CEED Seminar Proceedings 2018 Gong: Magnesium Recovery from Desalination Brine

Magnesium Recovery from Desalination Brine

Maurice Hao Gong

Michael Johns, Einar Fridjonsson

Department of Chemical Engineering
The University of Western Australia

Phil Heckley
CEED Client: Water Corporation

Abstract

Recovering magnesium hydroxide Mg(OH)2 from the discharged brine from desalination
plants has the potential to generate economic, social and environmental benefits. While the
concept has not been practised in the industry, it is expected that it can be achieved by
adaptation of a conventional method of precipitating Mg(OH)2 from seawater through the
addition of lime. This project involves a series of experiments to investigate the feasibility
of producing commercial grade Mg(OH)2 from desalination brine utilising this process.
The key parameters of interest include the settling rates, the yield of products, and
Mg(OH)2 purity. These parameters are being assessed under varying experiment
conditions, such as increasing the lime dosage, screening of lime slurry and the use of
flocculants. Initial results showed that a major problem with the process is a low product
purity, generally below 80% Mg(OH)2. In a recent experiment, however, where the slaked
lime slurry was screened, a considerable increase in the product purity has been achieved.
Further research is to be undertaken to assess the effectiveness of screening lime slurry
and other experimental conditions. These results will provide valuable information as to
the product quality and the performance of this process for Water Corporation.

1. Introduction
With a declining rainfall and a rising population, Western Australia has become increasingly
reliant on desalination for its potable water supply. The desalination process extracts fresh water
from seawater and produces concentrated seawater, also known as brine, as a waste stream. The
two desalination plants in Western Australia have a combined capacity of 150 billion litres of
fresh water per year (Water Corporation, 2017), accompanied by a massive amount of waste
brine discharged back into the ocean. This waste stream, however, contains magnesium ion
(Mg2+) up to 2.51 kg/m3, which could be converted into valuable products such as magnesium
hydroxide (Mg(OH)2). Mg(OH)2 has a versatile range of applications within environmental
protection, fire retardant, and pharmaceutical fields. (QYR Chemical & Material Research
Center, 2017). Recovering magnesium hydroxide from brine has the potential to generate
additional economic, social, and environmental benefits. For these reasons, this concept has
recently attracted interest from Water Corporation.

1.1 Magnesium Hydroxide Production from Seawater

Seawater precipitation is one of the conventional methods of producing Mg(OH)2 currently
used in industry (QYR Chemical & Material Research Center, 2017). This method uses

49
CEED Seminar Proceedings 2018 Gong: Magnesium Recovery from Desalination Brine

‘naturally occurring’ seawater as a raw material, with the addition of an alkaline reagent, to
precipitate magnesium hydroxide particles. The alkaline reagents commonly used are caustic
soda (NaOH), quicklime (CaO), or ammonia (NH3) (QYR Chemical & Material Research
Center, 2017). The main reaction that occurs in this process is:

𝑀𝑔2+ + 2𝑂𝐻 − → 𝑀𝑔(𝑂𝐻)2 ↓

The solid products that subsequently form can be separated from the liquids by a series of
settling thickeners, washing tanks, and finally by filtration. Depending on its intended
application, the filter cake can then either be processed into a slurry form of approximately 60
wt% solid, or into its powder form through a drying operation. Commercial grade Mg(OH)2
generally have a purity of between 95% to 100% (QYR Chemical & Material Research Center,
2017).

Because of the relative maturity of this technology, it is expected that the same process will be
implemented to produce Mg(OH)2 from brine. Furthermore, low-cost sources of quicklime are
readily available to Water Corporation through local WA suppliers. Therefore, this study will
only investigate the case where such quicklime is used as the alkaline reagent.

1.2 Desalination Brine Characterisation and Implications

Desalination brine is comprised of the same major constituents as the seawater, but at
concentrations approximately twice as high. Table 1 shows the comparative compositions of
the brine and seawater. Because brine is significantly richer in Mg2+ than seawater, an inherent
advantage of this process will be the decreased equipment volume requirements.

Component Concentration in brine, g/L Concentration in seawatera, g/L

Na+ 18.6b 10.5

Mg2+ 2.54c 1.29
Ca2+ 0.857c 0.410
SO42- 5.70c 2.72
HCO3- 0.273b 0.150
Cl- 34.0b 18.7
a.Sampled in the seawater intake of the Perth Seawater Desalination Plant.
b.Source of data: preliminary experiments by inductively coupled plasma atomic emission spectroscopy
(ICP-AES) (ChemCentre WA, 2018).
c. Source of data: hand calculations based on the designed recovery ratio of the desalination plant.
Source of the rest of the data: Water Corporation existing test results (SGS Ltd., 2017)
Table 1 Concentrations of major ions relevant to this study in desalination
brine, and in seawater.

As a result of these higher concentrations, a problem associated with the use of brine in place
of seawater is the increased tendency to form solid impurities. Impurities that might precipitate
with the addition of lime include gypsum (CaSO42H2O) and calcium carbonate (CaCO3). To
eliminate CaCO3 impurity, the brine is normally pretreated by acidification to 𝑝𝐻 = 4 and
purged of CO2 (Shand, 2006). An analysis of the brine system with a thermodynamic modelling
package JESS V8.7 indicated that the starting brine solution is already supersaturated with
CaSO4 (Rowland, 2018), and that this is most likely to precipitate out along with Mg(OH)2
based on solution equilibria.

50
CEED Seminar Proceedings 2018 Gong: Magnesium Recovery from Desalination Brine

1.3 Quicklime Characterisation and Slaking Process

The study used quicklime supplied by Cockburn Cement, commercially known as Quicklime
Munster. The main constituent of Quicklime Munster is calcium oxide (CaO), at ~83 wt%.
Other notable constituents of the quicklime include silicon oxide (SiO4), aluminum oxide
(Al2O3), iron oxide (Fe2O3), magnesium oxide (MgO), and sulfur trioxide (SO3).
As a pretreatment step, quicklime is generally slaked with water to produce a slurry of hydrated
lime, where the main constituent is Ca(OH)2. Slaking is vital as water serves as a medium for
easy transportation and release of the hydroxide ions (Oates, 2008). The reactions that occur
are:

𝐶𝑎𝑂 + 𝐻2 𝑂 → 𝐶𝑎(𝑂𝐻)2

𝐶𝑎(𝑂𝐻)2 → 𝐶𝑎2+ + 2𝑂𝐻 −

The effectiveness of the slaked lime may be indicated by the particle size distribution of the
Ca(OH)2 particles. A previous study found that the utilisation efficiency of quicklime, defined
as the mass portion of quicklime that participates in the reaction, correlates most strongly with
the particle size of the quicklime after slaking (Ineich et al., 2017). It is thus, highly desirable
to produce a lime slurry with finer particles, which is associated with higher reactivity.
Among all of the factors that might affect the particle size of slaked lime, temperature during
slaking is reported to be the most critical (Hassibi, 1999). While higher temperatures are
desirable to produce finer particles, the temperature should not reach the limit where local hot-
spots can form causing detrimental effects on slaking (Oates, 2008). Furthermore, the particle
size of the lime slurry can be reduced by ‘sieving’ (size reduction) of either the hydrated lime
slurry or the dry quicklime.

1.4 Present State of the Art

There is currently no commercial application of producing Mg(OH)2 from desalination brine
cited in the literature. In the case of using lime or Ca(OH)2 as the precipitating agent, the few
studies found mentioned the main risks of gypsum impurity and poor filterability (Barba et al.,
1980, Cipollina et al., 2014). Barba (Barba et al., 1980) recommended an alternative process
which essentially produces MgCO3 solid and further converts it into the desired product.
However, within our client’s context, it is desirable to utilise an existing mature technology.
Hence, this study will investigate the feasibility of recovering commercial grade Mg(OH)2 from
desalination brine using industrial quicklime, with no significant changes to the traditional
seawater precipitation process.

2. Methodology
This study is being executed through a series of experiments which aim to replicate the
industrial process of Mg(OH)2 production. A general experimental procedure is illustrated in
Figure 1. Brine has been sampled directly from the Perth Seawater Desalination Plant in March
2018, and a sample of Quicklime Munster was supplied by Cockburn Cement. The lime slaking
procedure was developed with reference to the Australian Standard AS4489.3.1. This procedure
essentially involves mixing quicklime with water at a ratio of 1:4 in an Erlenmeyer flask, and
allowing the reaction to occur under agitation at 400 rpm until the reaction is completed.

51
CEED Seminar Proceedings 2018 Gong: Magnesium Recovery from Desalination Brine

Figure 1 Schematic illustration of a typical experiment procedure.

After adding the hydrated lime slurry to the pretreated brine, the reactions occurred under
agitation. When the reaction was complete, the solid was allowed to settle to the bottom, and
the settling rate was measured. A small portion of the mixture was also sampled for particle
size distribution analysis with a Malvern Mastersizer 2000.
The final product was then further separated from the liquid by filtration, washing to remove
soluble salts, and oven drying. To measure the purity of the resulting products, two methods
were employed:
1. A small portion of the solid product with a known mass was digested by hydrochloric
acid (HCl). The Mg2+ concentration of the digested solution was determined by a
benchtop spectrometer Hanna HI83300. This value was used to back-calculate the
Mg(OH)2 mass fraction in the solid. To verify the validity of this result, the Mg2+
concentration in the liquid filtrate was also determined with the same spectrometer,
enabling a mass balance to be performed.
2. The bulk solid product composition was also analysed with acid digestion of the solid,
followed by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES).

2.1 Experimental Framework

The following framework outlines two sets of experiments discussed in this paper, built on the
above procedure.

2.1.1 Effects of Lime Slurry Dosage

Studies have reported that the dosage ratio of the alkaline agent has considerable effects on the
composition of the solid products as well as sedimentation rates (Petric and Petric, 1980)
(Robinson et al., 1946). Hence, this set of experiments aims to investigate the quality of the
product when lime slurry dosage varies from 60% to 120% stoichiometrically equivalent to
Mg2+ in brine.

2.1.2 Effects of Slaking Condition

Slaking conditions are critical to the reactivity of slaked lime, which affects the utilisation
efficiency of quicklime and the Mg(OH)2 product quality. Slaking conditions to be varied in
this set of experiments include lime to water ratio, screening of slaked lime by sieves, slaking
water temperature, and the possibility of using the original brine as the slaking water.
Other experiments that have or will be undertaken will investigate the effects of addition of
seed material, and the effects of flocculant usage.

52
CEED Seminar Proceedings 2018 Gong: Magnesium Recovery from Desalination Brine

3. Preliminary Results and Discussions

In a set of experiments to study the effect of the Mg(OH)2 purity of the precipitate, the results
indicated that the purities of Mg(OH)2 were significantly generally low, as shown in Table 2.
Based on the results, a lower dosage of lime slurry appeared to associate with a higher product
purity, consistent with the trend reported in another study (Petric and Petric, 1980). Moreover,
gypsum contents tend to increase at higher lime slurry dosages, likely due to the increased
supersaturation of Ca2+ ions. The CaCO3 contents, however, were significantly higher than
expected. In theory, dissolved carbonates were removed under acidification pretreatment of
brine, and CaCO3 precipitation will be minimal. In a further experiment using NaOH instead
of lime as the precipitating agent, CaCO3 content remains ~10%, which eliminated the
hypothesis that the CO32- was introduced by the lime. It is suspected that the causes of the high
CaCO3 contents are 1) ineffective removal of CO2 in the acidification pretreatment step; 2)
atmospheric CO2 absorption during the reaction.

Lime slurry dosage, molar 60% 80% 100% 120%

equivalence to Mg2+ available in
the initial brine solution
Purity, mass fraction of Mg(OH)2 66.3% 65.4% 61.7% 62.3%
CaCO3, mass fraction 20.3% 12.9% 13.7% 15.3%
CaSO42H2O, mass fraction 7.30% 8.78% 11.3% 12.6%
Table 2 Purity of Mg(OH)2 at varying lime slurry dosages by ICP-OES.

In another set of experiments, significantly improved product purities were achieved by wet
screening of the lime slurry. As shown in Table 3, the resulting Mg(OH)2 has a purity reaching
91%, when the lime slurry is screened by a 100m sieve with before the addition to brine. These
results implied that a significant contribution to the impurities in products originates from
coarser particles in the lime slurry. Furthermore, high purity Mg(OH)2 may be recovered from
brine through the use of wet-screened lime slurry.

No Screening Underflow of a Underflow of a

250 m sieve 100 m sieve
Purity, mass fraction of Mg(OH)2 68% 79% 91%
Table 3 Purity of Mg(OH)2 when the lime slurry is directly dosed, wet
screened through a 250m sieve and a 100m sieve respectively.

4. Conclusions and Future Work

Much of the early work has been dedicated to developing an experimental methodology with
good reproducibility and consistent mass balance calculations. A functioning experimental
procedure with decent reproducibility has been developed, which enables Mg(OH)2 product
quality to be measured under varying experimental conditions. Following this framework, the
major shortcoming of the proposed process of recovering Mg(OH)2 was a low purity obtained.
However, it has been shown that wet-screening of lime slurry significantly improves the
product purity.

Along with wet screening, it is believed a higher water temperature during lime slaking can
further improve the purity of the product, which will be the next phase of the study. It is noted
that the Mg(OH)2 purity measured by the spectrometer has limited value. As a result, it is crucial
to analyse the elemental composition of these solid products with ICP-OES, which enables the

53
CEED Seminar Proceedings 2018 Gong: Magnesium Recovery from Desalination Brine

impurity contents to be identified. The experimental work will follow the framework outlined
above, to investigate the performance of the process with varying lime preparation conditions
and the addition of flocculants. In the long term, it will be worthwhile for further study,
investigating alternative processes to recover magnesium. For example, it has been argued that
recovering magnesium in the form of MgCO3 is more advantageous than the current process
(Barba et al., 1980).

5. Acknowledgements
I would like to express my sincere appreciation to my company mentor Phil Heckley and my
academic supervisor Michael Johns who have given me tremendous support and guidance
throughout this project. I am grateful to the team at CEED and Water Corporation, for the
opportunity I have been given to developing my professional skills over the course of this
project.

6. References
Barba, D., Brandani, V., Giacomo, G. D. and Foscolo, P. U. (1980) 'Magnesium oxide
production from concentrated brines', Desalination, 33, pp. 241-250.
ChemCentre WA (2018) Report of Examination, Perth, AU.
Cipollina, A., Bevacqua, M., Dolcimascolo, P., Tamburini, A., Brucato, A., Glade, H., Buether,
L. and Micale, G. (2014) 'Reactive crystallisation process for magnesium recovery from
concentrated brines', Desalination and Water Treatment, 55(9), pp. 2377-2388.
Hassibi, M. 'An overview of lime slaking and factors that affect the process'. Presentation to
3rd International Sorbalit Symposium.
Ineich, T., Degreve, C., Karamoutsos, S. and du Plessis, C. (2017) 'Utilization efficiency of
lime consumption during magnesium sulfate precipitation', Hydrometallurgy, 173, pp.
241-249.
Oates, J. A. (2008) Lime and limestone: chemistry and technology, production and uses. John
Wiley & Sons.
Petric, B. and Petric, N. (1980) 'Investigations of the rate of sedimentation of magnesium
hydroxide obtained from seawater', Industrial & Engineering Chemistry Process
Design and Development, 19(3), pp. 329-335.
QYR Chemical & Material Research Center (2017) Global Magnesium Hydroxide Market
Research Report 2017.
Robinson, H. A., Friedrich, R. E. and Spencer, R. S. 1946. Magnesium hydroxide from sea
water. Google Patents.
Rowland, D. 2018. (Unpublished) Saturation Results of Lime Addition to Brine with JESS
V8.7.
SGS Ltd. (2017) Water Quality Report of Seawater Intake in PSDP (000233779, JW164560).
Shand, M. A. (2006) The chemistry and technology of magnesia. John Wiley & Sons.
Water Corporation (2017) 2017 Annual Report.

54