Cement and Concrete Research 173 (2023) 107301

Contents lists available at ScienceDirect

Cement and Concrete Research

journal homepage: www.elsevier.com/locate/cemconres

Carbonation mechanisms and kinetics of lime-based binders: An overview

Carlos Rodriguez-Navarro *, Teodora Ilić , Encarnación Ruiz-Agudo , Kerstin Elert
Dept. Mineralogy and Petrology, University of Granada, Fuentenueva s/n, 18002 Granada, Spain

A R T I C L E I N F O A B S T R A C T

Keywords: The reaction of slaked lime with atmospheric CO2 in the presence of humidity leads to the formation of
Mortar cementing carbonate phases in traditional aerial lime mortars and plasters. This carbonation reaction also affects
Carbonation the setting and degradation of hydraulic lime mortars and modern cement. Here, we present an overview of the
Kinetics
existing knowledge on carbonation of lime-based binders, which are experiencing a revival as compatible ma­
Acceleration
Reaction
terial for the conservation of the built heritage and new sustainable construction. First, the carbonation reaction
Amorphous material is defined and its importance in a range of technical and natural processes is outlined. This sets the ground for
Pore solution presenting a review of existing mechanistic models for the carbonation of lime-based materials, including the
Microstructure recent interface-coupled dissolution-precipitation model, and the understanding of carbonation in terms of non-
Ca(OH)2 classical crystallization theory. Kinetics models and experimental results for carbonation of lime-based binders
CaCO3 (crystals and powder, as well as mortars/plasters) and its acceleration are presented and discussed. Finally,
conclusions and future research directions are indicated.

1. Introduction Portland cement [11–17]. On the one hand, the highly alkaline pH
reached during the hydration of cement can lead to the early precipi­
Carbonation refers to the reaction of a mineral including mono or tation of CaCO3, which aids in early strength gain [18]. Accelerated
divalent metal cations with CO2 to form a solid carbonate phase. In the carbonation curing of concrete is thus considered as an effective means
case of lime-based binders, carbonation is defined as the reaction of an for gaining early strength while at the same time contributing to CO2
alkaline-earth metal hydroxide with (atmospheric) CO2, resulting in the sequestration [19–21]. On the other hand, portlandite, CSH and ettrin­
crystallization of a carbonate phase as follows: gite (Ca6Al2(SO4)3(OH)12⋅26H2O) formed upon cement hydration can
undergo carbonation when the set and hardened cement is exposed to
M(OH)2 + CO2 = MCO3 + H2 O (1)
atmospheric CO2 or carbonate-bearing fluids [15,22–24]. Such a sec­
where M is either Ca2+ or Mg2+. This is a highly exothermic reaction ondary carbonation process can be highly deleterious, resulting in the
(ΔH = +74 kJ mol− 1 for the case of Ca(OH)2 carbonation) that spon­ degradation (cracking and spalling) of cement and concrete structures,
taneously takes place under Earth surface P-T conditions [1]. particularly in the case of reinforced concrete, as the reaction results in a
Carbonation is largely responsible for the setting and hardening of pH decrease from ~12.5 in uncarbonated cement, to ≤10 in carbonated
lime-based binders used in old and modern constructions (Fig. 1a). It is cement, facilitating corrosion of the steel reinforcing elements [25].
the main process responsible for strength development of aerial lime Corrosion products with high molar volume generate internal stresses
binders, both calcitic (high-calcium) and dolomitic (or magnesian) limes and cause cracking of the cover concrete aligned in the direction of
[2,3]. Carbonation also occurs to a significant extent in hydraulic limes reinforcing iron bars. Fig. 1b shows a dramatic example of such a
[4,5], where the main setting and strengthening mechanism is however degradation.
the hydration of calcium silicate (and aluminate) phases (i.e., natural Carbonation of cement can result in the formation of deleterious,
hydraulic limes), or the reaction of portlandite (Ca(OH)2) with a highly soluble alkali carbonate salts such as trona (Na3(CO3)(HCO3)⋅
pozzolanic material (i.e., artificial hydraulic limes or ancient Roman 2H2O), natrite (Na2CO3), thermonatrite (Na2CO3⋅H2O), natron
concrete, i.e., Opus caementicium), forming a matrix of calcium silicate (Na2CO3⋅10H2O), kalicinite (KHCO3) and potash (K2CO3). Alkalis (Na
(and aluminate) hydrate phases (CSH, CAH, CASH) [6–10]. and K) are commonly present as impurities in cement clinker, and upon
Carbonation is also a common and thoroughly studied process in chemical weathering (dissolution) of set cement, such alkalis can be

* Corresponding author.
E-mail address: carlosrn@ugr.es (C. Rodriguez-Navarro).

https://doi.org/10.1016/j.cemconres.2023.107301
Received 17 March 2023; Accepted 3 August 2023
Available online 17 August 2023
0008-8846/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

leached [26]. The resulting alkali-rich, high-pH pore solutions foster the (TGA-DSC), differential thermal analysis (DTA), and X-ray diffraction
dissolution and hydration of atmospheric CO2, forming CO2− 3 -rich so­ (XRD), which enable to measure the content of CaCO3 along a carbon­
lutions and leading to the precipitation of Na and K (along with Ca) ation profile. A simple and widely used method to evaluate the progress
carbonates. While formation of secondary CaCO3 might have a protec­ of the carbonation front with depth involves phenolphthalein spraying
tive effect [26], crystallization of alkali carbonates within the pore on freshly cut mortars specimens. Other techniques are, for instance,
system of cement or adjacent building materials such as stone or bricks scanning electron microscopy (SEM), optical microscopy, Raman spec­
can produce substantial damage. It is recognized that sodium carbonates troscopy, and micro-computed X-ray tomography (micro-CT) [43,44].
are, along with sodium sulfates, the most deleterious salts affecting All of them have been used to analyze the carbonation of lime mortars
building materials [27,28] as they are able to precipitate at relatively and plasters [45–47].
high supersaturations, resulting in high crystallization pressures, typi­ From an environmental perspective, carbonation is of global signif­
cally higher than the tensile strength of most building materials icance as it contributes to the draw-down of atmospheric CO2 and the
[29–34]. As an example, Fassina et al. [35] report degradation of marble regulation of Earth's climate over geologic timescales (>106 years)
at the S. Mª dei Miracoli church in Venice due to the crystallization of [48,49]. Carbonation takes place in nature following Earth's surface
sodium carbonate associated with the improper use of (chemically and chemical weathering of primary silicates including divalent metals such
mechanically incompatible) Portland cement during a conservation as Ca2+, Mg2+ or Fe2+. This is the case of minerals such as olivine
intervention. Problems like this one, which underline the in­ (FeMgSiO4), serpentine (Mg3Si2O5(OH)4), pyroxenes (MgSiO3, CaSiO3,
compatibility and deleterious effects of the use of modern cement in the FeSiO3), and Ca-plagioclase (CaAl2Si2O4), which upon reaction with
conservation of the built heritage, have contributed to the revival of CO2 in an aqueous environment result in the precipitation of carbonate
traditional lime-based binders for the conservation of ancient masonry phases such as calcite (CaCO3), hydrated magnesium carbonates (at low
[3,36,37]. T) or magnesite (MgCO3) (at high T) or siderite (FeCO3) (under condi­
Carbonation has been demonstrated to be an effective mechanism for tions of low oxygen fugacity), and SiO2 or other secondary aluminosil­
the setting and hardening of low-lime calcium silicate cements icate phases [50,51]. The overall carbonation of primary silicates is
[13,38,39]. As opposed to traditional hydration setting and hardening of given by the following Urey-type reaction [51,52]:
high-lime calcium silicates (i.e., belite, 2CaO⋅SiO2, and alite, 3CaO⋅SiO2)
(Mg, Ca, Fe)x Siy Ox+2y− t (OH)2t + xCO2 = x(Mg, Ca, Fe)CO3 + ySiO2 + tH2 O
in Portland cement (which takes place following a dissolution-
precipitation process, see Juilland et al. [40]), low-lime cements (2)
include non-hydraulic phases such as wollastonite (CaSiO3) and/or Such a natural process contributes to the safe and stable geological
rankinite (Ca3Si2O7) that react with CO2 (at relatively high CO2 pressure mineral storage of C on the Earth surface and subsurface for millions of
and T) to form calcium carbonates plus silica that provide a compact and years. This type of reaction has drawn significant attention in the last
mechanically sound structure [38,39,41]. This class of cements with few decades because carbonation of primary silicates has emerged as a
reduced carbon footprint might find numerous applications, such as the technology for both in situ and ex situ mineral carbon capture and
fabrication of pre-cast concrete elements or the sealing of CO2 injection storage (CCS) aimed at reducing anthropogenic CO2 emission and the
wells during geologic carbon sequestration (GCS) [42]. draw-down of the concentration of atmospheric CO2, a greenhouse gas
Due to its importance in cementitious materials, especially in claimed responsible for the on-going global warming [53,54]. Related to
cement, carbonation has been experimentally studied and quantified CCS, the search for more sustainable ways to dispose industrial alkaline
using a range of analytical techniques [43]. Among them, we can list wastes have led to their use as supplementary cementitious materials in
thermogravimetric analysis coupled to differential scanning calorimetry cement as well as for mineral carbon capture [55]. It has been recently

Fig. 1. Lime carbonation: positive and negative effects. a) Masonry structure of the XIIth century Colegiata of Santillana del Mar (Spain) built using sandstone blocks
and aerial lime mortar; b) the deleterious effect of delayed carbonation of a reinforced concrete structure (II World War submarine bunker Valentine, Bremen,
Germany). Massive loss of concrete scales associated with steel bar corrosion upon carbonation is observed, along with carbonate efflorescence/encrustation (white
stains). See detail in inset.

2
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

estimated that carbonation of Ca- and Mg-rich alkaline industrial resi­

dues such as fly ash, cement kiln dust, steel slag, and red mud can result
in the permanent and safe storage of about 200–300 Mt of CO2 annually
by forming Ca and/or Mg carbonates [56]. Such waste materials (also
including carbide lime and paper mill sludge) are of high interest for
their potential use as additives (e.g., pozzolanic additives) or as Ca(OH)2
sources for lime-based binders, enabling to reduce the carbon footprint
of the cement industry. Also, low T direct air capture of CO2 using CaO
and/or Ca(OH)2 for carbonation/decarbonation cycling is drawing
attention in recent years [57].
Due to its many technological applications and environmental im­
plications, the carbonation reaction has been extensively studied, not
just in the case of cement, but also in the case of lime-based building
materials (e.g., refs. [1,36,45,47,58–75]), which are experiencing a
revival, especially in the field of heritage conservation due to their high
compatibility with ancient structures [3,37,76–78]. However, despite
significant recent progress, there are still several aspects of the mecha­
nism and kinetics of carbonation that are not well understood.
Here we present a general overview of the carbonation of lime-based Fig. 2. The lime cycle of calcitic lime.
binders, focusing both on the mechanism(s) underlying this reaction, as Modified from [72].
well as its kinetics. Particular attention is paid to the understanding of
carbonation in terms of the recently proposed non-classical crystalliza­
tion theory. We also pay attention to the interface-coupled dissolution-
precipitation (ICDP) model for mineral replacement reactions [71] to
explain how, from a mechanistic point of view, carbonation takes place
at the individual portlandite crystal level. Additionally, we focus here on
progress made regarding the better understanding of this reaction and
its implications in the performance of lime mortars and plasters. Finally,
general conclusions about the mechanism and kinetics of carbonation of
lime-based binders, as well as possible research directions to further
advance the understanding of the carbonation of lime mortars and
plasters, and the optimization of this fundamental process are outlined.

2. Carbonation within the frame of the lime cycle

Lime mortars and plasters have been used for decorative and build­
ing purposes since the origin of pyrotechnology in the Levant ca.
10,000–12,000 BCE [79–82]. Lime was the binder of choice until the
invention of Portland cement back in the XIX century, which phased-out Fig. 3. The lime cycle of dolomitic lime. Note that the cycle is not closed,
lime as the primary binder in building and construction [3,83]. In recent because the end-product of carbonation is not the starting dolomite, but a full
decades, however, lime-based binders have experienced a revival as range of calcium and magnesium carbonate phases.
compatible and more environmentally friendly materials, as compared
with Portland cement, for the conservation of the built heritage and are
CaCO3 = CaO + CO2 (5)
finding applications in modern sustainable construction [37,84,85]. The
preparation and setting/hardening of lime-based binders involve a series Note, however, that Rodriguez-Navarro et al. [72] have demon­
of sequential steps collectively known as the lime cycle [2,3,37,72], strated that this reaction involves the initial decomposition of dolomite
schematically shown in Fig. 2 for the case of calcitic lime and in Fig. 3 for at relatively low T (~500 ◦ C) into a mixed Ca–Mg oxide, which rapidly
dolomitic lime. Below, we present a brief overview of the main features undergoes spinodal decomposition into CaO plus MgO. The newly
of each step of the lime cycle. formed CaO is highly reactive and re-carbonates in the kiln (via a gas-
solid reaction) to form CaCO3 that further decomposes at a higher T
2.1. Calcination into CaO and CO2. In contrast, the newly formed MgO is less reactive
towards CO2, which prevents its re-carbonation, so that it only experi­
The first step of the lime cycle involves the calcination of carbonate ences further growth (sintering) as T increases during calcination. This
rocks. If the calcined rock is a limestone or a calcitic marble, i.e., basi­ decomposition mechanism has profound effects on the reactivity of the
cally made of calcite, the material resulting from its calcination is CaO resulting oxides (i.e., highly reactive CaO and poorly reactive MgO), and
(quicklime), produced via the reaction, determines the properties of the products obtained in the subsequent
steps of the (dolomitic) lime cycle. Note that industrially, the products of
CaCO3 = CaO + CO2 (3)
the calcination of dolomite rocks are known as “dolime”, and in some
This is a highly-endothermic, solid-state topotactic reaction, which cases the dolomitic lime cycle is known as the “dolime cycle” [88].
starts (under atmospheric P) at ~600 ◦ C [86] and is typically performed There is also the possibility of using rocks made of magnesite for the
in lime kilns at T between 750 and 900 ◦ C [2,3]. In contrast, if the preparation of magnesian limes [88]. These rocks, however, are rare, as
carbonate rock used for calcination is a dolostone (made up of dolomite, compared with limestones, calcitic and dolomitic marble, or dolostones.
CaMg(CO3)2), the reaction progresses according to what is commonly Nonetheless pure magnesian limes can be produced by the calcination of
known as “two-step calcination” via the reactions [87], for instance, magnesite-rich bodies in serpentinite rock outcrops or
magnesite marble, via the reaction,
CaMg(CO3 )2 = MgO + CaCO3 + CO2 (4)

3
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

MgCO3 = MgO + CO2 (6) ash or from limestone calcination, the resulting hydrated lime carbon­
ates producing a fraction of aragonite. It could be argued that aragonite
2.2. Hydration or slaking forms because of the presence of Mg in the calcined material. It is well-
known that Mg inhibits the precipitation of calcite and favors the crys­
The second step of the lime cycle involves the hydration (or slaking) tallization of aragonite [101] as it has been observed in the case of
of the oxide(s) to form either portlandite in the case of calcitic limes, or dolomitic lime [102]. Toffolo [100], in contrast, has confirmed that
portlandite plus brucite Mg(OH)2 in the case of dolomitic limes, via the aragonite forms in Mg-free lime plasters, suggesting that (likely for ki­
following reactions: netic reasons) aragonite precipitates upon carbonation in the highly
alkaline and reactive solution derived from the hydration of pyrogenic
CaO + H2 O = Ca(OH)2 (7) CaO. Identification of original pyrogenic carbonates in archaeological
sites is key not just to disclose the presence of lime plasters and mortars,
MgO + H2 O = Mg(OH)2 (8)
but also for an accurate 14C dating [100].
Hydration is a highly exothermic process which can be performed by In the case of dolomitic limes, the end product of carbonation (in
adding to the oxide an amount of water slightly in excess of the stoi­ addition to Mg-calcite and metastable vaterite and/or aragonite), is not
chiometric amount or adding a significantly higher amount of water dolomite or magnesite, but a number of hydrated magnesium carbonate
than the stoichiometric amount. In the first case the hydrated lime forms and hydroxycarbonate phases such as nesquehonite (MgCO3⋅3H2O),
a more or less dry powder, whereas in the second case, lime putty (a landsfordite (MgCO3⋅5H2O), dypingite (Mg5(CO3)4(OH)2⋅5H2O), artin­
dispersion of portlandite crystals in water) is obtained [2,89]. Powder ite (Mg2CO3(OH)2⋅3H2O), giorgiosite (Mg5(CO3)4(OH)2⋅5H2O), and
lime, also known as dry hydrated lime, can be readily bagged and is the hydromagnesite (Mg5(CO3)4(OH)2⋅4H2O), or even a mixed Ca–Mg
main product of current industrial lime production, typically performed carbonate such as huntite (CaMg3(CO3)4) [102–108]. Neither dolomite
using steam hydrators [2]. Conversely, lime putties were the most nor magnesite forms upon carbonation of dolomitic limes because such
common traditional products used in construction since the origins of phases do not precipitate under standard P-T conditions on the Earth
pyrotechnology [3]. One of the main advantages of the use of slaked surface due to kinetics reasons. This is the root-cause of the so-called
lime putties is the fact that their rheological properties and reactivity “dolomite problem”, which is based on the fact that, for yet unknown
improve upon long term storage under water. This so-called “aging” reasons, dolomite was very common in the geologic past but extremely
process results in the development of sub-micrometer sized plate-like scarce in the recent geologic record, and its low T (abiotic) synthesis in
portlandite crystals [89] that are highly reactive (i.e., have a high spe­ the laboratory has systematically failed [101].
cific surface area) and impart a high plasticity to the lime paste [90]. In
turn, drying is avoided (as it occurs during the preparation of a dry 2.3.2. Role of aggregate and additives
hydrate) thereby preventing detrimental particle coarsening via drying- For practical building or decorative applications, hydrated limes are
induced irreversible oriented aggregation [91]. typically mixed with an aggregate (carbonatic or silicic in nature) to
prepare plasters and mortars once they are mixed with water [77]. Such
2.3. Carbonation products are applied in the fresh state by masons either as structural
mortars or as finishing plasters or renders. Upon application, they
The third, and final, step of the lime cycle involves the carbonation of experience drying and their subsequent carbonation. As a result of this
Ca and/or Mg hydroxides to form calcium and/or magnesium carbon­ reaction, an interconnected microstructure of calcium carbonate crystals
ates. This way the end-product of the cycle is similar (compositionally, forms which is responsible for the strength development of the mortar or
but not texturally/structurally) to the starting phase(s) in the uncalcined plaster [65]. For a full understanding of their carbonation both from a
raw carbonate rock used to produce lime. However, this is only mechanistic and a kinetic point of view, the effects of the mixing water
(partially) true in the case of calcitic limes, as the end-product is typi­ and the aggregate, as well as the binder:aggregate ratio, on the devel­
cally calcite, the same mineral present in the starting limestone or opment of the set and hardened mortars or plasters phase composition
calcitic marble. It is, however, not true in the case of dolomitic limes (see and (micro)structure (i.e., porosity, pore size distribution, and pore
below). network connectivity) have to be considered [77,109,110].
There is another important aspect to consider when studying the
2.3.1. Phases formed upon carbonation carbonation of lime mortars and plasters: natural (and more recently
In the case of calcitic lime, the stable end-product of carbonation is artificial) organic and inorganic additives or admixtures were/are often
calcite [36,69,73]. However, other anhydrous metastable polymorphs added to the lime-based mortars and plasters (either during lime slaking
have been observed to develop during the carbonation of portlandite or mortar/plaster preparation) to improve the fresh and set state prop­
[75,92]. Indeed, one interesting aspect of the carbonation of calcitic erties of the mixes [80,111,112]. In ancient Europe, India and other
lime mortars and plasters is the formation of metastable CaCO3 phases. regions, plant extracts, fruit juices, oils, animal fats and even blood or
While the Ostwald rule of stages predicts that both vaterite and arago­ beer were added to slaked lime to improve the properties of mortars and
nite could precede the formation of stable calcite [93], it is rather un­ plasters [111,112,113]. Chinese builders traditionally used sticky rice
clear why in some fully carbonated lime mortars and plasters, even in (amylopectin-rich) as additive, which enables the formation of a more
the case of historical buildings, vaterite and aragonite are still present compact and durable calcite matrix after carbonation, imparting a high
[94–96]. It is likely that in the case of vaterite, its formation and sta­ durability to lime mortars [113]. Ethnohistoric, archaeological and
bilization is due to the presence of organic additives in the original analytical evidence shows that pre-Columbian civilizations in Meso­
mortar mix [96–98]. It has been widely reported that organics foster the america (e.g., ancient Maya and Aztec) used plant extracts to improve
formation and stabilization of vaterite [99]. Conversely, it is observed the properties of lime mortars, plasters and stuccoes [115–119]. How­
that in lime mortars without organics vaterite forms (in concentrations ever, how such additives affect the dynamics and kinetics of the
< 5 wt%) during the early stage of carbonation (both under normal and carbonation process undergone by the lime binder is not well known,
accelerated carbonation curing conditions) but readily transforms into currently being the subject of intensive research.
more stable calcite over time, i.e., weeks or months [92]. However, it is
not that clear why aragonite forms and is preserved in lime mortars. The 3. Carbonation chemistry
presence of this phase is of special relevance to differentiate (man-made)
pyrogenic and (natural) geogenic CaCO3 in archaeological sites. Toffolo Once a fresh lime paste (i.e., a lime plaster or mortar) is applied in
[100] has shown that upon hydration of CaO derived either from wood place, it will first undergo drying. According to Van Balen and Van

4
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

Gemert [109] this results in an initial strength gain (basically due to All these reaction steps are interrelated and altering the kinetics of
capillary forces). During this early stage, incipient carbonation can take one of them influences the others [69]. The order of these individual
place. However, carbonation of the saturated material is very limited reaction steps offers a complex mechanism that is pH-dependent because
because the diffusion rate of CO2 in an aqueous solution is ~10,000 of its strong effect on the speciation of carbonic species [120,123,124].
lower than in air [64]. Therefore, only a thin surface layer of the fresh At pH > 10, which represents the situation of the pore solution in
saturated paste will be carbonated during this early stage. Subsequently, carbonating lime mortars, HCO−3 ions can readily form by the direct
upon further drying an accelerated event of carbonation takes place. reaction of carbon dioxide and OH− ions at a forward reaction rate 8 ×
However, full carbonation may only be reached after long exposure to 103 L mol− 1 s− 1 (Eq. (15)) [122]. Because the rate of Eq. (16) is faster
atmospheric CO2 (months, years or even centuries) [47]. This is so than that of Eq. (15), and Eq. (17) is known to be instantaneous, Eq. (15)
because the carbonation process is complex and includes several rate- therefore is the rate-controlling step at this highly alkaline pH range. At
limiting steps that control how CO2 is dissolved in the pore water and lower pHs (8 < pH < 10), representative of a more advanced stage of the
reach the portlandite crystals surface, how it diffuses through the carbonation reaction, the direct hydration of carbon dioxide with water,
(saturated or open) pore system of the lime mortar or plaster from the CO2 + H2O = H2CO3, and the subsequent dissociation of carbonic acid to
(carbonated) surface to the (uncarbonated) interior, and how the form bicarbonate ions, will compete with Eq. (15), whose rate decreases
carbonation reaction progresses from the surface of the portlandite as pH decreases in a carbonate-bicarbonate buffer. At pH < 8, repre­
crystals to their core through a carbonate product layer [1,69]. To these sentative of the final stage of the carbonation process, the direct hy­
steps, which can be rate-limiting, it must be added the effect of the dration of carbon dioxide with water has a forward reaction rate of 1.1
diffusion of H2O product, from the reaction interface out to the external × 10− 3 L mol− 1 s− 1 at 25 ◦ C, which is six orders of magnitude lower than
surface of the lime plaster or mortar, followed by its evaporation [109]. that of Eq. (15) [122]. This slow reaction leads to the formation of HCO3ˉ
Fig. 4 graphically shows these effects. and H+, making the solution more acidic. As a result, precipitated cal­
cium carbonate may undergo (partial) dissolution at such acidic con­
ditions until the pore solution reaches saturation with respect to CaCO3,
3.1. Reactions and rate-controlling steps
thus increasing the pH up to ~8.5–9 [125]. This will enable further
restructuring/regrowth of the calcite crystals, which typically change
Carbonation of calcium hydroxide involves the chemical reaction
their shape from scalenohedral to rhombohedral [69]. Cizer et al. [69]
between atmospheric carbon dioxide and Ca(OH)2 dissolved in the pore
have shown that the high [Ca2+]/[CO2− 3 ] ratio during the early
water of the mortar, resulting in the precipitation of calcium carbonate
carbonation of lime plasters favors the formation of calcite crystals with
according to the general eq. [47,61]:
scalenohedral morphology. However, upon further progress of the
Ca(OH)2 (s) + CO2 (g) + H2 O (aq) = CaCO3 (s) + 2H2 O (aq) (9) carbonation reaction, such crystals tend to evolve into rhombohedral
shaped calcite crystals via a dissolution-reprecipitation process [69].
Note that H2O is included in the left and right terms of this equation
Such a morphological evolution can have a significant impact on the
to highlight the crucial role water plays in the reaction (see below). The
structure and physical-mechanical properties of the set and hardened
reaction is exothermic and proceeds spontaneously under standard P-T
lime mortars [92].
conditions [1].
In general terms, and irrespectively of the pH, the slow conversion of
In the case of dolomitic limes, there are several possible overall
carbon dioxide into HCO3‾ is the rate-controlling step of the carbonation
carbonation reactions for magnesium hydroxide, such as:
reaction [123,124,126] and significantly limits the yield of the overall
Mg(OH)2 + CO2 + 2H2 O→MgCO3 ⋅3H2 O (nesquehonite) (10) reaction.

5Mg(OH)2 + 4CO2 →Mg5 (CO3 )4 (OH)2 ⋅4H2 O (hydromagnesite) (11) 3.2. Solid-state vs. through-solution reaction: the role of H2O

5Mg(OH)2 + 4CO2 + H2 O→Mg5 (CO3 )4 (OH)2 ⋅5H2 O (dypingite) (12) There is another critical rate-limiting factor during carbonation of Ca
If we focus on the carbonation reaction of high-Ca lime, the most (OH)2: water availability. It has been known for more than a century
common in building applications, the overall carbonation reaction in Eq. that little or no carbonation of portlandite takes place in dry or very low
(9) involves the following steps [1,69,120–122]: (i) continuous disso­ relative humidity (RH = 100(pH2O/pH2Osat), where pH2O and pH2Osat
lution of calcium hydroxide in the pore water with dissociation of Ca2+ are the partial pressure of water under actual and saturation conditions,
and OH‾ ions (Eq. (13)); (ii) dissolution of gaseous CO2 into the alkaline respectively) conditions at room T [127]. This observation suggests that
pore solution to form a loosely hydrated aqueous form (Eq. (14)); (iii) the formation of CaCO3 during portlandite carbonation is a through-
hydration of CO2 with OH‾ ions to form carbonic acid (H2CO3) followed solution process at room T [61,128]. It has been shown, however, that
by its (nearly instantaneous) dissociation into bicarbonate (HCO3ˉ) (Eq. in nominally dry (RH < 0.01 %), low pCO2 and room T conditions,
(15)) and carbonate (CO2− powder Ca(OH)2 can undergo carbonation by physisorption of CO2
3 ) ions (Eq. (16)) and, finally, (iv) the reaction
between Ca2+ and CO2− followed by chemisorption, with further progress of the reaction cata­
3 ions forming a calcium carbonate precipitate
through nucleation and subsequent growth resulting in an inter­ lyzed by product H2O [129]. Yet, Potinga et al. [129] state that water
connected microstructure (Eq. (17)) [1,69,70]. molecules present in the system would already be physisorbed on por­
tlandite crystals to kick-start the carbonation process. In any case, the
Ca(OH)2 (s) = Ca2+ (aq) + 2OH− (aq) (13) authors reported that the carbonate yield under such dry conditions was
almost negligible. Shih et al. [62] and Montes-Hernandez et al. [130]
CO2 (g) = CO2 (aq) (14) also pointed to the autocatalytic role of H2O released during the
carbonation reaction. In all these cases carbonation was assumed to be a
CO2 (aq) + OH− (aq) = HCO3 ¬ (aq) (15) gas-solid reaction, that is, a solid-state reaction, which might be relevant
in terms of reaction rate for the carbonation of Ca(OH)2 at high T
HCO3 − (aq) + OH− (aq) = CO3 = (aq) + H2 O (aq) (16)
(>200 ◦ C) when ion mobility is sufficiently rapid as to facilitate sub­
stitution of OH− by CO2− 3 , resulting in a change of d001-spacing of por­
Ca2+ (aq) + CO3 = (aq) = CaCO3 (s) (17)
tlandite and the progress of carbonation [131]. A solid-state model,
The same equations describe the carbonation of brucite present in however, does not explain why a few adsorbed monolayers of H2O on
dolomitic limes (i.e., replacing Ca with Mg in Eqs. (13) and (17), but portlandite particles are necessary to enable the progress of carbonation
considering that the end product is a hydrated phase). at room T at a significant rate [132], because the presence of water is not

5
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

Fig. 4. The carbonation of lime plasters: Scheme of the progress of the carbonation front across a lime render. Note the counter diffusion of CO2 (reactant) and H2O
(product). The FESEM photomicrograph shows a detail of the calcite crystals structure formed after carbonation of Ca(OH)2.

6
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

required during a solid-state reaction. Conversely, water is critical This equation shows that the radius of the critical cluster in equi­
during a reaction involving dissolution followed by or coupled to pre­ librium with a supersaturated solution is proportional to its surface
cipitation [133]. Further analyses have clearly demonstrated that, energy and inversely proportional to the system's supersaturation and T.
indeed, for carbonation of Ca(OH)2 at room P-T conditions to progress at This means that it is thermodynamically more favorable to nucleate a
a significant rate, the presence of H2O even in very low quantities (i.e., new phase (with smaller rc) at high rather than at low supersaturation.
adsorbed monolayers) is required [75,132]. Water adsorbed on por­ From Eqs. (18) and (19) the free energy barrier for nucleation, ΔG* is
tlandite enables its (partial) dissolution according to Eq. (13). As a given by,
result, CO2 can readily dissolve and hydrate forming bicarbonate and
16πvm 2 γ3
carbonate ions in the high pH aqueous solution in contact with the ΔG* = [ ( ) ]2 (20)
portlandite substrate, thus enabling CaCO3 precipitation. Note, howev­ 3 kTln aa0
er, that the high-T (>200 ◦ C) carbonation of Ca(OH)2, as well as CaO,
which has been thoroughly studied both from a mechanistic and kinetic Eq. (20) shows that the free energy barrier varies with the cube of γ,
point of view due to its relevance for CO2 mineral capture and ther­ underlining that the surface energy of the crystal nuclei is a critical
mochemical energy storage, is considered a true gas-solid (solid-state) parameter during nucleation. However, this parameter is very difficult
reaction [134–136]. to determine with accuracy, limiting the applicability of CNT [141,142].
In the following sections we will further explore the mechanism of Fig. 5 shows the evolution of ΔG considering a CNT scenario (i.e.,
carbonation involving dissolution, with a focus on the carbonation of Ca calculated using Eq. (18) and considering the variation of cluster size as
(OH)2 (and Mg(OH)2) at room T and atmospheric pCO2, conditions most the reaction coordinate), evidencing that there is a maximum, ΔG*,
relevant for the setting and hardening of lime-based binders. given by Eq. (20), that must be overcome for the system to reach the
minimum free energy, represented by the stable crystalline product
4. Carbonation mechanism: a non-classical crystallization view phase. Note that Eqs. (19) and (20) are directly applicable for homo­
geneous nucleation. In most cases, and particularly for the case of lime
The actual mechanism resulting in the conversion of Ca (or Mg) carbonation, there are however pre-existing surfaces that favor hetero­
hydroxide into a product carbonate phase has been the subject of geneous nucleation resulting in lower values of ΔG* than those deter­
extensive research. Early studies on the carbonation of hydrated lime are mined for homogeneous nucleation [143–145].
summarized by Boynton [2], and more recent ones are presented in refs.
[1,10,36,37,45,47,64–75,77,102]. They focused on the analysis of the 4.2. Carbonation in the light of non-classical crystallization theory
reaction of Ca (or Mg) hydroxide with CO2 in an aqueous solution,
observing the evolution of the reaction and determining the product CNT shows that there is a high energy barrier for the nucleation of a
phases. Yet, it is not fully clear how the carbonate phase nucleates and solid phase to occur. However, there is the possibility of reaching the
grows in such an aqueous solution. energy minima of the system without necessarily overcoming the energy
barrier determined by CNT using Eq. (20). As shown by Fig. 5, if the
transition from the solution to the stable crystalline phase proceeds via a
4.1. Carbonation and classical nucleation theory series of (meta)stable states, it is possible to bypass such a high nucle­
ation barrier [146,147]. This possibility, which considers crystallization
It has been commonly assumed that carbonation leads to the direct not just under thermodynamic but also under kinetic control [146], is at
precipitation of a particular crystalline carbonate phase, typically stable the root of the so-called non-classical crystallization (NCC) theory.
calcite in the case of high calcium lime plasters [2]. Nucleation and The CNT view of crystal formation is currently challenged by an
growth of this product phase is explained in terms of classical nucleation increasing body of theoretical, computational, and experimental studies
theory (CNT), which is rooted on the seminal works by Gibbs [137,138], disclosing alternative crystallization processes and routes, including the
further developed by Volmer and Weber [139] and Becker and Döring existence of stable pre-nucleation clusters (PNC) [148], liquid [149,150]
[140], among others, for the specific case of the nucleation of a solid and solid amorphous precursor phases [74,75,119,151,152] during the
phase from solution. According to CNT the formation of a solid in a pre- and post-nucleation stages. Moreover, it has been shown that
solution is a first-order phase transition occurring once an aggregate or growth of a newly formed solid phase (amorphous or crystalline) does
cluster of monomers (atoms, ions or molecules) that continuously form not necessarily progress via monomer addition as predicted by CNT. It
and disintegrate, reaches a critical size rc, that enables spontaneous can take place via the addition or aggregation of nanoparticles (amor­
growth to a macroscopic size via monomer-by-monomer incorporation phous or crystalline), including oriented-aggregation (OA) and meso­
into the crystal lattice [141]. For this to happen, a free energy barrier ΔG crystal formation [152]. All these processes are considered as “non-
must be overcome, which involves two competing factors: the free en­ classical” crystallization routes, which are schematically presented in
ergy consumption needed for the creation of a new surface, ΔGs, and the Fig. 6.
energy released by the formation of the bulk solid phase, ΔGv, NCC plays a crucial role in the formation of carbonates, being
4
πr3
( )
a particularly relevant for the case of lime-based binders undergoing
ΔG = ΔGv + ΔGs = − 3 kTln + 4π r2 γ (18) carbonation. Gebauer et al. [148] demonstrated that the formation of
vm ac
CaCO3 is preceded by the formation of “stable prenucleation clusters”
where r is the size of the cluster, vm its molar volume, k the Boltzmann (PNCs) with size below the critical radius (<1–2 nm). Upon aggregation,
constant, T the temperature, a and ac are the ionic activity in solution such clusters, which are described as “dynamically ordered liquid-like
and the (equilibrium) ion activity of the newly formed solid phase, oxyanion polymers” (DOLLOPs) [153], can lead to liquid-liquid phase
respectively, a/ac is the supersaturation of the system, and γ is the sur­ separation forming highly hydrated liquid-like entities, which upon
face energy of the solid phase. Taking the derivative of Eq. (18) and water exclusion and densification result in the formation of an amor­
setting it equal to zero, the value of rc, that is, the radius of a nuclei phous solid: amorphous calcium carbonate (ACC) [154]. Note that the
where the surface and volume terms of Gibbs energy are equal, is formation of an amorphous solid precursor does not fit within the
obtained, classical CNT crystallization picture where densification and long-range
order must emerge simultaneously during precipitation [155].
2vm γ ACC in turn can undergo growth and/or aggregation before trans­
rc = ( ) (19)
kTln aa0 formation into a more stable crystalline phase. Controversy exists,
however, regarding how such an amorphous-to-crystalline transition

7
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

Fig. 5. Free energy landscape for crystallization pathways under thermodynamic (classical pathway) and kinetic control (solution → DOLLOP/PNCs → ACC→ stable
crystal). The free energy barrier for nucleation depends on whether the system follows the classic pathway (red line) according to CNT, or the non-classical pathway
(black line) according to NCC. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 6. Classical vs. nonclassical crystallization:

Possible alternative routes for the formation of a
crystalline phase and its growth via classical
(grey arrowed lines) and non-classical crystalli­
zation (solid black arrowed curves). The route
involving an amorphous precursor (most rele­
vant for lime carbonation) is indicated (red
arrow). (For interpretation of the references to
color in this figure legend, the reader is referred
to the web version of this article.)
Reprinted from [152], with permission by AAAS.

takes place in solution or in humid air [74,75]. In an aqueous solution, it anhydrous CaCO3 polymorph, this aqueous film will be supersaturated
has been reported that the easiest and energetically favored mode of with respect to any one of these crystalline phases, enabling their pre­
transformation would be the dissolution of ACC and the subsequent cipitation and the subsequent replacement of ACC as the dissolution of
nucleation and growth of a crystalline CaCO3 phase [74,156]. the parent phase (ACC) and the precipitation front advances to the core
Conversely, at high T, where ion diffusion within a solid is favored, ACC of this amorphous phase [74,75]. Because of the tight coupling between
transforms into calcite (at T ~ 330 ◦ C) via a solid-state mechanism dissolution of ACC and precipitation of a crystalline CaCO3 phase, a
[157]. However, at low T, in the absence of a bulk aqueous solution, for pseudomorphic replacement takes place [74].
instance upon exposure to humidity or under (nearly) dry conditions (i. The general observation that in the CO2-CaO-H2O system (as well as
e., condition relevant for lime-binder carbonation), it is unclear how this in the CO2-MgO-H2O system), crystallization of calcium (or magnesium)
transformation takes place. Two schools of thought exist. One proposes carbonate phases is non-classical [74,75,152,159] has direct and pro­
that ACC can transform into any of the CaCO3 anhydrous polymorphs found consequences for the carbonation of lime mortars and plasters, as
(vaterite, aragonite or calcite) via a solid-state mechanism [157,158]. well as for any carbonation reaction.
The other school of thought proposes that ACC can transform into a It has been unambiguously demonstrated that carbonation of Ca
crystalline phase via an ICDP mechanism [74,133]. In this latter case (OH)2 proceeds via a NCC pathway [1,69,73,75,160]. After reaction of
there is an adsorbed aqueous fluid at the ACC-atmosphere interface dissolved and hydrated CO2 in a saturated Ca(OH)2 solution, the systems
saturated with respect to ACC. Because ACC is more soluble than any follows a downhill energy landscape where PNCs first form, giving way

8
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

to the nucleation and growth of ACC (after a dense liquid precursor), and well as the final stable calcite product (Fig. 8c–d). Importantly, the
its transformation into crystalline CaCO3 (Fig. 5). Phases formation and initial calcite crystals display scalenohedral form (Fig. 8c) and, over
transition(s) follow the Ostwald rule of stages with the sequence (from time, in the presence of humidity, they transform into calcite crystals
less to most stable phases): ACC → vaterite → aragonite → calcite with rhombohedral form (Fig. 8d) [69].
[74,75]. As an alternative for the formation of ACC as a precursor to crys­
Such a NCC pathway is not limited to a (super)saturated Ca(OH)2 talline CaCO3 phases during the carbonation of Ca(OH)2, Matsushita
solution in contact with atmospheric CO2, thus representing the initial et al. [161] proposed the formation of an amorphous hydroxycarbonate
carbonation of a freshly prepared lime mortar or plaster (i.e., fully or precursor with formula Ca1+xCO3(OH)2x⋅yH2O (x > 0.05, y = 0.6–0.8)
partially saturated). The same sequence of events has been observed in based on a X-ray photoelectron spectroscopy (XPS) study. Further
dry Ca(OH)2 powders [75] and dry lime pastes [69] exposed to a suffi­ studies by Wang et al. [162] confirmed that at high pH conditions,
ciently high RH as to enable adsorption of water on the surface of the relevant for the carbonation of Ca(OH)2, the formation of an amorphous
portlandite particles [132] and the formation of ACC. basic calcium carbonate (ABCC) takes place. It is therefore quite possible
Gillott [59] was the first to report on the presence of X-ray amor­ that rather than ACC, ABCC is the relevant amorphous phase preceding
phous calcium carbonate following carbonation of Ca(OH)2 crystals, the formation of crystalline CaCO3 in lime-binders undergoing
which was responsible for the exothermic band at 310 ◦ C (determined by carbonation.
differential thermal analysis) corresponding to the transformation of In the case of dolomitic limes, recent results by Oriols et al. [102]
ACC into crystalline calcite. More recent transmission electron micro­ reveal that the carbonation of calcium hydroxide also involves the for­
scopy (TEM) studies have shown that portlandite exposed to atmo­ mation of amorphous and metastable crystalline phases prior to the
spheric CO2 in a humid environment starts to dissolve, typically at the formation of stable crystalline ones. Furthermore, the authors show that
center of the (001) faces of the crystals where a high defect density leads magnesium, which is known to play a role in the stabilization of ACC
to more reactive surfaces [75]. Subsequently, a pseudomorphic [163], delays the formation of crystalline calcium carbonates. Yet, while
replacement of the remaining portlandite hexagonal plates by ACC takes the authors demonstrate that the carbonation of Mg(OH)2 saturated
place via an ICDP mechanism [133] as shown in Fig. 7a–c [75]. ACC can solutions involves the formation of an amorphous magnesium carbonate
also precipitate as individual spherical particles or as aggregates in the (AMC) phase, they do not show whether the carbonation of Mg(OH)2
bulk (highly alkaline) solution formed upon dissolution of portlandite crystals in the dolomitic slaked lime paste also follows the same non-
(Fig. 7d–f). As the transformation continues, ACC undergoes dissolution classical pathway observed for the carbonation of Ca(OH)2 crystals.
while anhydrous crystalline CaCO3 phases form. The resulting pre­ However, Mg(OH)2 crystals have been observed to undergo carbonation
cipitates include metastable vaterite (Fig. 8a) and aragonite (Fig. 8b), as via a NCC route involving the formation of amorphous magnesium

Fig. 7. Metastable amorphous precursor CaCO3 phases formed during carbonation of Ca(OH)2. TEM images of portlandite crystals before carbonation (a) and after
carbonation in air at room T and 80 % RH for 3 h (b). Note the hollow cores indicative of the dissolution of parent hydroxide crystals; c) detail of the hollow core
structure, identified as amorphous ACC (see absence of diffraction spots in the SAED pattern in inset), which pseudomorphically replaced portlandite via an interface-
coupled dissolution-precipitation reaction; d) FESEM photomicrograph of ACC nanoparticles formed following carbonation in air of a saturated Ca(OH)2 solution; e)
TEM image of ACC nanoparticles formed in a saturated Ca(OH)2 solution exposed to atmospheric CO2; f) SAED pattern of the ACC nanoparticles in (e). The diffuse
rings are indicative of the amorphous nature of ACC nanoparticles.
Figure parts a) to c) reprinted from [75] with permission by RSC.

9
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

Fig. 8. FESEM and TEM observations of crystalline CaCO3 phases formed after ACC following carbonation of Ca(OH)2. a) Spherulitic vaterite structures identified by
their TEM-SAED pattern (right); b) spindle-like aragonite structures (SAED pattern on the right, showing arced diffraction spots confirming its mesocrystal structure);
c) scalenohedral calcite (SAED pattern on the right); d) rhombohedral calcite (formed after scalenohedral calcite) (SAED pattern on the right).

10
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

carbonate (AMC) on the brucite surface prior to the crystallization of (atoms, ions, or molecules) of sub-nanometer dimensions. Ultimately,
nesquehonite [159,164]. It so appears that irrespectively of the reactant this colloidal-like nanoparticle growth mechanism, schematically
alkaline earth metal hydroxide, NCC is the ruling mechanism during depicted in Fig. 9b, would contribute to the rapid growth of calcite (and
carbonation. likely aragonite and vaterite) during the carbonation of portlandite
In summary, recent research has shown that the carbonation of (and, possibly, brucite) linking the formation of ACC observed during
calcitic (and dolomitic) limes is non-classic. As a result, the formation of lime mortar carbonation, to the actual development of the cementing
amorphous and crystalline (metastable) precursors prior to the forma­ CaCO3 crystals. Interestingly, such a non-classical growth mechanism
tion of the stable carbonate phases must be considered when modeling involving attachment of ACC nanoparticles leads to a nanogranular
the kinetics of carbonation (see below) and interpreting the evolution of structure in the final calcite crystals, which is preserved if organics ad­
the physical-chemical, structural/textural and mechanical properties of ditives are present [165,166]. The organics prevent fusion of nano­
lime plasters and mortars undergoing carbonation. particles during the transformation of ACC into crystalline CaCO3
(Fig. 9c–d). The presence of a nanogranular structure in calcite formed
4.3. Carbonation mechanism at the nanoscale: a NCC view following carbonation of portlandite may thus aid in the identification of
organics addition in ancient lime mortars and plasters.
Nanoscale investigations of the carbonation process within the frame It is currently unknown if the growth of hydrated magnesium car­
of NCC have been recently performed. One important aspect was to bonate phases formed upon carbonation of dolomitic limes also takes
elucidate how the calcite crystals formed upon carbonation of Ca(OH)2 place via this non-classic colloidal attachment mechanism. Further
could further grow in the presence of ACC precursor nanoparticles. research is warranted to explore this possibility.
Rodriguez-Navarro et al. [165] performed in situ atomic force micro­
scopy (AFM) analysis of the grow of calcite crystals in the presence of a 5. Carbonation kinetics
saturated Ca(OH)2 solution. It was observed that after an initial disso­
lution of the calcite substrate and the corresponding release of Ca2+ and The carbonation of lime mortars and plasters is recognized as a slow
CO2−3 ions, supersaturation with respect to ACC was rapidly achieved, process with sluggish kinetics. This view is supported by analytical re­
leading to the precipitation of ACC nanoparticles. Remarkably, in situ sults of lime mortars and plasters in some ancient buildings still showing
AFM imaging showed that the newly formed ACC nanoparticles attached incomplete carbonation. Marchese [167] reported that 12th century
to the (104) faces of the calcite substrate in an ordered manner, non-hydraulic lime-based mortars in mosaics at the Museum of the
contributing to layer growth via advancement of macrosteps formed by Duomo in Salerno (Italy) included uncarbonated amorphous (or poorly
the attached ACC nanoparticles (Fig. 9a). Upon further aging, the ACC crystalline) Ca(OH)2. These results were, however, challenged by
nanoparticles converted into calcite, preserving the overall rhombohe­ Newton and Sharp [168]. Nonetheless, further evidence has shown that
dral morphology of the calcite substrate. These results demonstrate that a fraction of Ca(OH)2 in lime mortars can remain uncarbonated for long
ACC nanoparticles actively contribute to the growth of calcite via a non- periods of time, typically months or years [36], and even centuries
classical attachment mechanism that mimics at the nanoscale the [47,169]. XRD and TGA/DTA analyses performed by Adams et al. [169]
growth mechanism observed at the atomic scale during classical crystal confirmed the presence of significant amounts of uncarbonated por­
growth, that is, the incorporation at kink and steps of growth units tlandite in 13th–14th century mortars collected from Salisbury

Fig. 9. Non-classical growth of calcite via attachment of ACC nanoparticles. a) Time sequence of AFM deflection images (topographic images in insets) showing the
attachment of ACC nanoparticles onto (104) calcite in the AFM fluid cell. Note how they arrange in an orderly manner following the contour of calcite rhombohedral
pits (1, 2 and 3) (90 s after injection of a saturated Ca(OH)2 solution) and fully cover the calcite surface after 15 min; b) model for the growth of calcite via attachment
of ACC nanoparticles. Attachment can lead to growth spirals or 2D islands. If attachment occurs at a step, nanoparticles first adsorb on the terrace (1), migrate to
macrosteps (2) and finally to kinks (3). The result is the ordered attachment of ACC nanoparticles along specific calcite crystallographic directions, as shown in the
AFM deflection image on the right; c) and d) FESEM images of calcite crystals overgrown via the attachment of ACC nanoparticles formed in the presence of
polyacrylate. Once ACC particles transformed into crystalline calcite, the surface overgrowth preserved the nanogranular structure imprinted by the ACC precursor
particles.
Modified from [165] with permission by ACS.

11
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

Cathedral, UK. Uncarbonated Ca(OH)2 in mortars ca. 800 years old and pCO2 conditions during standard application of lime mortars,
present in the meters-thick walls of the Civic Tower of Pavia (Italy) has carbonation involves a dissolution-precipitation process [1,65,75], the
been claimed to be partly responsible for its collapse in 1989 [170]. kinetic models used for the analysis of this reaction are based on solid-
Regarding the identification of uncarbonated Ca(OH)2 fractions in state reactions, summarized in Table 1 [86,179]. Despite this apparent
ancient lime mortars and plasters, studies on the (nano)structural fea­ inconsistency, such models have been successfully applied for the ki­
tures of carbonated lime plasters show lattice distortion in calcite that netic analysis of the carbonation process, yet with disparate results
has been associated with the presence of uncarbonated Ca(OH)2, not [75,80–183]. It should be kept in mind, however, that while they can
easily detected using conventional spectroscopic or XRD analysis [171]. yield accurate kinetic parameters, the mechanistic interpretation of the
This suggests that the presence of uncarbonated Ca(OH)2 in ancient air carbonation reaction using a particular kinetic model is not straight
lime mortars and plasters might be more common than previously forward.
thought. As a result of the observed partial carbonation, the ultimate The rate (k) of a solid-state reaction, that is dα/dt, can be generally
strength of aerial lime mortars and plaster might not be reached, even described by,
after several years. The slow carbonation kinetics might compromise the ( )
stability of masonry structures including lime-binders, at least in the dα − Ea
RT

short term, and possibly in the long term too. But may also help increase k= = Ae f (α) (22)
dt
the durability of ancient lime mortars and plasters, as the uncarbonated
Ca(OH)2 can contribute to the self-healing of fractures (it might also where A is the pre-exponential (frequency) factor, Ea is the apparent
improve the plastic behavior of the mortar/plaster) [37,172]. Fractures activation energy, T is absolute temperature, R is the gas constant, t is
would enable an easier access of CO2 (and H2O) to the uncarbonated time, f(α) is the reaction model, and α is the fractional conversion
areas, enabling precipitation of CaCO3 that could seal the fractures. This defined here by,
self-healing effect has been claimed to contribute to the high durability
X0 − Xt
of ancient Roman concrete prepared using the so-called hot mixing α= (23)
X0
technology [173].
In the case of dolomitic limes, it has been observed that the where X0 and Xt are the amount of reactant phase at time 0 and at time t.
carbonation of Mg(OH)2 is even slower than that of Ca(OH)2 [174,175]. The Arrhenius parameters are A, Ea and the kinetic model f(α) [179].
Lanas et al. [174] observed that while after 1-year exposure to atmo­ They can be obtained from isothermal kinetic data by applying the
spheric CO2 full carbonation of Ca(OH)2 was achieved in dolomitic lime above rate law (Eq. (22)). For convenience, the integral form of Eq. (22)
mortars, a negligible fraction of Mg(OH)2 had carbonated. It is not
known why brucite shows such a slow rate of carbonation as compared
with portlandite. In fact, the mechanisms and kinetics of carbonation of Table 1
Mg(OH)2 crystals in dolomitic limes have not been thoroughly studied, Rate equations for the analysis of the kinetics of solid-state reactions.
although some progress has been made recently [102].
Rate model f(α) = 1/k g(α) = kt
In this respect, it is important to indicate that Mg(OH)2 can form in dα / dt
situ within calcitic lime mortars when a dolomitic aggregate is used. The
1. Sigmoid α-T (a) Prout-Tompkins's Eq. (B1) α(1 − α) 1 − (1 −
reaction between Ca(OH)2 and dolomite results in the formation of curves α)
calcite plus brucite [176], 1.1 Nucleation (a) Random nucleation—Avrami- 2(1 − α) [1 − ln(1
and nuclei Erofeev Eq. (I) (A2) [− ln(1 − − α)]1/2
Ca(OH)2 + MgCa(CO3 )2 = Mg(OH)2 + 2CaCO3 (21) growth α)]1/2
(b) Random nucleation—Avrami- 3(1 − α) [1 − ln(1
Typically, porous calcite reaction rims around partially replaced Erofeev Eq. (II) (A3) [− ln(1 − − α)]1/3
dolomite grains are observed in lime mortars with dolomite aggregate α)]2/3
[177]. Because the carbonation of the resulting Mg(OH)2 is so slow, (c) Random nucleation—Avrami- 4(1 − α) [1 − ln(1
there is the possibility of its leaching (Mg(OH)2 being a relatively soluble Erofeev Eq. (III) (A4) [− ln(1 − − α)]1/4
α)]3/4
phase) and the creation of porosity in the mortar, ultimately leading to
2. Acceleratory (a) Exponential law (E1) α ln α
degradation. α-T curves
Another important consideration regarding the very slow kinetics of 3. Deceleratory
the carbonation of dolomitic (or magnesian) limes is the fact that the α-T curves
uncarbonated Mg(OH)2 crystals are prone to react with pollutant (acid) 3.1 Reaction (a) Zero order (F0/R1) 1 α
order (b) First order—Unimolecular (1 − α) − ln(1 −
gases such as SO2, resulting in the formation of highly deleterious sol­ decay (F1) α)
uble salts such as epsomite (MgSO4⋅7H2O) and/or hexahydrite (c) Second order (F2) (1 − α)2 [1/(1 −
(MgSO4⋅6H2O) [106]. A similar sulfation process might also occur in the α)] − 1
case of the relatively soluble hydrated magnesium carbonate phases (d) Third order (F3) (1 − α)3 (1/2)[1
− α)− 2 −
formed upon carbonation of brucite in dolomitic mortars or lime mortars
1]
with a dolomite aggregate suffering a dedolomitization reaction (Eq. 3.2 Diffusion (a) One dimensional transport 0.5 α− 1
α2
(21)). Note that calcitic lime mortars, once carbonated, can also expe­ mechanism (D1)
rience sulfation in polluted environments forming calcium sulfate salts. (b) 2D transport (cylindrical [− ln(1 − ((1 − α)
However, the effects of the crystallization of calcium sulfate phases are geometry) (D2) α)]− 1 ln(1 −
α)) + α
much less dramatic than those due to the crystallization of magnesium
(c) 3D diffusion, spherical 1.5(1 − (1 − (1 −
sulfates [27,106,178]. symmetry—Jander Eq. (D3) α)2/3[1 − α)1/3)2
(1 − α)1/
3 − 1
5.1. Kinetics models ]
(d) 3D diffusion, spherical 1.5[(1 − 1 − (2/
symmetry—Ginstling- α)− 1/3 − 3)α − (1
Considering the implications of the above-described effects, much Brounshtein Eq. (D4) 1]− 1 − α)2/3
research has been dedicated to understanding the kinetics of lime 3.3 Phase- (a) 2D (cylindrical geometry) 2(1 − α)1/2 1 − (1 −
carbonation to determine rate controlling steps and rate constants, boundary (R2) α)1/2
Arrhenius parameters, and reaction model(s) best fitting the observed reaction (b) 3D (spherical geometry) (R3) 3(1 − α)2/3 1 − (1 −
α)1/3
kinetics. Although it is now recognized that, under the prevailing low T

12
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

is used to determine Arrhenius parameters, role in determining the rates of Ca(OH)2 carbonation in air at low T.
( ) Carbonation rates close to zero have been reported for RH < 8 %,
− Ea
RT
exponentially increasing with RH [62]. Beruto and Botter [132] indi­
g(α) = Ae t (24) cated that at RH > 70 % carbonation rates increase exponentially due to
multilayer water adsorption. The authors argued that adsorbed liquid-
where g(α) is defined by, like water played a catalytic role in this reaction, considered as a gas-
∫α
dα liquid-solid reaction. Such an effect was directly observed at the nano­
g(α) = (25) scale by Yang et al. [191] during the carbonation of portlandite crystal
0 f (α)
in air using AFM. The authors showed the formation of nanogranular
Experimental data representing the variation of α with carbonation CaCO3 precipitates on the (001) basal plane of portlandite, presumably
time t can be fitted to the integral form g(α) of the different models in ACC, only when RH ≥ 30 % (no precipitates formed at lower RH). Water
Table 1. The best fitting model is thus selected as the most appropriate vapor sorption isotherms on both calcite and portlandite experimentally
kinetic model for this reaction. Such a best-fitting kinetic model can be obtained by Beruto et al. [65] demonstrated that portlandite was more
used for the determination of the rate constant k, and Arrhenius pa­ hydrophilic than calcite. As a result, portlandite experienced multilayer
rameters. It can also provide some insights on the mechanistic model H2O adsorption at RH > 70 % favoring calcium carbonate precipitation.
best describing a particular reaction. However, as indicated above, a Because the latter phase was less hydrophilic, carbonation rates tended
particular mechanistic model deduced from the kinetic analysis, cannot to reduce as the carbonation degree increased. Dheilly et al. [190]
be considered as the “actual” mechanism of any reaction. Additional pointed out that at a high RH (≫30 %), Ca(OH)2 could dissolve in the
tests and analyses are necessary to properly determine the “actual” re­ adsorbed water film, and such a highly alkaline film would foster CO2
action mechanism. adsorption and its subsequent hydration to form carbonate ions (i.e.,
Eqs. (14) and (15)), ultimately facilitating CaCO3 precipitation onto
5.2. Kinetics of Ca(OH)2 carbonation: rate-controlling steps portlandite [62]. Note, however, that using Raman spectroscopy Dubina
et al. [192] demonstrated the formation of ACC on Ca(OH)2 powders
Any model describing the kinetics of a reaction such as carbonation (after CaO hydration) at lower RH values of 10–20 % (80 ◦ C). The study
(assumed to be a gas-solid reaction) needs to consider the contribution by Pesce et al. [193] using 18O isotope-labeled Ca(OH)2 unambiguously
of several “resistances” corresponding to reactant/by-product mass demonstrated that, at the molecular level, carbonation in air involves
transfer and chemical reaction. In general, for the relevant systems the dissolution of reactant hydroxide and atmospheric CO2 in an
considered here, the following resistances, which can eventually become aqueous-film formed on portlandite crystals. Remarkably, the isotopic
the rate-controlling steps of a carbonation reaction, are [62,69]: (1) fingerprint of newly formed carbonates showed that a significant frac­
mass transfer and diffusion of the gas towards and within the interpar­ tion of oxygen came from portlandite, demonstrating that OH− groups
ticle pores (for powder samples) or porous structure (pastes and mor­ from the latter phase are directly involved in the hydration of CO2 via
tars); (2) adsorption/hydration of the gas in aqueous surface films or Eq. (15). The authors concluded that dissolution of portlandite at the
pore solution; (3) diffusion of the dissolved/hydrated gas to the solid- solid-liquid interface was the rate-limiting step for carbonation at such
liquid interface; (4) dissolution of the solid at the solid-liquid inter­ an early stage. At latter stages, the isotopic study showed that most of
face; (5) surface reaction; (6) diffusion of reactant (and counter- the oxygen in newly formed carbonates came from H2O in ambient
diffusion of by-product) through the solid product layer. The influence humidity, suggesting that isotopic re-equilibration was due to dissolu­
of heat transfer on the kinetics of such a gas-solid reaction must also be tion of metastable precursor phases (i.e., ACC, or even ABCC) and pre­
considered because the carbonation reaction is highly exothermic. cipitation of stable ones (e.g., calcite) following the Ostwald rule of
However, only when speeding up the reaction significantly (e.g., by stages [75,189]. Chen et al. [194] performed carbonation of levitated Ca
forced carbonation in high pCO2 conditions) significant heat transfer (OH)2 aerosol particles showing that carbonation was only progressing
takes place [61]. for RH > 70 %, in excellent agreement with the RH at which multilayer
The kinetics (and mechanism) of Ca(OH)2 carbonation in solution water adsorption onto portlandite occurs according to Beruto and Botter
(aqueous phase or aqueous dispersion) have been extensively studied [132]. The product H2O released during portlandite carbonation could
due to the relevance of this process in the synthesis of precipitated in turn self-catalyze further carbonation until completion, or until a
calcium carbonate (PCC), which has important industrial applications product (CaCO3) layer formed that hampered the advancement of the
(e.g., as filler for plastics, drugs, paper, rubber, paints) reaction front to the core of portlandite particles. According to Shih et al.
[74,130,184–186]. [62], who underlined the critical role of H2O adsorption on portlandite
Carbonation of Ca(OH)2 slurries via injection of CO2 for PCC pro­ particles, the rate controlling step for carbonation was assumed to be the
duction involves the formation of metastable precursor phases both dissolution of Ca(OH)2 at the water-adsorbed surface layer, as also
amorphous (ACC) and crystalline (vaterite and aragonite) and their pointed out by Van Balen [64] and Pesce et al. [193].
partial or complete transformation into stable calcite [184–186]. The A recent study by Park et al. [195] suggests that during the inter­
same applies for the homogeneous precipitation of CaCO3 in solution action of portlandite and water, there is intercalation of H2O molecules
[74,93,156,187–189]. The main parameters controlling the kinetics of within the (001) planes of the portlandite crystal lattice, a process that
CaCO3 formation and solid phase evolution/polymorph selection are occurs along the 〈100〉 directions (Fig. 10a). As a result, an increase in
pH, T, [Ca2+]/[CO2− 3 ], supersaturation, pCO2 (i.e., concentration of CO2 the d001-spacing of ~0.39 % (i.e., interplanar distance in the [001] di­
in the gaseous phase, typically expressed in %), gas pressure P, and the rection) compared to completely dry portlandite samples was detected
presence of impurities or additives, both organic (e.g., polycarboxylates) using XRD (Fig. 10b). This result could have important implications for
or inorganic (e.g., phosphate or magnesium ions) [1]. the carbonation mechanism and kinetics because it leaves in question
In contrast, little research has focused on the understanding of the the rate-determining step of this process. According to the authors the
kinetics and mechanisms of Ca(OH)2 crystals/powders carbonation in intercalation of water molecules could be the very first step enabling
air at room T, conditions that are most relevant during the setting and CO2 molecules to interact with the portlandite structure thus triggering
hardening of lime mortars and plasters [1,64,71]. So far, it has been initial carbonation along the <100> directions, which according to
experimentally shown that parameters such as RH, T, reactant size and Ruiz-Agudo et al. [71] display a higher carbonation rate as compared
surface area, and pCO2, in addition to impurities/additives, affect the with the [001] direction.
carbonation rate and polymorph selection [62,64,75,130,132,182,190]. The pioneering study by Aono [58] indicated that for a given expo­
As indicated above, it is generally recognized that RH plays a critical sure time, t, the carbonation degree Wt of hydrated lime powder (large

13
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

Fig. 10. Effect of water molecules intercalation

within the portlandite structure on its carbonation
kinetics. a) Scheme showing the intercalation of water
molecules within the (001) planes of the portlandite
structure. This could lead to an acceleration of the
carbonation of this phase; b) XRD pattern of the 001
Bragg peak of un-hydrated and hydrated portlandite
crystals. Note the left shift in the peak position after
hydration, implying an increase in the d001-spacing
after H2O intercalation.
Reprinted from [195] with permission by Elsevier.

sample mass of 10 g), revealed an exponential relationship with CO2 water vapor saturated atmosphere. Although the determination of the
concentration, C (i.e., pCO2), given by the following (empirical) carbonation degree by measuring the weight increase might be ques­
equation, tionable, as an unknown amount of product water will be absorbed on
( ) the carbonated reactant, thus contributing to the value of Wt, and the
Wt = 30 1 − e− 0.0135Ct (26)
large sample mass might impose diffusion-controlled kinetics, these
results could suggest that carbonation rates are dependent on CO2
where Wt is determined as weight increase following carbonation in a

14
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

concentration. Contrary to this common assumption, more recent

studies have shown that carbonation rates are independent of pCO2
[1,62,64]. Van Balen [64] demonstrated that the carbonation rate of
portlandite powders is zeroth order with respect to CO2 concentration
(in the range 15–50 %, at room T), suggesting that the controlling factor
in the carbonation process might be the dissolution of portlandite at the
water-adsorbed surface. As a results, Ca(OH)2 crystals with a higher
surface area show a higher carbonation rate (per unit mass of reactant)
[64].
Accurate carbonation rates are typically measured using pure Ca
(OH)2 powder samples, where transport-limited CO2 supply to reaction
sites is negligible. In practice, however, it is observed that the carbon­
ation rate in lime-based mortars and plasters is accelerated by increasing
pCO2 [63,92,196]. Such an acceleration is not likely due to an increase
in the rate of the actual carbonation reaction, but to the diffusion of CO2
within the pore system of the mortars towards the reaction interface
(portlandite-solution interface), which should be faster the higher the
CO2 concentration gradient (see below).

5.3. Proposed kinetic models for the carbonation of Ca(OH)2 crystals/

powders

Several kinetic models for the carbonation of portlandite in air at low

T have been proposed. Note that below we refer to pure Ca(OH)2 crystals
or powders, not to mortars or plasters. This clarification is important
because their carbonation kinetics are different. Carbonation has been
described as: (i) a deceleratory (or asymptotic) process displaying no
induction time [62,190], which can be fitted to a (pseudo)second order
model (F2) eq. [130,186]; (ii) a diffusion-controlled reaction involving
3D diffusion (spherical symmetry) according to the Jander eq. (D3)
[180]; (iii) a process that follows sigmoidal-type Avrami-Erofeev ki­
netics with an induction time before nucleation and growth [181]; (iv) a
(pseudo)first order deceleratory reaction (F1) with no induction time
and only dependent on the amount of untransformed reactant [75,183];
and (v) a reaction which can be fitted to a boundary nucleation and
growth model (BNGM) where the rate depends on the nucleation of
particles at the portlandite-solution interface and subsequent growth
over the substrate surface (i.e., unreacted portlandite) after coalescence
and full coverage [182]. In the general case, the BNGM shows an s-shape
α vs. t evolution (i.e., denoting the presence of an induction time).
However, in the limiting case scenario, the rate limiting step is the
growth of the product surface layer, which is described by first order
kinetics (F1). Fig. 11 shows a comparison of α vs. t curve fittings ob­
tained from two of the above-mentioned kinetic models, exemplifying Fig. 11. Carbonation degree vs. time for Ca(OH)2 (nano)particles exposed to
the disparity of models used to fit experimental results. The disagree­ humid air. Values are fitted to two different kinetic models: a) Boundary
ment among the proposed kinetic (and underlaying mechanistic) nucleation and growth model (BNGM). Reprinted from [182] with permission
models, may lay in the fact that most of the previous studies did not by Elsevier; and b) first order (F1) model. Reprinted from [75] with permission
consider the possible role of metastable precursors phases (ACC in by RSC.
particular) in the carbonation process. For instance, quantification of
portlandite conversion into CaCO3 using XRD or FTIR neglects the for­ capture or the value of 7.5 kJ mol− 1 reported for the carbonation of
mation of ACC, underestimating the rate of conversion at earlier stages portlandite at low T using liquid and supercritical CO2 [182,198]. Note,
and leading to the assumption of an induction time [1], yielding a s- however, that Yu et al. [199] report a higher Ea value of 40 kJ mol− 1 for
shaped (e.g., Avrami-Erofeev) kinetic model. However, analysis using the high T (≥500 ◦ C) capture of CO2 by Ca(OH)2 powders forming an
TG, which quantifies the full amount of CaCO3 (amorphous plus crys­ intermediate bicarbonate phase (Ca(HCO3)2) and pointing to a reaction-
talline) yields no induction time and a (pseudo)first order kinetic model controlled process. The discrepancy in the obtained Ea values likely re­
[75]. The choice of one model or the other strongly affects k values (i.e., flects that the carbonation mechanism (and/or the rate controlling step)
reaction rates) and Arrhenius parameters. is different in the studied systems (as the experimental conditions are
From the above-mentioned kinetic analyses, Arrhenius parameters different too). Nonetheless, the relatively high Ea values obtained by
are obtained. However, there is a large scattering in their values. Camerini et al. [182] suggest that the carbonation process at room T,
Camerini et al. [182] reported apparent activation energy values of atmospheric pCO2 and in the presence of humidity (75 % RH), which
~31–60 kJ mol− 1 for different types of hydrated limes and nanolimes (i. most closely reflects the carbonation conditions of lime plasters and
e., Ca(OH)2 nanoparticles, typically with size < 200 nm, used as alco­ mortars applied in the field, is reaction-controlled (i.e., surface-
holic dispersions for the consolidation of cultural heritage materials controlled), whereas for the lower Ea values reported, carbonation is
[197]). These values contrast with values of activation energy of 6–12 transport-controlled (i.e., diffusion-controlled) [200]. Interestingly, the
kJ mol− 1 reported for the dry solid-gas carbonation at relatively high T values of Ea reported by Camerini et al. [182] are in excellent agreement
(>250 ◦ C) of portlandite micrometric aggregates used for acid gas

15
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

with those reported for the dissolution of common silicate and carbonate center of portlandite (001) faces, where a higher defect density is pre­
minerals (Ea values ranging from 35 kJ mol− 1 for calcite to 75 kJ mol− 1 sent (Fig. 8). However, these areas of high defect density do not act as
for quartz) when dissolution is reaction-controlled [200]. It is thus very nucleation sites for CaCO3. Instead, being highly reactive, they are the
likely that, as pointed out by Shih et al. [62], Van Balen [64] (2005), and first areas to dissolve, leading to the formation of ACC which eventually
Pesce et al. [193], the rate controlling step for portlandite carbonation is transforms into crystalline CaCO3.
the actual dissolution of this phase in the aqueous layer adsorbed on its
surface. 6. Carbonation kinetics of lime mortars and plasters
There is another aspect that needs to be considered when analyzing
the kinetics of carbonation of portlandite crystals: the presence of de­ The above kinetic studies were performed using samples of pure
fects. Pisu et al. [183] detected an anomalous Raman emission band at (typically thin) powders of Ca(OH)2 crystals. In contrast, in the case of a
780 cm− 1 using near-infrared (NIR) excitation in portlandite crystals, lime plaster or a mortar, where a 3D porous structure exists, carbonation
which increased in intensity upon thermal treatment (max. at proceeds in three different stages with clearly differentiated kinetic re­
200–300 ◦ C). This luminescence effect was associated with the creation gimes [1].
of point defects. A kinetic analysis of the carbonation of heat-treated
portlandite showed first order kinetics (F1) with faster conversion for
the samples annealed at 200–300 ◦ C, i.e., those with the highest defect 6.1. Carbonation stage 1
density. The authors concluded that defects acted as nucleation sites for
CaCO3, playing a key role in speeding carbonation in annealed por­ During the initial stage, CO2 diffusion to reacting sites (i.e., por­
tlandite. These results partially agree with the observations by tlandite surface) is limited due to saturation by capillary water in the
Rodriguez-Navarro et al. [75] showing that carbonation starts at the porous system of the plaster or mortar [64,109]. The carbonation ki­
netics of this initial stage are therefore strongly dependent on the drying

Fig. 12. ACC particles formation during carbonation.

a) ACC nanoparticles formed at the air-solution
interface forming fractal-like aggregates observed
under a polarized light microscope. Plane light image
(left) and corresponding crossed polars image (right).
Note the absence of birefringence under crossed po­
lars, demonstrating the amorphous nature of the
particles; b) ACC nanoparticles formed on the surface
of portlandite hexagonal plates (squared area) in a
lime paste during early carbonation (24 h) in air, at
room T and 60 % RH.
Part (b) reprinted from [69] with permission by
Springer.

16
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

rate of the mortar as well as on the rate of H2O production following Ca 6.3. Carbonation stage 3
(OH)2 carbonation [64]. Carbonation first occurs on the surface of the
plaster/mortar where pore solutions saturated with respect to Ca(OH)2 During the final stage, the overall carbonation rate is reduced as the
are in direct contact with atmospheric CO2. Rodriguez-Navarro et al. reaction front moves further away from the surface of the porous plaster
[74] have shown that under these conditions very rapid carbonation or mortar. In this case the reaction is controlled by diffusion of CO2 to
takes place via the formation of ACC at the air-solution interface the reaction front through the carbonated mortar layer. In turn, the
(Fig. 12a). The precipitated ACC undergoes self-organization [201] reaction rate is also affected by the counter-diffusion of product H2O (as
displaying a fractal-like structure associated with diffusion-limited water vapor) from the reaction interface towards the exterior of the
colloid-aggregation [202]. During this initial stage, the surface of por­ structure.
tlandite crystals in the plaster matrix can also start to carbonate, forming
an ACC phase. While the smaller, more reactive portlandite crystals 6.4. An explanation for the transition between stage 2 and 3
could fully dissolve and transform into ACC during this stage, as shown
in Fig. 8, ACC can cover the less reactive, larger portlandite crystals While the transition between stage 1 and 2 is explained by the
(Fig. 12b). This ACC surface covering can hinder further portlandite transformation of (partially passivating) ACC formed on Ca(OH2) crys­
dissolution [1], explaining why after an initial fast carbonation the rate tals into crystalline CaCO3 [1], it is not so clear why/how a transition
decreases sharply, leading to a dormant period at the end of stage 1. from stage 2 to stage 3 occurs. As indicated above, the formation of a
passivating CaCO3 product layer on the reactant Ca(OH)2 surface has
6.2. Carbonation stage 2 been suggested as a rate limiting effect that could contribute to
explaining why the carbonation of portlandite crystals [75,182] as well
During the second stage, conversion of metastable ACC into stable as lime plasters and mortars shows deceleratory kinetics [61].
calcite (or vaterite and/or aragonite) takes place, a phase transition that Ruiz-Agudo et al. [71] studied at the nanoscale how the carbonation
according to Rodriguez-Navarro et al. [74] occurs via a dissolution/ process evolved in the case of portlandite single crystals, providing
precipitation mechanism. Also, further drying of the mortar paste con­ additional insights on the mechanism of carbonation at the nanoscale
tributes to creating an open pore network that facilitates CO2 access to and showing that full passivation might not occur. The authors per­
reaction sites (i.e., the surface of unreacted portlandite). These two ef­ formed a detailed in situ AFM analysis of the dissolution of portlandite
fects trigger another fast carbonation period, until the rate starts to single crystals and their subsequent carbonation, complemented by ex
decrease and eventually becomes asymptotic during the following stage situ FESEM and 2DXRD analyses of portlandite single crystals subjected
3 [1]. to carbonation in air at room T and 93 % RH. The authors showed that
the dissolution and carbonation reactions are strongly anisotropic, tak­
ing place at a faster rate along the <100> than along the 〈001〉 di­
rections of portlandite. Moreover, it was observed that initial

Fig. 13. Carbonation of portlandite single crystals. a) Topographic (height) AFM image of the (001) surface of portlandite covered by pyramidal calcite crystals
following initial carbonation at 93 % RH in air at room T; b) scheme of the formation of calcite on the portlandite (001) basal plane; c) atomic structure of portlandite
projected along [001]; d) atomic structure of calcite projected along [001]. Note the similarity of the structures with just a difference in the Ca–Ca bond length of
~20 %, which enables epitaxial crystallization of calcite on portlandite; e) fractional conversion (α) of portlandite into calcite over time (the inset shows and FESEM
image of the basal plane of carbonated portlandite); f) FESEM image of a cross-section of partially carbonated portlandite showing the detachment of a thin surface
layer of calcite from the {100} faces of portlandite; g) pervasive fracturing in calcite surface layers formed on the basal plane of partially carbonated portlandite.
Reprinted from [71] with permission by ACS.

17
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

carbonation following dissolution of the (001) faces resulted in the the carbonate product layer formed on portlandite crystals, but rather to
epitaxial precipitation of numerous calcite crystals (most likely after other(s) rate-controlling process(es) that will be discussed below.
ACC formation) with their c-axis parallel to the c-axis of portlandite
(Fig. 13a). Such crystals nucleated as randomly distributed islands on 6.5. Carbonation stages in dolomitic lime mortars
such surfaces, not forming a continuous carbonate layer. Importantly,
2DXRD analysis showed that the individual calcite crystals displayed a The previous discussion refers to portlandite crystals in high calcium
random orientation along their <100> directions (i.e., they were lime pastes, plasters, and mortars based on the large body of published
rotated around their c-axis). This demonstrated that the formation of research. In contrast, little information exists on the kinetics of the
calcite on portlandite (001) faces was epitaxial, not topotaxial (and a carbonation of brucite in dolomitic limes. Nonetheless, there are some
similar epitaxial effect was expected for the (100) faces). These results experimental results on the carbonation of this hydroxide indicating
also demonstrated that the process did not proceed by a solid-state that, as in the case of portlandite, negligible carbonation occurs in the
replacement mechanism, which would be topotaxial. Rather, the reac­ absence of water, because carbonation takes place via a dissolution-
tion involved a tight ICDP mechanism [133]. However, it was postulated precipitation mechanism [164]. Precipitation of a poorly crystalline
that because the transformation of portlandite into calcite implies an carbonate phase in the early stages of the reaction does not significantly
increase in volume of ~12 % (i.e., the molar volume of portlandite is hinder brucite dissolution, as the carbonate coating remains sufficiently
32.81 cm3/mol, and the molar volume of calcite is 36.90 cm3/mol) an permeable, but Harrison et al. [164] postulate that the conversion of this
impervious (passivating) calcite layer could form eventually. Interest­ phase to substantially less porous, crystalline nesquehonite could result
ingly, FESEM observation of split sections of partially carbonated por­ in passivation of the brucite surface. However, no experimental proof for
tlandite crystals showed fracturing at the carbonate/hydroxide the later has been presented. In any case, there is no reason to exclude
interface, resulting in the detachment of carbonate layers along the the possibility of the existence of three distinct stages during the
(001) planes (Fig. 13b) [71]. This was caused by strain accumulation at carbonation of the Mg(OH)2-component of dolomitic limes as observed
this interface associated with the volume increase in the product layer, for the case of the Ca(OH)2-component. Yet it is expected that the
resulting in stresses high enough to cause the observed fracturing. carbonation of the portlandite component in dolomitic limes will be
Fracturing and detachment of carbonated layers exposed fresh por­ affected by Mg2+ ions in the pore solution. As indicated above, Mg2+ is
tlandite surfaces for carbonation to progress towards the core of the known to stabilize ACC [163], which would likely prolong the dormant
parent phase. This study thus showed that the portlandite crystals were period of stage 1, as the conversion of ACC into crystalline (Mg)calcite
not fully passivated by the carbonate product, in contrast to what was will be delayed.
stated by Galan et al. [203] (see below). According to the results by Ruiz-
Agudo et al. [71], the transition from stage 2 to 3 is due to the formation 6.6. Key parameters affecting carbonation rates
of a product layer on the portlandite crystals surface, which does not
lead to full passivation. The latter explains why during stage 3 the Studies on the carbonation kinetics of cementitious materials focused
carbonation rate does not reach a zero value if a fraction of uncarbo­ mostly on Portland cement, but in some cases, they also dealt with lime
nated portlandite still persists [1]. Ultimately, Ruiz-Agudo et al. [71] plasters and mortars. These studies show that there are three main
showed that even though the carbonate product layer did not result in factors affecting the overall carbonation rates: (i) T, (ii) RH, and (iii) CO2
full passivation, it acted as a diffusional barrier, inducing an exponential concentration [17,62,205]. Other factors that have a less pronounced
reduction of the carbonation rate over time (Fig. 13e). effect on the carbonation rate are the pressure of the reacting gas, grain
Galan et al. [203] also observed epitaxial films of CaCO3 on mm- size and surface area, and porosity and pore-size distribution [206].
sized portlandite crystals subjected to carbonation at room T in labo­ Below we focus on the evaluation of the three main factors affecting
ratory air as well as 100 % CO2 atmosphere under variable RH condi­ carbonation rates.
tions (25 to 90 %). By performing dissolution tests, the authors
concluded that the CaCO3 surface shells formed at 75 % RH were 6.6.1. Temperature
passivating, but those formed at 90 % RH were not. It seems likely that at The solubility of portlandite decreases with increasing T [2,61]. Such
90 % RH, the adsorbed water layer was too thick to enable a pseudo­ a retrograde solubility means that there will be less Ca2+ in solution
morphic replacement via a tightly coupled ICDP process (i.e., CaCO3 (from Ca(OH)2 dissolution) to react with dissolved CO2 as T increases.
could nucleate both in the bulk aqueous films as well as at the Similarly, the solubility of CO2 is inversely proportional to T [61], and
portlandite-solution interface), so a non-continuous permeable calcite there will be less dissolved CO2 at high T than at low T. Thus, lower
layer developed. Conversely at 75 % RH, the few adsorbed H2O mono­ carbonation rates are expected as T increases [17]. Conversely,
layers facilitated a tightly coupled ICDP process (i.e., CaCO3 nucleation carbonation of portlandite shows Arrhenius behavior (see previous
only occurred at the portlandite-solution interface), resulting in a section). This means that the reaction rate will increase with T. Simi­
pseudomorphic replacement, and yielding a non-porous impervious larly, an increase in T will also increase the diffusion rate of reactant and
product surface layer. It is interesting to consider two aspects of the product species (i.e., CO2 and H2O) thereby speeding up carbonation. It
study by Galan et al. [203]: (i) while the carbonate shell formed around follows that there should be an optimal T at which these two opposite
the portlandite crystals reduced the dissolution rate of the unreacted Ca effects balance out, resulting in a maximum carbonation rate. In the case
(OH)2, it is unclear whether this shell would be able to fully passivate the of cement, increasing rates of carbonation are observed with increasing
hydroxide against further carbonation, and (ii), what is more important, T (for constant RH conditions) up to ~60 ◦ C, which marks an inflection
the thickness of the calcium carbonate shell formed on mm-sized por­ point in the carbonation rate [17]. It is expected that such an inflection
tlandite crystals reached values of up to ~100 μm after 6 months point would also exist in the case of aerial lime mortars. For the latter
carbonation in air (laboratory conditions). Considering that Ca(OH)2 case, however, this inflection point has not been determined, although
crystals in slaked lime typically have sizes up to a few micrometers Van Balen and Van Gemert [109] state that the optimum carbonation
[36,91], complete carbonation of such crystals under standard condi­ speed is found at ~20 ◦ C. There is another important factor to consider:
tions (room T, atmospheric pCO2 and medium-high ~75 % RH) can be increased T (leading to lower RH for a given pH2O) will speed up the
expected within a few weeks. In conclusion, and in contrast to what has evaporation of water from the carbonating lime mortar. If the evapo­
been stated by several researchers regarding the carbonation of Ca(OH)2 ration rate is too high, premature drying might occur, stopping the
[12,203,204], we can state that the slow kinetics of lime binders' carbonation process. Conversely, if T is too low, the evaporation rate
carbonation, leading to the observed incomplete carbonation of old would be lower, and product H2O might accumulate in the pores,
(medieval) mortars and plasters is not due to a full passivation effect by hampering CO2 diffusion and reducing the carbonation rate.

18
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

Sanchez-Moral et al. [207] explored the effect of T on the kinetics of further CO2 dissolution will lead to an acidification of the mortar's pore
lime mortars carbonation, and observed faster and more thorough solution and foster dissolution of scalenohedral calcite crystals. The
carbonation upon curing in air at 17 ◦ C than at 30 ◦ C. In contrast, their resulting pore solution saturated with respect to calcite will buffer the
thermodynamic simulation using the geochemical computer code pH (i.e., pH of the bicarbonate‑carbonate buffer). Upon further drying,
PHRQPITZ showed that the carbonation rate should be higher at the precipitation of calcite will take place under close to equilibrium con­
higher T. The authors claimed that the slower carbonation at the higher ditions (i.e., low supersaturation and [Ca2+]/[CO2− 3 ] ≈ 1) favoring the
T was due to rapid crystallization of numerous small calcite crystals formation of a limited number of calcite seeds (i.e., low nucleation
blocking surface pores, which led to a reduction in the diffusion rate of density) that will grow as large rhombohedral crystals with a compact
CO2 and did not occur at the lower T. It follows that the effect of T on interlocked structure. This morphology evolution will progress from the
carbonation rates is not as simple or predictable as previously thought, surface to the interior, as the surface is directly exposed to CO2, and will
and further research is necessary to better understand the effect of this be faster and more thorough at high CO2 concentrations. The latter helps
variable on the carbonation of lime plasters. to explain why forced carbonation results in plasters and mortars with
higher strengths. The above-described calcite morphology evolution
6.6.2. Humidity also helps to explain why after full carbonation in air, lime mortars show
As indicated above, water is critical for the progress of the carbon­ a steady increase in compressive strength over time [92,211].
ation reaction and will be very slow at very low RH. The rate of Similar effects are expected in the case of dolomitic limes. A study by
carbonation will also be minimum at very high RH, when capillary De Silva et al. [107] shows that carbonation at a high CO2 pressure (20
condensation occurs inside the pore system of the mortar [109]. In the MPa) of compacts prepared with magnesium and calcium hydroxide and
latter case, the excess water will hamper diffusion of CO2 to the reaction their mixtures yielded maximum values of compressive strength for the
sites. As a result, an optimal intermediate RH value must exist which Mg-rich compacts (>60 MPa). Such carbonated compacts showed large
maximizes the carbonation rate. According to Van Balen and Van and interlocked nesquehonite and Mg–Ca carbonate phases.
Gemert [109] this optimal RH ranges from 40 to 80 % for lime mortars. Conversely, no reaction or strength gain was observed in Mg-rich com­
In the case of cement, and in good agreement with the previous RH pacts when carbonated for 28 days in air at 1 atm. Only the Ca-rich
values, maximum carbonation rates are observed at RH ranging from 50 compacts showed an increase in compressive strength under the latter
to 70 % [17]. carbonation conditions. These results underline that magnesium hy­
droxide is poorly reactive at standard atmospheric curing conditions and
6.6.3. CO2 concentration emphasize that the crystal morphology and texture evolution of the
CO2 concentration (i.e., pCO2) appears to be a ruling factor deter­ carbonated paste at high CO2 concentrations is critical for strength gain
mining carbonation rates both in aerial lime mortars [77] as well as in in both magnesian/dolomitic and calcitic limes. These results are also
cement [17]. relevant for the understanding of the carbonation curing of MgO ce­
Carbonation rates of lime mortars and plaster increase with pCO2. ments [20]. However, it should be noted that nesquehonite can evolve
Nevertheless, very high CO2 concentrations (i.e., 100 % [CO2]) can into dypingite or hydromagnesite over time (inducing changes in crystal
result in extremely fast heat release by the highly exothermic carbon­ morphology and volume), which might jeopardize the structural integ­
ation reaction. As a result, premature drying might take place, rity of carbonated Mg-rich lime binders [212].
hampering further carbonation [61]. Numerous studies show, however,
that a less extreme increase in CO2 concentration (5–20 % [CO2]) not 7. Kinetics of the carbonation of lime mortars and plasters
only accelerates carbonation but also improves the mechanical proper­
ties of the carbonated lime mortars [92,196]. Indeed, several authors 7.1. Carbonation models
observed a significant increase (up to one order of magnitude) in the
compressive strength [92,208], as compared with typical values ranging Several models have been proposed for the analysis and prediction of
from ~1 up to ~3 MPa for pastes and mortars carbonated in air (i.e., the advancement of the carbonation reaction of hardened cement [43]
~0.04 % [CO2]) [46,108,209,210]. These results are striking, as one and lime mortars and plasters [109,213]. These models can be grouped
would expect similar strengths after similar degrees of carbonation for into the following main categories [16]: (i) empirical models, where the
the same pastes/mortars mixes. Apparently, accelerated (or forced) relationship between the carbonation depth and its influencing factors is
carbonation curing using CO2-rich atmospheres does not only speed up derived from experiments; (ii) statistical models, where the dependent
the carbonation process, but also affects the structure of the resulting and independent variables are related by mathematical functions, such
CaCO3 binder and, in turn, the strength of the material. This was as multiple linear regression; (iii) numerical models, where several
elegantly demonstrated by De Silva et al. [67] who observed that physico-chemical equations, including reaction rates, mass conserva­
maximum compressive strength of compacted hydrated lime pastes tion, dissolution and diffusion of CO2 in pore solutions, and energy
subjected to forced carbonation (2 MPa CO2 pressure) was associated conservation are computer solved; (iv) machine learning (ML)-based
with the formation of large, interconnected rhombohedral calcite crys­ models, which have been recently applied to solve complex non-linear
tals. In contrast, carbonation of the same pastes at lower CO2 pressure relationships among different variables and parameters involved in
led to the formation of smaller calcite crystals, not showing an inter­ the carbonation process.
connected structure. Considering that the samples carbonated at higher According to the first type of models (empirical), the progress of the
CO2 pressure showed lower portlandite to calcite conversion than those carbonation front in 1D, that is, along the x direction normal to the
carbonated for the same period of time at lower CO2 pressure, it can be plaster/mortar surface, can be modeled by the second Fick's law of
concluded that the morphology of the calcite crystals and the structure diffusion [47,109]:
of the carbonate matrix formed upon carbonation had a more significant √̅
effect on strength than the degree of carbonation. Cizer et al. [69] pre­ x=k t (27)
sented microscopic evidence demonstrating that calcite crystals formed
where t is the carbonation reaction time and k is a rate constant related
after ACC in mortars, evolved from poorly interlocked scalenohedral to
to the physical-chemical characteristics of the system. Based on this
highly interlocked rhombohedral calcite crystals via a dissolution-
diffusion equation, Van Balen and Van Gemert [109] developed a
precipitation mechanisms at a faster rate in 100 % CO2 atmospheres
mathematical model for the evaluation of the uptake of CO2 by a mortar
as compared to ambient pCO2. A plausible explanation for this trans­
undergoing carbonation while drying progresses. They observed a very
formation, which progressed over time from the surface to the interior of
fast CO2 uptake during the early stages of carbonation, followed by a
lime mortar, is the following: upon consumption of available Ca(OH)2,

19
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

monotonic decrease in the rate of CO2 uptake. The initial fast uptake is (dissolved) CO2 diffusion through this product layer the rate controlling
consistent with rapid CO2 hydration and subsequent CaCO3 precipita­ step. The obtained ion diffusivity across the product carbonate layer was
tion in the highly alkaline capillary solution (saturated with respect to 10− 13 m2 s− 1, in good agreement with the values of 3.1 × 10− 12 to 8.6 ×
Ca(OH)2) present in the porous lime mortar. The subsequent monotonic 10− 12 m2 s− 1 reported by Galan et al. [203] for accelerated carbonation
reduction in the carbonation rate is explained by the combined effect of of portlandite crystals in 100 % CO2 atmosphere. Importantly, the for­
pore volume reduction following conversion of Ca(OH)2 into CaCO3, mation of a gap between reactant and product as the reaction progresses,
with a higher molar volume than the former phase, as well as the pro­ might act as a source of fractures that could detach the product layer
duction of H2O, which hampers the diffusion of CO2 to the reaction from the reactant surface, as experimentally observed by Ruiz-Agudo
front. The reduction in carbonation rate was also assumed to be related et al. [71]. This would enable further unrestricted progress of the re­
to the passivation of the portlandite crystals following surface precipi­ action. In any case, while the results show a diffusion-controlled
tation of CaCO3 (see, however, Ruiz Agudo et al. [71]). This model as­ mechanism, full passivation is not observed. However, the carbonation
sumes that there is a sharp reaction interface between the uncarbonated rate is reduced as the thickness of the carbonate layer increases.
and carbonated areas. However, it has been experimentally shown that Importantly, the model shows that an increase in the concentration of
the carbonation front is not a sharp interface, but a diffuse layer with CO2 does increase the carbonation rate, but not so dramatically as could
areas of low and high carbonation degree [12,85,204]. The fact that the be expected, and this increase only takes place up to a concentration of
wake of the carbonation front is not homogeneous, or even continuous, about 30 % CO2. Yet, higher concentrations of CO2 modify the grow of
is demonstrated by the observation of the formation Liesegang rings in the carbonate layer: at low CO2 it grows towards the CO2 source,
carbonating lime mortars [73]. Such periodic structures develop in whereas at high CO2 concentrations it also grows in the opposite di­
porous systems undergoing diffusion-reaction precipitation [214]. In the rection (i.e., towards the reactant portlandite). This would modify the
case of lime mortars, especially those prepared with aged slaked lime, it morphology of the carbonate structure, contributing to explaining why
is observed that there are alternating bands with high concentration of lime pastes carbonated at high CO2 concentrations show higher
either Ca(OH)2 or CaCO3, which are clearly distinguished upon appli­ strength, even if the degree of carbonation is not very high. Another
cation of phenolphthalein (a pH indicator, colorless al pH close to important outcome of this computational study is the following: (i) as
neutrality and with a deep magenta color at pH > 10). Fig. 14a shows an the product layer grows, a reduction in the size of the pores existing
example of the development of such a Liesegang pattern in lime mortars among the calcite particles in this carbonate layer will take place; and
prepared with aged slaked lime putty and quartz sand aggregate (1:3 (ii) as a result, the solubility of this product phase in the pore solution
binder aggregate ratio). The prisms (4 × 4 × 16 cm) were split in half at will be increased, hampering further calcite growth. This so-called
different elapsed times of carbonation (in air at room T) and the fracture “pore-size dependent solubility” effect [213] would prevent closure of
surfaces sprayed with a phenolphthalein solution [36,73]. The presence the pores between calcite grains of the product layer, enabling a
of Liesegang patterns in carbonating lime mortars, evidenced by dif­ continuous progress of carbonation at reduced rates, even if no cracks
ferential weathering in historic lime mortar structures [215], is a direct develop at the reactant-product interface.
proof of the complexity of this process linked to the formation of In addition to the previous models specific for lime mortars, several
amorphous precursors (ACC) before the formation of stable calcite [73]. models have been proposed to evaluate the kinetics of carbonation of
A more detailed numerical model for the carbonation of portlandite cement which are reviewed in Qiu [43]. One of them, based on Fick's
at the nano- and micro(pore)scale was recently developed by Varzina second law, is the so-called unreacted core-model. It states that the two
et al. [213]. The authors used a lattice Boltzmann framework to model possible rate-controlling steps in the carbonation of portlandite (in
the coupled portlandite dissolution and calcite precipitation during cement) are: (i) diffusion of CO2 to the reaction interface, and (ii) the
carbonation in air in the presence of an aqueous phase, and modeled kinetics of the actual carbonation reaction. Castellote and Andrade
reactant diffusion using Fick's second law. Their results show that [216] demonstrated that experimental carbonation data fit the first case
initially reaction rates are controlled by the dissolution of portlandite, but not the second. The authors conclude that the carbonation of cement
but later, the formation of a porous carbonate layer makes Ca2+ and is controlled by the diffusion of CO2 along a product shell formed around

Fig. 14. Carbonation evolution of lime mortars. a)

Development of Liesegang rings in lime mortars pre­
pared with aged (7 years) slaked lime putty after
curing at 60 % RH in air for different periods of time.
Standard 4 × 4 × 16 cm mortar prisms (1:3 binder:
aggregate ratio) were split and sprayed with phenol­
phthalein; b) time-evolution of the carbonation depth
for the mortars samples depicted in (a); c) the data
presented in (b) now represented considering the
square root of time. Note the excellent linear fitting to
Eq. (27) with a k value of 1.043 mm⋅day− 0.5.
Modified from [73] with permission by Royal Society
Publishing.

20
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

the unreacted core. Based on Fick's second law, their diffusion- Table 2
controlled model allows to calculate the fractional conversion of reac­ Carbonation rate (k) values of lime mortars calculated from published values of
tant s, Xs, at a given time t, for the case of a cylindrical sample of radius carbonation depth over time determined using phenolphthalein.
R, by using the following equation, Mortar type, aggregate, binder/ Curing conditionsa k value Ref.
1/2
( r )2 aggregate (B/A) ratio, water/ (mm⋅day− )
Xs = 1 − (28) binder (W/B) ratio
R
Lime mortar n.a.b 1.00 [221]
Lime-cement mortar n.a. 0.25 [221]
where r is the radius of the unreacted core. The model also enables to
Hemp-lime mortar 60 % RH 5.24 [221]
calculate the time τ for achieving full carbonation knowing the frac­ Lime mortar, silicate sand, 1:3 v/ 25 ◦ C, 60 % RH 0.95–1.37 [222]
tional conversion at a time t, given by, v B/A, 0.5–0.54 wt/wt W/B
Lime mortar (4 months aged 90 % (7 days)/60 % 1.16 [68]
t
= Xs + (1 − Xs )ln(1 − Xs ) (29) putty), oolitic stone aggregate, RH
τ 1:3 v/v B/A
Lime mortar (4 months aged 90 % (7 days)/60 % 1.56 [68]
Knowing τ, it is also possible to calculate the effective diffusion co­
putty), silicate sand, 1:3 v/v B/ RH
efficient of CO2 across the product layer, D, by, A
Lime mortar (4 months aged 90 % (7 days)/60 % 1.23 [68]
ρ s R2
τ= (30) putty), crushed bioclastic stone RH
4bDCCO2 aggregate, 1:3 v/v B/A
Lime mortar, silicate sand, 1:2 v/ 60 % RH, 0.05 % 1.53–2.00 [92]
where ρs is the molar fraction of reactant s, in the solid, b is the stoi­ v B/A, water content 16–18 wt CO2
chiometric coefficient for the carbonation reaction (in the case of por­ %
Lime mortar, silicate sand, 1:2 v/ 60 % RH, 5 % CO2 6.32–6.66 [92]
tlandite b is 1), and CCO2 is the concentration of CO2 in the gas phase.
v B/A, water content 16–18 wt
Importantly, once D is determined for a particular system, the model can %
be applied to other conditions including variations in CO2 concentration Lime mortar, silicate sand, 1:2 v/ 60 % RH 1.58–1.96 [217]
and exposure time. However, this model does not take into account v B/A, added air-entraining
other relevant parameters such as T or RH, which are considered in agents
Lime mortar, silicate sand, 1:3 v/ 60 % RH 1.25–1.30 [46]
other, more complex models [43,167]. Nonetheless, Castellote and
v B/A, 1.3 v/v W/B
Andrade [216] model, or the more complex ones listed in Qiu [43], have Lime mortar, fine aggregate, 1:3 60 % RH 1.13 [219]
not been used yet to evaluate the kinetics of carbonation of aerial lime wt/wt B/A, 0.8 wt/wt W/B
mortars. Lime mortar (putty), silicate 50 % or 90 % RH 2.29 [220]
sand. 1:3 v/v B/A
Lime mortar (aerial+hydraulic), 50 % or 90 % RH 1.82 [220]
silicate sand, 1:3 v/v B/A
7.2. Experimental carbonation rates of lime mortars Lime mortar (putty), silicate 50 % or 90 % RH 2.10 [220]
sand/crushed rock aggregate,
There are several studies that experimentally determined the prog­ 1:3 v/v B/A
ress of the carbonation front during setting and hardening of aerial lime Lime mortar (aerial), silicate 50 % or 90 % RH 1.40 [220]
sand/crushed rock aggregate,
mortars. However, only a few report the actual value of k, that is, the
1:3 v/v B/A
overall carbonation rate according to Eq. (27). Yet many include Lime mortar, silicate sand 60 % RH, 30 ◦ C 0.16–0.23 [223]
numeric values that enable its calculation. Once k is known, it is possible (grinding 0′-15′), 1:3 wt/wt B/
to evaluate the time evolution of the carbonation front using Eq. (27). A, 0.75 wt/wt W/B
Table 2 presents k values from several studies on lime mortars and Lime mortar, silicate sand 60 % RH, 30 ◦ C 0.40–0.71 [223]
(grinding 0′-15′), 1:3 wt/wt B/
plasters carbonation in air at standard P-T conditions (a few k values A, 0.75 wt/wt W/B, 4 wt%
corresponding to forced carbonation in CO2-rich atmosphere are organics
included for comparison) [46,47,66,68,92,211,217–223], evidencing Lime mortar, crushed limestone 97 % (7 days, 0.20–0.47 [211]
that there is a significant scattering in carbonation rate values. Note that aggr., 1:3 v/v B/A, 0.9 v/v W/ 20 ◦ C)/51 % (up to
B 448 days, 17 ◦ C) RH
except for Ardigoyen and Alvarez [66], k values presented in Table 2 are
Lime paste, no aggregate, 60 % RH 0.66–0.91 [66]
calculated from the analysis of the carbonation depth in split samples 0.8–1.3 W/B
sprayed with phenolphthalein (in most cited papers only the depth of the a
Unless stated otherwise, all values correspond to mortars cured in air
carbonation front at different time intervals is reported). In the case of
(~0.03–0.04 % CO2 concentration) at room T (20 ◦ C).
the lime mortars showing Liesegang pattern development depicted in b
Not available.
Fig. 14a, a value of k of 1.04 mm⋅day− 1/2 is calculated (Fig. 14b–c).
Overall, carbonation rate values presented in Table 2 show that full
It is important to indicate that most experimental results on the
carbonation of an air lime mortar or render with a thickness of about 5
carbonation of aerial lime mortars show that at the selected testing times
cm can take up to two and a half centuries in the worst-case scenario (k
(typically <90 days) incomplete carbonation was observed. For
= 0.16 mm⋅day− 1/2), and for the average k value of 1.23 (±0.89, N = 28)
instance, Oliveira et al. [46] report a limit for the carbonation level of
mm⋅day− 1/2 from the previous studies, the full carbonation of such a
70 % irrespective of curing time up to 90 days. This is commonly
mortar/plaster element would take ~4.5 years. This must be considered
associated with a passivating effect by the carbonate product layer
when designing lime-based masonry units or when preparing laboratory
formed on portlandite crystals. As indicated above, this is very unlikely
samples for testing, as curing times, for instance, need to be adjusted to
[71]. It is more likely that the limited carbonation observed is due to low
the kinetics of lime carbonation. The latter is critical because in some
RH conditions and excessive drying, in addition to the limited span of
cases test specimen properties, such as compressive and tensile strength,
the carbonation curing.
are determined at 28 days, which might be appropriate for cement, but
Indeed, from the data reported by Ruiz-Agudo et al. [71] for the
not for slow-setting (carbonating) aerial lime mortars.
carbonation of mm-sized single crystals (room T, in air and 93 % RH) a k
The large scattering of k values from different studies is likely due to
value of 0.0034 mm⋅day− 1/2 is obtained, whereas from the model by
variations in mortar preparation (i.e., different lime or aggregate type
Varzina et al. [213], simulating carbonation of portlandite single crys­
and water:binder:aggregate ratios) and curing conditions, all of these
tals at room T in air, a k value of 0.012 mm⋅day− 1/2 is obtained (i.e., a
factors having an influence on the carbonation rate.

21
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

carbonated layer of 15 μm is developed within 20 or 1.5 days, respec­ where R is C2H5 for ethyl carbamate and NH+ 4 for ammonium carbamate,
tively). These k values are about two to three orders of magnitude and R′ is H.
smaller than the average k value determined for lime mortars and As a result, in depth carbonation is achieved speeding up the kinetics
plasters carbonated under similar conditions (see above). It is therefore of lime mortars setting and hardening. The authors observed that
evident that the rate controlling step for carbonation in a lime mortar or carbonation followed a 3D-diffusion model (i.e., D3 model in Table 1).
plaster is not the diffusion of ions across the carbonate product layer The good fitting of experimental carbonation data of lime grouts dosed
formed on individual portlandite crystals. Conversely, it is very likely with carbamates shows that 3D diffusion of CO2 to the reaction sites (i.
that the small (micrometer) size/thickness of individual portlandite e., portlandite-pore solution interfaces) was the rate-determining step.
crystals in slaked lime leads to their full conversion (within a few days) Medici and Rinaldi [225] tested the addition of poly-amino-phenolic
before a sufficiently thick layer of product carbonate would have (PAP) compounds, known for their capacity to capture CO2, to accel­
developed on such crystals to act as an effective diffusional barrier. This erate the carbonation of lime mortars. The authors observed a dose-
shows that the most likely overall rate-limiting factor for the carbon­ dependent (0.1–0.3 %) acceleration of carbonation in saturated Ca
ation of lime mortars and plasters is the actual diffusion of CO2 along the (OH)2 solutions. In parallel they also observed a significant increase in
(open and/or partially saturated) pore network in the carbonated the compressive strength of the carbonated mortars with PAP as
profile. compared with the additive-free control. However, dosing the additive
A detailed understanding of the mechanisms and kinetics of lime at a high concentration (1 %) led to a lower strength than the control
mortars carbonation is not only relevant for predicting the time evolu­ mortar. Despite the promising results obtained by using these additives
tion of these materials' physical-mechanical properties, but also for (carbamates and PAP) no follow-up studies or on-site applications have
predicting their capacity to capture atmospheric CO2 and enable a been reported so far to our knowledge.
proper Life Cycle Assessment (LCA) [224]. Note that it has been reported Ergent et al. [226] proposed the use of diethyl carbonate (DEC,
that if a lime-based structure achieves full carbonation, it will contribute commercially known as DiloCarB®) as a carbonation accelerator. The
to an average capture of ~33 % of the amount of process CO2 emitted authors stated that a faster carbonation was achieved in lime mortars
during the production of this binder for permanent and safe mineral dosed with this additive. However, their TG analysis of the consumption
storage as carbonate phases (EULA, https://www.eula.eu/down of portlandite over time does not show a clear effect on the carbonation
load/capturing-co2-with-lime/). It is therefore necessary to determine rate of mortars with added DEC as compared with the control without
the carbonation rate of a particular lime-based material to evaluate if it additive.
will achieve its maximum CO2 capture capacity over its service life. The addition of photoactive anatase (TiO2) nanoparticles was pro­
posed by Karatasios et al. [227] to speed up the carbonation of lime
8. Accelerating carbonation plasters. The authors stated that light irradiation would enable the
photooxidation of organic deposits on the mortars surface by the TiO2
One of the main handicaps for the current use of aerial lime-based nanoparticles in the mortar mix, resulting in the release of CO2 and an
mortars and plasters, both in built heritage conservation and modern acceleration of the carbonation of the lime-binder. However, it is rather
construction, is the fact that their carbonation is very slow. This makes it unclear if this approach could lead to an acceleration of the carbonation
difficult to meet building requirements where early strength is needed. in-depth (i.e., where it is most needed), as the photo-activity of TiO2
To solve this problem, several approaches have been explored to would be limited to a very thin surface layer (i.e., the depth that light
accelerate carbonation and by-pass one or more of the main rate- could penetrate in the mortar).
controlling factors above indicated, namely: (i) T, (ii) moisture con­ Another way of speeding up carbonation by increasing CO2 con­
tent/RH, and (iii) CO2 concentration, diffusion through the pore system, centration within a mortar mix involves the use of organic additives,
and dissolution in pore water. The two first factors can be more or less fermented plant extracts in particular [223,228–230]. Upon fermenta­
easily controlled. In general, room T is considered appropriate for tion of the plant extract, CO2 is produced and released within the pore
optimal carbonation and optimal RH values range from 40 to 80 % system of the fresh mortar mix. Ultimately, faster and more in-depth
[109]. The optimization of T and RH has an impact on carbonation rates, carbonation is observed, resulting in a higher compressive strength at
but in practical terms the effects are not so dramatic. Much more early age (28 days). Other natural organic additives have been proposed
important effects are observed by altering the concentration of CO2 and to speed up lime mortar carbonation. Following the ancient Meso­
in what manner this gas dissolves in the pore water and reaches the american tradition of adding cactus aqueous extracts to lime mortars
reaction interfaces. and plasters [118], León-Martínez et al. [231] observed a slight increase
As indicated in the previous sections, increasing pCO2 has a dramatic in the carbonation rate of lime mortars prepared with Oppunctia ficus-
effect on carbonation rates [61,67]. However, it is not always feasible to indica and Acanthocereus tetragonus mucilage dosed in a concentrations
increase the CO2 concentration on site where a plaster or mortars is of 0.25 wt%, but mortars prepared with a higher dose of additive (0.5 wt
applied. And it is even more difficult to increase the concentration of %) showed little acceleration or even a decrease in the rate of carbon­
CO2 where it is more needed, that is, in-depth, within the pore system of ation. Addition of up to 0.4 wt% chitosan to pure lime mortars slightly
a lime mortars or plaster. There are however several studies that present accelerated carbonation at 100 % CO2 curing conditions, leading to a
feasible strategies to increase pCO2 within the pore system of a mortar minor increase in compressive strength [232]. It is unclear though
without the need for an external (pressurized) CO2 supply. Compounds whether the latter acceleration effect will also be observed upon
such as NH4HCO3 or (NH4)2CO3 lead to the rapid release of CO2 and carbonation under ambient conditions.
NH3, fostering carbonation if dosed in the mortar mix. However, they Arizzi and Cultrone [108] indicate that carbonate aggregate can
show significant problems for practical application as they induce pre­ accelerate carbonation because its angular-shape and roughness favor
mature CaCO3 precipitation upon mixing, increasing the fresh lime paste the formation of communicated pores as opposed to smooth/rounded
viscosity, and drastically reducing workability [180]. As an alternative silicate aggregate (quartz sand). However, although porosity and pore
to this approach, Baglioni et al. [180] proposed the addition of ethyl or size have an impact on carbonation rates, it is likely that the calcite
ammonium carbamates to a lime-based conservation grout. During what aggregate will act as a template for CaCO3 heterogeneous nucleation,
the authors call “autogenous” setting, the carbamate slowly decomposes further speeding up carbonation. Moreover, the newly formed carbonate
at room T releasing CO2 via the reaction, cement likely grows in crystallographic continuity (i.e., self-epitaxy),
R − O − CO − NR′2 + H2 O = ROH + CO2 + HNR′2 (31) and no discontinuity between the aggregate and the carbonate cement
is to be expected. As a result, higher strength should be achieved as

22
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

compared with mortars prepared using siliceous aggregate (where no and pore size of the mortar either using a nanosized aggregate, such as
crystallographic continuity between the aggregate and the carbonate ball-milled quarry waste nanoparticles [236], or using additives, such as
cement exists), which is in accordance with findings by Arizzi and surfactants with the capacity of acting as air entraining agents (AEA)
Cultrone [108] and others [233,234]: a higher strength is achieved as [217]. The increase in porosity and pore size enables a faster and easier
compared with mortars prepared using siliceous aggregate (where no access of diffusing CO2 to reaction sites. However, Silva et al. [217]
crystallographic continuity between the aggregate and the carbonate reported that while an acceleration of the carbonation process was
cement exists). Hay et al. [208] also reported increased carbonation observed with the addition of AEAs, in some cases it was observed that
rates and strength of lime mortars prepared adding ground limestone. As the resistance to weathering (freeze-thaw) was reduced as compared
in the previous case, it is assumed that such effects are due to the tem­ with AEA-free reference aerial lime mortars. Moreover, it should be
plate action of the ground calcite in the limestone. However, not all noted that organic additives such as proteins or polysaccharides in some
carbonate aggregates speed up carbonation: Martinez-Garcia et al. [235] cases are reported to speed up carbonation [223], but in other cases tend
report a reduction in the carbonation rate of lime mortars prepared to slow down the carbonation process, yet they typically improve the
using ground mussel shell aggregate, as compared with mortars pre­ mechanical properties of the carbonated lime mortars or plaster
pared using standard limestone aggregate. Nonetheless, once carbon­ [97,160,237].
ated, the former mortars reached higher compressive strength than the An interesting possibility to enhance in-depth carbonation of lime
latter, which is explained considering that the organics in the shell mortars and plasters could involve the use of silica aerogel particles
biomineral played a role in the formation of the carbonate binder. loaded with CO2, or silica aerogels functionalized with amine com­
Other possibility to speed up carbonation is to increase the porosity pounds for a more effective CO2 capture during CO2 loading, as reported

Fig. 15. Effect of carbonic anhydrase (CA) enzyme on the carbonation kinetics and physical mechanical properties of lime-based binders. Time-resolved XRD
patterns of Ca(OH)2 saturated solution droplets without (a) and with CA (b) exposed to atmospheric CO2. Note the increase in the background intensity due to the
formation of ACC and its decrease upon calcite precipitation (best seen in individual XRD patterns, in inset). The 104 Bragg peak of calcite (Cc) is indicated. (c) Time
evolution of the fractional amount (α) of ACC (solid lines) and calcite, Cc (dashed lines) in the control (CA-free) and runs with CA dosed in different concentrations
(μM); d) time evolution of carbonation for lime pastes with CA and without CA (control, LP); Time evolution of the compressive (e) and flexural (f) strength (σc and σf,
respectively) of lime pasted with CA and without CA (control, LP).
Figure parts a–c reprinted from [151] with permission by Elsevier; parts d–f reprinted from [70] with permission by Taylor & Francis.

23
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

by Jassam et al. [238] for the case of cement carbonation curing. Upon techniques have unambiguously demonstrated that it is a dissolution-
mixing with lime during mortar preparation and subsequent setting, CO2 precipitation reaction, where the presence of adsorbed water is critical.
would be released from the silica aerogel within the mortar matrix, Regarding the kinetics of carbonation, a clear distinction has to be
contributing to enhanced lime carbonation. made between the carbonation kinetics of Ca(OH)2 crystals, powders
For the case of nanosized portlandite particles (nanolimes), Zhu et al. and thin films, and the carbonation kinetics of lime-based plasters and
[239] reported that the addition of graphene quantum dots (GQD) mortars. The kinetics of the former can be modeled using classical “solid-
during homogeneous nanolime synthesis, enabled a faster carbonation, state” kinetic models, with deceleratory kinetic models (e.g., F1 and F2)
not only linked to the high surface area of the nanoparticles, but also to yielding excellent fittings. Conversely, in the case of plasters and mor­
the effect of the adsorbed/incorporated GQD. The latter speeded up the tars, diffusion-limited kinetics rule, and modeling of the carbonation
absorption and hydration of CO2 in the presence of humidity, facilitating evolution using Fick's second law yields very good fittings, demon­
the conversion of Ca(OH)2 into CaCO3. strating that the carbonation front advances with the square root of time.
Recently, a biomimetic approach has been proposed to accelerate Yet, numerical models using Fick's second law enable to obtain a more
carbonation of lime mortars and plasters. It involves the use of natural accurate picture of the (nano)micro-scale evolution of the carbonation
enzymes such as carbonic anhydrase (CA) that catalyzes the hydration of reaction. Importantly, in the case of lime plasters and mortars, rate
CO2 [70,151], as well as their biomimetics, in this case molecular constants (k values) display a large scattering. This is likely due to the
organic framework (MOF), compounds with active metal sites for such a fact that different researchers used different lime-based materials, with
catalysis [50]. In the case of CA, which is key for the biomineralization variation in the type of binder and aggregate, binder/aggregate ratios,
of CaCO3 by different organisms (e.g., sea-urchins and sea shells), it is water/binder ratio, and curing conditions. Nonetheless, reported k
observed that this enzyme accelerates the formation of ACC and its values show that the carbonation of lime mortars and plasters is very
transformation into calcite following carbonation in air of a saturated Ca slow, with carbonation of cm-thick sections taking several years.
(OH)2 solution (Fig. 15a–c) [151]. Similarly, aerial lime pastes dosed It is also shown that the formation of a carbonate product layer
with 1.5 μM CA display a faster carbonation (in air) than reference around portlandite crystals via an ICDP mechanism, does not necessarily
pastes without this enzyme (Fig. 15d) [70]. Remarkably, both the result in passivation. Upon reaching a limiting thickness, strain and
compression and flexural strength of CA-including lime pastes were associated stress due to the molar volume difference between reactant
higher than those of the CA-free reference pastes regardless of their and product phases leads to cracking and detachment of carbonate
carbonation time (Fig. 15e-f). It follows that the enzyme not only ac­ layers, exposing fresh portlandite surfaces and facilitating the
celerates carbonation, but also affected the microstructure of the calcite advancement of the carbonation front towards the core of the hydroxide
binder, leading to an improvement in the mortars' mechanical proper­ crystals. It is also observed that the carbonate layer is porous, and
ties. Ultimately, this biomimetic approach by-passes one of the most modeling predicts that pore-closure is unlikely due to pore size solubility
significant rate-limiting step of the carbonation reaction, which is the effects. As a result, carbonation rates become diffusion controlled by the
conversion of CO2 into bicarbonate ions (Eq. (15)). It also leads to slow diffusion of reactant/product ions/molecules across this diffusive
binders with improved physical-mechanical properties and nano/ carbonate barrier.
microstructural features that somehow resemble those of calcite bio­ It is also shown that dolomitic limes undergo NCC during carbon­
minerals (e.g., shells), whose hybrid inorganic-organic nature and hi­ ation, with the presence of both ACC after the carbonation of the Ca
erarchical (nano-micro-meso)structure, impart them a higher hardness (OH)2 fraction and AMC following the carbonation of Mg(OH)2. How­
and toughness than their individual components [166]. ever, while the Ca(OH)2 fraction carbonates in a similar way as in
calcitic limes, the Mg(OH)2 fraction is basically unreactive, and un­
9. Conclusions and outlook dergoes negligible or very little carbonation over time under standard P-
T conditions.
Here a general overview of what is known as the carbonation reac­ Future research on the carbonation of lime-based binders should
tion is presented, showing that it is highly relevant not just for the better address several key issues that are still not completely understood. One
understanding of the setting and hardening of lime-based mortars and of them refers to fully disclosing the differences in the mechanism and
plasters, but also for a range of technological and natural processes, kinetics of carbonation of Ca(OH)2 and Mg(OH)2. In particular, it would
spanning from PCC production and the degradation of reinforced con­ be important to fully understand why brucite is so poorly reactive as
crete to the long-term control of the climate on Earth. compared with portlandite. Attention needs also to be paid to under­
It is shown that the carbonation reaction is complex from a (geo) stand the role of magnesium ions on the kinetics of both calcium and
chemical point of view, with several interrelated reactions taking place magnesium carbonate phases formation. Another aspect that should be
nearly simultaneously that control both the mechanism and the kinetics further studied is the (nano-micro)structure-property relationship of
of this fundamental reaction. Importantly, it is observed that from a carbonated calcitic and dolomitic limes, and how this is affected by
chemical (reaction) point of view the rate controlling step is the hy­ additives, particularly organic (bio)macromolecules. This will provide
dration of CO2 to form bicarbonate and carbonate ions. However, in key information about the secret of ancient masons from different civ­
practice, for systems that are not diffusion-limited, the rate controlling ilizations which produced lime mortars and plasters of outstanding
step of the carbonation reaction is the dissolution of the hydroxide properties and durability using different natural organic additives such
parent phase. This is supported by the fact that calculated activation as sticky rice (ancient China) or bark extracts (ancient Maya) [240]. It
energy values for carbonation of portlandite match those of reaction- will also be relevant to evaluate the effects of such additives (natural or
controlled kinetics for the dissolution of common minerals. synthetic) on the carbonation kinetics of lime mortars and plasters. Ul­
It is underlined that the carbonation reaction progresses via non- timately, we could learn from Nature and strive to produce, by a bio­
classical crystallization routes, involving several precursors including mimetic approach, improved lime-based carbonate binders with
PNC and amorphous (liquid-like and solid) phases. Moreover, it is now structure and properties matching those of carbonate biominerals. This
clear that carbonate crystal growth does not necessarily follow a clas­ could be achieved by using (natural or synthetic) organic additives
sical route but involves aggregation and attachment of solid nano­ similar to those involved in CaCO3 biomineralization.
particles (e.g., ACC). Finally, a better understanding of the mechanism and kinetics of
The analysis of the literature reveals that the idea regarding the carbonation of aerial calcitic and dolomitic limes will offer the possi­
mechanism of carbonation of lime binders has changed over time. While bility of modifying key (rate controlling) parameters to speed up the
in the past it was assumed that it was a gas-solid reaction (i.e., solid-state carbonation reaction and improve the physical-chemical and mechani­
reaction), more recent results obtained using advanced analytical cal properties of the set and hardened lime mortars and plasters. This

24
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

could be achieved by using a biomimetic approach involving natural processes in Imperial Roman architectural mortar, Proc. Natl. Acad. Sci. 111
(2014) 18484–18489.
and/or synthetic additives, such as carbonic anhydrase enzyme or MOF-
[9] P. Hewlett, M. Liska (Eds.), Lea's Chemistry of Cement and Concrete,
based compounds to accelerate carbonation. Another related area of Butterworth-Heinemann, New York, 2019.
research should focus on disclosing the role of organic additives in [10] J. Elsen, M.D. Jackson, E. Ruiz-Agudo, Historic concrete science: opus
modifying the structure of the final CaCO3 crystals in the set and caementicium to “natural cements”, Elements 18 (2022) 301–307.
[11] W.F. Cole, B. Kroone, Carbonate minerals in hydrated Portland cement, Nature 84
carbonated lime mortars/plasters. It is hypothesized that organics could (1959) BA57.
be occluded in the CaCO3 crystals of carbonated lime mortars, as occurs [12] A. Morandeau, M. Thiery, P. Dangla, Investigation of the carbonation mechanism
in CaCO3 biominerals such as mollusk shells, thereby imparting the of CH and CSH in terms of kinetics, microstructure changes and moisture
properties, Cem. Concr. Res. 56 (2014) 153–170.
carbonated matrix similar physical-mechanical properties as those [13] W. Ashraf, Carbonation of cement-based materials: challenges and opportunities,
observed in calcite biominerals. The latter, being organic-inorganic Constr. Build. Mater. 120 (2016) 558–570.
hybrid materials, show toughness and damage tolerance orders of [14] B. Šavija, M. Luković, Carbonation of cement paste: understanding, challenges,
and opportunities, Constr. Build. Mater. 117 (2016) 285–301.
magnitude higher than their individual components. Learning from [15] S. Steiner, B. Lothenbach, T. Proske, A. Borgschulte, F. Winnefeld, Effect of
nature, we could strive to develop lime-based materials with improved relative humidity on the carbonation rate of portlandite, calcium silicate hydrates
properties for their use as compatible materials for built heritage con­ and ettringite, Cem. Concr. Res. 135 (2020), 106116.
[16] Z. Chen, J. Lin, K. Sagoe-Crentsil, W. Duan, Development of hybrid machine
servation, as well as for their use as sustainable binders in modern learning-based carbonation models with weighting function, Constr. Build. Mater.
construction [240]. 321 (2022) 126359.
[17] Z. Xu, Z. Zhang, J. Huang, K. Yu, G. Zhong, F. Chen, W. Yang, Y. Wang, Effects of
temperature, humidity and CO2 concentration on carbonation of cement-based
CRediT authorship contribution statement materials: a review, Constr. Build. Mater. 346 (2022), 128399.
[18] S. Monkman, Y. Shao, Assessing the carbonation behavior of cementitious
Carlos Rodriguez-Navarro: Conceptualization, Methodology, materials, J. Mater. Civ. Eng. 18 (2006) 768–776.
[19] S. Kashef-Haghighi, S. Ghoshal, CO2 sequestration in concrete through
Investigation, Validation, Writing – original draft, Supervision, Project accelerated carbonation curing in a flow-through reactor, Ind. Eng. Chem. Res. 49
administration, Funding acquisition. Teodora Ilić: Investigation, (2010) 1143–1149.
Writing – review & editing. Encarnación Ruiz-Agudo: Investigation, [20] D. Zhang, Z. Ghouleh, Y. Shao, Review on carbonation curing of cement-based
materials, J. CO2 Util. 21 (2017) 119–131.
Validation, Writing – review & editing, Project administration, Funding [21] Z. Liu, W. Meng, Fundamental understanding of carbonation curing and
acquisition. Kerstin Elert: Conceptualization, Investigation, Methodol­ durability of carbonation-cured cement-based composites: a review, J. CO2 Util.
ogy, Validation, Writing – review & editing. 44 (2021), 101428.
[22] V.G. Papadakis, C.G. Vayenas, M.N. Fardis, Fundamental modeling and
experimental investigation of concrete carbonation, ACI Mater. J. 88 (1991)
363–373.
Declaration of competing interest [23] J. Cowie, F.P. Glasser, The reaction between cement and natural waters
containing dissolved carbon dioxide, Adv. Cem. Res. 4 (1992) 119–134.
[24] A.V. Saetta, B.A. Schrefler, R.V. Vitaliani, The carbonation of concrete and the
The authors declare that they have no known competing financial
mechanism of moisture, heat and carbon dioxide flow through porous materials,
interests or personal relationships that could have appeared to influence Cem. Concr. Res. 23 (1993) 761–772.
the work reported in this paper. [25] L.J. Parrott, Damage caused by carbonation of reinforced concrete, Mater. Struct.
23 (1990) 230–234.
[26] C. Dow, F.P. Glasser, Calcium carbonate efflorescence on Portland cement and
Data availability building materials, Cem. Concr. Res. 33 (2003) 147–154.
[27] A. Goudie, H.A. Viles, Salt Weathering Hazard, Wiley, New York, 1997.
All data in this paper are included in the main text. [28] E. Doehne, C.A. Price, Stone Conservation: An Overview of Current Research,
Getty Conservation Institute, Los Angeles (CA), 2010.
[29] C. Rodriguez-Navarro, E. Doehne, Salt weathering: influence of evaporation rate,
Acknowledgements supersaturation and crystallization pattern, Earth Surf. Process. Landf. 24 (1999)
191–209.
[30] G.W. Scherer, Crystallization in pores, Cem. Concr. Res. 29 (1999) 1347–1358.
We acknowledge funding by the Spanish Government grant [31] G.W. Scherer, Stress from crystallization of salt, Cem. Concr. Res. 34 (2004)
PID2021-125305NB-I00 funded by MCIN/ AEI /10.13039/ 1613–1624.
[32] L.A. Rijniers, H.P. Huinink, L. Pel, K. Kopinga, Experimental evidence of
501100011033 and by ERDF A way of making Europe, Junta de Anda­
crystallization pressure inside porous media, Phys. Rev. Lett. 94 (2005), 075503.
lucía research group RNM-179 and grant P20_00675, University of [33] M. Schiro, E. Ruiz-Agudo, C. Rodriguez-Navarro, Damage mechanisms of porous
Granada, Unidad Científica de Excelencia UCE-PP2016-05, and Euro­ materials due to in-pore salt crystallization, Phys. Rev. Lett. 109 (2012), 265503.
[34] R.J. Flatt, N. Aly Mohamed, F. Caruso, H. Derluyn, J. Desarnaud, B. Lubelli, R.
pean Commission, H2020 Programme, Marie Skłodowska-Curie Action
M. Espinosa Marzal, L. Pel, C. Rodriguez-Navarro, G.W. Scherer, N. Shahidzadeh,
European Training network – Innovative Training Network (ETN-ITN), M. Steiger, Predicting salt damage in practice: a theoretical insight into
SUBlime (grant agreement # 955986). Funding for open access charge: laboratory tests, RILEM Tech. Lett. 2 (2017) 108–118.
Universidad de Granada/CBUA. [35] V. Fassina, R. Arbizzani, A. Naccari, Salt efflorescences on the marble slabs of S.
Maria dei Miracoli church: a survey on their origin and the methodology of their
removal, in: Proc. 8th Int. Congr. Deterioration and Conservation of Stone, Berlin,
References 1996, pp. 523–534.
[36] O. Cazalla, C. Rodriguez-Navarro, E. Sebastian, G. Cultrone, M.J. De la Torre,
Aging of lime putty: effects on traditional lime mortar carbonation, J. Am. Ceram.
[1] Ö. Cizer, K. Van Balen, J. Elsen, D. Van Gemert, Real-time investigation of
Soc. 83 (2000) 1070–1076.
reaction rate and mineral phase modifications of lime carbonation, Constr. Build.
[37] E.F. Hansen, C. Rodríguez-Navarro, K. Van Balen, Lime putties and mortars, Stud.
Mater. 35 (2012) 741–751.
Conserv. 53 (2008) 9–23.
[2] R.S. Boynton, Chemistry and Technology of Lime and Limestone, Wiley, New
[38] W. Ashraf, J. Olek, S. Sahu, Phase evolution and strength development during
York, 1980.
carbonation of low-lime calcium silicate cement (CSC), Constr. Build. Mater. 210
[3] K. Elert, C. Rodriguez-Navarro, E.S. Pardo, E. Hansen, O. Cazalla, Lime mortars
(2019) 473–482.
for the conservation of historic buildings, Stud. Conserv. 47 (2002) 62–75.
[39] X. Wang, M.Z. Guo, T.C. Ling, Review on CO2 curing of non-hydraulic calcium
[4] J. Lanas, J.P. Bernal, M.A. Bello, J.A. Galindo, Mechanical properties of natural
silicates cements: mechanism, carbonation and performance, Cem. Concr.
hydraulic lime-based mortars, Cem. Concr. Res. 34 (2004) 2191–2201.
Compos. 133 (2022), 104641.
[5] A. El-Turki, R.J. Ball, G.C. Allen, The influence of relative humidity on structural
[40] P. Juilland, E. Gallucci, R. Flatt, K. Scrivener, Dissolution theory applied to the
and chemical changes during carbonation of hydraulic lime, Cem. Concr. Res. 8
induction period in alite hydration, Cem. Concr. Res. 40 (2010) 831–844.
(2007) 1233–1240.
[41] W. Ashraf, J. Olek, Carbonation behavior of hydraulic and non-hydraulic calcium
[6] H.F. Taylor, Cement Chemistry, Thomas Telford, London, 1997.
silicates: potential of utilizing low-lime calcium silicates in cement-based
[7] Ö. Cizer, K. Van Balen, D. Van Gemert, Competition between hydration and
materials, J. Mater. Sci. 51 (2016) 6173–6191.
carbonation in hydraulic lime and lime-pozzolana mortars, Adv. Mater. Res. 133-
[42] K. Svensson, A. Neumann, F. Feitosa Menezes, C. Lempp, H. Pöllmann,
134 (2010) 241–246.
Carbonation of natural wollastonite at non-ambient conditions relevant for
[8] M.D. Jackson, E.N. Landis, P.F. Brune, M. Vitti, H. Chen, Q. Li, M. Kunz, H.
R. Wenk, P.J.M. Monteiro, A.R. Ingraffea, Mechanical resilience and cementitious

25
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

CCS—the possible use as cementitious material in wellbores, Appl. Sci. 9 (2019) [75] C. Rodriguez-Navarro, K. Elert, R. Ševčík, Amorphous and crystalline calcium
1259. carbonate phases during carbonation of nanolimes: implications in heritage
[43] Q. Qiu, A state-of-the-art review on the carbonation process in cementitious conservation, CrystEngComm 18 (2016) 6594–6607.
materials: fundamentals and characterization techniques, Constr. Build. Mater. [76] R. Veiga, Air lime mortars: what else do we need to know to apply them in
247 (2020), 118503. conservation and rehabilitation interventions? A review, Constr. Build. Mater.
[44] K. Wan, Q. Xu, Y. Wang, G. Pan, 3D spatial distribution of the calcium carbonate 157 (2017) 132–140.
caused by carbonation of cement paste, Cem. Concr. Compos. 45 (2014) 255–263. [77] J.I. Alvarez, R. Veiga, S. Martínez-Ramírez, M. Secco, P. Faria, P.N. Maravelaki,
[45] R.M.H. Lawrence, A Study of Carbonation in Non-hydraulic Lime Mortars, M. Ramesh, I. Papayanni, J. Valek, RILEM TC 277-LHS report: a review on the
University of Bath, Bath, UK, 2006 (PhD Thesis). mechanisms of setting and hardening of lime-based binding systems, Mater.
[46] M.A. Oliveira, M. Azenha, P.B. Lourenco, A. Meneghini, E.T. Guimaraes, Struct. 54 (2021) 75.
F. Castro, D. Soares, Experimental analysis of the carbonation and humidity [78] C. Groot, R. Veiga, I. Papayianni, R. Van Hees, M. Secco, J.I. Alvarez, P. Faria,
diffusion processes in aerial lime mortar, Constr. Build. Mater. 148 (2017) 38–48. M. Stefanidou, RILEM TC 277-LHS report: lime-based mortars for restoration–a
[47] E. Despotou, A. Shtiza, T. Schlegel, F. Verhelst, Literature study on the rate and review on long-term durability aspects and experience from practice, Mater.
mechanism of carbonation of lime in mortars, Mauerwerk 20 (2016) 124–137. Struct. 55 (2022) 245.
[48] J.C. Walker, P.B. Hays, J.F. Kasting, A negative feedback mechanism for the long- [79] W.D. Kingery, P.B. Vandiver, M. Prickett, The beginnings of pyrotechnology, part
term stabilization of Earth’s surface temperature, J. Geophys. Res. 86 (1981) II: production and use of lime and gypsum plaster in the Pre-Pottery Neolithic
9776–9782. Near East, J. Field Archaeol. 15 (1988) 219–243.
[49] R.A. Berner, A.C. Lasaga, R.M. Garrels, Carbonate-silicate geochemical cycle and [80] D. Carran, J. Hughes, A. Leslie, C. Kennedy, A short history of the use of lime as a
its effect on atmospheric carbon dioxide over the past 100 million years, Am. J. building material beyond Europe and North America, Int. J. Archit. Herit. 6
Sci. 283 (1983) 641–683. (2012) 117–146.
[50] F. Di Lorenzo, C. Ruiz-Agudo, A. Ibañez-Velasco, R. Gil-San Millán, J.A. [81] G. Artioli, M. Secco, A. Addis, The Vitruvian legacy: mortars and binders before
R. Navarro, E. Ruiz-Agudo, C. Rodriguez-Navarro, The carbonation of and after the Roman world, EMU Notes Miner. 20 (2019) 151–202.
wollastonite: a model reaction to test natural and biomimetic catalysts for [82] D.E. Friesem, I. Abadi, D. Shaham, L. Grosman, Lime plaster cover of the dead
enhanced CO2 sequestration, Minerals 8 (2018) 209. 12,000 years ago–new evidence for the origins of lime plaster technology, Evol.
[51] L. Monasterio-Guillot, A. Fernandez-Martinez, E. Ruiz-Agudo, C. Rodriguez- Hum. Sci. 1 (2019), e9.
Navarro, Carbonation of calcium-magnesium pyroxenes: physical-chemical [83] A.D. Cowper, Lime and Lime Mortars, Building Research Station, HM Stationary
controls and effects of reaction-driven fracturing, Geochim. Cosmochim. Acta 304 Office, London, 1927.
(2021) 258–280. [84] A. Moropoulou, A. Bakolas, K. Bisbikou, Investigation of the technology of
[52] D. Daval, I. Martinez, J. Corvisier, N. Findling, B. Goffe, F. Guyot, Carbonation of historic mortars, J. Cult. Herit. 1 (2000) 45–58.
Ca-bearing silicates, the case of wollastonite: experimental investigations and [85] R.M. Lawrence, T.J. Mays, S.P. Rigby, P. Walker, D. D’Ayala, Effects of
kinetic modeling, Chem. Geol. 265 (2009) 63–78. carbonation on the pore structure of non-hydraulic lime mortars, Cem. Concr.
[53] W. Seifritz, CO2 disposal by means of silicates, Nature 345 (1990) 486. Res. 37 (2007) 1059–1069.
[54] K.S. Lackner, C.H. Wendt, D.P. Butt, E.L. Joyce Jr., D.H. Sharp, Carbon dioxide [86] C. Rodriguez-Navarro, E. Ruiz-Agudo, A. Luque, A.B. Rodriguez-Navarro,
disposal in carbonate minerals, Energy 20 (1995) 1153–1170. M. Ortega-Huertas, Thermal decomposition of calcite: mechanisms of formation
[55] A. Sanna, M. Uibu, G. Caramanna, R. Kuusik, M.M. Maroto-Valer, A review of and textural evolution of CaO nanocrystals, Am. Mineral. 94 (2009) 578–593.
mineral carbonation technologies to sequester CO2, Chem. Soc. Rev. 43 (2014) [87] C. Rodriguez-Navarro, K. Kudlacz, E. Ruiz-Agudo, The mechanism of thermal
8049–8080. decomposition of dolomite: new insights from 2D-XRD and TEM analyses, Am.
[56] G. Gadikota, Carbon mineralization pathways for carbon capture, storage and Mineral. 97 (2012) 38–51.
utilization, Commun. Chem. 4 (2021) 23. [88] M.A. Shand, The Chemistry and Technology of Magnesia, John Wiley & Sons,
[57] M. Erans, S.A. Nabavi, V. Manović, Carbonation of lime-based materials under Hoboken, NJ, 2006.
ambient conditions for direct air capture, J. Clean. Prod. 242 (2020), 118330. [89] C. Rodriguez-Navarro, E. Hansen, W.S. Ginell, Calcium hydroxide crystal
[58] T. Aono, Studies on the reactions between gas and solid. Part II. Absorption of evolution upon aging of lime putty, J. Am. Ceram. Soc. 81 (1998) 3032–3034.
CO2 by CaO and Ca(OH)2, Bull. Chem. Soc. Japan 6 (1931) 319–324. [90] E. Ruiz-Agudo, C. Rodriguez-Navarro, Microstructure and rheology of lime putty,
[59] J.E. Gillott, Carbonatation of Ca(OH)2 investigated by thermal and x-ray Langmuir 26 (2010) 3868–3877.
diffraction methods of analysis, J. Appl. Chem. 17 (1967) 185–189. [91] C. Rodriguez-Navarro, E. Ruiz-Agudo, M. Ortega-Huertas, E. Hansen,
[60] O. Matsuda, H. Yamada, On the carbonation of slaked lime, Gyps. Lime 97 (1968) Nanostructure and irreversible colloidal behavior of Ca(OH)2: implications in
3–10. cultural heritage conservation, Langmuir 21 (2005) 10948–10957.
[61] D.R. Moorehead, Cementation by the carbonation of hydrated lime, Cem. Concr. [92] B.A. Silva, A.F. Pinto, A. Gomes, A. Candeias, Effects of natural and accelerated
Res. 16 (1986) 700–708. carbonation on the properties of lime-based materials, J. CO2 Util. 49 (2021),
[62] S.M. Shih, C.S. Ho, Y.S. Song, J.P. Lin, Kinetics of the reaction of Ca(OH)2 with 101552.
CO2 at low temperature, Ind. Eng. Chem. Res. 38 (1999) 1316–1322. [93] T. Ogino, T. Suzuki, K. Sawada, The formation and transformation mechanism of
[63] G. Cultrone, E. Sebastián, M.O. Huertas, Forced and natural carbonation of lime- calcium carbonate in water, Geochim. Cosmochim. Acta 51 (1987) 2757–2767.
based mortars with and without additives: mineralogical and textural changes, [94] S. Martinez-Ramirez, S. Sanchez-Cortes, J.V. Garcia-Ramos, C. Domingo,
Cem. Concr. Res. 35 (2005) 2278–2289. C. Fortes, M.T. Blanco-Varela, Micro-Raman spectroscopy applied to depth
[64] K. Van Balen, Carbonation reaction of lime, kinetics at ambient temperature, profiles of carbonates formed in lime mortar, Cem. Concr. Res. 33 (2003)
Cem. Concr. Res. 35 (2005) 647–657. 2063–2068.
[65] D.T. Beruto, F. Barberis, R. Botter, Calcium carbonate binding mechanisms in the [95] M. Singh, S.V. Kumar, S.A. Waghmare, P.D. Sabale, Aragonite–vaterite–calcite:
setting of calcium and calcium–magnesium putty-limes, J. Cult. Herit. 6 (2005) polymorphs of CaCO3 in 7th century CE lime plasters of Alampur group of
253–260. temples, India, Constr. Build. Mater. 112 (2016) 386–397.
[66] M. Arandigoyen, J.I. Alvarez, Carbonation process in lime pastes with different [96] R. Ravi, S. Thirumalini, N. Taher, Analysis of ancient lime plasters–reason behind
water/binder ratio, Mater. Constr. 56 (2006) 5–18. longevity of the Monument Charminar, India a study, J. Build. Eng. 20 (2018)
[67] P. De Silva, L. Bucea, D.R. Moorehead, V. Sirivivatnanon, Carbonate binders: 30–41.
reaction kinetics, strength and microstructure, Cem. Concr. Compos. 28 (2006) [97] S. Thirumalini, R. Ravi, M. Rajesh, Experimental investigation on physical and
613–620. mechanical properties of lime mortar: effect of organic addition, J. Cult. Herit. 31
[68] R.M.H. Lawrence, T.J. Mays, P. Walker, D. D’ayala, Determination of carbonation (2018) 97–104.
profiles in non-hydraulic lime mortars using thermogravimetric analysis, [98] K.A. Gour, R. Ramadoss, T. Selvaraj, Revamping the traditional air lime mortar
Thermochim. Acta 444 (2006) 179–189. using the natural polymer–Areca nut for restoration application, Constr. Build.
[69] Ö. Cizer, C. Rodriguez-Navarro, E. Ruiz-Agudo, J. Elsen, D. Van Gemert, K. Van Mater. 164 (2018) 255–264.
Balen, Phase and morphology evolution of calcium carbonate precipitated by [99] C. Rodriguez-Navarro, C. Jimenez-Lopez, A. Rodriguez-Navarro, M.T. Gonzalez-
carbonation of hydrated lime, J. Mater. Sci. 47 (2012) 6151–6165. Muñoz, M. Rodriguez-Gallego, Bacterially mediated mineralization of vaterite,
[70] Ö. Cizer, E. Ruiz-Agudo, C. Rodriguez-Navarro, Kinetic effect of carbonic Geochim. Cosmochim. Acta 71 (2007) 1197–1213.
anhydrase enzyme on the carbonation reaction of lime mortar, Int. J. Archit. [100] M.B. Toffolo, The significance of aragonite in the interpretation of the
Herit. 12 (2018) 779–789. microscopic archaeological record, Geoarchaeology 36 (2021) 149–169.
[71] E. Ruiz-Agudo, K. Kudłacz, C.V. Putnis, A. Putnis, C. Rodriguez-Navarro, [101] F. Lippmann, Crystal Chemistry of Sedimentary Carbonate Minerals, Springer,
Dissolution and carbonation of portlandite [Ca(OH)2] single crystals, Environ. Berlin, 1973.
Sci. Technol. 47 (2013) 11342–11349. [102] N. Oriols, N. Salvadó, T. Pradell, N. Jiménez, M. Cotte, V. Gonzalez, S. Butí,
[72] C. Rodriguez-Navarro, Binders in historical buildings: traditional lime in Carbonation of fresco mural paintings with a dolomitic mortar, Cem. Concr. Res.
conservation, Semin. SEM 9 (2012) 91–112. 157 (2022), 106828.
[73] C. Rodriguez-Navarro, O. Cazalla, K. Elert, E. Sebastian, Liesegang pattern [103] R.M. Dheilly, A. Bouguerra, B. Beaudoin, J. Tudo, M.I. Queneudec,
development in carbonating traditional lime mortars, Proc. R. Soc. London A 458 Hydromagnesite development in magnesian lime mortars, Mater. Sci. Eng. A 268
(2002) 2261–2273. (1999) 127–131.
[74] C. Rodriguez-Navarro, K. Kudłacz, Ö. Cizer, E. Ruiz-Agudo, Formation of [104] E. Königsberger, L.C. Königsberger, H. Gamsjäger, Low temperature
amorphous calcium carbonate and its transformation into mesostructured calcite, thermodynamic model for the system Na2CO3–MgCO3–CaCO3–H2O, Geochim.
CrystEngComm 17 (2015) 58–72. Cosmochim. Acta 63 (1999) 3105–3119.
[105] J. Lanas, J.L. Alvarez, Dolomitic lime: thermal decomposition of nesquehonite,
Thermochim. Acta 421 (2004) 123–132.

26
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

[106] G. Cultrone, A. Arizzi, E. Sebastián, C. Rodriguez-Navarro, Sulfation of calcitic [139] M. Volmer, A. Weber, Keimbildung in übersättigten Gebilden, Z. Phys. Chem. 119
and dolomitic lime mortars in the presence of diesel particulate matter, Environ. (1926) 277–301.
Geol. 56 (2008) 741–752. [140] R. Becker, W. Doring, Kinetische Behandlung der Keimbildung in übersättigen
[107] P. De Silva, L. Bucea, V. Sirivivatnanon, Chemical, microstructural and strength Dämpfen, Ann. Phys. 24 (1935) 719–752.
development of calcium and magnesium carbonate binders, Cem. Concr. Res. 39 [141] B. Mutaftschiev, The Atomistic Nature of Crystal Growth, Springer, Berlin, 2001.
(2009) 460–465. [142] D. Gebauer, M. Kellermeier, J.D. Gale, L. Bergström, H. Cölfen, Pre-nucleation
[108] A. Arizzi, G. Cultrone, The influence of aggregate texture, morphology and clusters as solute precursors in crystallisation, Chem. Soc. Rev. 43 (2014)
grading on the carbonation of non-hydraulic (aerial) lime-based mortars, Q. J. 2348–2371.
Eng. Geol. Hydrogeol. 46 (2013) 507–520. [143] J.W. Mullin, Crystallization, 4th edition, Butterworth, Oxford, 2001.
[109] K. Van Balen, D. Van Gemert, Modeling lime mortar carbonation, Mater. Struct. [144] J.J. De Yoreo, P.G. Vekilov, Principles of crystal nucleation and growth, Rev.
27 (1994) 393–398. Mineral. Geochem. 54 (2003) 57–93.
[110] M. Arandigoyen, J.P. Bernal, M.B. López, J.I. Alvarez, Lime-pastes with different [145] Q. Li, Y.S. Jun, The apparent activation energy and pre-exponential kinetic factor
kneading water: pore structure and capillary porosity, Appl. Surf. Sci. 252 (2005) for heterogeneous calcium carbonate nucleation on quartz, Commun. Chem. 1
1449–1459. (2018) 56.
[111] L.B. Sickels, Organic additives in mortars, Edinb. Architect. Res. 8 (1981) 7–20. [146] H. Cölfen, S. Mann, Higher-order organization by mesoscale self-assembly and
[112] L. Rampazzi, M.P. Colombini, C. Conti, C. Corti, A. Lluveras-Tenorio, transformation of hybrid nanostructures, Angew. Chem. Int. Ed. 42 (2003)
A. Sansonetti, M. Zanaboni, Technology of Medieval mortars: an investigation 2350–2365.
into the use of organic additives, Archaeometry 58 (2016) 115–130. [147] J.H. Harding, C.L. Freeman, D.M. Duffy, Oriented crystal growth on organic
[113] L. Ventola, M. Vendrell, P. Giraldez, L. Merino, Traditional organic additives monolayers, CrystEngComm 16 (2014) 1430–1438.
improve lime mortars: new old materials for restoration and building natural [148] D. Gebauer, A. Völkel, H. Cölfen, Stable prenucleation clusters, Science 322
stone fabrics, Constr. Build. Mater. 25 (2011) 3313–3318. (2008) 1819–1822.
[114] F. Yang, B. Zhang, Q. Ma, Study of sticky rice–lime mortar technology for the [149] A.F. Wallace, L.O. Hedges, A. Fernandez-Martinez, P. Raiteri, J.D. Gale, G.
restoration of historical masonry construction, Acc. Chem. Res. 43 (2010) A. Waychunas, S. Whitelam, J.F. Banfield, J.J. De Yoreo, Microscopic evidence for
936–944. liquid-liquid separation in spersaturated CaCO3 solutions, Science 341 (2013)
[115] E.H. Morris, J. Charlot, A.A. Morris, The Temple of the Warriors at Chichen Itza, 885–889.
Yucatan, Carnegie Institution, Washington D.C., 1931. [150] E. Ruiz-Agudo, A. Burgos-Cara, C. Ruiz-Agudo, A. Ibañez-Velasco, H. Cölfen,
[116] E.R. Littmann, Ancient Mesoamerican mortars, plasters, and stuccos: the use of C. Rodriguez-Navarro, A non-classical view on calcium oxalate precipitation and
bark extracts in lime plasters, Am. Antiq. 25 (1960) 593–597. the role of citrate, Nat. Commun. 8 (2017) 768.
[117] D. Magaloni, R. Pancella, Y. Fruh, J. Cañetas, V. Castaño, Studies on the Mayan [151] C. Rodriguez-Navarro, Ö. Cizer, K. Kudłacz, A. Ibañez-Velasco, C. Ruiz-Agudo,
mortars technique, MRS Proc. 352 (1995) 483–489. K. Elert, A. Burgos-Cara, E. Ruiz-Agudo, The multiple roles of carbonic anhydrase
[118] C. Rodriguez-Navarro, E. Ruiz-Agudo, A. Burgos-Cara, K. Elert, E.F. Hansen, in calcium carbonate mineralization, CrystEngComm 21 (2019) 7407–7423.
Crystallization and colloidal stabilization of Ca(OH)2 in the presence of nopal [152] J.J. De Yoreo, P.U.P.A. Gilbert, N.A.J.M. Sommerdijk, R.L. Penn, S. Whitelam,
juice (Opuntia ficus indica): implications in architectural heritage conservation, D. Joester, H. Zhang, J.D. Rimer, A. Navrotsky, J.F. Banfield, A.F. Wallace, F.
Langmuir 33 (2017) 10936–10950. M. Michel, F.C. Meldrum, H. Cölfen, P.M. Dove, Crystallization by particle
[119] C. Rodriguez-Navarro, A. Burgos-Cara, F. Di Lorenzo, E. Ruiz-Agudo, K. Elert, attachment in synthetic, biogenic and geologic environments, Science 349
Nonclassical crystallization of calcium hydroxide via amorphous precursors and (2015), aaa6760-1.
the role of additives, Cryst. Growth Des. 20 (2020) 4418–4432. [153] R. Demichelis, P. Raiteri, J.D. Gale, D. Quigley, D. Gebauer, Stable prenucleation
[120] D.M. Kern, The hydration of carbon dioxide, J. Chem. Educ. 37 (1960) 14–23. mineral clusters are liquid-like ionic polymers, Nat. Commun. 2 (2011) 590.
[121] V.A. Juvekar, M.M. Sharma, Absorption of CO2 in a suspension of lime, Chem. [154] F. Sebastiani, S.L. Wolf, B. Born, T.Q. Luong, H. Cölfen, D. Gebauer, M. Havenith,
Eng. Sci. 28 (1973) 825–837. Water dynamics from THz spectroscopy reveal the locus of a liquid–liquid binodal
[122] J.R. Burns, R.J.J. Jachuck, Monitoring of CaCO3 production on a spinning disc limit in aqueous CaCO3 solutions, Angew. Chem. Int. Ed. 56 (2017) 490–495.
reactor using conductivity measurements, AICHE J. 51 (2005) 1497–1507. [155] M. Sleutel, A.E. Van Driessche, Role of clusters in nonclassical nucleation and
[123] W. Dreybrodt, J. Lauckner, Liu Zaihua, U. Svensson, D. Buhmann, The kinetics of growth of protein crystals, Proc. Natl. Acad. Sci. 111 (2014) E546–E553.
the reaction CO2 + H2O → H+ + HCO−3 as one of the rate limiting steps for the [156] M.H. Nielsen, S. Aloni, J.J. De Yoreo, In situ TEM imaging of CaCO3 nucleation
dissolution of calcite in the system H2O-CO2-CaCO3, Geochim. Cosmochim. Acta reveals coexistence of direct and indirect pathways, Science 345 (2014)
60 (1996) 3375–3381. 1158–1162.
[124] W. Dreybrodt, L. Eisenlohr, B. Madry, S. Ringer, Precipitation kinetics of calcite in [157] J. Ihli, W.C. Wong, E.H. Noel, Y.Y. Kim, A.N. Kulak, H.K. Christenson, M.J. Duer,
the system CaCO3-H2O-CO2: the conversion to CO2 by the slow process H+ + F.C. Meldrum, Dehydration and crystallization of amorphous calcium carbonate
HCO−3 → CO2 + H2O as a rate limiting step, Geochim. Cosmochim. Acta 61 (1997) in solution and in air, Nat. Commun. 5 (2014) 3169.
3897–3904. [158] A. Gal, W. Habraken, D. Gur, P. Fratzl, S. Weiner, L. Addadi, Calcite crystal
[125] P. Somasundaran, G.E. Agar, The zero point of charge of calcite, J. Colloid growth by a solid-state transformation of stabilized amorphous calcium carbonate
Interface Sci. 24 (1967) 433–440. nanospheres in a hydrogel, Angew. Chem. Int. Ed. 52 (2013) 4867–4870.
[126] Z. Liu, W. Dreybrod, Dissolution kinetics of calcium carbonate minerals in H2O- [159] X. Zhang, A.S. Lea, A.M. Chaka, J.S. Loring, S.T. Mergelsberg, E. Nakouzi,
CO2 solutions in turbulent flow: the role of the diffusion boundary layer and the O. Qafoku, J.J. De Yoreo, H.T. Schaef, K.M. Rosso, In situ imaging of amorphous
slow reaction H2O + CO2 → H+ + HCO−3 , Geochim. Cosmochim. Acta 61 (1997) intermediates during brucite carbonation in supercritical CO2, Nat. Mater. 21
2879–2889. (2022) 345–351.
[127] V.H. Veley, The inertness of quicklime, J. Chem. Soc. Trans. 63 (1893) 821–833. [160] K. Zhang, Y. Zhang, Y. Liu, L. Wang, L. He, T. Dong, R. Lu, Y. Zhang, F. Yang,
[128] J.R. Johnstone, F.P. Glasser, Carbonation of single crystals of portlandite in Influence of polar amino acids on the carbonation of lime mortars, Herit. Sci. 10
cement paste, in: 9th ICCCC, vol. 5, 1992, pp. 370–376 (New Dehli). (2022) 190.
[129] F. Pontiga, J.M. Valverde, H. Moreno, F.J. Duran-Olivencia, Dry gas–solid [161] I. Matsushita, T. Suzuki, T. Moriga, T. Ashida, I. Nakabayashi, J. Metson, XPS
carbonation in fluidized beds of Ca(OH)2 and nanosilica/Ca(OH)2 at ambient study on the carbonation process of Ca(OH)2, J. Ceram. Soc. Japan 101 (1993)
temperature and low CO2 pressure, Chem. Eng. J. 222 (2013) 546–552. 725–727.
[130] G. Montes-Hernandez, A. Pommerol, F. Renard, P. Beck, E. Quirico, O. Brissaud, [162] H.W. Wang, L.L. Daemen, M.C. Cheshire, M.K. Kidder, A.G. Stack, L.F. Allard,
In situ kinetic measurements of gas-solid carbonation of Ca(OH)2 by using an J. Neuefeind, D. Olds, J. Liub, K. Page, Synthesis and structure of synthetically
infrared microscope coupled to a reaction cell, Chem. Eng. J. 161 (2010) pure and deuterated amorphous (basic) calcium carbonates, Chem. Commun. 53
250–256. (2017) 2942–2945.
[131] E.T. Stepkowska, Hypothetical transformation of Ca(OH)2 into CaCO3 in solid- [163] Y. Politi, D.R. Batchelor, P. Zaslansky, B.F. Chmelka, J.C. Weaver, I. Sagi,
state reactions of portland cement, J. Therm. Anal. Calorim. 80 (2005) 727–733. S. Weiner, L. Addadi, Role of magnesium ion in the stabilization of biogenic
[132] D.T. Beruto, R. Botter, Liquid-like H2O adsorption layers to catalyze the Ca(OH)2/ amorphous calcium carbonate: a structure− function investigation, Chem. Mater.
CO2 solid–gas reaction and to form a non-protective solid product layer at 20◦ C, 22 (2010) 161–166.
J. Eur. Ceram. Soc. 20 (2000) 497–503. [164] A.L. Harrison, G.M. Dipple, I.M. Power, K.U. Mayer, Influence of surface
[133] A. Putnis, Mineral replacement reactions, Rev. Mineral. Geochem. 70 (2009) passivation and water content on mineral reactions in unsaturated porous media:
87–124. implications for brucite carbonation and CO2 sequestration, Geochim.
[134] V. Nikulshina, M.E. Gálvez, A. Steinfeld, Kinetic analysis of the carbonation Cosmochim. Acta 148 (2015) 477–495.
reactions for the capture of CO2 from air via the Ca(OH)2–CaCO3–CaO solar [165] C. Rodriguez-Navarro, A. Burgos Cara, K. Elert, C.V. Putnis, E. Ruiz-Agudo, Direct
thermochemical cycle, Chem. Eng. J. 129 (2007) 75–83. nanoscale imaging reveals the growth of calcite crystals via amorphous
[135] V. Manovic, E.J. Anthony, Lime-based sorbents for high-temperature CO2 nanoparticles, Cryst. Growth Des. 16 (2016) 1850–1860.
capture—a review of sorbent modification methods, Int. J. Environ. Res. Public [166] C. Rodriguez-Navarro, E. Ruiz-Agudo, J. Harris, S.E. Wolf, Nonclassical
Health 7 (2010) 3129–3140. crystallization in vivo et in vitro (II): nanogranular features in biomimetic
[136] V. Materic, S.I. Smedley, High temperature carbonation of Ca(OH)2, Ind. Eng. minerals disclose a general colloid-mediated crystal growth mechanism, J. Struct.
Chem. Res. 50 (2011) 5927–5932. Biol. 196 (2016) 260–287.
[137] J.W. Gibbs, On the equilibrium of heterogeneous substances (first part), Trans. [167] B. Marchese, Non-crystalline Ca(OH)2 in ancient non-hydraulic lime mortars,
Connecticut Acad. Arts Sci. 3 (1876) 108–248. Cem. Concr. Res. 10 (1980) 861–864.
[138] J.W. Gibbs, On the equilibrium of heterogeneous substances (concluded), Trans. [168] R.G. Newton, J.H. Sharp, The chemical composition of lime plasters, Cem. Concr.
Connecticut Acad. Arts Sci. 16 (1878) 343–524. Res. 17 (1987) 77–80.

27
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

[169] J. Adams, D. Dollimore, D.L. Griffiths, Thermal analytical investigation of [200] A.C. Lasaga, Chemical kinetics of water-rock interactions, J. Geophys. Res. 89
unaltered Ca(OH)2 in dated mortars and plasters, Thermochim. Acta 324 (1998) (1984) 4009–4025.
67–76. [201] H.H. Wickman, J.N. Korley, Colloid crystal self-organization and dynamics at the
[170] D. Ferretti, Z.P. Bažant, Stability of ancient masonry towers: moisture diffusion, air/water interface, Nature 393 (1998) 445–447.
carbonation and size effect, Cem. Concr. Res. 36 (2006) 1379–1388. [202] M.Y. Lin, H. Lindsay, D.A. Weitz, R.C. Ball, R. Klein, P. Meakin, Universality in
[171] B. Xu, M.B. Toffolo, E. Boaretto, K.M. Poduska, Assessing local and long-range colloid aggregation, Nature 339 (1989) 360–362.
structural disorder in aggregate-free lime binders, Ind. Eng. Chem. Res. 55 (2016) [203] I. Galan, F.P. Glasser, D. Baza, C. Andrade, Assessment of the protective effect of
8334–8340. carbonation on portlandite crystals, Cem. Concr. Res. 74 (2015) 68–77.
[172] B. Lubelli, T.G. Nijland, R.P.J. Van Hees, Self-healing of lime based mortars: [204] M. Thiery, G. Villain, P. Dangla, G. Platret, Investigation of the carbonation front
microscopy observations on case studies, Heron 56 (2011) 75–91. shape on cementitious materials: effects of the chemical kinetics, Cem. Concr.
[173] L.M. Seymour, J. Maragh, P. Sabatini, M. Di Tommaso, J.C. Weaver, A. Masic, Hot Res. 37 (2007) 1047–1058.
mixing: mechanistic insights into the durability of ancient Roman concrete, Sci. [205] Y. Pu, L. Li, Q. Wang, X. Shi, C. Luan, G. Zhang, L. Fu, A.E.F. Abomohra,
Adv. 9 (2023), eadd1602. Accelerated carbonation technology for enhanced treatment of recycled concrete
[174] J. Lanas, J.P. Bernal, M.A. Bello, J.I. Alvarez, Mechanical properties of masonry aggregates: a state-of-the-art review, Constr. Build. Mater. 282 (2021) 122671.
repair dolomitic lime-based mortars, Cem. Concr. Res. 36 (2006) 951–960. [206] S. Sánchez Moral, J. García-Guinea, L.D. Luque Ripoll, R. González-Martín,
[175] A. Arizzi, G. Cultrone, The difference in behaviour between calcitic and dolomitic P. López-Arce, Cinética de carbonatación de morteros experimentales de cal de
lime mortars set under dry conditions: the relationship between textural and tipo romano, Mater. Constr. 54 (2004) 23–37.
physical–mechanical properties, Cem. Concr. Res. 42 (2012) 818–826. [207] D. Wang, J. Xiao, Z. Duan, Strategies to accelerate CO2 sequestration of cement-
[176] X. Zhang, F.P. Glasser, K.L. Scrivener, Reaction kinetics of dolomite and based materials and their application prospects, Constr. Build. Mater. 314 (2022),
portlandite, Cem. Concr. Res. 66 (2014) 11–18. 125646.
[177] G. Ponce-Antón, M.C. Zuluaga, L.A. Ortega, J.A. Mauleon, Multi-analytical [208] R. Hay, G. Kashwani, K. Celik, Carbonation, strength development, and
approach for chemical-mineralogical characterization of reaction rims in the lime characterization of calcined limestone as a potential construction material, Cem.
mortars from Amaiur Castle (Navarre, Spain), Microchem. J. 152 (2020), 104303. Concr. Res. 139 (2021), 106263.
[178] E. Ruiz-Agudo, F. Mees, P. Jacobs, C. Rodriguez-Navarro, The role of saline [209] A. Moropoulou, A. Bakolas, P. Moundoulas, E. Aggelakopoulou,
solution properties on porous limestone salt weathering by magnesium and S. Anagnostopoulou, Strength development and lime reaction in mortars for
sodium sulfates, Environ. Geol. 52 (2007) 269–281. repairing historic masonries, Cem. Concr. Compos. 27 (2005) 289–294.
[179] A. Khawam, D.R. Flanagan, Solid-state kinetic models: basics and mathematical [210] E. Ontiveros-Ortega, R. Rodríguez-García, A. González-Serrano, L. Molina,
fundamentals, J. Phys. Chem. B 110 (2006) 17315–17328. Evolution of mechanical properties in aerial lime mortars of traditional
[180] P. Baglioni, L. Dei, F. Piqué, G. Sarti, E. Ferroni, New autogenous lime-based manufacturing, the relationship between putty and powder lime, Constr. Build.
grouts used in the conservation of lime-based wall paintings, Stud. Conserv. 42 Mater. 191 (2018) 575–589.
(1997) 43–54. [211] L. Garijo, X. Zhang, G. Ruiz, J.J. Ortega, Age effect on the mechanical properties
[181] P. Baglioni, D. Chelazzi, R. Giorgi, E. Carretti, N. Toccafondi, Y. Jaidar, of natural hydraulic and aerial lime mortars, Constr. Build. Mater. 236 (2020)
Commercial Ca (OH)2 nanoparticles for the consolidation of immovable works of 117573.
art, Appl. Phys. A Mater. Sci. Process. 114 (2014) 723–732. [212] S.A. Walling, J.L. Provis, A discussion of the papers “Impact of hydrated
[182] R. Camerini, G. Poggi, D. Chelazzi, F. Ridi, R. Giorgi, P. Baglioni, The carbonation magnesium carbonate additives on the carbonation of reactive MgO cements” and
kinetics of calcium hydroxide nanoparticles: a Boundary Nucleation and Growth “Enhancing the carbonation of MgO cement porous blocks through improved
description, J. Colloid Interface Sci. 547 (2019) 370–381. curing conditions”, by C. Unluer & A. Al-Tabbaa, Cem. Concr. Res. 79 (2016)
[183] F.A. Pisu, D. Chiriu, P.C. Ricci, C.M. Carbonaro, Defect related emission in 424–426.
calcium hydroxide: the controversial band at 780 cm− 1, Crystals 10 (2020) 266. [213] A. Varzina, Ö. Cizer, L. Yu, S. Liu, D. Jacques, J. Perko, A new concept for pore-
[184] H. Yagi, A. Iwazawa, R. Sonobe, T. Matsubara, H. Hikita, Crystallization of scale precipitation-dissolution modelling in a lattice Boltzmann
calcium carbonate accompanying chemical absorption, Ind. Eng. Chem. Res. 23 framework–application to portlandite carbonation, Appl. Geochem. 123 (2020)
(1984) 153–158. 104786.
[185] J. García-Carmona, J. Gómez-Morales, R. Rodríguez-Clemente, Morphological [214] H.K. Henish, Crystal Growth in Gels, Pennsylvania State University Press,
control of precipitated calcite obtained by adjusting the electrical conductivity in University Park, Pennsylvania, USA, 1970.
the Ca(OH)2-H2O-CO2 system, J. Cryst. Growth 249 (2003) 561–571. [215] J. Delgado Rodrigues, Liesegang rings in differential deterioration patterns of
[186] G. Montes-Hernandez, F. Renard, N. Geoffroy, L. Charlet, J. Pironon, Calcite lime mortars, J. Cult. Herit. 21 (2016) 819–822.
precipitation from CO2-Ca(OH)2 slurry under high pressure of CO2, J. Cryst. [216] M. Castellote, C. Andrade, Modelling the carbonation of cementitious matrixes by
Growth 308 (2007) 228–236. means of the unreacted-core model, UR-CORE, Cem. Concr. Res. 38 (2008)
[187] J.R. Clarkson, T.J. Price, C.J. Adams, Role of metastable phases in the 1374–1384.
spontaneous precipitation of calcium carbonate, J. Chem. Soc. Faraday Trans. 88 [217] B.A. Silva, A.F. Pinto, A. Gomes, A. Candeias, Suitability of different surfactants as
(1992) 243–249. air-entraining admixtures for lime mortars, Constr. Build. Mater. 256 (2020)
[188] C. Tai, F.-B. Chen, Polymorphism of CaCO3 precipitated in a constant-composition 118986.
environment, AICHE J. 44 (1998) 1790–1798. [218] S.H. Kang, Y.H. Kwon, J. Moon, Quantitative analysis of CO2 uptake and
[189] J.D. Rodriguez-Blanco, S. Shaw, L.G. Benning, The kinetics and mechanisms of mechanical properties of air lime-based materials, Energies 12 (2019) 2903.
amorphous calcium carbonate (ACC) crystallization to calcite, via vaterite, [219] S.H. Kang, Y.H. Kwon, J. Moon, Controlling the hydration and carbonation in
Nanoscale 3 (2011) 265–271. lime-based materials: advantage of slow carbonation in CO2 curable construction
[190] R.M. Dheilly, J. Tudo, Y. Sebaïbi, M. Quéneudec, Influence of storage conditions materials, Constr. Build. Mater. 249 (2020) 118749.
on the carbonation of powdered Ca(OH)2, Constr. Build. Mater. 16 (2002) [220] F.G. Branco, M.D.L. Belgas, C. Mendes, L. Pereira, J.M. Ortega, Characterization
155–161. of fresh and durability properties of different lime mortars for being used as
[191] T. Yang, B. Keller, E. Gagyari, K. Hametner, D. Günther, Direct observation of the masonry coatings in the restoration of ancient constructions, Sustainability 13
carbonation process on the surface of calcium hydroxide crystals in hardened (2021) 4909.
cement paste using an atomic force microscope, J. Mater. Sci. 38 (2003) [221] F.P. Campo, C. Tua, L. Biganzoli, S. Pantini, M. Grosso, Natural and enhanced
1909–1916. carbonation of lime in its different applications: a review, Environ. Technol. Rev.
[192] E. Dubina, L. Korat, L. Black, J. Strupi-Šuput, J. Plank, Influence of water vapour 10 (2021) 224–237.
and carbon dioxide on free lime during storage at 80 ◦ C, studied by Raman [222] M.I. Romero-Hermida, A.M. Borrero-López, V. Flores-Alés, F.J. Alejandre, J.
spectroscopy, Spectrochim. Acta A 111 (2013) 299–303. M. Franco, A. Santos, L. Esquivias, Characterization and analysis of the
[193] G.L. Pesce, I.W. Fletcher, J. Grant, M. Molinari, S.C. Parker, R.J. Ball, Carbonation carbonation process of a lime mortar obtained from phosphogypsum waste, Int. J.
of hydrous materials at the molecular level: a time of flight-secondary ion mass Environ. Res. Public Health 18 (2021) 6664.
spectrometry, Raman and density functional theory study, Cryst. Growth Des. 17 [223] S. Jayasingh, T. Selvaraj, S. Raneri, Evaluating the impact of organic addition and
(2017) 1036–1044. aggregate gradation on air lime mortar: new compatible green material for
[194] B. Chen, M.L. Laucks, E.J. Davis, Carbon dioxide uptake by hydrated lime aerosol heritage application, Int. J. Architect. Herit. 16 (2022) 681–691.
particles, Aerosol Sci. Technol. 38 (2004) 588–597. [224] J. Diaz-Basteris, J.C.S. Rivero, B. Menéndez, Life cycle assessment of restoration
[195] J.H. Park, S.H. Kim, J.C. Kim, B.Y. Choi, S.K. Kwak, O.H. Han, Y.I. Kim, S.W. Lee, mortars and binders, Constr. Build. Mater. 326 (2022) 126863.
Role of intercalated water in calcium hydroxide interlayers for carbonation [225] F. Medici, G. Rinaldi, Poly-amino-phenolic additives accelerating the carbonation
reaction, Chem. Eng. J. 420 (2021) 130422. of hydrated lime in mortar, Environ. Eng. Sci. 19 (2002) 271–276.
[196] D. Ergenç, R. Fort, Accelerating carbonation in lime-based mortar in high CO2 [226] D. Ergenç, R. Fort, A. Santos Silva, R. Veiga, D. Sanz Arauz, The effects of
environments, Constr. Build. Mater. 188 (2018) 314–325. DiloCarB as carbonation accelerator on the properties of lime mortars, Mater.
[197] C. Rodriguez-Navarro, E. Ruiz-Agudo, Nanolimes: from synthesis to application, Struct. 51 (2018) 10.
Pure Appl. Chem. 90 (2018) 523–550. [227] I. Karatasios, M.S. Katsiotis, V. Likodimos, A.I. Kontos, G. Papavassiliou,
[198] K. Vance, G. Falzone, I. Pignatelli, M. Bauchy, M. Balonis, G. Sant, Direct P. Falaras, V. Kilikoglou, Photo-induced carbonation of lime-TiO2 mortars, Appl.
carbonation of Ca(OH)2 using liquid and supercritical CO2: implications for Catal. B 95 (2010) 78–86.
carbon-neutral cementation, Ind. Eng. Chem. Res. 54 (2015) 8908–8918. [228] S. Jayasingh, T. Selvaraj, Effect of natural herbs on hydrated phases of lime
[199] J. Yu, X. Zeng, G. Zhang, J. Zhang, Y. Wang, G. Xu, Kinetics and mechanism of mortar, J. Archit. Eng. 26 (2020), 04020021.
direct reaction between CO2 and Ca(OH)2 in micro fluidized bed, Environ. Sci. [229] M. Shivakumar, T. Selvaraj, M.P. Dhassaih, Preparation and characterization of
Technol. 47 (2013) 7514–7520. ancient recipe of organic lime putty-evaluation for its suitability in restoration of
Padmanabhapuram Palace, India, Sci. Rep. 11 (2021) 13261.

28
C. Rodriguez-Navarro et al. Cement and Concrete Research 173 (2023) 107301

[230] A. Manoharan, C. Umarani, Properties of air lime mortar with bio-additives, [236] I. Rigopoulos, L. Kyriakou, M.A. Vasiliades, T. Kyratsi, A.M. Efstathiou,
Sustainability 14 (2022) 8355. I. Ioannou, Improving the carbonation of air lime mortars at ambient conditions
[231] F.M. León-Martínez, P.D.J. Cano-Barrita, F. Castellanos, K.B. Luna-Vicente, via the incorporation of ball-milled quarry waste, Constr. Build. Mater. 301
S. Ramírez-Arellanes, C. Gómez-Yáñez, Carbonation of high-calcium lime mortars (2021), 124073.
containing cactus mucilage as additive: a spectroscopic approach, J. Mater. Sci. [237] H. Liu, Y. Zhao, C. Peng, S. Song, A. López-Valdivieso, Lime mortars–the role of
56 (2021) 3778–3789. carboxymethyl cellulose on the crystallization of calcium carbonate, Constr.
[232] M. Carmona-Carmona, P. Acedo-Fuentes, A. Romero-Casado, J.M. Meneses- Build. Mater. 168 (2018) 169–177.
Rodríguez, M. Trujillo-Gómez, J.J. Tejado-Ramos, Chitosan as a carbonation [238] T.M. Jassam, K. Kien-Woh, B. Lau, M.M.M. Yaseer, Novel cement curing
catalyst in lime mortars, Res. Eng. Des. 17 (2023), 100912. technique by using controlled release of carbon dioxide coupled with nanosilica,
[233] J. Lanas, J.I. Alvarez-Galindo, Masonry repair lime-based mortars: factors Constr. Build. Mater. 223 (2019) 692–704.
affecting the mechanical behavior, Cem. Concr. Res. 33 (2003) 1867–1876. [239] J. Zhu, X. Li, Y. Zhang, J. Wang, B. Wei, Graphene-enhanced nanomaterials for
[234] S. Scannell, M. Lawrence, P. Walker, Impact of aggregate type on air lime mortar wall painting protection, Adv. Funct. Mater. 28 (2018) 1803872.
properties, Energy Procedia 62 (2014) 81–90. [240] C. Rodriguez-Navarro, L. Monasterio-Guillot, M. Burgos-Ruiz, E. Ruiz-Agudo,
[235] C. Martinez-Garcia, B. Gonzalez-Fonteboa, D. Carro-Lopez, F. Martinez-Abella, K. Elert, Unveiling the secret of ancient Maya masons: Biomimetic lime plasters
Carbonation evolution of lime putty coatings with mussel shell aggregate, Constr. with plant extracts, Sci. Adv. 9 (2023) eadf6138.
Build. Mater. 264 (2020) 120165.

29