This document evaluates three weak organic acids (formic, acetic, and DL-lactic acid) for extracting calcium from wollastonite at temperatures between 22°C and 80°C. Experiments found that formic acid achieved the highest calcium extraction rate of 26(±7) × 10−5 mol m−2 s−1 at 80°C, followed by acetic acid and DL-lactic acid. Activation energies indicated the initial dissolution in formic acid was diffusion controlled, while kinetic limitations controlled dissolution in acetic and DL-lactic acids. Formic acid showed the best performance for calcium extraction and potential for mineral carbonation.

1. Selection of acid for weak acid processing of wollastonite
for mineralisation of CO2
Manisha Ghoorah a
, Bogdan Z. Dlugogorski b,⇑
, Reydick D. Balucan c
, Eric M. Kennedy a
a
Priority Research Centre for Energy, Faculty of Engineering and Built Environment, ATC Building, The University of Newcastle, Callaghan, NSW 2308, Australia
b
School of Engineering and Information Technology, Murdoch University, Murdoch, WA 6159, Australia
c
School of Chemical Engineering, The University of Queensland, St. Lucia, QLD 4072, Australia
h i g h l i g h t s
Results report dissolution of wollastonite in formic, acetic and DL-lactic acids.
Formic acid extracted 96% of Ca at 80 °C at a rate of 26(±7) Â 10À5
mol mÀ2
sÀ1
.
Activation energy for formic acid corresponds to 11 ± 3 kJ molÀ1
.
Ca2+
dissolution appears to be mass-transfer controlled with formic acid.
a r t i c l e i n f o
Article history:
Received 13 September 2013
Received in revised form 6 January 2014
Accepted 6 January 2014
Available online 21 January 2014
Keywords:
Mineral carbonation
Formic acid
Dissolution
Wollastonite
a b s t r a c t
Typically, mineral carbonation comprises aqueous phase reactions involving the dissolution of naturally
occurring magnesium and calcium silicate rocks, such as olivine, serpentinites and wollastonite, followed
by the precipitation of magnesium and calcium carbonate minerals. In this report, we evaluated the effect
of formic, acetic and DL-lactic acids on the calcium-leaching process from wollastonite between 22 °C and
80 °C and at atmospheric pressure. OLI Analyzer Studio 3.0 predicted equilibrium conversions of calcium
and its speciation in the aqueous phase. Additionally, we measured dissolution rates, for a constant pH
system, as a function of temperature for the three organic acids. All experiments involved the reaction
of 17 ± 1 lm (volume mean diameter) ground rock samples with the acids in a stirred batch reactor
equipped with in situ pH measurements. Inductively coupled plasma-optical emission spectrometry
(ICP-OES) analysed the concentration of calcium ions in the leaching medium while scanning electron
microscopy/energy dispersive spectroscopy (SEM/EDS) examined the morphology and surface chemical
composition of the residual solid phase from dissolution experiments. We estimated the maximum dis-
solution rates of wollastonite in the limit of low but achievable pH and in the absence of diffusion lim-
itation in pores and cracks of the SiO2 skin. At 80 °C, these rates correspond to 26(±7) Â 10À5
,
14(±3) Â 10À5
and 17(±4) Â 10À5
mol mÀ2
sÀ1
for formic, acetic and DL-lactic acids, respectively. The
apparent activation energies amount to 11 ± 3, 47 ± 13 and 52 ± 14 kJ molÀ1
for dissolution in formic, ace-
tic and DL-lactic acids, respectively. These values indicate the initial diffusion limitation in the film
around wollastonite particles for formic acid, and kinetic limitation for acetic and DL-lactic acids. The
rates of dissolution rapidly decline for acetic and DL-lactic acids, but remain high for formic acid. The
findings are altogether indicative of high performance of formic acid for extraction of Ca2+
for storing
CO2. Further experiments are needed to assess the recycling of formic acid to determine its overall suit-
ability as a Ca2+
carrier for the weak acid process.
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
Although thermodynamic analyses indicate that carbonate
formation reactions can proceed spontaneously due to their
exoergicity, the observed reaction rates are extremely slow under
mild conditions. Therefore the engineering challenge primarily
hinges on improving throughput rates as well as minimising capi-
tal and energy expenses by speeding up the reaction kinetics by
several orders of magnitude thus allowing the entire process to
take place on a large-scale basis. Decreasing particle size through
pulverisation, raising reaction temperature and pressure, changing
http://dx.doi.org/10.1016/j.fuel.2014.01.015
0016-2361/Ó 2014 Elsevier Ltd. All rights reserved.
⇑ Corresponding author. Tel.: +61 8 9360 6770.
E-mail address: B.Dlugogorski@murdoch.edu.au (B.Z. Dlugogorski).
Fuel 122 (2014) 277–286
Contents lists available at ScienceDirect
Fuel
journal homepage: www.elsevier.com/locate/fuel

2. solution chemistry and using catalysts/additives, altogether accel-
erate reaction rates [1–4]. In addition, heat activation of serpentine
minerals (Mg3Si2O5(OH)4) between 600 and 750 °C removes part of
hydroxyl groups and amorphises the mineral structure, signifi-
cantly increasing their reactivity to CO2 [5].
The most comprehensive studies so far outline two thermody-
namically feasible approaches to CO2 mineralisation in the aque-
ous phase. The first process, developed by the Albany Research
Centre, involves direct carbonation in aqueous solutions of
0.64 M NaHCO3 and 1 M NaCl conducted at 150 atm CO2 and
155 °C and 185 °C for heat pre-treated serpentinite and finely-
ground olivine, respectively [1,6]. The second approach is based
on two principal steps namely silicate dissolution, usually by acids,
and carbonate precipitation [7]. Since the former mechanism is
generally assumed to be rate-limiting with respect to the overall
carbonation process, many studies have focused on the extraction
of calcium or magnesium from native minerals [8–10].
Moreover, extensive research has been undertaken to reveal
and subsequently enhance dissolution kinetics. While Golubev
and co-workers (2005) [11] inferred that the presence of bicarbon-
ate ions, at concentrations of 0.01–0.1 M and pH 7–8, would en-
hance dissolution rates, Krevor and Lackner [12] concluded that
sodium salts of citrate, oxalate and EDTA considerably increased
the degree of dissolution. Increasing NaHCO3 concentration report-
edly promotes brucite dissolution, measured in terms of etch pit
spreading rates – from 0.038 ± 0.004 nm sÀ1
at 10À5
M (pH 7.2)
to 0.38 ± 0.07 nm sÀ1
at 1 M (pH 9.3) [13]. Similarly, Pokrovsky
et al. [14] observed a catalysing effect of HCOÀ
3 on brucite dissolu-
tion owing to the formation of surface complexes that weaken
Mg–O bonds and water coordination to Mg atoms at the surface.
Experimental investigations coupled with kinetic modelling have
also been performed with a view to estimating dissolution rates
of basic silicates at conditions relevant to geologic CO2
sequestration [15,16].
Lackner et al. [17] initially investigated the hydrochloric acid-
aided dissolution, to leach out magnesium ions. This scenario
was highly energy intensive and was thus phased out by a novel
process using acetic acid as an accelerating medium for the artifi-
cial weathering of wollastonite. The thermodynamic consideration
that the extraction acid must not only be stronger than silicic acid
but also weaker than carbonic acid, such that the precipitation of
carbonates occurs spontaneously, contributed towards the selec-
tion of this acid [4]. Other weak acids have been subject to less
detailed studies [12,18–20] until 2013 when Zhao et al. [21]
explored the effect of a series of chelating agents (with concentra-
tions ranging between 0.002 and 0.006 M) including ascorbate,
acetate, gluconate, glutamate, phthalate, oxalate, iminodiacetate,
picolinate, nitrilotriacetate, citrate and EDTA. For experiments per-
formed at 22 °C, acetic acid achieved the highest extraction (about
17%) in 6 min. The authors pointed out that oxalic acid, which is
reportedly the best Mg-extracting agent, did not enhance dissolu-
tion of wollastonite.
The present study investigates the extent of calcium extraction
from wollastonite when the latter was treated with three weak
organic acids, viz. acetic, formic and DL-lactic acids, under the
effect of increasing reaction temperature. Moreover OLI Analyzer
Studio 3.0, a thermodynamic prediction software employing the
Helgeson–Kirkham–Flowers–Tangers (HKFT) equation of state
and the Bromley equation for solution non-ideality, predicted the
equilibrium calcium extraction [22]. The RCO2 value of wollastonite
stands at about 2.8 (RCO2 denotes the mass of ore required to
convert a unit mass of CO2 to a carbonate [2]), which is higher than
that of serpentine (RCO2 2.1 for antigorite), a widely studied
contender for carbonation. The former was selected because it is
a useful model mineral which offers more reactive features
towards dissolution and carbonation compared to magnesium
silicates. Experiments, performed at constant pH, determined the
maximum initial (t = 0) kinetics of acid digestion of the rock, at
low but achievable pH and in the absence of a silica layer that
builds on particle surfaces as a result of incongruent dissolution
of wollastonite. Finally, we examined the morphology and surface
chemical composition of wollastonite particles before and after
reaction.
2. Major deposits of wollastonite in New South Wales (NSW)
and Queensland (QLD)
Even though a holistic evaluation of the content of wollastonite
in many skarns have never been undertaken, records have exposed
occurrences of this mineral in significant amounts within the Lach-
lan and New England Orogens; territories situated on the eastern
coast of the Australian continent and shared mainly by the states
of QLD and NSW. Wollastonite deposits in these two states can
transform into a possible destination for carbon dioxide captured
from the major coal-fired power stations, as mapped out in
Fig. 1. Deposits of the mineral include those at Browns Creek and
Doradilla while Jeremiah Creek, Attunga Creek and Yetholme rep-
resent minor sites.
The deposit at Doradilla, the site showing the most promise,
averages 50–80 m wide with a depth of at least 200 m and extends
over 16 km [23]. Skarn-hosted deposits found in NSW usually con-
sist of about 70% garnet and 30% wollastonite [23]. Therefore,
assuming a purity of 30%, the deposit at Doradilla would offer a po-
tential source of feedstock equivalent to an estimated 178 Mt of
wollastonite. Based on an RCO2 value of 2.8, the latter is expected
to store roughly 63 Mt of CO2, which is equal to the amount of
CO2 emitted from electricity generation in NSW, in 2007 [24]. Judg-
ing from its low purity and limited quantity, wollastonite would
only provide short-term solution to CO2 storage in QLD and NSW.
Allinson et al. [25] reported that pipelining compressed CO2
gases to Eromanga Basin, another possible storage site situated
within 1000 km from the cluster of coal-fired power stations in
NSW and southern QLD, would entail a cost of 35 AU$/t of CO2.
Assuming that transport costs depend solely on the distance be-
tween the deposits and emission hubs, it can be deduced that a
cost of about 25 AU$/t would be incurred to pipeline the com-
pressed gases towards the deposit at Doradilla, located at 700 km
from the power stations. A number of smaller yet exploitable
sources of wollastonite, for instance Attunga Creek found within
300 km, are therefore not to be neglected.
3. Materials and methods
3.1. Characterisation of wollastonite sample
We procured the wollastonite specimen used for dissolution
reactions from New South Wales Pottery Supplies, Australia. Laser
particle sizing of the ground and sieved sample, which was per-
formed in aqueous media on a Malvern Mastersizer ‘‘E’’, indicated
a volume mean diameter (VMD or D[v, 0.5]) of 17 ± 1 lm, with D[v,
0.9], D[v, 0.1] and D[3, 2] of 56 ± 1, 2 ± 1 and 6 ± 1 lm, respectively.
Fig. 2 illustrates the cumulative particle size distribution of wollas-
tonite particles before and after acid dissolution. The average den-
sity of the starting material was 2.86 g cmÀ3
while its specific
surface area amounted to 0.1 m2
gÀ1
based on a low temperature
N2 adsorption BET analysis (Micromeritics Gemini).
X-ray diffraction (XRD) analysis confirmed the presence of
wollastonite as the major phase (Fig. 3). Diopside and pectolite,
appearing as minor phases, are both metasilicates like wollastonite
and crystallise in the monoclinic and triclinic systems respectively.
Table 1 lists the elemental composition derived from X-ray
278 M. Ghoorah et al. / Fuel 122 (2014) 277–286

3. fluorescence (XRF). Distributing the elemental abundances among
minerals leads to an approximate composition of 81.8% wollaston-
ite (CaSiO3), 9.2% diopside (MgCaSi2O6), 4.6% silica (SiO2), 1.9%
pectolite (NaCa2Si3O8(OH)), and possibly 0.7% hedenbergite (CaFe-
Si2O6); with diopside and hedenbergite end members forming a so-
lid solution. The remainder of about 0.5%, after accounting for the
loss on ignition, seems to include mostly aluminosilicate minerals.
3.2. Experimental procedure
We performed wollastonite (CaSiO3) dissolution experiments in
a 250 mL three-neck glass reactor, immersed in a temperature-
controlled water bath, equipped with a condenser to minimise
solution losses due to evaporation, as illustrated in Fig. 4. Two ser-
ies of experiments, incorporating a non-pH controlled and a pH
controlled system, provided the basis to determine the extent of
calcium extraction and the rates of dissolution, respectively.
The first set of experiments included reactions conducted at
temperatures ranging from 22 °C to 80 °C in an acidic leaching
medium with a concentration of 0.1 M, for a total reaction time
of 3 h. Sigma Aldrich (Australia) supplied analytical reagent grade
formic and DL-lactic acids while acetic acid was purchased from
Ajax Finechem Pty Ltd., (Australia). We prepared acid solutions in
ultrapure deionised water with electrical resistivity of 18.2 MX/
cm, by standard volumetric dilution techniques. A Hanna pH probe
and meter registered in situ pH measurements while a water bath,
mounted on a hot plate, maintained the temperature at the set
point. Continuous stirring of the slurry, accomplished through a
magnetic stirrer, ensured dispersion of the particles.
Each run consisted of charging 0.58 g of ground samples of
CaSiO3 into the batch reactor after heating 100 mL of the diluted
acid to the desired temperature. The ratio of acid to CaSiO3 was
fixed according to stoichiometry. Eqs. (1)–(3) illustrate the overall
reaction for extraction of calcium from CaSiO3 using formic acid
(HCOOH – pKa 3.75), acetic acid (CH3COOH – pKa 4.76) and DL-lac-
tic acid (CH3CHOHCOOH – pKa 3.86).
CaSiO3 þ 2HCOOH ! Ca2þ
þ 2HCOOÀ
þ H2O þ SiO2 ð1Þ
CaSiO3 þ 2CH3COOH ! Ca2þ
þ 2CH3COOÀ
þ H2O þ SiO2 ð2Þ
CaSiO3 þ2CH3CHOHCOOH ! Ca2þ
þ2CH3CHOHCOOÀ
þH2OþSiO2
ð3Þ
In solution, other ions will also exist, such as, for Reaction 1, cal-
cium monoformate Ca(HCOO)+
and calcium formate Ca(HCOO)2,
and, Reaction 2, calcium monoacetate Ca(CH3COO)+
and calcium
acetate Ca(CH3COO)2. OLI Analyzer Studio 3.0 [22] predicted only
calcium ion (Ca2+
) for Reaction 3, due to the absence of thermody-
namic data for calcium monolactate and calcium lactate in the
database of the software.
Fig. 1. Map showing the relative distance of the major carbon dioxide emitters and wollastonite deposits in eastern NSW and QLD. Different icon sizes have been used to
contrast small and large deposits [23,26].
Fig. 2. Particle size distribution of particles before and after reaction at 80 °C,
within 3 h and without pH control.
M. Ghoorah et al. / Fuel 122 (2014) 277–286 279

4. At the end of the desired test time, the suspension was filtered
through a 0.45 lm PVDF membrane and ICP-OES, which was cali-
brated using multielement standard solution that matched the fil-
trate composition, served to measure the calcium ion
concentrations in the filtrate. The extent of calcium extraction cor-
responds to the ratio of calcium concentration in the filtrate solu-
tion to the initial fraction of calcium in the feed. The filter cake was
washed with deionised water prior to drying overnight in an oven
set at 105 °C. Subsequent SEM/EDS and Malvern Mastersizer anal-
yses revealed the properties as well as surface morphology/chem-
ical composition and the particle size distribution of the reacted
particles, respectively.
The analysis of the filter cake assisted in the closure of elemen-
tal balance on calcium, silica, magnesium, iron sodium, and alu-
minium. Volumes of 4.5 mL of 65% HNO3, 4.5 mL of 37% HCl and
3 mL of 50% HBF4 were added to 0.1 g of the dried solid residue
prior to digestion in a Milestone start D microwave unit, which
yielded complete digestion after 1 h at 160 °C. ICP-OES evaluated
the metal ion composition of the clear liquid thus formed.
Another series of experiments, performed in a constant-pH sys-
tem, by employing the setup described in Fig. 4, allowed us to ob-
tain estimates of the maximum dissolution rates in the absence of
diffusional resistance in pores and cracks of the silica skin, as a
function of temperature. The investigated reaction temperatures
were 40 °C, 60 °C and 80 °C. In order to obtain maximum possible
rates, we performed the dissolution in excess amounts of 5 M
acids. Although this buffered the system’s pH, we added small
amounts of acid (not exceeding 10 mL) to the leaching medium
from time to time to adjust the pH. We rapidly injected about
0.4 g of powdered CaSiO3 in the batch stirred-vessel containing
100 mL of acid solution. A syringe afforded withdrawal of about
0.6 mL of slurry samples that immediately underwent syringe-fil-
tration through a 0.22 lm membrane at intervals of 5 min within
the total reaction time of 1 h. The combined volume of the aliquots
from any given experiment represented less than 10% of the total
volume. We minimised changes in acid concentration during
experiments by ensuring that the volumes of aliquots and added
acid for pH adjustment were kept within the stated limits. The ini-
tial dissolution rates, normalised to the specific surface area of the
feedstock, were determined from the change in calcium concentra-
tion in the sampled solution, as evaluated by ICP-OES.
4. Results and discussion
Figs. 5–7 illustrate typical pH profiles as wollastonite dissolu-
tion proceeds in acidic medium. For all runs, the pH varied in the
range of 2.0–4.5. The consumption of protons and the release of
cations from the silicate mineral characterise this process, result-
ing in the alkalisation of the reaction mixture thus increasing solu-
tion pH. Since dissolution is incongruent at such low pH [27], we
expect the concentration of calcium to be much higher compared
to other ions that could leach out, hence allowing an initial moni-
toring of the reaction course via the pH of the system.
Fig. 8, where we plot the extraction of Ca2+
at the end of 3 h of
the process, graphically summarises the measurements of Figs. 5
and 6. Evidently, the process depends strongly on temperature,
as both the diffusion and chemical-reaction rates increase with
temperature. A linear relationship was observed within the
Fig. 3. XRD spectrum of the raw material.
Table 1
Chemical composition (by weight) of wollastonite derived from XRF, excluding oxides
of less than 0.1% in abundance; the total composition in the table corresponds to
99.7%.
SiO2 MgO CaO Fe2O3 Al2O3 Na2O LOIa
53.5 1.71 42.7 0.234 0.176 0.222 1.2
a
Loss on ignition.
Heating mantle
Magnetic stirrer
Water bath
Outlet for batch addition
and solution sampling
Water-cooled condenser
Three-neck glass reactor
pH
Temperature
probes
Water inlet
Water outlet
Fig. 4. Schematic drawing of the experimental apparatus for calcium extraction
from wollastonite.
280 M. Ghoorah et al. / Fuel 122 (2014) 277–286

5. investigated temperature interval where the amount of calcium in
solution has risen by more than 55%.
In comparison to acetic and DL-lactic acids, formic acid demon-
strated a higher calcium-extracting capability attaining 96% after
3 h at 80 °C (Figs. 5 and 8). Elemental balance on calcium, silica,
magnesium, iron sodium, and aluminium was closed by analysing
the amount of these elements in solution and in the residual solid
phase at the end of the experiment. Table 2 lists the results.
Furthermore, we modelled the dissolution in formic acid at
80 °C, neglecting the presence of mineralogical impurities, on OLI
Analyzer Studio 3.0 [22] (Figs. 9 and 10). The input to the software
was similar to the initial experimental conditions; a temperature
of 80 °C, pressure of 1 atm, and wollastonite/formic acid (0.58 g/
0.46 g) in stoichiometric ratio, representing 0.2 g of calcium.
The equilibrium aqueous phase consists of the following: 0.1 g
Ca2+
, 0.14 g calcium monoformate equivalent to 0.07 g Ca2+
and
0.06 g calcium formate equivalent to 0.02 g Ca2+
. The total mass
of calcium in solution amounts to 0.19 g, which represents 95%
of the input mass. The software also predicted a solid phase con-
sisting of only silicon dioxide and pH range of 2.5–5. Experimental
data are therefore in good agreement with results from the simu-
lation, except for the measurements of pH, which appear to be sig-
nificantly lower in the experimental measurements than in
thermodynamic predictions. The difference is as high as 1 pH unit
at the end of an experiment.
Fig. 5. pH profile for dissolution in 0.1 M formic acid.
Fig. 6. pH profile for dissolution in 0.1 M acetic acid.
Fig. 7. pH profile for dissolution in 0.1 M lactic acid.
Fig. 8. Extent of Ca extraction with increasing temperature in 3 h (pH 2.0–4.5).
Table 2
Elemental balance for dissolution in formic acid at 80 °C and 3 h.
Input (g) Output (g)
Filtrate Filter cake
Ca 0.20 ± 0.03 0.192 ± 0.020 0.006 ± 0.001
Si 0.14 ± 0.02 0.022 ± 0.003 0.11 ± 0.02
Mg 0.006 ± 0.001 (0.020 ± 0.004) Â 10À2
(0.55 ± 0.08) Â 10À2
Fe (0.1 ± 0.02) Â 10À2
(0.020 ± 0.004) Â 10À2
(0.074 ± 0.011) Â 10À2
Na (0.1 ± 0.02) Â 10À2
(0.022 ± 0.004) Â 10À2
(0.072 ± 0.014) Â 10À2
Al (0.1 ± 0.02) Â 10À2
(0.018 ± 0.004) Â 10À2
(0.078 ± 0.020) Â 10À2
Fig. 9. Concentration of aqueous species as calculated at thermodynamic equilib-
rium, at 80 °C and 1 atm (OLI Analyzer Studio 3.0); symbols are used only to identify
each plot.
M. Ghoorah et al. / Fuel 122 (2014) 277–286 281

6. As revealed by XRD results, the wollastonite sample involved in
the experiments comprised other minor minerals such as diopside,
pectolite and silica. The fact that the software treats wollastonite,
available from its databank, as pure CaSiO3 can lead to failure in
establishing an accurate pH model for the real system. Further-
more, it was observed that the two thermodynamic frameworks
implemented in the OLI Analyzer, namely Aqueous (H+
ion) and
Mixed Solvent Electrolyte/MSE (H3O+
ion) [22], yielded slightly dif-
ferent predictions, as illustrated in Fig. 11.
Moreover, ion chromatography provided the composition of the
liquid phase obtained after filtering the suspension, to investigate
the presence of anions, other than formate, that accumulated in
solution during dissolution and lead to lower pH of the system.
Only chloride ion was detected by ion chromatography. ICP-OES
results showed that the sample also contained Al, Zn, Mn, Na, Ni,
Sr, Cu, Fe and Co in small concentration. It is likely that, some of
these minor metal elements exist as chlorides in the mineral and
react with the acid to form organic salts, hence leaving chloride
ions in solution. As a result, the measured pH and thermodynamic
predictions bear slight discrepancies. Additionally, the studied pH
range lies above the isoelectric point for both silica and wollaston-
ite. As a result, the surface potential of the particles assumes a neg-
ative value leading to a slightly lower pH close to the surface. This
effect may explain some discrepancy between measured and pre-
dicted pH.
As indicated by pH measurements, the reaction rate is initially
high but it plateaus with time. This occurs as a consequence of
slower kinetic and mass transfer rates. At the beginning of each
experiment, both H+
and anions (e.g., HCOOÀ
) diffuse only through
the liquid film surrounding each particle. However, as SiO2 skin
thickens on the particle surfaces, the diffusion proceeds through
the cracks and pores of the skin, significantly slowing down the
dissolution process. The other reason for slowing down of the dis-
solution rate is the dependence of breaking of Ca–O bonds on the
activity of protons. The lower the activity (e.g., the higher the
pH), the slower is the reaction process. The dissolution may pro-
ceed via the following steps, where –O–Ca–OH denotes surface cal-
cium atoms terminated with hydroxyl groups:
—O—Ca—OHðsÞ þ Hþ
ðaqÞ ! —O—Ca—OHþ
2ðsÞ ð4Þ
—O—Ca—OHþ
2ðsÞ ! —O—Caþ
ðsÞ þ H2OðlÞ ð5Þ
—O—Caþ
ðsÞ þ Hþ
ðaqÞ ! —OHþ
À Caþ
ðsÞ ð6Þ
—OHþ
—Caþ
ðsÞ ! —OHðsÞ þ Ca2þ
ðaqÞ ð7Þ
Clearly, Reactions 4 and 6 are pH dependent. In addition, the an-
ions themselves may assist in the dissolution of wollastonite via
the chelating pathway as illustrated in the following reaction
—OHþ
—Caþ
ðsÞ þ HCOOÀ
ðaqÞ ! —OHðsÞ þ CaHCOOþ
ðaqÞ ð8Þ
The effectiveness of anions depends both on the nature of their
functional groups, molecular structure and thermodynamic stabil-
ity of the transitional surface complexes they form [28]. Organic
anions such as acetate, lactate and formate are known to form
monodentate complexes on oxides, which upon polarisation, labi-
lise the Ca–O bonds thereby facilitating the removal of calcium
atoms from the crystal lattice [29]. As the calcium-ligand com-
plexes detach from the surface, the underlying layers are exposed
to further contact with the solvent.
The higher extraction yield of formic acid can also be justified in
terms of H+
activity, as the so-called pH pathway. Among the three
weak acids studied, formic acid is the strongest (pKa 3.75) and
hence it dissociates to a larger extent to produce H+
ions when in
solution. Increased protonation of the lone pairs of electrons in
oxygen atoms (Reactions 4 and 6) weakens the O–Ca bond. Under
the pH conditions considered in this contribution, the pH pathway
will dominate the chelating effects in extracting Ca2+
from
wollastonite.
The next part of the article aims at assessing the dissolution
kinetics as a function of temperature in acidic medium in the limit
of low pH achievable for these acids. Consequently, we conducted
experiments at constant pH or H+
activity, for a period of 1 h, at
40 °C, 60 °C and 80 °C. Maintaining the mass ratio of wollastonite
Fig. 10. Solid phase composition and pH profile for the results presented in Fig. 8
(OLI Analyzer Studio 3.0).
Fig. 11. Comparison of experimental and simulation results on the aqueous and
MSE frameworks.
Fig. 12. Dissolution in formic acid at 80 °C and pH 1.04. The open symbols denote a
repeat.
282 M. Ghoorah et al. / Fuel 122 (2014) 277–286

7. to acid as low as 0.02 ensured constant pH throughout the runs.
Figs. 12–14 depict representative examples of the temporal evolu-
tion of the leaching solution composition at 80 °C and in aqueous
solutions of formic, acetic and DL-lactic acids, respectively. In addi-
tion, we were particularly interested in estimating the maximum
initial dissolution rates, i.e., the rates in the absence of the silica
layer present on the particle surfaces.
The kinetic behaviour of the dissolution reactions displays a fast
initial rate during the first 10 min followed by a slowdown, ob-
served in all cases. During the early stages of the reaction, the dis-
solution rates can be considered to be surface controlled (i.e., film-
diffusion or reaction-rate controlled) but the levelling off of cal-
cium concentration in filtrates indicates a diffusion limitation
(i.e., pore/crack-diffusion controlled) at the later part of the process
[30,31]. This limitation can be attributed to the fact that wollaston-
ite dissolution is strongly incongruent at acidic pH leading to the
formation of a passivating, amorphous silica rich layer, which
could partly reduce the transport of aqueous species and eventu-
ally hinder further dissolution of the mineral [11,27,32–36]. The
low concentration of dissolved silicon (5–10%) in our experiments
also suggests the build-up of silica coating on the rock particles
during reaction. However, referring to experimental and simula-
tion data (Figs. 7 and 8), we deduce that running the experiment
for longer reaction times (in this case 3 h) compensated for the
passivating effect of the silica layer. In other words, calcium will
continue to diffuse out of the crystal lattice, albeit at a slower rate.
We applied the method of initial rates to estimate the dissolu-
tion rates in the absence of a passivating layer of amorphous silica.
The rate of reaction can be computed by plotting the concentration
of calcium in the leaching medium as a function of time and then
evaluating the gradient of the curve at time t = 0 min. As we were
unable to measure Ca2+
concentration at a very short time (our
shortest measurement interval was 5 min), we fitted a sixth degree
polynomial to all measurements and applied that polynomial to
estimate the rates at t = 0 min. Table 3 summarises the rates of dis-
solution, which have been normalised to the specific surface area
at the investigated temperatures. Formic acid achieved the highest
initial rate at 80 °C.
We based our calculations on an estimate of the mineral surface
area which is equal to the total mass of reacting material multi-
plied by the specific surface area per unit mass of material, as
determined by the BET method. While some researchers have as-
sumed that the surface area of the leached layer grows linearly
with time [37], others found that the surface area of their reacted
wollastonite grains increased according to a power law function
[33]. The surface area certainly changes as the reaction proceeds
but for this study, during the very onset of the reactions, when
we made our measurements, it is the same as the BET surface area
of the fresh particles.
The observed increase in extent of dissolution with growing
temperature can be interpreted by the empirical Arrhenius equa-
tion given by
r ¼ A expðÀEa=RTÞ
ð9Þ
where r designates the rate of reaction in mol mÀ2
sÀ1
, A refers to a
pre-exponential factor in mol mÀ2
sÀ1
(which is here a function of
pH) and Ea represents the activation energy, in kJ molÀ1
, defined by
Ea ¼ À2:303R½@ log r=@ð1=TÞŠpH ð10Þ
Fig. 15 illustrates an Arrhenius plot with the logarithm of the
measured wollastonite dissolution rates on the ordinate and the
reciprocal of temperature on the abscissa. The activation energies
Fig. 13. Dissolution in acetic acid at 80 °C and pH 1.61. The open symbols denote a
repeat.
Fig. 14. Dissolution in DL-lactic acid at 80 °C and pH 1.19. The open symbols denote
a repeat.
Fig. 15. Arrhenius plot of the initial dissolution rates.
Table 3
Initial rates of dissolution in a constant pH system under the effect of increasing
temperature.
Acid Temperature
(°C)
% Ca extraction
within 3 h
Initial rate  105
(mol mÀ2
sÀ1
)
Formic acid 40 60 ± 7 16 ± 4
60 78 ± 9 23 ± 6
80 96 ± 10 26 ± 7
Acetic acid 40 33 ± 5 1.9 ± 0.5
60 63 ± 9 4.7 ± 1.4
80 85 ± 10 14 ± 3
DL-lactic acid 40 37 ± 6 1.8 ± 0.5
60 71 ± 8 6.3 ± 1.6
80 90 ± 10 17 ± 4
M. Ghoorah et al. / Fuel 122 (2014) 277–286 283

8. are derived from the slopes of the straight lines that best fit the
points, given by ÀEa/2.303 R. Table 4 presents the calculated acti-
vation energies and pre-exponential factors. It should be noted that
the term for temperature dependence denotes an apparent global
activation energy because the dissolution of minerals is not a sin-
gle elementary reaction but rather involves a complex series of
reactions, each carrying their own activation energy [38]. The
low activation energy for the formic acid seems to imply mass
transfer control for diffusion of H+
and HCOOÀ
in the aqueous film
surrounding the reacting particles. Approximate mass transfer cal-
culations, based on diffusion in electrolyte solutions, yield the ini-
tial mass-transfer-limited rate of dissolution of wollastonite by
formic acid of (20–40) Â 10À5
mol mÀ2
sÀ1
, in agreement with the
experimental measurement of (16 ± 4) Â 10À5
mol mÀ2
sÀ1
. The re-
sults for acetic and DL-lactic acid point to kinetic control of
removing Ca2+
via protonation of the reacting surface and hetero-
geneously breaking of O–Ca bonds. For comparison, apparent acti-
vation energies of 68, 70 and 74 kJ molÀ1
, for the dissolution of
serpentinite in H2SO4, HCl and HNO3 respectively, reported in liter-
ature [39], highlight a lower reactivity of serpentine minerals for
the dissolution.
SEM analysis revealed that fresh wollastonite particles consist
primarily of fibrous needle-like structures (Fig. 16a). The dissolu-
tion products displayed distinct microstructural features including
fractures, cracks and surface unevenness (Fig. 16b and c). The
greater impact of formic acid can be discerned through more pro-
nounced crazing of the grains. This may be one of the reasons for
the dissolution of wollastonite proceeding faster, even in the pres-
ence of a silica layer; compare Fig. 14 with Figs. 12 and 13. How-
ever, in both cases, the particles preserved their original
morphology, indicating the formation of amorphous silica deposits
on the particle surface while its core gradually disappeared. SEM
analysis used in conjunction with EDS further supports this state-
ment by demonstrating that the outer layer of fresh wollastonite
particles is mainly made up of calcium while that of reacted parti-
cles consist mostly of silica.
To confirm the formation of the silica layer, we measured the
particle size distribution of the reacted wollastonite grains. The
distribution shifted to a slightly smaller average particle size. From
the results shown in Fig. 2, it can be observed that the volume
20 µm
0 5 10 15 20
Energy (keV)
0
50
100
150
200
250
cps
Si
Ca
Ca
100 µm
0 5 10 15 20
Energy (keV)
0
50
100
150
200
250
cps
Si
Ca
100 µm
20 µm
0 5 10 15 20
Energy (keV)
0
50
100
150
200
250
cps
Si
Ca
20 µm
100 µm
(a) (b) (c)
(a) (b) (c)
(a) (b) (c)
Fig. 16. SEM microphotographs of unreacted wollastonite particles (a) and after treatment with acetic acid (b) and formic acid (c) at 80 °C and 3 h.
Table 4
Apparent kinetic parameters calculated from the Arrhenius plots in Fig. 8.
Kinetic parameters Solvent
Formic
acid
Acetic
acid
DL-lactic
acid
Activation energy (kJ molÀ1
) 11 ± 3 47 ± 13 52 ± 14
Pre-exponential factor
(mol mÀ2
sÀ1
)
0.01 9000 1070
284 M. Ghoorah et al. / Fuel 122 (2014) 277–286

9. mean diameter of the acetic acid-treated particles decreased by 6%
(16 ± 1 lm) whereas those treated with formic acid have
diminished by 12% (15 ± 1 lm), when compared to the original
particles (17 ± 1 lm). It can thus be inferred that formic acid was
more capable of removing or crazing the silica layer.
XRD analysis of the reacted particles confirmed the amorphicity
of the silica layer formed due to the interaction of wollastonite
with organic acids and subsequent dissolution. Fig. 17 depicts
the XRD pattern obtained for the case of formic acid. For compar-
ison, unreacted wollastonite sample (Fig. 3) consists of a major
phase of wollastonite and minor phases of diopside and pectolite.
Phases similar to the original material co-exist with an additional
quartz phase. Pectolite was not detected as it is likely to have
undergone complete dissolution. The bump or broad peak occur-
ring at about 22° in both cases indicates the presence of an amor-
phous phase. According to Azizi and Yousefpour [40], a broad peak
centered at 2h angle of 22° is typical for amorphous silica.
5. Conclusions
We have observed significant differences in the dissolution of
Ca2+
from wollastonite with formic acid on one hand, and acetic
and DL-lactic acids on the other. Even in the absence of the amor-
phous silica layer on particle surfaces, the dissolution of Ca2+
ap-
pears to be mass-transfer controlled in the case of formic acid
and kinetically controlled in the case of acetic and DL-lactic acids.
All rates decrease with time as the silica layer accumulates. How-
ever, the decrease in rate is significantly less pronounced for formic
than for acetic and DL-lactic acids. The SEM microphotographs and
particle-size measurements suggest enhanced crazing in silica
layer formed in the presence of formic acid and partial dissolution
of silica. For comparison, at 80 °C and similar pH, for the particle
distribution investigated in this study (D[v, 0.1] = 2 ± 1 lm, D[v,
0.5] = 17 ± 1 lm, D[v, 0.9] = 56 ± 1 lm, and D[3, 2] = 6 ± 1 lm), it
takes less than 20 min for the extraction of Ca2+
by formic acid to
reach completeness, whereas acetic and DL-lactic acids only
achieve Ca2+
60–70% extraction in the same time. Further dissolu-
tion of Ca2+
proceeds extremely slowly. As expected, the rates of
dissolution attain their maximal values under the lowest achiev-
able pH (between 1 and 1.6 depending on acid) and the highest
temperature (80 °C). Overall, these findings lead us to conclude
that formic acid may constitute the preferred reactant for extract-
ing Ca2+
, subject to its suitability for recycling in the second step of
the mineralisation process.
Acknowledgements
This study was funded by an internal grant from the University
of Newcastle. The authors are grateful to Dr. L. Elliot for performing
BET measurements as well as D. Phelan and J. Zobec (EM-X-ray
Unit, The University of Newcastle) for their assistance with SEM,
XRD and XRF analyses. Special thanks go to J. Hamson for helping
in the operation of ICP-OES and microwave unit. Discussions with
Professor J. Bałdyga of the Technical University of Warsaw are
acknowledged with gratitude. M. Ghoorah is thankful to the Uni-
versity of Newcastle for the postgraduate research scholarship.
References
[1] O’Connor WK, Dahlin DC, Nilsen DN, Rush GE, Walters RP, Turner PC. Carbon
dioxide sequestration by direct mineral carbonation: results from recent
studies and current status. Proc. 1st Annual DOE carbon sequestration
conference 2001; DOE/ARC-2001-029.
[2] Gerdemann SJ, O’Connor WK, Dahlin DC, Penner LR, Rush H. Ex situ aqueous
mineral carbonation. Environ Sci Technol 2007;41:2587–93.
[3] Maroto-Valer MM, Fauth DJ, Kuchta ME, Zhang Y, Andrésen JM. Activation of
magnesium rich minerals as carbonation feedstock materials for CO2
sequestration. Fuel Process Technol 2005;86:1627–45.
[4] Kakizawa M, Yamasaki A, Yanagisawa Y. A new CO2 disposal process via
artificial weathering of calcium silicate accelerated by acetic acid. Energy
2001;26:341–54.
[5] Balucan RD, Kennedy EM, Mackie JM, Dlugogorski BZ. An optimised antigorite
heat pre-treatment strategy for CO2 sequestration by mineral carbonation. In:
3rd International conference on accelerated carbonation for environmental
and materials engineering, Turku; 2010.
[6] O’Connor WK, Dahlin DC, Nilsen DN, Walters RP, Turner PC. Carbon dioxide
sequestration by direct mineral carbonation with carbonic acid. In: Proc. 25th
International technical conference on coal utilisation fuel systems, Coal
Technology Assoc, Clear Water FL 2000; DOE/ARC-2000-008.
[7] Oelkers EH, Schott J. Geochemical aspects of CO2 sequestration. Chem Geol
2005;217:183–6.
[8] Park AA, Fan L. CO2 mineral sequestration: physically activated dissolution of
serpentine and pH swing process. Chem Eng Sci 2004;59:5241–7.
[9] Carroll SA, Knauss KG. Dependence of labradorite dissolution kinetics on CO2
(aq), Al (aq) and temperature. Chem Geol 2005;217:213–25.
[10] Hänchen M, Prigiobbe V, Storti G, Seward TM, Mazzotti M. Dissolution kinetics
of forsteritic olivine at 90–150 °C including effects of the presence of CO2.
Geochimica et Cosmochimica Acta 2006:4403–16.
[11] Golubev SV, Pokrovsky OS, Schott J. Experimental determination of the effect
of dissolved CO2 on the dissolution kinetics of Mg and Ca silicates at 25 °C.
Chem Geol 2005;217:227–38.
[12] Krevor SC, Lackner KS. Enhancing process kinetics for mineral carbon
sequestration. Energy Procedia 2009;1:4867–71.
[13] Hövelmann J, Putnis CV, Ruiz-Aguda E, Austrheim H. Direct nanoscale
observations of CO2 sequestration during brucite [Mg(OH)2] dissolution.
Environ Sci Technol 2013;46:5253–60.
[14] Pokrovsky OS, Schott J, Castillo A. Kinetics of brucite dissolution at 25 °C in the
presence of organic and inorganic ligands and divalent metals. Geochimica et
Cosmochimica Acta 2005;69:905–18.
Fig. 17. XRD pattern for wollastonite after reaction with formic acid.
M. Ghoorah et al. / Fuel 122 (2014) 277–286 285

10. [15] Prigiobbe V, Hänchen M, Werner G, Baciocchi R, Mazzotti M. Mineral
carbonation process for CO2 sequestration. Energy Procedia 2009;1:4885–90.
[16] Daval D, Martinez I, Corvisier J, Findling N, Goffé B, Guyot F. Carbonation of Ca-
bearing silicates, the case of wollastonite: experimental investigations and
kinetic modelling. Chem Geol 2009;265:63–78.
[17] Lackner KS, Wendt CH, Butt DP, Joyce EL, Sharp DH. Carbon dioxide disposal in
carbonate minerals. Energy 1995;20:1153–70.
[18] Teir S, Eloneva S, Fogelholm CJ, Zevenhoven R. Dissolution of steelmaking slags
in acetic acid for precipitated calcium carbonate production. Energy
2007;32:528–39.
[19] Carey JW, Ziock H, Guthrie Jr G. Reactivity of serpentine in CO2-bearing
solutions: application to CO2 sequestration. Am Chem Soc, Div Fuel Chem
2004;49:371–2.
[20] Bałdyga J, Henczka M, Sokolnicka K. Utilization of carbon dioxide by
chemically accelerated mineral carbonation. Mater Lett 2010;64:702–4.
[21] Zhao H, Park Y, Lee DH, Park AA. Tuning the dissolution kinetics of wollastonite
via chelating agents for CO2 sequestration with integrated synthesis of
precipitated calcium carbonates. Phys Chem Chem Phys 2013;15:15185–92.
[22] OLI Systems Inc. A guide to using OLI Analyzer Studio version 3.2. Morris
Plains, New Jersey; 2011.
[23] NSW Department of Industry and Investment, 2010.
[24] NSW State of the Environment 2009. www.environment.nsw.gov.au. Accessed
18.10.10.
[25] Allinson G, Cinar Y, Hou W, Neal PR. The Costs of CO2 transport and injection in
australia, CO2TECH Consultancy, Report 2009.
[26] Google earth. www.google.com/earth/index.html, Accessed 24.09.10.
[27] Schott J, Pokrovsky OS, Spalla O, Devreux F, Mielczarski JA. The mechanism of
altered layer formation on wollastonite revisited: a combined spectroscopic/
kinetic study. Geochimica et Cosmochimica Acta 2002;66:A686.
[28] Pokrovsky OS, Shirokova LS, Bénézeth P, Schott J, Golubev SV. Effect of organic
ligands and heterotrophic bacteria on wollastonite dissolution kinetics. Am J
Sci 2009;309:731–72.
[29] Swaddle TW. Inorganic chemistry: an industrial and environmental
perspective. Academic Press; 1997.
[30] Park AA, Jadhav R, Fan L. CO2 mineral sequestration: chemically enhanced
aqueous carbonation of serpentine. Can J Chem Eng 2003;81:885–90.
[31] Alexander G, Maroto-Valer MM, Gafarova-Aksoy P. Evaluation of reaction
variables in the dissolution of serpentine for mineral carbonation. Fuel
2006;86:273–81.
[32] Xie Z, Walther JV. Dissolution stoichiometry and adsorption of alkali and
alkaline earth elements to the acid-reacted wollastonite surface at 25 °C.
Geochimica et Cosmochimica Acta 1994;58:2587–98.
[33] Weissbart EJ, Rimstidt JD. Wollastonite: incongruent dissolution and leached
layer formation. Geochimica et Cosmochimica Acta 2000;64:4007–16.
[34] Béarat H, McKelvy MJ, Chizmeshya AVG, Gormley D, Nunez R, Carpenter RW,
et al. Carbon sequestration via aqueous olivine mineral carbonation: role of
passivating layer formation. Environ Sci Technol 2006;40:4802–8.
[35] Huijgen WJJ, Witkamp G-J, Comans RNJ. Mechanisms of aqueous wollastonite
carbonation as a possible CO2 sequestration process. Chem Eng Sci
2006;61:4242–51.
[36] Green E, Lüttge A. Incongruent dissolution of wollastonite measured with
vertical scanning interferometry. Am Mineral 2006;91:430–4.
[37] Stillings LL, Brantley SL. Feldspar dissolution at 25 °C and pH 3 – reaction
stoichiometry and the effect of cations. Geochimica et Cosmochimica Acta
1995;59:1483–96.
[38] Lasaga AC. Fundamental approaches in describing mineral dissolution and
precipitation rates. Rev Mineral 1995;31:23–86.
[39] Teir S, Revitzer H, Eloneva S, Fogelholm C-J, Zevenhoven R. Dissolution of
natural serpentinite in mineral and organic acids. Int J Mineral Process
2007;83:36–46.
[40] Azizi SN, Yousefpour M. Spectroscopic studies of different kind of rice husk
samples grown in North of Iran and the extracted silica by using XRD, XRF, IR,
AA and NMR techniques. Eurasian J Anal Chem 2008;3:298–306.
286 M. Ghoorah et al. / Fuel 122 (2014) 277–286