Influence of natural water composition on reactivity of quicklime derived from Ca-rich and Mg-rich limestone: implications for sustainability of lime manufacturing through geochemical modeling

Abstract

We studied the effects of water quality on the hydration of quicklimes prepared from calcitic limestone and dolomitic limestone during calcination at 1050 °C. At a low total chloride and sulfate content, reactivity is insensitive to concentration and no effect on quicklime slaking is seen. At higher concentrations, however, we observe a chromatographic effect for chlorides and sulfates. Concentrated chlorides react with the slaking water, forming products such as CaCl₂ that are much more soluble than Ca(OH)₂, which facilitates transformation of quicklime to slaked lime. In the range of sulfate concentrations in natural waters we observed no effect of sulfates during quicklime hydration. Consequently, in natural waters and their mixtures (conductivity < 2600 μS cm⁻¹), only chloride concentrations are found to promote quicklime hydration. These results suggest that water treatment technologies, applied to natural light brackish water compositions, will not lead to quality improvement for lime-industry products.

---

1. Introduction

Limestones of various types are among the most frequently used rocks in industry and the global consumption, which has been increasing roughly linearly with time, is projected to reach 5.7 billion tonnes by 2020.¹ In fact, one of the most important applications of limestone is the production of quicklime, CaO, by calcination and subsequently slaked lime, Ca(OH)₂, by hydration following the reactions:

CaCO₃(s) → CaO(s) + CO₂(g) (1)

CaO(s) + H₂O(l) → Ca(OH)₂(aq). (2)

The calcination reaction (1) is endothermic and proceeds at very high temperatures (>900 °C), whereas the slaking reaction (2) is strongly exothermic. Despite the widespread and longstanding use of quicklime and slaked lime throughout the world, there has been rather little in-depth scientific study of slaking.² In particular, the crucial role of slaking water has received less attention than it should. Many sources have claimed that pure (distilled) water is the ideal reactant for hydration of quicklime, but the effect of impure slaking water is in fact not well known.^3–6

Water treatment includes three major methods to control water chemistry: deionization, desalination, and softening. Softening is unimportant because it affects only the cation concentrations. The methods of deionization and desalination produce more or less the same result in terms of water chemistry, despite following different pathways. Deionization is a very old water treatment technology, in use for several decades; it applies two different resin columns (anionic and cationic) to produce water with very low dissolved salts. Desalination, on the other hand, is a newer and more sophisticated water treatment that takes advantage of reverse osmosis (RO) in a series of membranes to reduce the overall content of dissolved salts. Techniques using membranes have been rapidly developed since the 1960s, and now surpass thermal processes in new plant installations.^7,8 The cumulative costs of desalination projects include the expenses of complex materials needed to form highly salt-retentive membranes, of pre-treatment equipment, and of ongoing energy consumption. The benefit of RO water treatment for a particular application needs to be weighed against these substantial costs. It is important to test whether high purity water is in fact optimal for slaking because of the substantial economic costs and environmental impact of water treatment plants for production of pure water to be used in the hydration of quicklime.

One notable study is Potgieter et al.,⁹ which focused on the role of chloride, sulfate and carbonate ions dissolved in simplified synthetic aqueous solutions on the slaking rate of lime. That study concluded that chlorides in the slaking water increase the lime reactivity, whereas sulfates and carbonates retard the hydration reaction.

The present study focuses instead on the effect on slaking of a range of dissolved ions in different natural waters. Two limestones used in the present work, one calcitic and one dolomitic, were previously studied by Leontakianakos et al.² Both materials have optimal reactivity when calcined at 1050 °C. Here, the quicklimes produced from these two source rocks were used to investigate the effects of anions in the slaking water on the reactivity and slaking rate at laboratory scale. Six different water samples were analyzed and used as additives for the hydration of quicklimes. Four samples were prepared by modification of a natural well water and cover a range from distilled to light brackish water (<2600 μS cm⁻¹). Furthermore, two synthetic waters were produced by spiking the well water with Cl⁻ or with SO₄²⁻. The results demonstrate the effectiveness of brackish water in quicklime hydration, which may lead to substantial cost and environmental impact reduction through elimination of unnecessary water treatment operations.

2. Materials and methods

The current study is the second part of an integrated project examining both steps of slaked lime production from calcitic (sample SL) and dolomitic (sample SD) limestones (Fig. 1). The starting rocks and the quicklimes resulting from their calcination (SLQ and SDQ, respectively) were analyzed in detail by ref. 2 using petrography, scanning electron microscopy, X-ray diffraction, Raman spectroscopy, thermogravimetry, differential thermal analysis, and Brunauer–Emmett–Teller (BET) surface area analysis. Details of the equipment used for each analytical method can be found in ref. 2. Freshly prepared samples of quicklimes SDQ and SLQ were analysed by thermogravimetry under dry air flow from room temperature up to 1080 °C. Here we discuss in detail only the reactivity tests of the two quicklimes in the various natural and spiked waters. The chemical analyses and other properties (including residual material, organic material, available lime content, total lime, and BET) of the quicklimes are presented in Table 1.

Fig. 1 (a) Dolomitic limestone (SD): well crystallized dolomite crystals intersect at triple junctions, showing a mosaic texture, (b) SDQ quick lime calcined at 1050 °C is characterized by fine dispersed globules of CaO, (c) fossiliferous calcitic limestone (SL) consists of fine-grained calcite crystals, (d) SLQ quick lime calcined at 1050 °C is characterized by a dense structure and larger grain size than SDQ. (a and c) Optical microscope images (plane polarized light); (b and d) scanning electron microscope images.

Table 1 Summary of chemical composition (in wt%) and properties of raw materials (SD and SL) and derived quicklimes (SDQ and SLQ) calcined at 1050 °C (from ref. 2)^a

	SD	SL	SDQ	SLQ
a b.d.: below-detection value.
SiO2 (in wt%)	0.20	1.71	0.41	3.04
Al2O3 (in wt%)	0.04	0.53	0.07	0.93
MgO (in wt%)	17.5	0.63	35.5	1.12
CaO (in wt%)	30.4	52.5	61.6	93.1
CO2 (in wt%)	45.8	43.5	0.81	0.48
Residual material (in wt%)	0.11	0.35
Organic material (in wt%)	0.05	0.36
Available lime content			56.7	69.3
Total lime			97.1	94.2
% of total available CaO			92.1	74.4
(CaO + MgO)lime			97.1	94.2
C/M			1.001	1.002
BET (m^2 g^−1)	b.d.	b.d.	11.4	1.8

---

2.1 Chemical analyses of waters

The chemical parameters of the natural and treated waters were determined as follows. The pH was measured using a WTW pH meter according to ISO 4316. Conductivity was determined according to EN 27888: 1993 using a conductivity electrode. Total dissolved solids (TDS) were determined gravimetrically according to APHA 2540 (ref. 10) by filtering through a standard glass fiber filter, evaporating the filtrate to dryness in a weighed dish and drying to constant weight. Total hardness (TH; given here in parts per million-ppm CaCO₃) was determined by typical titration based on a colorimetric reaction according to the APHA 2340 method. Alkalinity was determined by the APHA 2320 method via titration with a standard HCl solution to an end-point pH of 3.7. Typical anions (Cl⁻, SO₄²⁻, NO₃⁻) were determined by ion chromatography, EN 480-10 E2: 2010 using a DIONEX ICS-900 based on APHA method 4110.¹¹

Six different water samples were analyzed and used as additives for the hydration of quicklimes. Natural light brackish well water (denoted here as W3) contains a variety of dissolved solids. Sample W2 is the result of deionization of the raw well water using a mixed-bed column. W1 is a 50 : 50 mixture and W4 a 25 : 75 mixture of the well and distilled waters. Finally, the influence of Cl⁻ and SO₄²⁻ on the reactivity tests was evaluated with spiked addition of ∼1000 ppm of each anion to well water, using CaCl₂ for Cl⁻ (identified as sample W5) and NaSO₄ for SO₄²⁻ (identified as sample W6). We could also have added NaCl to modify Cl⁻ concentration, but tests show that the identity of the cations in solution are not significant during slaking processes.¹² The chemical analyses and physico-chemical properties of all six waters are given in Table 2.

Table 2 Physico-chemical properties of the analysed natural, mixed and synthetic waters^a

a *: 50 : 50 mixture of W2 and W3; **: 75 : 25 mixture of W2 and W3, b.d.: below detection limit.
Sample	W1	W2	W3	W4	W5	W6
Type	Mixed*	Distilled	Well water	Mixed**	Cl-rich	SO4-rich
pH	6.7	7.7	7.3	6.8	7.3	7.3
Conductivity (in μS cm^−1)	571	31	792	274	2591	2447
Total hardness (in ppm CaCO3)	192	17.8	398	106.7	1513	370
m-Alkalinity (in ppm CaCO3)	170	30	350	85	330	350
Chloride (ppm Cl^−)	55	7.5	62.5	25.7	920	60
Sulfate (ppm SO4^2−)	12	b.d.	26	8	25	842
Bicarbonate (ppm HCO3^−)	203	5	409	102	403	427
Nitrate (ppm NO3^−)	10	b.d.	30	5	b.d.	b.d.

---

2.2 Reactivity test

Following the European standard procedure (NF EN 549-2 (ref. 13)), a lime/water ratio of 1 : 4 was used at ambient temperature (∼20 °C) to produce Ca(OH)₂ through reaction (2). This reaction is highly exothermic, increasing the temperature of the added water. The hydration experiment takes place under adiabatic conditions, with continuous stirring and monitoring of temperature at a sampling rate of 1 Hz until temperature reaches a maximum, after which is may stabilize or decline. The rise in temperature due to the addition of water is an accepted criterion for chemical “reactivity” of the quicklime. This was determined for each of the two quicklimes for each of the 6 different water types.

The reactivity can be estimated using three different approaches, which are described respectively in ref. 11, 14 and 15. According to the EN 459-2: 2001 standard,¹¹ in order to define the reactivity, it is necessary to measure the maximum temperature of the reaction reached by the water–lime system. Then, the reactivity is equal to the time required from the start of the reaction for the sample to reach 80% of its maximum temperature, that is when T = 0.8 × T_max + 0.2 × T_st, where T_max is the maximum measured temperature and T_st is the starting temperature of both the water and the introduced quicklime. In the second approach, the reactivity parameter named “RDIN” results from the division of 2400 (40 °C × 60 s min⁻¹) by the time (in sec) required until temperature reaches 60 °C.¹⁴ Based on the RDIN values, lime reactivity is divided into three categories: highly reactive lime RDIN > 30, reactive lime 10 < RDIN < 30, and unreactive lime RDIN < 10. Both these methods measure the rate of the hydration reaction, as opposed to the extent to which the reaction approaches completion. In the third approach, which was suggested by ref. 15, reactivity is estimated as the difference between the maximum measured temperature and the starting temperature of the water and quicklime before the reaction. For evaluating the quality of the final product, we argue that the extent to which the hydration reaction reaches completeness is a more important measurement than the rate at which hydration proceeds. Therefore, we prefer the method suggested by ref. 15 for reactivity estimation. However, reactivity rate, as opposed to total reactivity, may also be an important parameter in the production process. We obtain a reactivity rate by dividing the reactivity (temperature increase) in each trial by the time required to reach T_max.

2.3 Geochemical modeling

The geochemical reaction code PHREEQC (v 3.3.3.10424) was employed to simulate the outcome of reactions between lime samples and water solutions using the minteq.dat database.¹⁶ Models were carried out for each of the six water compositions of the experiment and for each of the two quicklime starting materials. All calculations were run with initial temperature of 20 °C and quicklime : water ratio of 1 : 4 by weight (i.e. 0.45 moles of SLQ or 0.5 moles of SDQ and 100 g water). In each case, we constructed both an equilibrium model and a kinetic model.

In the equilibrium model the amount of slaked lime formed is evaluated from a time-independent thermodynamic equilibrium calculation. At equilibrium, each water reached saturation with respect to quicklime (i.e., saturation index SI = 0), whereas slaked lime, carbonates, Mg oxides and gypsum were considered as possible precipitates if oversaturated (SI > 0). The thermodynamic data used are shown in Table 3.

Table 3 Thermodynamic data of the reacted mineral phases at 25 °C (minteq.dat)

Phase	Reaction	log K	ΔHr (kcal mol^−1)
a Calculated from thermodynamic data of 60% lime and 40% periclase.b Thermodynamic data by Smith et al.^21
SLQ	CaO + 2H^+ = Ca^2+ + H2O	32.57	−46.29
SDQ^a	Ca0.6Mg0.4O + 2H^+ = 0.6Ca^2+ + H2O + 0.4Mg^2+	28.2	−42.07
Slaked lime	Ca(OH)2 + 2H^+ = Ca^2+ + 2H2O	22.67	−22.67
Brucite	Mg(OH)2 + 2H^+ = Mg^2+ + 2H2O	16.7	−25.84
Periclase	MgO + 2H^+ = Mg^2+ + H2O	21.51	−36.13
Calcite	CaCO3 = CO3^2− + Ca^2+	−8.48	−2.58
Aragonite	CaCO3 = CO3^2− + Ca^2+	−8.36	−2.61
Dolomite	CaMg(CO3)2 = Ca^2+ + Mg^2+ + 2CO3^2−	−17.09	−8.29
Magnesite	MgCO3 = Mg^2+ + CO3^2−	−8.03	−8.03
Gypsum	CaSO4·2H2O = Ca^2+ + SO4^2− + 2H2O	−4.58	−0.11
Vaterite^b	CaCO3 + H^+ = Ca^2− + HCO3^−	2.42	−29.6

---

The kinetic model computes the rates of dissolution and precipitation reactions, which allows the water chemistry and mineral abundances to be tracked over time. The kinetic calculation considers the rates of both dissolution of quicklime and precipitation of slaked lime and/or carbonate minerals, gypsum and Mg oxides.

The dissolution/precipitation rates used are based on the transition state theory, which leads to expressions for the reaction rate r_n of a mineral n such as Lasaga:¹⁷

r_n = ±k_nS_n|1 − Ω_n^p_n|^q_n (3)

where k_n is the kinetic rate constant (mol m⁻² s⁻¹), S_n is the reactive surface area (m²), Ω_n is the saturation ratio and p_n, q_n are constants. The positive and negative signs denote the dissolution and precipitation of the phase respectively.

The kinetic constant k_n for mineral n depends on the temperature of the reaction (Arrhenius law) as well as the pH of the environment. In acidic environments, the reaction is governed by the H⁺ activity and, in alkaline environments, by the OH⁻ activity. There may also be a dependence on the activity of dissolved inorganic carbon species such as HCO₃⁻, especially in the case of carbonate minerals. The detailed expression for k_n as a function of temperature T is given by:

where R is the gas constant and in each term k_0,n,i is a rate constant for mineral n at 25 °C, E_a,i (kJ mol⁻¹) is an activation energy, and a_i is the activity of the respective ion.

The reactive surface area S_n of the mineral depends on its specific surface area S_A,n (m² g⁻¹), the number of moles of the mineral present m_n, its molar weight M_n, and the reactive fraction λ (equal to 1 if the whole surface area is reactive, which we will assume throughout this work for simplicity):

S_n = λ_nm_nM_nS_A,n. (5)

Finally, the saturation ratio Ω_n depends on the free energy ΔG_r (kJ mol⁻¹) of reaction and the temperature T of the reaction.

According to Ritchie et al.,¹⁸ precipitation of slaked lime has a different form because the rate-limiting step is diffusion of calcium species away from the lime surface. The main calcium species present in water are Ca²⁺ and CaOH⁺. The modified rate constant k_{slaked lime} can be described by the sum of the fluxes (J, mol m⁻² s⁻¹) of these species from the slaked lime surface as:

J(Ca²⁺) = 0.62D(Ca²⁺)^2/3 × v^−1/6 × {[Ca²⁺]_s − [Ca²⁺]_b} (6)

J(CaOH⁺) = 0.62D(CaOH⁺)^2/3 × v^−1/6 × {[CaOH⁺]_s − [CaOH⁺]_b} (7)

k_{slaked lime} = J(Ca²⁺) + J(CaOH⁺)/2 (8)

where s and b subscripts denote the surface and bulk concentrations respectively, D is the diffusion coefficient of the respective species (m² s⁻¹) and v is the kinematic viscosity of water (m² s⁻¹). The rest of the rate law for slaked lime precipitation has the same form as eqn (3) and (5) above.

The kinetic parameters used in eqn (3) through (8) are presented in Table 4. Although there is a well-developed database regarding the dissolution parameters for the majority of the minerals,¹⁹ data for many of the precipitation parameters have not yet been compiled or are of poor quality. In our case we used the precipitation parameters of selected sulfate, oxides, hydroxides and carbonates to describe the formation of gypsum, periclase, brucite and magnesite respectively.²⁰

Table 4 Dissolution and precipitation parameters of the selected phases (at 25 °C)

Phase	SA (m^2 g^−1)	M	m (mol)	Acid mechanism			Neutral mechanism		Base mechanism			p	q
				K^acid0 (mol m^−2 s)	E^acida (J mol^−1)	n^acid	K^neu0 (mol m^−2 s)	E^neua (J mol^−1)	K^base0 (mol m^−2 s)	E^basea (J mol^−1)	n^base
a Parameters of base mechanism in respect to [HCO3^−].b Siderite rate parameters used for magnesite precipitation.c Quartz rate parameters used for periclase precipitation.d Gibbsite rate parameters used for brucite precipitation.e Celestite rate parameters used for gypsum precipitation.
Dissolution rate parameters^18
SLQ	1.8	56.08	0.45	—	—	—	12 × 10^−5	13 600	—	—	—	1	1
SDQ	11.4	49.77	0.5	—	—	—	12 × 10^−5	13 600	—	—	—	1	1
Precipitation rate parameters^20
Calcite^a	1	100	0	—	—	—	1.810 × 10^−7	66 000	1.9 × 10^−3	67 000	1.63	0.5	2
Aragonite^a	1	100	0	—	—	—	1.810 × 10^−7	66 000	1.9 × 10^−3	67 000	1.63	0.5	2
Vaterite^a	1	100	0	—	—	—	1.810 × 10^−7	66 000	1.9 × 10^−3	67 000	1.63	0.5	2
Magnesite^b	1	84	0	—	—	—	1.61 × 10^−11	10 800	—	—	—	1	1
Dolomite	1	184.4	0	—	—	—	9.5 × 10^−15	103 000	—	—	—	1	1
Periclase^c	1	40	0	—	—	—	3.2 × 10^−12	50 000	—	—	1	4.58	0.54
Brucite^d	1	78	0	—	—	—	—	—	3.6 × 10^−6	—	1	1	1
Gypsum^e	1	172	0	—	—	—	5.1 × 10^−8	34 000	—	—	1	0.5	2

---

Slaking parameters^18
Phase	SA (m^2 g^−1)	M	m (mol)	v (m^2 s^−1)	DCa^2+ (m^2 s^−1)	DCaOH^+ (m^2 s^−1)	p	q
Slaked lime	1	74.09	0	1 × 10^−6	0.79 × 10^−9	1.58 × 10^−9	1	1

---

3. Results

3.1 Quicklime

We verified the purity of the quicklime samples with respect to rehydration and recarbonation after calcination and before the measured slaking tests. Total mass loss was quite limited for both samples (SDQ: 1.98%, SLQ: 1.25%; Fig. 2). In both cases the recorded decomposition occurs in various discrete steps indicating different volatile species. Below 180 °C an adsorbed humidity loss of <0.05% is found for SLQ and ≤0.13% for SDQ. In the range 280–400 °C, significant mass losses occur for both samples (SDQ: 1.0%, SLQ: 0.7%), attributed to dehydroxylation of hydrated lime. This reaction is followed in both samples by a gradual decomposition and mass loss step (SDQ: 0.85%, SLQ: 0.4%) extending up to 600 °C and attributed to gradual decomposition of residual carbonates from dolomite and limestone. For SLQ, a third minor mass loss (0.15%) is completed below 650 °C, which may indicate a more strongly bound carbonate species. In both cases the decarbonation mass losses are in accordance with the independently measured CO₂ contents (Table 1). The degree of rehydration and recarbonation of the two samples is a function both of specific surface area and intrinsic reactivity, as discussed by ref. 2. The specific surface area of SLQ is more than six times lower than that of SDQ (Table 1), but the extent of rehydration and recarbonation are only lower by about a factor of two. This indicates higher reactivity of SLQ.²

Fig. 2 Comparison of TG thermal analysis results of samples SLQ (top) and SDQ (bottom).

3.2 Water quality

The conductivity of the analyzed waters range from 31 to 2591 μS cm⁻¹, increasing in sequence from distilled water W2 (31 μS cm⁻¹), 75 : 25 mix W4 (274 μS cm⁻¹), 50 : 50 mix W1 (572 μS cm⁻¹), natural well water W3 (792 μS cm⁻¹), SO₄²⁻-rich spiked water W6 (2447 μS cm⁻¹), and Cl⁻-rich spiked water W5 (2591 μS cm⁻¹). The TH (17.8 to 1513 ppm CaCO₃) and m-alkalinity (30 to 350 ppm CaCO₃) values increase consistently in the same sequence (Table 2).

The dissolved anion concentrations Cl⁻, SO₄²⁻, HCO₃⁻ and NO₃⁻ are significant for gaining quantitative insights into the chemical processes that control the hydration of quicklime. Cl⁻ content ranges from 7.5 (distilled water; sample W2) to 62.5 ppm (well water; W3), with intermediate values for samples W1, W4 and W6; the maximum Cl⁻ content (920 ppm) is for spiked sample W5. SO₄²⁻ content ranges from below detection value in sample W2 to 8 ppm (sample W4) to 12 ppm (sample W1) to 25–26 ppm (samples W5 and W3, respectively), except for the SO₄²⁻ spiked sample W6 (842 ppm). Bicarbonate and nitrate ions range are low in distilled water W2 (5 ppm – below detection value) and increase in the expected order up to the natural water W3 (427 ppm and 30 ppm).

3.3 Hydration tests

The variation of reactivity due to the chemistry of the quicklimes SDQ (Mg-rich) and SLQ (Ca-rich) can be understood by examining the parameters (CaO + MgO)_lime and (denoted here as C/M¹⁵). A systematic increase of reactivity with increasing (CaO + MgO)_lime and decreasing C/M for the samples calcined at 1050 °C was previously observed² (Table 1). This work, however, focused on the differences in total reactivity of both quicklimes due to water chemistry, as shown in Fig. 3. At high concentrations, chloride anions (sample W5) promote higher reactivity, while sulfates ions (sample W6) do not show an obvious effect.

Fig. 3 Comparison of reactivity variations (expressed as peak temperature increase in °C) with the different water samples (W1 to W6, see Table 1) and quicklime samples.

3.4 Geochemical modeling

3.4.1 Equilibrium modeling. In the equilibrium model used, the quicklimes SLQ and SDQ were set to dissolve, whereas slaked lime saturation was imposed and periclase/brucite, carbonates and gypsum were considered the other possible precipitates. According to the model, SLQ produces ∼99% slaked lime and SDQ produces ∼59% slaked lime and 40% brucite or periclase. Both quicklimes also produce up to 0.15% calcite (depending on the initial bicarbonate concentration of the waters) (Fig. 6). The largest calcite fractions are precipitated upon reaction with W3 and W6 waters, which show the highest alkalinity values. Note, the model is unable to consider the possibility of vaterite precipitation and no other carbonate phase seems to precipitate. According to the saturation indices, the final waters are saturated with quicklime, calcite and the Mg-phases, whereas dolomite, magnesite and gypsum are undersaturated (SI = −6, −1.5 and −2, respectively).

Fig. 4 Variation of reactivity values (expressed as peak temperature increase in °C) against chloride and sulfate concentrations for quicklimes (a) SDQ and (b) SLQ. The lines are calculated using linear least squares fitting to chloride- and sulfate-spiked data. Error bars correspond to the standard uncertainty of the peak temperature values.

Fig. 5 Variation of reactivity rate (in °C s⁻¹) of quicklimes with concentration of the anions (a) chloride and (b) sulfate.

Fig. 6 Results of equilibrium modeling. SDQ and SLQ produce the same amount of slaked lime, regardless of the water chemistry. Water chemistry affects the formation of calcite only.

The hydration of both limes releases cations into the solutions. [Ca] reaches 700–950 mg L⁻¹ in the waters reacted with either quicklime. The equilibrated solutions become strongly alkaline (pH ∼ 12) and show high conductivity values (5500–7500 μS cm⁻¹).

3.4.2 Kinetic modeling. Kinetic reactions were investigated in a range from 0 to 1150 minutes (19 h). We used such long time in order to check whether will reproduce the experimentally needed interval for slaking (1 h). In all cases, quicklime dissolution was accompanied by precipitation of slaked lime and/or portlandite and minor calcite. Other secondary phases were limited to less than 2.3 × 10⁻⁷ moles in the whole system and so are considered absent from the system. All the waters remained undersaturated with respect to gypsum throughout the kinetic evolution.

In the models, the slaking reaction proceeds at the same rate in all six waters, regardless of their composition, thus only one representative graph is shown for each quicklime. The reaction kinetics do depend on composition and surface area of the source material. SDQ, with higher reactive surface area reacts faster. Equilibrium in SLQ is reached within 16 minutes of contact. At that time calcite formation also begins. On the other hand, when SDQ dissolves, periclase is formed faster than slaked lime. The system is equilibrated after 7 minutes of reaction (Fig. 7).

Fig. 7 Representative graphs describing the kinetic models of SQD and SLQ slaking. Quicklime, slaking lime and periclase mass fraction (%) are plotted on the right axis. Calcite mass fraction (%) is plotted on the left axis of the graphs. Equilibrium is the same in all waters.

The model also investigated the influence of carbonates or sulfates in the slaking process. SLQ slaking may produce up to 0.15% (in total mass) carbonates, in the waters with high alkalinity (W3, W5, W6). Calcite equilibrium is attained after 46 minutes in SDQ and 100 minutes in SLQ when reacting with W3, W5, or W6. Calcite saturation never occurs in the deionized water W2. No gypsum is formed from either quicklime, even in the sulfate-rich water (W6).

4. Discussion

4.1 Reviewing the role of ions in the slaking water

It has been previously reported in studies of slaking water with simple one-salt experiments^3–6,9 that the effect of each anion on the slaking process can be understood in terms of the solubility of its Ca salt: insoluble Ca salts may interfere with slaking, whereas Ca salts more soluble than Ca(OH)₂ should promote slaking. For example, high sulfate concentration may retard the hydration of quicklime due to gypsum (CaSO₄·2H₂O) formation (at the reaction surface), which may block the infiltration of water to the quicklime mass. In contrast, the addition of Cl⁻ in the slaking water accelerates the hydration process through a common-ion effect whereby the high solubility of CaCl₂ increases the dissolution rate of Ca ions, facilitating the transformation of quicklime to slaked lime. Therefore, as a general rule, the anions CO₃²⁻, SO₄²⁻ and PO₄³⁻ should slow the hydration of quicklime whereas Cl⁻ and NO₃⁻ should enhance it.^9,14

4.2 The role of natural waters in the slaking process

Conventional lime production methods are based on the assumption that high-purity water leads to successful slaking.^3–6 However, close inspection of our data demonstrates an absence of any statistical trends, positive or negative, between reactivity or hydration rate of slaking lime and concentration of anions in the slaking water, except for a positive effect on reactivity at due to Cl⁻ at the highest tested concentration only. In the concentration range from distilled up to natural well water, the reactivity is independent of the anion concentrations. No benefit in terms of reactivity or slaking rate from distillation of the slaking water was detected (Fig. 4 and 5).

Fig. 4 shows the dependence reactivity values on anion concentrations. The effect of high Cl⁻ concentrations is captured by the slope of regression lines through the distilled water and Cl⁻ spiked samples for each quicklime composition. The slope is higher for the Ca-rich quicklime (SLQ, Fig. 4a) than for the Mg-rich one (SDQ, Fig. 4b). In fact, the enhancement in reactivity due to chloride is only well-resolved for SLQ, while SDQ may not show a significant effect.

Reactivity rate vs. anion concentration is plotted on logarithmic axes, for chloride in Fig. 5a and for sulfate in Fig. 5b. Mg-rich quicklime has much lower reactivity rate compared with Ca-rich quicklime and furthermore the positive effect of high chloride concentrations on the reactivity rate is clearly visible in SLQ, while it is minor in SDQ. A slight negative effect of sulfate on reactivity rate is visible when the spiked sample is included in SLQ. However, within the natural concentration range (62.5 ppm Cl⁻ and 26 ppm for SO₄²⁻), there is no significant correlation between anion concentration and reaction rate for either quicklime.

Overall, for slaking waters within the chemical range considered in this study (e.g., conductivity < 2600 μS cm⁻¹, Cl⁻ and SO₄²⁻ <1000 ppm), we expect results consistent with the observed reactivity trends. In accordance with previous studies of simpler waters at lower conductivity,^3–6 no benefits for the reactivity of quicklime were observed when using very low conductivity waters.

4.3 Evaluation of modeling

The equilibrium and kinetic models both reproduce the slaking processes of the selected quicklimes rather well. The precipitation of slaked lime depends on the initial calcium concentration of the lime. Thus, although the whole of the quicklime dissolves, SDQ produces about half as much slaked lime (∼60%) as SLQ (∼100%) due to the initial low CaO concentration. The rest of the mass forms MgO or Mg(OH)₂ phases (∼40%). A minor amount (∼0.15%) may precipitates as carbonates if the slaking water is alkaline enough.

The models confirm the experimental observation that slaking is a fast reaction.² In particular, the kinetic model that was constructed showed that quicklime quickly reacts and disappears within the first 7 minutes in SDQ and 10 minutes in SLQ. The reaction rate is proportional to the reactive surface of the source material, and, in the models, is independent of the water composition.

When comparing the slaked lime with the water conductivity, no significant differences were observed in the quality of the final product. The chemical composition of the initial waters affects the formation of byproducts (such as calcite), but in all cases considered the precipitated mass can be considered insignificant (<0.15%). Consequently, the equilibrium and kinetic models emphasize that use of desalinized water increases the manufacturing cost without improving significantly either the production rate of slaked lime, the quantity of slaked lime formed, or the quality of the final product.

5. Conclusions

We have investigated the effect of water chemistry on the reactivity and slaking rate of laboratory-produced lime within the compositional range from distilled to lightly brackish (<2600 μS cm⁻¹) water. We infer that high chloride concentrations positively influences the quicklime hydration reaction for calcitic quicklimes, sulfate has no significant effect, and no difference is observed between distilled and natural water. As suggested by Leontakianakos et al.,² the slaking process depends dominantly on the composition of the source material, its surface area and the temperature of the reaction rather than on water chemistry. Furthermore, our observed data and modeling results suggest that slaking is not influenced by the water composition, proceeding with the same rate in all six waters for each lime. Although the kinetic model does not reproduce the experimental observation of a rate and reactivity increase at high chloride concentration, this discrepancy does not affect the conclusion that distillation of slaking waters yields no evident benefit.

Finally, we suggest the possibility of a major cost-saving benefit for the lime industry. We showed that natural waters—broadly within the light-brackish range (<2600 μS cm⁻¹)—give the same total reactivity and reactivity rate during slaking of quicklime as do low-conductivity distilled water (<35–50 μS cm⁻¹). Therefore, it is not necessary to apply any specific water treatment technology, if a light-brackish natural water source is available, before using the water in a slaking process.

Acknowledgements

Sincere thanks are due to Onassis Foundation for the financial support to G. Leontakianakos (Sponsoring information). Also, Professor Paul Asimow from California Institute of Technology is greatly acknowledged for the substantial comments and editorial improvement made on an earlier and the final version of the manuscript. We sincerely thank Dr Zhong-Yong Yuan for editorial handling and three anonymous reviewers for extremely thorough and helpful reviews.

References

1. G. I. Analysts and I. Imaging Agents, A Global Strategic Business Report, San Jose, CA, 2010.
2. G. Leontakianakos, I. Baziotis, A. Papandreou, D. Kanellopoulou, V. Stathopoulos and S. Tsimas, Mater. Struct., 2014, 1–19,  DOI:10.1617/s11527-014-0436-y.
3. R. S. Boynton, Chemistry and Technology of Lime and Limestone, John Wiley & Sons, New York, 1966.
4. G. F. Dornap, Zem.-Kalk-Gips, 1977, 30, 34–39.
5. M. Hassibi, in Presentation to 3rd International Sorbalit Symposium, 1999.
6. S. Holmes and M. Wingate, Building With Lime: A Practical Introduction. Intermediate Technology Publications, 1997.
7. S. Loeb and S. Sourirajan, Adv. Chem. Ser., 1963, 38, 117–132.
8. L. F. Greenlee, D. F. Lawler, B. D. Freeman, B. Marrot and P. Moulin, Water Res., 2009, 43, 2317–2348.
9. J. H. Potgieter, S. S. Potgieter and D. DeWaal, Water SA, 2003, 29(2), 157–160.
10. APHA, Standard Methods for the Examination of Water and Waste Water, American Public Health Association, American Water Works Association, Water Pollution Control Federation, Washington DC, USA, 2007.
11. European Committee for standardization, EN 459-2, Building Lime, Test Methods, 2010.
12. G. S. Kumar, A. Ramakrishnan and Y. T. Hung, in Advanced Physicochemical Treatment Technologies, Humana Press, 2007, pp. 611–633.
13. NF EN 549-2, 2012, Norme française et européenne, Chaux de construction, partie 2: méthode d'essai.
14. J. H. Potgieter, S. S. Potgieter, S. J. Moja and A. Mulaba-Bafubiandi, Miner. Eng., 2002, 15, 201–203.
15. I. Baziotis, G. Leontakianakos, A. Proyer, H. S. Lee and S. Tsimas, Int. J. Chem., 2011, 3(2), 187–197.
16. D. Parkhust and C. Appelo, U.S. Geological Survey, Denver, Colorado, 2013.
17. A. Lasaga, Kinetic Theory in the Earth Sciences, Princeton University Press, Princeton, NJ, 1998.
18. I. Ritchie and B.-A. Xu, Hydrometallurgy, 1990, 23, 377–396.
19. J. Palandri and Y. Kharaka, U.S. Geological Survey, Mento Park, California, 2004.
20. N. C. Marty, F. Claret, A. Lassin, J. Tremosa, P. Blanc, B. Made, E. Giffaut, B. Cochepi and C. Tournassat, Appl. Geochem., 2015, 55, 108–118.
21. R. M. Smith, A. E. Martell and R. J. Motekaitis, Users Guide, NIST Standard Reference Database, National Institute of Standards and Technology, Gaithersburg, Maryland, 2004, vol. 46.

---