.5
0.7
11.1
63.1
22.9

Table 5
Struvite purity of the precipitates obtained from synthetic and real urine using different magnesium sources.
Mg2+ sources

Crystal: Mg/Ca ratio

Struvite purity, %

MgCl2

99.7

Synthetic urine

Real urine

Seawater

Brine

MgCl2

Seawater

Brine

4.7/1
92.8

6.8/1
94.3

99.9

8.0/1
95.6

10.4/1
96.6

Mg/P molar ratio at the rst analyzed area was nearly 1:1 that is
consisted with the stoichiometric proportion of struvite. According
to previous FT-IR and XRD results, it can be concluded that the
feather shaped crystal was struvite. At the second analyzed area,
the Mg/P molar ratio reveals that Mg was in excess compared to
P, which partially incorporated with calcite crystals. This suggests
that the ne crystal was Mg-calcite.
For real urine, the rst analyzed area was at the cofn-shaped
crystal (Fig. 7b-1). The main elements were Mg, P, C and O and
the relative concentration of calcium was low at 0.4%. The second
analyzed area was an inter-growth rhombohedral crystal with
rounded morthologies indicating calcite formation (Fig. 7b-2).
The main elements were Ca, C and O and the relative concentration

of calcium was 11.4%. Combined with FT-IR and XRD, the cofnshaped was struvite and the ne crystal was CaCO3. In each case
the formation of ne crystals on the surface of struvite could explain the multitude of background noise observed previously on
the XRD patterns.
3.2.3. Struvite quantication
The struvite content in the precipitates (purity) is very important for the successful and economical implementation of precipitation process. Precipitates obtained from MgCl2, seawater and brine
addition to synthetic and real urine were analyzed by XRF. The main
elements of these precipitates were Mg, P and O with an Mg/P/O
ratio of about 1:1:4, which complied with the stoichiometric

2746

B. Liu et al. / Chemosphere 93 (2013) 27382747

proportion of struvite. When MgCl2 was added to synthetic and real

urine, struvite purity was 99.7% and 99.9%, respectively (Table 5).
When seawater and brine were added, the precipitates contained
calcium crystals. It seems that the higher Mg/Ca ratio in the crystals,
the higher struvite purity was derived. According to this, precipitates with Mg/Ca molar ratio of 4.7:1 and 6.8:1 from seawater
and brine addition in synthetic urine had 92.8% and 94.3% struvite
purity, respectively. For real urine, precipitates with Mg/Ca molar
ratio of 8.0:1 and 10.4:1 from seawater and brine addition had
95.6% and 96.6% struvite purity, respectively. This study conrmed
that the initial calcium concentration affected struvite purity. Similar ndings were observed in previous studies (Song et al., 2007).
As indicated in Table 1, the Mg/Ca ratio in seawater and brine
was 3:1 and 3.3:1, respectively. Using seawater or brine as additional magnesium sources, calcium is inevitably added into the
system. Due to the high carbonate concentration in urine solution,
the addition of calcium could lead to the formation of calcite or
Mg-calcite. Besides, calcium precipitation was affected by Ca2+
concentration in the solution. The lower the Ca2+ concentration
was, the lesser the precipitate formed. Since phosphate concentration in real urine was relatively lower compared to synthetic urine,
the required amount of Mg2+ was equally lower. Therefore, the
amount of Ca2+ inevitably added was accordingly lower, which
led to less formation of calcite. This might explain the higher struvite purity in real urine. In this view, bittern could be another
promising magnesium source since the remaining calcium in bittern was very limited. Hence, the formation of calcite or Mg-calcite
impurities would not signicantly compromise the struvite purity.
However, bittern is not available in local context, therefore, it was
not considered as a cheap alternative magnesium source.
It is proven that both seawater and brine are low-cost magnesium sources in the local context. The struvite comprised more
than 95% of the total precipitates under all conditions in this study,
which still make the precipitate mixture benecial in the actual
fertilizer application. This indicated the great potential of engaging
seawater or desalination brine for large scale struvite harvesting.
The impurities, mainly the calcite or Mg-calcite, might affect the
fertilizing efciency of the precipitates comparing to the struvite
harvested by adding MgCl2. It is still considered economical and
benecial given the lower cost of the raw materials. In addition,
Sakthivel et al. (2012) have observed high calcite production when
wood ash was used as magnesium source. This kind of impurity
can be used as soil conditioner in acidic soils.

4. Conclusions
The present study demonstrated that both seawater and desalination brine can be considered as low-cost magnesium source to
recover phosphorus from hydrolyzed urine. High phosphate recovery efciency (>99%) was achieved at Mg/P molar ration of 1.1
without further pH adjustment (pH 9.3). The induced precipitate
was mainly composed of cofn-shaped struvite in real urine and
feather-shaped crystal in synthetic urine. For real urine, ne calcite
crystals were formed on the surface of struvite and the nal product contained roughly 90% of struvite and 10% of CaCO3. For synthetic urine, Mg-rich calcite was formed along with the struvite.
It indicated the calcium in the magnesium sources could compromise the purity of struvite product. Although the presence of calcite or Mg-calcite in the fertilizer might make it slightly
ineffective, the nal products can be used as phosphorus enhanced
conditioner in acidic soils.
The outcomes of this project can be used for the development of
P-recovery process in industrial-scale. Both seawater and brine are
low-cost magnesium sources to be incorporated into sourceseparating sanitation systems. However, there are implications of

the P-recovery process such as the high ionic concentrations in

the efuent. Among these ions, ammonium can be further recovered by air stripping to form of ammonium sulphate as proposed
by other studies (Saracco and Genon, 1994). The rest can be discharged to the sewer system or further treated. Considering the
small volume of urine stream compared to the total domestic
wastewater, the efuent from nutrient recovery process might
not impose restriction for it application.
Acknowledgements
This study is supported by the National Research Foundation,
Singapore; program number NRF-CRP5-2009-02, for the School of
Civil and Environmental Engineering/Residues and Resource Reclamation Centre, Nanyang Technological University, Singapore. We
appreciate Mr. Bernard Ng and Ms. Peijun Zhao for their assistance
in the experimental work.
References
Ahmed, M., Arake, A., Hoey, D., Thumarukudy, M.R., Goosen, M.F.A., Al-Haddabi, M.,
Al-Belushi, A., 2003. Feasibility of salt production from inland RO desalination
plant reject brine: a case study. Desalination 158 (13), 109117.
Ali, M.I., Schneider, P.A., Hudson, N., 2003. Nutrient recovery from piggery efuents.
Int. J. Eng. Model. 16 (12), 7176.
Berner, R.A., 1975. The role of magnesium in the crystal growth of calcite and
aragonite from sea water. Geochim. Cosmochim. Ac. 39 (4), 489494.
Bhuiyan, M.I.H., Mavinic, D.S., Koch, F.A., 2008. Thermal decomposition of struvite
and its phase transition. Chemosphere 70 (8), 13471356.
Buchanan, J.R., Mote, C.R., Robinson, R.B., 1994. Thermodynamics of struvite
formation. T. Am. Soc. Agr. Eng. 37 (2), 617621.
Case, D.H., Wang, F., Giammar, D.E., 2011. Precipitation of magnesium carbonates as
a function of temperature, solution composition, and presence of a silicate
mineral substrate. Environ. Eng. Sci. 28 (12), 881889.
Celen, I., Buchanan, J.R., Burns, R.T., Robinson, R.B., Raman, R.D., 2007. Using a
chemical equilibrium model to predict amendments required to precipitate
phosphorus as struvite in liquid swine manure. Water Res. 41 (8), 16891696.
Chauhan, C.K., Joshi, M.J., 2013. In vitro crystallization, characterization and growthinhibition study of urinary type struvite crystals. J. Cryst. Growth 362, 330337.
Cheng, W., Li, Z., 2010. Controlled supersaturation precipitation of hydromagnesite
for the MgCl2Na2CO3 system at elevated temperatures: chemical modeling and
experiment. Ind. Eng. Chem. Res. 49 (4), 19641974.
Cordell, D., Drangert, J.O., White, S., 2009. The story of phosphorus: global food
security and food for thought. Global Environ. Chan. 19 (2), 292305.
El Diwani, G., El Rae, S., El Ibiari, N.N., El-Aila, H.I., 2007. Recovery of ammonia
nitrogen from industrial wastewater treatment as struvite slow releasing
fertilizer. Desalination 214 (13), 200214.
Etter, B., Tilley, E., Khadka, R., Udert, K.M., 2011. Low-cost struvite production using
source-separated urine in Nepal. Water Res. 45 (2), 852862.
Ferguson, J.F., McCarty, P.L., 1971. Effects of carbonate and magnesium on calcium
phosphate precipitation. Environ. Sci. Technol. 5 (6), 534540.
Gadekar, S., Pullammanappallil, P., 2010. Validation and applications of a chemical
equilibrium model for struvite precipitation. Environ. Model. Assess. 15 (3),
201209.
Grifth, D.P., Musher, D.M., Itin, C., 1976. The primary cause of infection-induced
urinary stones. Invest. Urol. 13 (5), 346350.
Gustafsson, J.P., 2011. Visual MINTEQ, ver. 3.0. <http://www2.lwr.kth.se/English/
OurSoftware/vminteq/index.html>.
Harada, H., Shimizu, Y., Miyagoshi, Y., Matsui, S., Matsuda, T., Nagasaka, T., 2006.
Predicting struvite formation for phosphorus recovery from human urine using
an equilibrium model. Water Sci. Technol. 54 (8), 247255.
Johnston, A.E., Richards, I.R., 2003. Effectiveness of different precipitated
phosphates as phosphorus sources for plants. Soil Use Manage. 19 (1), 4549.
Jnsson, H., Stintzing, A.R., Vinneras, B., Salomon, E., 2005. Guidelines on the use of
urine and faeces in crop production. J. Indian Wat. Work. Assoc. 37 (4), 293
295.
Kangal, O., Gney, A., 2006. A new industrial mineral: huntite and its recovery.
Miner. Eng. 19 (4), 376378.
Kumashiro, K., Ishiwatari, H., Nawamura, Y., 2001. A pilot plant study on using
seawater as a magnesium source for struvite precipitation. In: Second
International Conference on Recovery of Phosphates from Sewage and Animal
Wastes. Noordwijkerhout, Netherlands.
Larsen, T.A., Gujer, W., 1996. Separate management of anthropogenic nutrient
solutions (human urine). In: Proceedings of the 1996 18th Biennial Conference
of the International Association on Water Quality. Part 2. Singapore. pp. 8794.
Larsen, T.A., Alder, A.C., Eggen, R.I.L., Maurer, M., Lienert, J., 2009. Source separation:
will we see a paradigm shift in wastewater handling? Environ. Sci. Technol. 43
(16), 61216125.
Le Corre, K.S., Valsami-Jones, E., Hobbs, P., Parsons, S.A., 2005. Impact of calcium on
struvite crystal size, shape and purity. J. Cryst. Growth 283 (34), 514522.

B. Liu et al. / Chemosphere 93 (2013) 27382747

Lee, S.I., Weon, S.Y., Lee, C.W., Koopman, B., 2003. Removal of nitrogen and
phosphate from wastewater by addition of bittern. Chemosphere 51 (4), 265
271.
Li, X.Z., Zhao, Q.L., 2002. MAP precipitation from landll leachate and seawater
bittern waste. Environ. Technol. 23 (9), 9891000.
Loste, E., Wilson, R.M., Seshadri, R., Meldrum, F.C., 2003. The role of magnesium in
stabilising amorphous calcium carbonate and controlling calcite morphologies.
J. Cryst. Growth 254 (12), 206218.
Mamais, D., Pitt, P.A., Cheng, Y.W., Loiacono, J., Jenkins, D., 1994. Determination of
ferric chloride dose to control struvite precipitation in anaerobic sludge
digesters. Water Environ. Res. 66 (7), 912918.
Maurer, M., Pronk, W., Larsen, T.A., 2006. Treatment processes for source-separated
urine. Water Res. 40 (17), 31513166.
Muchi, M., Matsumoto, Y., 1979. Artinite from Hirotani, Hojo-machi, Fukuoka
Prefecture. Japan. Bull. Fukuoka Univ. Ed. III, 119126.
Musvoto, E.V., Wentzel, M.C., Loewenthal, R.E., Ekama, G.A., 2000. Integrated
chemical-physical processes modelling I. Development of a kinetic-based
model for mixed weak acid/base systems. Water Res. 34 (6), 18571867.
Rmer, W., 2006. Plant availability of P from recycling products and phosphate
fertilizers in a growth-chamber trial with rye seedlings. J. Plant Nutr. Soil Sc.
169 (6), 826832.

2747

Ronteltap, M., Maurer, M., Gujer, W., 2007. Struvite precipitation thermodynamics
in source-separated urine. Water Res. 41 (5), 977984.
Sakthivel, S.R., Tilley, E., Udert, K.M., 2012. Wood ash as a magnesium source for
phosphorus recovery from source-separated urine. Sci. Total. Environ. 419, 68
75.
Saracco, G., Genon, G., 1994. High temperature ammonia stripping and recovery
from process liquied wastes. J. Hazard. Mater. 37 (1), 191206.
Smil, V., 2000. Phosphorus in the environment: natural ows and human
interferences. Annu. Rev. Energ. Env. 25, 5388.
Song, Y., Yuan, P., Zheng, B., Peng, J., Yuan, F., Gao, Y., 2007. Nutrients removal and
recovery by crystallization of magnesium ammonium phosphate from synthetic
swine wastewater. Chemosphere 69 (2), 319324.
Zhang, J., Giannis, A., Chang, V.W.C., Ng, B.J.H., Wang, J.Y., 2013. Adaptation of urine
source separation in tropical cities: process optimization and odor mitigation. J.
Air Waste Manage. Assoc. 63 (1), 471481.
Zhou, J., Chang, V.W.-C., Fane, A.G., 2013. An improved life cycle impact assessment
(LCIA) approach for assessing aquatic eco-toxic impact of brine disposal from
seawater desalination plants. Desalination 308, 233241.