Impact of water-soluble cellulose ethers on polymer-modified mortars

Journal of Materials Science, Feb 2014

Cellulose ethers (CEs) are employed in many polymer-modified mortars, such as cement renders, masonry mortars, tile adhesives, repair mortars, skim coats, and self-levelling mortars. The addition of CEs to mortars causes the retardation of cement hydration and modifies the microstructural characteristics and the properties of these mortars. The present work attempts to provide a comprehensive review of the current state of knowledge on the effects of CEs and critically identifies gaps in the knowledge. A fundamental scientific understanding concerning the chemistry and hydration of cement, chemical natures, and relevant properties of CEs are discussed. The behaviours and mechanisms of CE adsorption on cement are assessed. The influences of CEs on the kinetics of cement hydration, mechanisms of retardation, and microstructural evolution of the mortars also are reviewed. Finally, the impact of CEs on the properties of fresh and hardened mortars as well as the approaches used to mitigate the negative impacts of CEs are discussed.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://link.springer.com/content/pdf/10.1007%2Fs10853-013-7732-8.pdf

Impact of water-soluble cellulose ethers on polymer-modified mortars

D. D. Nguyen 0 1 L. P. Devlin 0 1 P. Koshy 0 1 C. C. Sorrell 0 1 0 L. P. Devlin Bostik SA, D 319 Coubert, 77257 Brie-Comte-Robert, France 1 D. D. Nguyen (&) P. Koshy C. C. Sorrell School of Materials Science and Engineering, The University of New South Wales , Sydney, NSW 2052, Australia Cellulose ethers (CEs) are employed in many polymer-modified mortars, such as cement renders, masonry mortars, tile adhesives, repair mortars, skim coats, and self-levelling mortars. The addition of CEs to mortars causes the retardation of cement hydration and modifies the microstructural characteristics and the properties of these mortars. The present work attempts to provide a comprehensive review of the current state of knowledge on the effects of CEs and critically identifies gaps in the knowledge. A fundamental scientific understanding concerning the chemistry and hydration of cement, chemical natures, and relevant properties of CEs are discussed. The behaviours and mechanisms of CE adsorption on cement are assessed. The influences of CEs on the kinetics of cement hydration, mechanisms of retardation, and microstructural evolution of the mortars also are reviewed. Finally, the impact of CEs on the properties of fresh and hardened mortars as well as the approaches used to mitigate the negative impacts of CEs are discussed. - Polymer-modified mortars have assumed greater importance in construction applications due to their advantages over conventional mortars. Common polymer-modified mortars are currently being used as cement renders (to cover walls), masonry mortars (to join bricks or concrete blocks), tile adhesives (to adhere them to floors and walls), repair mortars (to fill cracks in buildings, pavements, and footpaths), self-levelling mortars (to level floor surfaces), and skim coats (to form thin smooth cover layers for floors and walls). The compositions of the polymer-modified mortars vary, depending on the applications and manufacturers. Beside Portland cement and sand, polymer-modified mortars can contain additives such as accelerators, retarders, plasticisers, superplasticisers, water-retaining agents, mineral constituents, bonding agents, air-entraining agents, defoamers, and viscosity enhancement agents. Cellulose ethers (CEs) were first introduced into tile adhesives in the 1960s [1, 2] and are commonly used in the production of many other polymer-modified mortars as listed above [35]. The addition of CEs is well known to improve the water retention of these mortars, and this property allows for the use of thin layers of the mortars, instead of thick layers of unmodified mortar, to compensate for the water loss [1, 2]. Moreover, the increase in water retention of the mortar achieved through the use of CEs allows for the application of mortars on substrates and tiles without pre-wetting. These advantages further help to decrease the material and labour requirements and associated costs. In addition to its effect on water retention, addition of CEs also impacts cement hydration, as well as the microstructural characteristics and properties of the mortars [6]. The CEs that are commonly employed in polymermodified mortars are methyl cellulose (MC), hydroxyethyl cellulose (HEC), methyl hydroxyethyl cellulose (MHEC), and methyl hydroxypropyl cellulose (MHPC) [6, 7]. The CEs may exist with varying degrees of substitution (DS), molar substitution (MS), and molecular weight (Mw) [8]. The interaction between substitution groups (viz., between methyl groups and hydroxyethyl groups in MHEC, or between methyl groups and hydroxypropyl groups in MHPC) results in highly complex chemical structures of the CEs. The chemical and structural characteristics of CEs affect their behaviours in the mortars [5]. Different applications of the polymer-modified mortars require properties specific for the applications. Therefore for the effective use of CEs for these applications, a comprehensive understanding of the impact of the CEs positive and negativeis essential. The critical factors which need to be considered with regard to CE selection and addition are the type of substitution, DS, MS, degree of polymerisation, and dosage. In contrast with the clarity in the effects of CE content, there exists ambiguity with regard to the impact of the chemical structure of CEs on their behaviour in the mortars and the resultant properties. In this review, the current knowledge of the influence of CE on the kinetics of cement hydration, microstructural development, and properties of the mortars will be reviewed in relation to their use as guidelines for formulators to design mortars using CEs, while the gap in the knowledge will be identified in order to suggest areas for future research and development. Portland cement The conventional nomenclature used in cement terminology is listed in Table 1. Portland cement generally is a fine powder that results from the grinding of a mixture of cement clinker and gypsum. Clinker consists of four main phases, viz. C3S, C2S, C3A, and C4AF [913]. C2S and C3S: Hydration of C2S and C3S produces calcium silicate hydrates (CSH) and portlandite (CH). Hydration of C2S is much slower and produces less CH compared to C3S [12, 13]. C3A: In the absence of gypsum, the reaction of C3A with water is very fast (known as flash set) [12, 13]. This firstly produces the metastable calcium aluminate hydrates, dicalcium aluminate hydrate (C2AH8), and Table 1 Conventional nomenclature for cement terminology tetracalcium aluminate hydrate (C4AH13), which then convert rapidly to the more stable product tricalcium aluminate hydrate (C3AH6) [9, 11, 14]. Gypsum (CaSO4 2H2O), hemihydrate (CaSO4 H2O), or anhydrite (CaSO4) is added in order to prevent flash setting of C3A. The reactions of C3A and calcium sulphate with water result in the formation of ettringite (C6AS3H32) [12, 13]. As calcium sulphate is consumed, ettringite reacts with excess C3A to produce tetracalcium aluminate monosulphate hydrate (C4ASH12). C4AF: The hydration behaviour of C4AF is similar to that of C3A [9, 1113]. However, C4AF is less reactive with water than C3A and, the higher the Al/Fe ratio in C4AF, the greater the reactivity of C4AF. Cellulose ethers Chemistry of CEs CEs are semi-synthesised polymers that are produced from natural cellulose. The molecules of CEs consist of monomers that are anhydroglucose units (AGUs) with substituted ether groups [8]. Each AGU has six carbon atoms which are marked from one to six, and it has three hydroxyl groups bonded to C2, C3, and C6 (Fig. 1). The ether groups can be alkyl groups (typically methyl (CH3), and ethyl (C2H5)) and hydroxyalkyl groups (typically hydroxyethyl (C2H4OH), and hydroxypropyl (C3H6OH)). The ether groups can be substituted randomly only into the (OH) groups bonded to C2, C3, and C6 (Fig. 2). The CEs that are commonly used in polymer-modified mortars contain one or two different types of ether groups. The CEs that contain only one type of ether group are MCs and HECs (Fig. 2a, b). The CEs that have two types of ether groups are MHECs and MHPCs. In MC, each (OH) group of AGU may be substituted by only one methyl group (Fig. 2a). In HEC, each (OH) group of AGU may be substituted by more than one hydroxyethyl group that can be written as (C2H4O)nH (with n C 1) (Fig. 2b) [8]. In MHEC and MHPC, the Fig. 1 Chemical structure of cellulose with the monomer being an AGU, adapted from [8] CH3 - O CH2 CH2 CH3 - O CH2 CH2 (CH2 CH2 O)3 - H HO CH2 CH2 (CH2 CH2 O)2 - H O CH3H Fig. 2 Chemical structure of different CEs, based on [8]. a MC [DS = MS = 1.5], b HEC [DS = 1.5, MS = 3], c MHEC [DS = 2, MS (CH3) = 1.67, MS (C2H4O) = 0.67], d MHPC [DS = 2, MS (CH3) = 1.67, MS (C3H6O) = 0.67] methyl group may be attached directly to the AGU or attached indirectly through hydroxyethyl group(s) in MHECs or through hydroxypropyl group(s) in MHPCs, while the hydroxyethyl groups (in MHECs) or hydroxypropyl groups (in MHPCs) can be attached directly to the AGU (Fig. 2c, d). There can be more than one hydroxyethyl group in MHECs or hydroxypropyl group in MHPCs that can be substituted into one hydroxyl group of AGU. The molecules of CEs may contain multiple monomers due to the difference in the type of ether groups, the position of the ether groups within the AGU, and it is also HO CH2 CH2 HO CH2 CH2 reported that a proportion of AGU may be unsubstituted by any ether group due to incomplete etherifying process (Fig. 2). The chemistry of CEs is characterised by: Type(s) of ether: The most commonly used CEs are methyl, hydroxyethyl, or hydroxypropyl CE. DS: It is the average number of hydroxyl groups of an AGU occupied by ether groups. The maximal value of DS is three. MS: It is the average number of ether groups (for CEs that contain only one type of ether) or of each type of ether groups (for CEs that contain more than one type of ether groups) per AGU. The maximal value of MS is not limited to three. Average Mw: It is the average molecular weight of CE molecules. Mw can be calculated using the following equation: Mw PPN nN MN2 ; 1 where N is number of molecules in the polymer sample and nN the number of molecules that have molecular weight of Mw. Relevant properties of CEs Solubility of CEs The solubility of CEs depends on their chemical nature. An increase in Mw reduces the solubility of CEs [8]. The influence of the DS on the solubility of CEs is shown in Fig. 3. The values of a and b (in Fig. 3) are dependent on the type of substitution. MC has higher values of a and b than HEC [8]. MC with DS (0.11.1) is soluble in 68 % NaOH solution, and MC with DS (1.42.0) is soluble in water. HEC with DS (0.110.31) is soluble in 7 % NaOH, and HEC with DS (0.661.66) is soluble in water. In cement mortars, the pore solution is an alkaline environment that mainly results from CH due to the dissolution of cement phases. The selection of CEs with appropriate chemical nature is critical to ensure their solubility in cement pore solution as it affects the impact of CEs on cement hydration and thereby the properties of the fresh and hardened mortars. Surface tension of CE solutions Surface tension at the airliquid interface of the CE solutions is related to the ability to entrain air into the mortar. The lower the surface tension of the solution, the greater the ability of the mortar to entrain air. Addition of CE reduces the surface tension of aqueous solutions at the interface between the solution and a pore (containing air) since CEs contain hydrophobic groups (such as CH3), which orient toward the air at the airwater interface. Increase in the CE concentration reduces the surface tension and hence increases the ability to entrain air due to an increase in the number of hydrophobic groups [15]. Rheology of CE solutions Rheology is the term used to describe the deformation and flow behaviour of the materials [1618]. The rheology of CE solutions is characterised by their viscosityshear rate or shear stressshear rate dependency. Viscosity of a CE solution at a specified shear rate is dependent on the CE concentration and the chemical nature, which includes the type of substitution, DS, MS, and distribution of substitutions on the cellulose chains [19]. Increase in the CE concentration increases the viscosity of the solution. The solution of CE shows shear thinning behaviour, i.e., the viscosity decreases as the shear rate increases (Fig. 4) [8]. Increase in the Mw of CE leads to more pronounced shear thinning behaviour. The CE solution shows thixotropic property, with the recovery of viscosity occurring as the shear stress is removed (Fig. 4). The formation of a threedimensional network structure by CE molecules in solution is believed to be responsible for this property. Under the application of shear stress, the network is disrupted, resulting in a decrease in viscosity. The viscosity becomes constant when the network is completely disrupted. When the shear stress is removed, the reconstruction of the three-dimensional network occurs, thereby increasing the viscosity. Shear rate (s-1) Fig. 3 Influence of the DS on the solubility of CEs, based on [19] Fig. 4 Rheology of the CE solution, based on [8] Ettringite C3S Fig. 5 Adsorption of HECs with different MS and Mw on cement hydrates (after 2-h exposure in CH solution of 20 mM), hydrating C3S, and hydrating Portland cement (after 1 h of contacting with water), initial CE/cement hydrate = 15 mg/g, water/solid ratio = 20, adapted from [20, 21] MHPC [DS = 1.9] MHPC [DS = 1.7] MHEC [DS = 1.7] MHEC [DS = 1.5] [DS = 1.1] Fig. 6 Adsorption of CEs with different DS on cement phases, adapted from [5] Adsorption of CEs on cement Cellulose ethers can be adsorbed on cement hydrates, hydrating clinker phases, and hydrating cement (Fig. 5, 6, 7). Adsorption of CEs on cement hydrates The adsorption of CEs on cement hydrates under a stable condition (i.e., under saturated calcium hydroxide solution, in order to prevent the dissolution of hydrates) reaches an equilibrium after a certain period of time, and the amount of HEC 1 [MS = 2.0, Mw = 175 kDa] HEC 2 [MS = 2.5, Mw = 175 kDa] HEC 3 [MS = 2.5, Mw = 1335 kDa] Conversion of C2AH8 and C4AH13 into C3AH6 Fig. 7 Adsorption of HEC with DS = 1.1 on hydrating Portland cement, pure C3S, pure C3A, and mixture of (C3A ? CSH0.5), initial HEC/cement ratio = 0.067 wt%, w/c ratio = 0.5), adapted from [5] adsorbed CEs becomes constant. The amounts of CE adsorbed on cement hydrates (weight of CE/surface area of cement hydrate) are in the order: calcium aluminate hydrates (e.g., C2AH8, C4AH13, and C3AH6) [ CH [ CSH [ ettringite (Fig. 5) [20, 21]. The study by Muller [5] shows that there is a significant adsorption of CEs on gypsum (Fig. 6). However, the use of different concentrations of CEs for adsorption on gypsum and other hydrates did not allow for any comparison between the adsorption of CEs on gypsum and other hydrates. There does not appear to be any published study that has focused on the adsorption of CEs on calcium aluminoferrite hydrates (e.g., C2(A,F)H8, C4(A,F)H13, and C3(A,F)H6), C6(A,F)S3H32 and monosulphate phases (viz., C6ASH12 and C4(A,F)SH12). However, it could be expected that the adsorption of CEs on calcium aluminoferrite hydrates is similar to that on calcium aluminate hydrates, and the adsorption of CEs on C6(A,F)S3H32 is similar to that on ettringite. Adsorption of CEs on hydrating clinker phases In contrast, the adsorption of CEs on hydrating cement or hydrating clinker phases is a dynamic process, and the amount of adsorbed CEs changes during time as the hydration of cement continues (Fig. 7) [5]. The amount of adsorbed CEs on hydrating C3S particles progressively increases with the time of hydration as the amounts of CSH and CH increases. There is a significant amount of adsorption of CEs on C3A particles after 1 h of mixing with water (Fig. 7) [5]. This is due to the adsorption of CEs on the hexagonal platy crystals of C2AH8 and C4AH13 that form rapidly in significant amounts on the surface of C3A during this period [14]. There is a reduction in the amount of absorbed CE between 1 h and 8 h of hydration and this is due to the desorption of CE from the surface of C3A particles during this period. This desorption of CE is believed to be the result of the conversion of hexagonal plates of C2AH8 and C4AH13 into cubic crystals of C3AH6 during this period. Moreover, the adsorption of CEs on the hexagonal plates of C2AH8 and C4AH13 is higher than the adsorption of CEs on cubic crystals of C3AH6, due to the higher surface area and greater affinity of C2AH8 and C4AH13 to CEs compared to C3AH6. As the hydration of C3A proceeds, the amount of C3AH6 formed around the C3A particles increases, and this results in an increase in the surface area available for CEs to be adsorbed. This explains the regain in CE adsorption on C3A after 24 h of hydration. It is clear that the presence of calcium sulphates greatly reduces the adsorption of CEs on hydrating C3A. This is due to the formation of ettringite that has significantly lower CE adsorption compared to calcium aluminate hydrates. There is no published study on the adsorption of CEs on hydrating C2S and C4AF. However, it could be expected that the adsorption behaviour of CEs on hydrating C2S increases gradually accompanied by the progressive formation of CSH and CH. The rate of increase in the CE adsorption on hydrating C2S is expected to be lower than that on C3S during early age of hydration as the rate of hydration of C2S and the amount of CH produced are lower than the corresponding values during the hydration of C3S. The trend of the adsorption of CEs on hydrating C4AF is believed to be similar to that of C3A as the kinetics of the hydration of C4AF is similar to C3A. However, the rate of changes (increase or decrease) in the adsorption of CEs on hydrating C4AF particles is lower than that of C3A. Adsorption of CEs on Portland cement Adsorption of CEs on OPC shows a gradual increase but occurs at a lower rate compared with that of C3S during the first 24 h (Fig. 7) [5]. This is due to OPC being composed on the following major phases: C3S (69.1 wt%), C2S (11.2 wt%), C3A (5.4 wt%), C4AF (11.5 wt%), CS (1.6 wt%), and CSH0.5 (1.3 wt%). During the first 24 h of hydration, adsorption of CEs on hydrating C3S contributes mainly to the adsorption of CEs on Portland cement, which in that study constitutes 69.1 wt% (instead of 100 wt%). Significant hydration of C3A also occurs, but the adsorption of CEs on ettringite contributes only to a limited amount of CE being adsorbed on Portland cement. Fig. 8 Mechanisms of adsorption of CEs on cement based on [6, 23 25] (R: hydrophobic groups, such as CH3) Mechanisms of adsorption of CEs on cement There are no studies that elucidate the mechanisms of adsorption of CEs on cement. However, several mechanisms have been proposed to explain the adsorption of CEs on cement as discussed below: CEs can be adsorbed on unhydrated cement and hydrates through the formation of hydrogen bonds between (OH) groups of CEs and oxygen on the surface of unhydrated cement and hydrates (Fig. 8a) [2224]. Adsorption of CE can occur through complex formation between (OH) groups of CE and metal ions (such as calcium, aluminium, and iron ions) on the surface of unhydrated cement and hydrates (Fig. 8b) [23, 25]. Ou et al. [23] stated that the hydrophobic interactions between the hydrophobic groups of CE (such as CH3) and the surface of unhydrated cement and hydrates may also contribute to the adsorption of CE (Fig. 8c). Adsorption of CE can occur by the formation of electrostatic forces (attractive forces) that are induced by the charged surfaces of unhydrated cement and hydrates and polar (OH) groups of CE (Fig. 8d) [23]. Influence of chemical nature of CEs on adsorption on cement Due to the complexity of CE chemistry, no published study has allowed for the direct comparison of the influence of the CE type on their adsorption on cement. Hydrophobic interaction Degree of substitution Muller [5] investigated the effect of DS on the adsorption of CE on ettringite, CH, gypsum, and hydrating C3S, and it was found that the effect of the DS on the adsorption depends on the substrate (Fig. 6). No significant effect of the DS on the adsorption of CEs on ettringite was found as the amounts of CE adsorbed were insignificant. The adsorption of CEs on CH was found to be independent on the DS. For gypsum and C3S, it was found that the lower the DS, the greater the adsorption of CE. For example, MHEC (DS = 1.5) was adsorbed on C3S and gypsum to a greater extent compared to MHEC (DS = 1.7). Similarly, MHPC (DS = 1.7) was adsorbed on C3S, and gypsum to a greater extent compared to MHPC (DS = 1.9). It should be noted that Muller did not provide the MS of methyl (CH3), hydroxyethyl (C2H4OH), and hydroxypropyl (C3H6OH) of MHEC and MHPC and this could have affected the results. The hydration of C3S produces CSH and CH. The strong dependence on the DS of the adsorption of CEs on C3S and the insignificant dependence of the DS on the adsorption of CEs on CH indicate that the DS has a significant effect on the adsorption of CEs on CSH (i.e., the lower the DS, the greater the adsorption). There is no published study that has reported the influence of DS on the adsorption behaviour of other types of CEs (e.g., MC and HEC), other hydrates (viz., calcium aluminate hydrates, tetracalcium aluminoferrite hydrates, C6(A,F)S3H32, C4ASH12, and C4(A,F)SH12), other clinker phases (viz., C2S, C3A, and C4AF), and Portland cement. Molar substitution Pourchez et al. [20, 21] studied the influence of MS of HEC on the adsorption on hydrates (viz., CSH, CH, calcium aluminate hydrates, and ettringite), C3S, and Portland cement (Fig. 5). These studies have shown that the lower the MS of CEs, the greater the adsorption on cement hydrates, except in the case of ettringite, where the adsorption was insignificant. The effect of MS on CE adsorption on cement hydrates could result from the fact that CEs with lower MS also may have lower DS, which results in greater adsorption on cement hydrates compared to the case of CEs with higher DS (as presented in Degree of substitution section). For the same Mw, HEC 1 (MS = 2.2) was absorbed on CSH, CH, and calcium aluminates, to a greater extent compared to HEC 2 (MS = 2.5). The significant errors in the results related to the adsorption of HECs on C3S and Portland cement do not allow for a strong conclusion regarding the effect of MS. However, the average values of adsorption of HECs on C3S and Portland cement show that the higher the MS, the greater the adsorption. This is in agreement with the effect of MS of HECs on the adsorption on hydrates (viz., CS H, CH, and calcium aluminate hydrates). It should be noted that Pourchez et al. [20, 21] did not provide DS values of these HECs, and the results could be affected by the variations in the DS values. There does not appear to be any study that has reported the influence of MS of other CEs on their adsorption behaviours and the influence of MS of CEs on adsorption on gypsum, others hydrates (viz., tetracalcium aluminoferrite hydrates, C6(A,F)S3H32, C4ASH12, and C4(A,F)SH12), and other clinker phases (C2S, C3A, and C4AF). Molecular weight Pourchez et al. [20, 21] studied the influence of Mw on the adsorption of HECs on hydrates (viz., CSH, CH, calcium aluminate hydrates, and ettringite), C3S, and Portland cement (Fig. 5). However, the results did not allow for any clear conclusion. The influence of the Mw of other CEs on their adsorption behaviour and the influence of MS of CEs on their adsorption on gypsum, others hydrates (viz., tetracalcium aluminoferrite hydrates, C6(A,F)S3H32, C4ASH12, and C4(A,F)SH12), and other clinker phases (C2S, C3A, and C4AF) have not been reported and need to be investigated. Influence of CEs on cement hydration Understanding the influence of CEs on cement hydration is critical to explain the observed changes in the properties of fresh and hardened mortars during curing. The understanding developed will help to design products which have the advantages of CE addition, while their disadvantages of their use are mitigated through the use of other additives [24]. Influence of CEs on the kinetics and phase generation during cement hydration Hydration of C3S Similar to the hydration of calcium silicates in the absence of CE, the hydration of calcium silicates in the presence of CE can be divided into four stages: initial reaction period, slow reaction period, acceleration period, and deceleration period (Fig. 9). The initial reaction period is characterised by the dissolution of C3S into water, with the evolution of a significant quantity of (exothermal) heat [14]. The presence of CE does not significantly affect this stage [21]. Limited amounts of CH and CSH are produced at this stage. Part Days Age of specimen Fig. 9 Typical heat flow curve of C3S hydration in the presence of CEs measured by isothermal calorimetry, based on [5] of the CE is adsorbed on the surface of C3S, while the rest is dissolved within the pore solution [5]. The formation of a silicate-rich shell is normally believed to be attributed to the end of this period [14]. The slow reaction period is characterised by the slow dissolution of C3S, with small amounts of heat released and gradual increase in Ca(OH)2 concentration in the pore solution [21]. This period ends when CSH and portlandite start to precipitate. Muller [5] measured the heat of hydration and found that the addition of CE significantly increased the length of this period (Fig. 10). By monitoring the electrical conductivity of the C3S suspension, Pourchez et al. [21] showed that the addition of CE increases the time required and the electrical conductivity value at which portlandite starts to precipitate (Fig. 11). The possible mechanisms for this retardation will be discussed in Mechanisms of retardation effect of CEs on cement hydration section. The acceleration period is characterised by the rapid precipitation of CSH and portlandite CH [14]. CSH forms on the surface of C3S particles, while CH precipitates within the pore solution. The formation of hydrate shell around the C3S particles resists the attack of water on the unhydrated surface and inhibits the transportation of dissolved ions through this shell and that could be attributed to the end of acceleration period. Muller believed that the CE addition inhibits CSH and CH precipitation and growth. This is because the heat flow curves of C3S pastes modified with CE addition have lower slopes and lower peak heat values compared to the heat flow curve of the unmodified C3S paste, except for the paste modified using HEC (DS = 2.0) (Fig. 10). The rate of heat produced depends on the rate of reactions within the paste that is controlled by the rate of precipitation of CSH and CH [26, 27]. CE addition produces two simultaneous opposite effects on the rate of precipitation of CSH and CH, viz., Fig. 10 Influence of the type and DS of CEs on the heat flow of C3S hydration (water/ C3S ratio = 0.5 (wt/wt), CE/C3S ratio = 0.5 wt%) adapted from [5] Start point of CH precipitation C3S + HEC 1 Fig. 11 Electrical conductivity of C3S suspension, and C3S suspension with addition of HEC 1 (Mw = 175 kDa, MS = 2), (water/C3S ratio = 20 (wt/wt), HEC/C3S ratio = 2 wt%), adapted from [21] retardation and dispersion-stabilisation. The retardation effect of CEs inhibits the precipitation and growth of CS H and CH; therefore, it reduces the rate of precipitation and growth of CSH and CH. In contrast, CE addition provides dispersion and stabilisation effects that prevent the agglomeration of cement particles; therefore, more surfaces of cement particles participate in the reactions compared to the unmodified cement paste. This increases the rate of the precipitation and growth of CSH and CH within the sample. The heat flow reflects the overall effects of CEs on the rate of precipitation and growth of CSH and CH. Addition of HEC (DS = 2.0) reduces the rate of precipitation and growth of CSH and CH by retardation effect, but the effect is lower in extent compared to the increase in the rate of precipitation and growth of CSH and CH by dispersion and stabilisation effect; thus, the heat curve is characterised by higher slope and higher peak heat value compared to the unmodified paste. In contrast, addition of MHPC (DS = 1.7), MHEC (DS = 1.5), or HEC (DS = 1.1) reduces the rate of precipitation and growth of CSH and CH by retardation effect more than the increase in the rate of precipitation and growth of CSH and CH by dispersion and stabilisation effect, and the heat curves are characterised by lower slopes and lower peak heat values. In the deceleration period, further formation of CSH on the surface of C3S particles results in a lower extent of reactions of C3S with water. The heat flow curves in the study by Muller [5] show the deceleration period was not affected by CE addition as the slope of the curves were similar to that of the unmodified C3S paste (Fig. 10). In summary, CE addition does not affect the initial reaction and deceleration periods, but extends the length of the slow reaction period and reduces the rate of precipitation of CH and CSH in the acceleration period. Hydration of C2S By measurement of the amount of free water in the C2S paste during curing, Ridi et al. [28] found that CE addition slightly retarded the hydration of C2S. The free water present in C2S paste modified by CE was higher than that of the unmodified C2S paste during the 365 days investigation. This indicates that CE addition reduces the dissolution of C2S and reduces the rates of formation of CSH and CH. Hydration of C3A C3A hydration in the absence of calcium sulphate In water: The study by Ridi et al. [28] shows that the hydration of C3A in water without gypsum can be divided into two stages, as described below: Stage 1: Formation of gel-like hydrates [observed by scanning electron microscopy (SEM)], which are believed to be metastable calcium aluminates, but may also consist of aluminium hydrates (Al(OH)3) Stage 2: Transformation of metastable forms into the stable form (cubic C3AH6) Ridi et al. [28] showed using thermogravimetric analysis-differential thermal analysis (TGA-DTA) and SEM that the addition of CE retards the formation of stable C3AH6 through the stabilisation of metastable forms. In lime water: Based on the electrical conductivity measurements of the C3A suspension containing HEC 1 Concentration of Al3+ Fig. 12 Electrical conductivity and concentration of Ca2? and Al3? ions in the C3A suspension with addition of HEC 1 (Mw = 175 kDa, MS = 2), (water/C3A ratio = 20 (wt/wt), HEC/C3A ratio = 2 wt%, initial [CH] = 20 mM), adapted from [20] (Mw = 175 kDa, MS = 2), Pourchez et al. [20] proposed that the hydration of C3A in lime water, in the absence of sources of calcium sulphate, is divided into two stages, the formation of C4AH13, followed by the progressive precipitation of C3AH6 (Fig. 12). Stage 1: This stage is characterised by a significant decrease in Ca2? concentration and a considerable increase in Al3? concentration that results in the reduction in the conductivity of the suspension (Fig. 12) [20]. Pourchez et al. proposed that there is a rapid dissolution of C3A and precipitation of calcium aluminate hydrates that have a calcium/aluminium ratio higher than 1.5 (i.e., Ca2?/Al3? ratio of dissolution of C3A). Although Pourchez did not specify the composition of the calcium aluminate hydrates, it can be assumed to be C4AH13 [11, 14]. Pourchez et al. [20] showed that CE modification reduces the dissolution rate of C3A and also inhibits the precipitation of C4AH13 as it decreases the rate of increase in Al3? concentration and lowers the rate of decrease in Ca2? concentration compared to the unmodified suspension of C3A (Fig. 13). The addition of CEs increases the concentration of Al3? at the end of this stage compared with the unmodified C3A suspension and that is believed to be due to inhibiting precipitation of C4AH13 by CEs [20]. Stage 2: This stage is characterised by a constant Ca2? concentration and a significant decrease in Al3? concentration that leads to a decrease in the conductivity of the suspension. Pourchez et al. proposed that there is a strong precipitation of C3AH6 in this stage. The addition of CEs results in a decrease in the rate of reduction in the concentration of Al3? ions, and this is Stage 2 Stage 3 Conversion of Formation C4AH13 into C3AH6 of C3AH6 Fig. 13 Conductivity and concentration of Ca2? and Al3? of C3A suspension (water/C3A ratio = 20, initial [CH] = 20 mM), adapted from [20] believed to be due to the retardation effect of CEs on the precipitation of C3AH6 [20]. However, this two-stage mechanism proposed by Pourchez et al., from the conductivity and concentrations of Ca2? and Al3? of C3A suspension modified by HEC 1 is not representative of the case for other CEs (Fig. 14; Table 2 clarifies details of the chemistry of CEs). The electrical conductivity curves of unmodified cement suspension and suspension with addition of HEC 2, MHPC 1, and MHPC 2 increase before becoming constant (Figs. 13, 14). This could be related to the conversion of C4AH13 into C3AH6. This conversion releases Ca2? back into the solution and this could be responsible for the increase in the conductivity of C3A suspension with the addition of HEC 2, MHPC 1 (illustrated in Fig. 14), and MHPC 2. The conversion of C4AH13 into C3AH6 with the increase in Ca2? is clearly shown in Stage 2 of Fig. 12 in the cement suspension without CE addition. The lower rate of increase in conductivity of the C3A suspension modified by CEs compared to the unmodified one indicates that CEs retard the transformation of C4AH13 into C3AH6. C3A hydration in the presence of calcium sulphate The early hydration of C3A in the presence of CaSO4, in both water and lime-saturated water, can be divided into three stages: (1) formation of first ettringite, (2) formation of second ettringite, and (3) formation of calcium aluminates (Fig. 15) [5, 20]. Stage 1: In this stage, there is a rapid formation of first ettringite [5], leading to the release of a significant amount of heat (Fig. 15). The amount of first ettringite depends on the amount of Ca2? and SO42- ions available in the pore solution. The hydration of C3A in the presence of calcium sulphate in saturated lime was Fig. 14 Influence of CE chemistry on the electrical conductivity of C3A suspension, in saturated lime water without calcium sulphate (relevant details of the chemistry of CEs in Table 2), CE/C3A ratio = 1.5 wt.%, water/C3A ratio = 20 (wt/wt), initial [CH] = 20 mM, adapted from [20] Table 2 Chemistry of CEs used in work by Pourchez et al. [7, 20, 21, 38, 39] Fig. 15 Typical curve showing the heat flow during hydration of [70 wt% C3A ? 15 wt% CSH0.5 ? 15 wt% CS] mixture with the addition of MHEC (DS = 1.5), adapted from [5] found to produce lower amounts of first ettringite than the hydration in tap water. This observation is explained through the reduction in the solubility of calcium sulphate sources (gypsum, hemihydrate, and Concentration Concentration of SO4 2- of Ca2+ Fig. 16 Electrical conductivity and concentration of Ca2?, Al3?, and SO42- ions in [C3A ? 2 wt% gypsum] suspension with addition of HEC 1 (Mw = 175 kDa, MS = 2), water/solid ratio = 20 (wt/wt), HEC 1/C3A ratio = 1.5 wt%, adapted from [21] anhydrite) [5]. The addition of CEs does not affect the formation of first ettringite. Due to the rapid formation of first ettringite, this stage may not be detected clearly in the study by Pourchez et al. [20]. The indicator for the first formation of ettringite could be the drop in Ca2? from the first point of measurement (approximate 15 mM) as it is lower than the Ca2? of the initial lime water employed (20 mM) (Fig. 16). Stage 2: This stage is characterised by a gradual formation of second ettringite with lower rate of heat production/flow (Fig. 15) [5]. The CE addition is found to retard the precipitation of second ettringite as it increases the length of this period when compared to the unmodified paste (Fig. 17). This retardation effect of CEs is also confirmed by Pourchez et al. [20] when the time for the consumption of all the SO42- ions (i.e., the length of stage 2, formation of second ettringite) is increased by CE addition (Figs. 16, 18). Stage 3: This stage is still a controversial issue. From the heat flow curve, Muller [5] proposed that the peak heat of stage 3 (Fig. 15) is due to the rapid formation of second ettringite. However, from the conductivity measurements, Pourchez et al. [20] proposed that stage 3 (Figs. 15, 16) is due to the formation of calcium aluminate hydrates (C4AH13 and C3AH6). This difference can be due to the use of different techniques and periods of time during that the hydration was investigated. Stage 2 in Fig. 16 (formation of second ettringite) in the work by Pourchez et al. consists of both stage 2 and 3 (slow and rapid formation of second ettringite) in Fig. 15 in the work by Muller. Stage 3 in Fig. 16 (formation of C4AH13 and C3AH6) in the work by Fig. 17 Influence of the DS of CEs on the heat flow of hydration of [70 wt% C3A ? 15 wt% CSH0.5 ? 15 wt% CS] mixture, adapted from [5] HEC 1 [MS = 2.0, Mw = 175 kDa] HEC 2 [MS = 2.5, Mw = 175 kDa] HEC 3 [MS = 2.5, Mw = 1335 kDa] Fig. 18 Electrical conductivity of C3A suspension with and without the addition of CEs in the presence of gypsum, water/solid ratio = 20 (wt/wt), CE/C3A ratio = 1.5 wt%, adapted from [20] Pourchez et al. was not investigated in the work by Muller. Although there is a difference in the products formed in this stage between these two studies, both studies show that the addition of CEs retards the formation of the hydrates. The peak heats of the pastes with CE addition are broader and lower than the ones without CE (Fig. 17). The rates of reduction in the conductivity of the suspension with CE addition are lower than that of the suspension without CEs (Fig. 18). Hydration of C4AF The study by Ridi et al. [28] shows that the hydration of C4AF in water in the presence of CEs produces firstly featureless, gel-like hydrates (possibly a mixture of C2(A,F)H8, C4(A,F)H13 and (A,F)H3 [11]) that then transforms into a more stable product comprised of cubic C3(A,F)H6. Although this mechanism is similar to the hydration of C4AF without CE addition, TGA-DTA and SEM analyses showed that the addition of CEs retarded the transformation of the gel-like products into C3(A,F)H6 [28]. Hydration of Portland cement From the heat flow curve, the hydration of Portland cement in the presence of CEs can be divided into four stages (Fig. 19): Stage 1: initial reaction period This stage is characterised by the significant production of heat [5, 6, 29]. The first ettringite is produced from the reaction between C3A and calcium sulphate dissolved into the pore solution [5]. The dissolution of C3S and free lime also occurs in this stage, and this increases the conductivity of the solution (Fig. 20) [6, 30]. A CSH gel layer is formed around the cement particles [30]. CE addition does not affect either the formation of first ettringite or the dissolution of C3S. The heat flow curve (Fig. 19), the amount of first ettringite formed, and the electrical conductivity curve (Fig. 20) of the CE-modified paste were found to be similar to those of the unmodified paste [5, 6, 30]. Stage 2: slow reaction period During this stage, very limited amounts of heat are produced compared to the first stage [5, 6, 29]. The second ettringite is gradually formed, while C3S dissolves slowly, gradually forming the CSH gel layer around the cement particles [5, 30]. CE addition is found to retard these processes, leading to an increase in the length of this stage compared to the case with the unmodified paste. By measurement of the heat of hydration, Muller [5], Knapen [6], and Betioli et al. [29] found that the time period of low heat production corresponding to this stage is extended by the addition of CEs (Fig. 19). Peschard et al. [30] and Pourchez et al. [7] monitored the conductivity of cement suspensions and found that the CE addition reduces the slope of the conductivity curve for this stage and increases the time required to reach the level where the precipitation of CH starts (Fig. 20). This observation is another indicator of the retardation effect of CEs on the dissolution of C3S and formation of CSH gel. Ohama [31], Khayat [32], Singh et al. [24], Paiva et al. [33], Betioli et al. [29], Lazau et al. [34] found that the setting time of the cement paste is extended by the addition of CEs. The setting time has a close correlation with the time at which the acceleration period begins. As the addition of CEs does not affect the initial reaction period, the increase in the setting time indirectly indicates an increase in the period of time for the slow reaction. Muller [5] used XRD analysis to study the development of cement hydrates in the hydration of Portland cement and found a retardation effect resulting from CE addition on the formation of second ettringite. Knapen [6] used TGA to measure the CH content in the cement paste during the early age and found that CE addition increases the time at which the CH content increases significantly (i.e. increases the time at which the acceleration period starts). Stage 3: acceleration period This stage is characterised by a sharp increase in the heat generated during cement Start point of CH precipitation Fig. 20 Influence of CE on the electrical conductivity of Portland cement suspension, CE is a mixture of MHEC and MHPC, Mw = 11600 kDa, w/c ratio = 8000, CE/cement = 0.5 wt%, adapted from [30] hydration [5, 6, 29]. This exothermal heat results mainly from the significant precipitation of CSH and CH from C3S hydration [6]. During this stage, second ettringite continues to form gradually. CE addition is found to inhibit the formation of CSH, CH, and second ettringite [5, 6, 29]. The measurement of heat flow from Portland cement hydration by Muller [5], Knapen [6], and Betioli et al. [29] shows that the rate of heat produced is reduced by the addition of CEs compared to the unmodified paste. This indicates that CE addition reduces the rate of precipitation of CSH and CH in the acceleration period. The TGA results in the study of Knapen [6] show that the rate of CH formation is retarded by CE addition. Similar to the case of stage 2 (slow reaction period), by quantitative XRD measurements, Muller [5] observed the retarding effect of CEs on the formation of second ettringite. Stage 4: deceleration period This stage is characterised by the reduction in the heat generation from Portland cement hydration [5, 6, 29]. The hydration of C3S to form CSH and CH is inhibited by the layer of cement hydrates formed around the cement particles. While the hydration of C3S is clear, the hydration of C3A in this stage is still a controversial issue. Synchrotron XRD results by Muller [5] show that there appears a formation of second ettringite, while the conventional XRD results show there is a transformation of ettringite into monosulphate. Similar controversial results were reported by Knapen [6]. The formation of calcium aluminates (C4AH13 and C3AH6) from the hydration of C3A is proposed to occur in this stage. The hydration of C4AF and C2S occurs at a late stage [6]. The addition of CEs is found to retard cement hydration in the deceleration period [6, 25, 34]. TGA analysis by Knapen [6], Ma et al. [25], Lazau et al. [34] shows that in the deceleration period, the CH content in the cement paste modified by CE is smaller than the unmodified one. However, this information provides only a general degree of hydration of the C3S and C2S, but does not give a clear indication of the effect due to hydration process of each phase. There does not appear to be any published study that focuses on the effect of CEs on the hydration of C4AF, on the formation of monosulphate and C4AH13, and on C3AH6 formation during hydration of Portland cement. Mechanisms of retardation effect of CEs on cement hydration Several theories have been proposed to explain the retardation effect of CE on cement hydration. These mechanisms are discussed and categorised on the basis of the presence of CE within the mortar: (a) the CE is adsorbed on the surface of cement particles, (b) the CE is adsorbed on hydrates, and (c) the CE is dissolved within the pore solution. Retardation induced by CE adsorption on cement particles The greater the CE adsorbed on cement, the greater its impact on the retardation of cement hydration [5]. The adsorbed CE layer can retard hydration of cement through different means. Firstly, the CE layer, due to the presence of (OH) groups, may hinder contact between the unhydrated surface of cement and water, thereby retarding cement hydration [24]. Moreover, Pourchez et al. [21] stated that the adsorbed CE retards cement hydration by lowering the number of C SH seeds formed on the surface of cement particles. This reduction is believed to be the reason for the reduction in the dissolution rate of C3S during the slow reaction period, which was in turn found to be dependent on the number of CSH nuclei. The decrease in the dissolution rate of C3S results in an increase in the time needed for the concentration of ions to reach the saturation level required for the growth of hydrates. This mechanism related to the retarding effect of CE has been supported by Ma et al. [25]. Forming a basis from the study by Thomas et al. [35] on the retardation of sugar, Knapen [6] proposed that the adsorbed CE may retard cement hydration through the formation of insoluble complexes between CE and the alkaline ions (ie. Ca2?, K?, and Na?). These insoluble complexes form a coating on the surface of the cement particles, thus hindering the water transport through this layer. Knapen [6] also proposed that the CE adsorbed on the surface of cement particles could retard cement hydration through its incorporation into the protective membrane that is rapidly formed on the surface of the cement particles as soon as these particles come into contact with water. This protective membrane is referred to as the metastable CS H layer as discussed by Bullard et al. [14]. Knapen [6] believed that the incorporation of CE into the protective membrane may reduce the permeability of this membrane; consequently, this inhibits the generation of hydrostatic pressure (by osmosis) required to rupture the protective membrane, which is needed for the initiation of the growth of hydrates. Retardation induced by CE adsorption on hydrates The adsorption of CEs on cement hydrates may retard the cement hydration through the inhibition of the growth of these hydrates [6, 21, 24]. Singh et al. [24] proposed that the adsorption of CE on portlandite (CH) is responsible for the retardation effect on the hydration of C3S. Knapen et al. [22] believed that the adsorption of CEs on CSH and CH nucleation sites retards the growth of these hydrates. Retardation induced by CE dissolved in pore solution The dissolved part of CE may form complexes with metal ions (such as Ca, Al, Fe), resulting in an increase to a saturation level of these ions in the pore solution [22]. The higher the saturation level, required for the precipitation of hydrates, the longer the time required for this to occur, and thus, the cement hydration is retarded. The application of this mechanism to Ca2? was doubted by Pourchez et al. [20] since the presence of complexes between CE and Ca2? was negligible. Moreover, the concentration of CE that is conventionally used is believed to be significantly lower than that required to induce the formation of considerable amounts of these complexes with Ca2? [20]. Therefore, the complexes between CE and Ca2? may not be responsible for the retarding effect of CE. The formation of complexes between CE and aluminates, silicates, or ferrites is believed to be more responsible for the retardation of CE, since the concentrations of these metal ions are significantly lower than the Ca2? levels [20, 25]. However, it is still controversial, since Pourchez et al [21]. monitored the electrical conductivity of the C3S suspension and noted an increase in the electrical conductivity before a massive amount of CH was formed in all CEmodified samples compared to the unmodified one. Furthermore, it may be noted that in the method used to study the complexation between CE and Ca2? by Pourchez et al [36], the CE may have been adsorbed on the excess CH particles [5], and this may significantly reduce the amount of CE in the solution that is available to form complexes with Ca2?; this could be responsible for the absence of any significant changes in the detected Ca2? concentration. The dissolved CE in the pore solution normally results in an increase in the viscosity of the solution that may decrease the dissolution rate of unhydrated cement [6, 22, 37]. Consequently, this retards cement hydration. However, Pourchez et al. [7] raised doubts on this mechanism since the addition of extra amount of water did not induce any considerable change in the C3S dissolution rate. Pourchez et al. [36] found that the degradation of CEs under the alkaline conditions of the pore solution was not responsible for the retardation of cement hydration, since the observed degradation levels were insignificant [5, 36]. Factors affecting retardation effect of CEs Concentration of CEs The increase in CE content enhances the retardation of cement hydration [24, 29, 38]. By monitoring the electrical conductivity of cement suspensions, Pourchez et al. [38] found that the increase in CE content (CE/cement) from 0.5 to 2.0 wt% resulted in an increase in the time taken for the precipitation of CH to occur. Further details of the correlation between the electrical conductivity of the cement suspension and the precipitation of CH can be found elsewhere [7]. Singh et al. [24] found that the higher the CE dosage, the longer the setting time of the mortar. From the heat flow curves of cement pastes, Betioli et al. [29] found the increase in CE dosage from 0.25 to 0.50 wt% caused an extension to the length of the [initial reaction ? slow reaction] period from 210 to 285 min, respectively. The increase in CE content also caused a greater decrease in the rate of heat evolved during the acceleration period, as well as the peak heat value. Chemical nature of CEs Type of CE Muller [5] used XRD technique to study the hydration of Portland cement and found that MHPCs have a more retarding effect. By monitoring the conductivity of the cement suspension, Pourchez et al. [7, 38] found that HECs have a greater retarding effect compared to MHPCs. The time at which the precipitation takes place in the sample modified by HECs was longer than the case of the precipitations modified by MHPCs (Fig. 21). Pourchez et al. [20] studied the influence of CE chemistry on the retardation of C3A hydration and found that HECs generally retarded C3A hydration to a greater extent compared to MHPCs (Figs. 14, 17). Knapen et al. [22] measured the heat flow of cement hydration and concluded that MC has a lower retarding effect compared to HEC: this was based on the observation that the length of slow reaction period for the paste using MC was much shorter, and peak heat was HEC 1 [MS = 2.0, Mw = 175 kDa] HEC 2 [MS = 2.5, Mw = 175 kDa] HEC 3 [MS = 2.5, Mw = 1335 kDa] Fig. 21 Influence of MS and Mw of HECs on the time at which CH starts to precipitate in Portland cement suspension in saturated lime water, detected by monitoring conductivity (w/c ratio = 20 (wt/wt), CE/cement ratio = 2 wt%, initial [CH] = 20 mM), adapted from [21] also sharper compared to the paste modified by HEC (Fig. 19). It should be noted that the comparison between CE types is not very accurate since different chemical parameters (DS, MS, distribution of substitutions) may have differing impacts. For example, Muller [5] studied the effect of the DS on the retarding effect of CEs and found that HEC with DS = 2.0 had a lower retarding impact compared with MHECs (DS = 1.5) and MHPCs (DS = 1.7). Degree of substitution The lower the DS of CE, the greater the retardation impact on cement hydration [5, 23]. Through the measurement of the heat of hydration of Portland cement, Ou et al. [23] found that MC with DS = 1.7 extended the slow reaction period to a greater extent than MC with DS = 1.8. By using calorimetry and conventional XRD analysis, Muller [5] found that HEC with DS = 1.1 extended the slow reaction period to a greater extent, and slowed down the consumption of C3S, and gypsum and the formation of ettringite and CH to a greater extent compared to HEC with DS = 2.0. Results from synchrotron XRD analysis showed that MHEC (DS = 1.5) and MHPC (DS = 1.9) inhibited the consumption of C3S, gypsum and the formation of ettringite and CH to a greater extent than MHEC (DS = 1.7) and MHPC (DS = 1.9), respectively. However, it should be noted that Muller did not provide any information related to the MS and Mw of the considered HECs; these factors could also have an impact on the results. Molar substitution The lower the MS of CE, the greater the retardation effect of the CE on cement hydration [7, 23, 38]. The retarding effects of MC and HEC are dependent on the MS of (OCH3) groups and (C2H4O) groups, respectively [7, 23]. Pourchez et al. [7] measured the electrical conductivity of the cement suspension and found that the decrease in MS of HEC from 2.5 (HEC 2) to 2 (HEC 1) increased the time taken for the precipitation of CH to occur (Fig. 21). MHEC and MHPC have more complicated chemical structures than MC and HEC, with two types of substitutions. Through the measurement of electrical conductivity of the cement suspensions, Pourchez et al. [38] found that the MS of (OCH3) group is the key parameter in the case of MHEC and MHPC with regard to the retardation effect, rather than the MS of (C2H4O) groups and (C3H6O) groups in MHEC and MHPC, respectively. With approximately similar MS of (C2H4 O) groups and (C3H6Ogroups in MHEC and MHPC respectively, the heat measurement results by Ou et al. [23] showed that the reduction in the MS of (CH3) led to an increase in the length of the slow reaction period. It is not clear whether the interactions between substitutions could affect those results; however, it should be noted that these CEs of the same type were from the same supplier and may have been fabricated through the same procedures, thus resulting in similar reactions between the substitutions. Molecular weight The effect of the Mw of CE is still ambiguous. Pourchez et al. [20] measured the electrical conductivity of C3A suspensions in the presence of HECs and found that lower Mw HEC produced a greater retardation effect on C3A hydration, inhibiting the formation of second ettringite and calcium aluminates. Through the use of the same technique, Pourchez et al. [7] found that longer times were required for initiating the precipitation of CH in the presence of lower Mw HEC compared to the higher Mw HEC. Ou et al. [23] using isothermal calorimetry found a greater extension of the slow reaction period in the paste modified using lower Mw HEC than the case of the higher Mw one. In contrast, Pourchez et al. [38] did not see any significant effect of the Mw of MHEC and MHPC on the time at which CH starts to precipitate. However, it should be noted that the very high water/cement ratio (w/c = 20) may not be appropriate for studying the effect of Mw of CEs on cement hydration as the increase in the Mw of CEs may not produce any significant increase in the viscosity of the pore solution. Stage 1: from mixing to the beginning of the acceleration period Development of cement hydrates As soon as contact with water occurs, CSH gel is formed on the surface of cement particles [6]. The first needleshaped ettringite crystals are formed on the surface of the cement particles. Although there are no studies that have focused on the formation of ettringite in the pore solution, there is a possibility of this phenomena occurring. The second ettringite formation gradually continues during this stage [5, 20]. Air entrapment During mixing, air is entrapped in the mortar. Entrapped air voids are clearly recognised by their irregular shape, which is distinct from the round shape of entrained air voids (Fig. 22). Two mechanisms have been proposed to explain air entrapment: vortex action and three-dimensional screen mechanisms [6]. During mixing, the vortex action leads to the infoldment of air that will then be dispersed and broken up into smaller bubbles by the shear force of the mixing Fig. 22 Development of pore structure in CE-modified mortar. a Resoluble air voids (i.e., voids are larger than resolution of the image) after mixing; b increase in size of resoluble voids with a constant number of resoluble air voids; c attachment between action. The three-dimensional screen is caused by the aggregates that infold air within their networks. Air entrainment and stabilisation The entrained air could be defined as the entrapped air that is further dispersed and broken down into smaller air voids during mixing [6]. The stabilisation effect of CE indicates their ability to mitigate the coalescence of entrained air voids and to prevent the large air voids from moving upwards to the surface and bursting [6, 39]. Knapen [6] proposed three mechanisms that may be attributed to air stabilisation by CE, including the formation of hydration sheet, decrease in surface tension, and viscosity enhancement. Due to high affinity to the airliquid interface as CEs contain both hydrophobic groups (such as CH3 which orient toward the air) and hydrophilic groups (such as OH, which orient toward the liquid phase), CE is adsorbed at the surface of the air voids during mixing. The work by Jenni et al. [40] and Pourchez et al. [39] supports this mechanism. The attachment of CE at the air void surface was confirmed by Jenni et al. [40] as CE film at the surface of the void was clearly observed in the environmental scanning electron microscopy (ESEM) image of a freeze-dried sample. The attached CE molecules can adsorb a certain amount of water and form hydration sheets. These hydration sheets are thought to be capable of separating air voids from each other and stabilising them. There are no studies that clearly identify how the cellulose chains and substitutions are oriented at the surface of air voids. This orientation could be the key to understanding the influence of CE chemistry on their air void stabilisation. The air void is also stabilised by the reduction in surface tension and surface elasticity resulting from CE addition [6]. Moreover, the viscosity enhancement inhibits the upward movement of air voids to the surface and acts as a barrier to prevent the coalescence of adjacent voids [6, 41]. resoluble air voids; d coalescence of adjacent air voids with a decrease in the number of resoluble voids; e formation of a 3D network of capillary pores, adapted from [39] The entrained air after mixing inherently undergoes destabilisation that results in an increase in void size during ageing time. The destabilisation tends to occur in order to reduce the total energy of the system due to the reduction in the surface energy between the air and the bulk phases (solid and liquid) [41]. From the results obtained from synchrotron X-ray tomography, Pourchez et al. [39] proposed two mechanisms for the coalescence of air voids. Mechanism 1: This mechanism is characterised by the increase in pore size without an increase in the number of pores (Fig. 22b) [39]. Pourchez et al. [39] believed that this is attributed to the coalescence between a bubble (around 50200 lm) and many air voids (smaller than 3 lm) that are not detected due to the limit of the resolution (2.8 lm) of the obtained images. Mechanism 2: The second mechanism is characterised by the reduction in the number of bubbles due to bubble coalescence (Fig. 22d) [39]. Due to the void coalescence during curing, the proportion of the large voids increases, while the proportion of the small voids decreases. These results appear to contradict the study by Alesiani et al. [42] where a shift in the pore size in the C3S paste modified by 0.27 wt% MHEC occurred from 140160 lm in the first hour to 100120 lm. Microstructural evolution across the thickness of the mortar layer After placement of the mortar layer on the substrate, the CE-modified mortar experiences migration of CE, ions of cement dissolution [mainly Ca(OH)2], and fine cement particles from the body of the mortar toward the exposed surface and to the interfacial layer between the mortar and the substrate [40, 4345]. This migration of CE, Ca(OH)2, and fine cement particles toward the exposed surface of the mortar results in the formation of a skin [40, 4345]. Skinning is critical to the performance of tile adhesives as the formation of a skin during the open time period (from time of placement of mortar to time of tiling) reduces wettability and adhesion strength of the mortar to the tile, thereby resulting in the detachment of tiles [40, 43, 44]. The enrichment by CE, Ca(OH)2, and fine cement particles at the exposed surface of the mortar is due to the migration of water molecules from the body of the mortar toward the surface of the mortar due to evaporation. This water movement transports dissolved CE and Ca(OH)2 from the body of the mortar toward the surface of the mortar, thereby increasing the concentration of CE and Ca(OH)2 at the surface of the mortar. The Ca(OH)2 at the surface subsequently undergoes carbonation to form CaCO3 [44, 45]. Although the viscosity of pore water containing dissolved CE is considered to facilitate the transport fine cement particles toward the surface of the mortar [43], actually, it is the surface tension and associated wetting that is responsible. When the rate of water transport to the surface is lower than the rate of evaporation, drying and densification at the surface occur and the skin starts to build up [44, 45]. The skin consists of the CE film, CaCO3, cement particles, and cement hydrates; the thickness of the skin increases with time. As the rate of evaporation and water transportation strongly depend on the concentration of CE (CE/water) and the chemistry of CE, the selection of these parameters of CE is critical to control skin formation in order to obtain the desirable wettability to the tile (i.e., open time) [43, 44]. Bentz et al. [43] found that an increase in CE content decreases the open time and that this occurs due to the reduction in the rate of water transport from the body of the mortar to the surface [44] owing to the impact of CE content on increasing the water retention of the mortar [46]. There do not appear to be any published studies on the influence of the chemistry of CE on the skinning of the mortar. The enrichment of both CE and Ca(OH)2 concentrations at the interface between the mortar and the substrate is due to water migration from the body of the mortar to this interface [40, 45]. This water migration is due to the greater water absorption by the substrate. This water movement transports dissolved CE and Ca(OH)2 from the body of the mortar toward the surface of the mortar, thereby resulting in an increase in the concentrations of CE and Ca(OH)2. The water movement also may cause transport of fine cement particles toward the surface of the mortar owing to the surface tension of the pore water containing dissolved CE [43]. It is believed that the addition of CE inhibits the enrichment of Ca(OH)2 at the exposed surface and at the interfacial layer between the mortar and the substrate. This would have the effect of giving a more even distribution of the mortar components across the thickness of the mortar layer [43]. This phenomenon would result from the water retention capacity of CE and the resultant inhibition of the water migration toward the surface and the substrate, thereby reducing the transportation of Ca(OH)2. Reduction in the enrichment of Ca(OH)2 at the surface layer lowers the bloom and subsequent efflorescence, which generally have an undesirable aesthetic effect on cement renders, skim coats, and self-levelling mortars. It is probable that increasing the CE concentration would cause a proportional reduction in the rate of enrichment of Ca(OH)2 at the surface and at the interfacial layer between the mortar and the substrate. However, at present, there do not appear to be any published studies on the effects of CE chemistry or concentration on the distribution of components of cement mortar across the thickness of the mortar layer. Stage 2: from the beginning of the acceleration period Growth of cement hydrates The amount of CSH and CH starts to increase significantly during the acceleration period and slows down in the deceleration period [6]. The second needle-shaped ettringite continues to precipitate and then transforms into hexagonal monosulphate when all calcium sulphate is consumed. The formation of hexagonal product of AFm phases (e.g., C4(A,F)H13) also occur at this stage. The SEM images show that addition of CE promotes inner CSH formation with a compact morphology rather than a fibril shape in unmodified mortars [28, 37, 42]. Addition of CE promotes the formation of inner products of C3S hydration, but does not produce a similar effect on C3A hydration. Integration of CE and cement hydrates There exists integration between cement hydrates and CE in the mortar where the CE chains and CE film are present within the microstructure of hydrates (Fig. 23). Singh et al. [24] believed that there are chemical interactions between Fig. 23 Incorporation of CE in cement matrix in CE-modified mortar, based on [6] hydrates and polar functional groups of CE to form amorphous materials that are scattered between the crystalline masses of cement paste. Knapen [6] observed polymer bridges between the layered CH crystals that glue the layers together and strengthen the microstructure. By using solid nuclear magnetic resonance (NMR) technique, Rottstegge et al. [47] found the immobility of CE in the hardened mortar, which suggested the incorporation of CE within the cement hydrates. The progressive adsorption of CE onto the cement hydrates when hydration proceeds may suggest the incorporation of CE into these cement hydrates [5, 20, 21]. The CE that is not incorporated into the structure of cement hydrates tends to form a film [5, 6, 40]. The formation of CE film is stimulated by the water consumption by cement hydration, water loss due to evaporation, and adsorption by the substrate [40]. This reduction in water enhances the entanglement of the CE chains, followed by gel formation, leading to the formation of a film. The CE film can appear in the void, at the surface of the void, and within the cement matrix at a submicron scale. Jenni et al. [40] observed the detachment of CE from the air void interface, resulting in the formation of sail-like structures in the void. The amount of CE that can be incorporated into the cement hydrate structure depends on the CE chemistry. In the study by Knapen [6], with the same polymer/cement ratio of 1 wt%, a film was clearly observed in the mortar modified by MC but not in the one modified by HEC. The difference in the ability of CE to be incorporated into cement hydrates can be explained in terms of the difference in their affinity to be adsorbed onto these hydrates. The integration of CE within the cement matrix is believed to be responsible for the reduction in cracking of the mortar due to the improvement of cohesiveness [5, 6]. Changes to the pore structure Similar to unmodified mortar, in the CE-modified mortar the water consumed by cement hydration results in capillary pores within the mortar structure [6]. The voids are connected through these capillary pores (Fig. 22e) [6, 21, 40]. The void instability continues to occur until the cement sets [48]. Besides the driving force due to the reduction in surface energy, Bentz et al. [48] proposed that the void instability is attributed to the movement of fine cement particles. This theory is based on the observation by X-ray tomography when the void size was stable during the first 24 h of measurement in mortar using coarse cement particles, while the instability of void size was found in the one using fine cement particles. In general, CE addition increases both capillary pores (100500 nm) and air voids (50250 lm) [39, 49]. Silva et al. [49] explained the increase in the amount of capillary pores in terms of the evaporation of excess water, while the increase in the amount of air voids is due to the air entrainment effect of CE. Very few studies have focussed on the influence of parameters, such as the concentration, the chemistry of CEs on the air entrainment and stabilisation behaviour. CE concentration: The increase in the CE concentration normally increases the air content and increases the size of the voids [41, 49]. This contradicts the results of Knapen [6] when the increase in water ratio led to an increase in the total porosity. This could be due to the lack of free water in the mortar in the study by Knapen [6] before more water was added, which resulted in micelle formation of CE chains [41]. The micelles do not contribute to air entrainment unless sufficient water is added to dissolve them within the pore solution. Type: Pourchez et al. [39] found that HPMC generally provides better air entrainment and void stabilisation than HEC. MS: Through 2D image analysis, the results in the study of Pourchez et al. [39] show that HEC with lower MS (MS = 2) showed greater numbers of air voids than the case of HEC with higher MS (MS = 2.5). Mw: The higher the molecular weight, the more the air content of the mortar [21]. There are no studies that directly compare the influence of DS of CE on their air entrainment behaviour. Microstructural evolution across the thickness of the mortar layer During the acceleration and deceleration periods, the enrichment of CE and Ca(OH)2 continues to occur at the drying front, so the drying front moves toward the body of the mortar, thus increasing the thickness of the skin [43]. The rate of enrichment decreases as the amount of free water reduces owing to water loss by evaporation, absorption by the substrate, and cement hydration. Also, the increasing thicknesses of the skin layer and that between mortar and substrate inhibit the evaporation and water absorption by the substrate. The addition of CE inhibits the enrichment of Ca(OH)2 at the surface and at the interfacial layer between the mortar and the substrate during this stage [44]. The increase in the CE concentration also can be expected to reduce the rate of enrichment of Ca(OH)2 at the surface layer and at the interfacial layer between the mortar and the substrate. This mechanism is discussed in more detail in Microstructural evolution across the thickness of the mortar layer section in Stage 1. There do not appear to be any published studies on the effects of CE chemistry or concentration on the development of the skin layer and the interfacial layer between the mortar and the substrate during this stage. Homogeneity of microstructure The microstructural homogeneity of the mortar is significantly enhanced by CE modification [6, 22]. By SEM, Knapen [6] observed the presence of well-distributed unhydrated cement particles that could be attributed to the steric stabilisation of CE. A considerable reduction in the porosity at the interfacial transient zone (ITZ) between the cement matrix and the surface of aggregates could be explained by the increase in the viscosity of the pore solution that inhibits the bleeding effect. Muller [5] and Knapen [6] found that a reduction in crack formation occurs within the mortar modified by CEs (Fig. 24). This enhancement was attributed to the increase in the cohesiveness between cement hydrates due to the integration with CE and CE film formation. Microstructural changes during moist and wet storage In many applications, the hardened mortars are subjected to moist (water vapour) or wet (liquid) conditions during their lifetimes. The render, masonry mortar, tile adhesive, repair mortars, and self-levelling mortar that are used for outdoor applications normally undergo cyclic changes in the moisturedry and wetdry conditions due to the weather changes. Understanding the effects of these conditions on the microstructural changes is critical to explain their impacts on the properties of the mortar. Both moisture and wet conditions stimulate cement hydration [6, 50]. In the moisture-laden condition, the water is adsorbed into the CE film, causing swelling, but without any CE migration within the mortar [6]. In contrast, the dissolution of CE film into the absorbed water causes the migration of CE within the mortar during wet dry cycles [50]. This may result in a distribution gradient of CE within the mortar. Jenni et al. [50] found an enrichment Fig. 24 a Reduction in crack formation in HEC-modified mortar compared with b unmodified mortar, adapted from [6] Fig. 25 Migration of polymer film after wet storage in mortar modified using 1 wt% MC (MC/cement): a bridges of MC film between ettringite crystals, b bridges of MC film between CH crystals, adapted from [62] Shear rate (s-1) Fig. 26 Shear stress versus shear rate dependency of CE-modified mortars, based on [32, 53] Increase in Mw of CEs Reduction in CE concentration if lower than the critical value Increase in CE concentration if higher than critical value Increase in Mw of CEs Reduction in CE concentration if lower than the critical value Increase in CE concentration if higher than critical value of CE content toward the surface of the substrates after wetdry cycles. The migration of CE occurs within the CE film, but not with the CE that is incorporated within the cement hydrates [6] (Fig. 25). Influence of CEs on the properties of cement mortar Rheological behaviour Rheology is the term used to describe the deformation and flow behaviour of the materials [1618]. Cement mortars exhibit viscoelastic behaviour, possessing the properties of solids and liquids [16, 17]. The parameters to consider with regard to the rheology of cement mortars are as follows: the yield stress value, shear stressshear rate dependency, viscosityshear rate dependency, and changes of these behaviours with time. Yield stress is defined as the minimum shear stress value that is to be applied to the mortar in order to initiate flow (i.e., shear stress value at shear rate of zero) (Fig. 26). High yield stress value is preferred for sag resistance of the mortar in cement render and for slip and slump resistance of tile adhesives. In contrast, low yield stress value is critical for self-levelling and self-compacting mortars which require the initiation of flow by itself (under gravity). The shear stressshear rate and viscosityshear rate dependency show the spreadability of the mortar. In Figs. 26 and 27, mortar 2 is considered to be harder to spread than mortar 1 as mortar 2 requires a higher shear stress (or has a higher viscosity) at a specified shear rate compared with mortar 1. The render and tile adhesives should be easy to spread on the wall or floor without the need of high forces applied by the worker. The viscosity of the mortar is important to the cohesiveness of components within the mortar, and this inhibits their separation [4]. Due to the complicated requirements of rheological properties for different mortars, the understanding of how CE influences the rheology is critical for its successful employment. The rheology of cement mortars is strongly dependent on the interactions between components within the mortar as summarised by Roussel et al. [51], and these interactions include surface forces (or colloidal interactions), Brownian forces, hydrodynamic forces, and contact forces between particles. The surface forces include van der Waals forces, electrostatic forces, and steric forces [51]. van der Waals forces act as attractive forces that enhance the flocculation of cement particles, and this inhibits the motion of cement particles. The electrostatic force arises from the charged surface of cement particles. If two cement particles have the same type of charge (i.e., negative or positive), the electrostatic force acts as repulsive force that enhances the Shear rate (s-1) Fig. 27 Viscosity versus shear rate dependency of CE-modified mortars, based on [32, 53] dispersion of cement particles. In contrast, if two cement particles have opposite charges, the electrostatic force between them acts as an attractive force that enhances the flocculation of these cement particles. The steric force results from CE adsorbed on the surface of cement particles, and this steric force always acts as a repulsive force to enhance the dispersion of cement particles. The Brownian force is important for small particles that may move through the suspension liquid by Brownian motions. It is believed that Brownian forces possess insignificant position among other interactions in normal cement system [51]. The hydrodynamic forces originate from the viscosity of the pore solution and that inhibit the movement of cement particles within the mortar [52]. The contact forces between the particles result from their friction during motion; therefore, these forces prevent the movement of cement particles. Due to the complexity of the system, current studies have only covered part of those interactions to explain the influence of CE on the rheological behaviour of the mortars. Effect of CE content on viscosity of the mortar The addition of CE increases the viscosity of the pore solution (Fig. 28) [32, 53]. The mechanisms of this viscosity enhancement were reviewed by Khayat et al. [32] and Paiva et al. [53]. The increase in viscosity of the solution is due to the adsorption of water molecules on the chains of CE. The attraction between adjacent CE molecules that block water mobility also results in an increase in the viscosity of the pore solution. At sufficiently high CE concentrations, the occurrence of intertwinement and entanglement between CE chains leads to a further viscosity enhancement of the pore solution. The viscosity enhancement effect does not only depend on the CE content, but also depend on the chemistry of CE. The increase in CE content increases the viscosity of the pore solution as a result of the increase in the degree of intraactions, interactions, and entanglement of CE chains [32, 53, 54]. There appear to be no published studies that have focussed on the influence of CE chemistry on the viscosity enhancement of the pore solution. The increase in the viscosity of the pore water results in an increase in the hydrodynamic force that reduces the motion of cement particles. However, hydrodynamic force is only one of many other forces that control the rheological behaviour of the mortar, and the increase in the viscosity of the pore solution by CE does not always result in the increase in the viscosity of the mortar. This is dependent on the CE content [53]. At the critical dosage of CE, the viscosity of the mortar reaches a minimal value. When the concentration falls below the critical value, the increase in CE concentration results in a decrease in the viscosity of the mortar. This could be explained in terms of the dispersing effect of CE. The dispersion and stabilisation of cement particles in cement systems can be explained according to extended DLVO theory (Fig. 29) [55]. Unmodified mortars inherently experience significant flocculation due to the dominant effect of the van der Waals attractive forces compared to the electrostatic repulsive forces (steric repulsive energy does not exist) [51, 55, 56] (Fig. 30). The addition of CE produces a steric hindrance that results from the CE adsorbed on the surface of cement particles [4, 53]. This steric repulsive force prevents the flocculation of the cement particles, therefore reducing the friction (i.e., contact forces) between them (Fig. 29). In this case, the reduction in viscosity of the mortar by the dispersing effect is dominant compared to the increase in the viscosity of the pore solution, and this impacts the measured viscosity of the mortar. When the concentration is higher than the critical value, the increase in CE concentration leads to an increase in the viscosity of the mortar [53]. The increase in the viscosity of the pore solution by an increase in CE concentration dominates any further viscosity reduction due to the dispersion effect of CE addition. The CE-modified mortar shows shear thinning behaviour, which implies that the viscosity of the mortar reduces as the shear rate increases (Fig. 31) [32]. This behaviour Electrosteric repulsive energy Van der Waals attractive energy Distance between particles CE concentration Fig. 28 Generic figure showing the influence of CE concentration on the viscosity of pore solution and on the viscosity/yield stress of cement mortars, based on [32, 53] Fig. 29 Interparticle forces according to extended DLVO theory in CE-modified mortars, adapted from [55] Cement Van der Waals attractive energy Fig. 30 Interparticle forces according to DLVO theory in cement mortars, adapted from [56] could arise from two phenomena; the first is the structural breakdown or deflocculation under shear stress (Fig. 31). From a practical point of view, cement mortars normally possess a certain degree of flocculation. The flocculation is reduced during the increase in the shear rate. Secondly, the reduction in the viscosity when the shear rate increases is characteristic of the pore solution with dissolved CE (Fig. 4) [32]. This behaviour of CE solution is due to a decrease in the degree of entanglement and association between adjacent polymer chains with increasing shear Cement particle Hydrated shell Shear thinning Structural break down Reduction in viscosity due to structural break down Thixotropy Shear rate (s-1) rate. Owing to this shear thinning behaviour of the mortar, the increase in CE dosage generally produces an increase in the viscosity at low shear rates which is more pronounced than the increase at high shear rates. Similar to viscosity, the effect of CE on the yield stress of the mortar is also dependent on the CE content [53]. At the critical content of CE, the yield stress reaches a minimum value. When the CE content is lower than the critical value, increase in the CE dosage results in a reduction in the yield stress. In contrast, when the concentration is higher than the critical value, the increase in CE content increases the yield stress. Similar explanations for the impact of CE content on viscosity could be applied to explain the effect of CE dosage on yield stress. It should be noted that the critical CE content for yield stress does not necessarily need to be equal to the critical CE dosage for the viscosity. CE-modified mortars experience hysteresis phenomenon (Fig. 31) [53]. At the same shear rate, the viscosity or shear rate of the up-curve which occurs due to the increases in the shear rate is normally higher than that of the downcurve which occurs with the decrease in the shear rate. This hysteresis effect is due to the structural breakdown which occurs under shear loading of the mortar that causes a reduction in the viscosity (Fig. 31) or shear stress. Effect of time The rheological behaviour of the CE-modified mortar changes with time (Fig. 32) [29]. The viscosity and yield stress of the mortars increase during curing. These increases are owing to the proceeding of hydration that results in a reduction in free water (i.e., an increase in solid/ water ratio), an increase in linkage of cement particles. The rate of change in the rheology of the mortar is critical as it affects the workability of the mortar. CE modification reduces the rate of increase in viscosity and yield stress of Unmodified Fig. 31 Rheological behaviour of CE-modified mortars, based on [17, 53] Fig. 32 Changes in rheology of unmodified and CE-modified mortar during curing, based on [29] the mortar, i.e., enhances the rheological consistency of the mortar. However, very few studies have focussed on the mechanisms related to this effect of CEs. Betioli et al. [29] correlated the changes in the rheological behaviour of the mortar modified by CEs with its hydration behaviour. The continuation of cement hydration increases the degree of flocculation of cement particles. In unmodified mortars, the flocculation of cement particles is due to the weak electrostatic forces that are produced by the double charge layer [56]. The increase in the ionic strength of pore solution further enhances the flocculation of cement particles. In CE-modified mortars, CE inhibits flocculation of cement particles by providing steric hindrance that arises from the CE adsorbed on cement particles. This steric hindrance is found to be less affected by the change in the ionic strength of the mortar. Moreover, the retardation of CE inhibits precipitation and growth of cement hydrates; therefore, CE slows down the process of formation of connections between the cement particles [29, 57]. Water retention Water retention is a critical property of cement mortars that tend to be subjected to evaporation and water absorption by the absorbent substrates, such as render, tile adhesives, self-levelling mortars, and repair mortars. The poor water retention of the mortars results in significant water loss due to evaporation and absorption by the tiles or substrates, and this causes improper hydration of cement, cracking, and poor adhesion between the mortar and the tiles or the substrates [1, 2]. The water loss could be compensated by the addition of excess water and by using thick layers of mortar; however, these two methods have disadvantages such as poor strength (due to increased pore content) or increased amount/costs of materials needed. Fig. 33 Influence of MS and Mw of HECs on the water retention of fresh mortars, adapted from [39] Addition of CE enhances significantly the water retention of the mortar (Fig. 33). The water retention is attributed to the non-adsorbed CE and not to the adsorbed part [46]. Several mechanisms have been proposed to explain the water retention effect of CEs. Increase in viscosity: Ohama [31] proposed that the viscosity improvement is responsible for the water retention. However, Patural et al. [58] disagreed with this and found that HECs increase significantly the water retention without causing any considerable increase in the viscosity of the mortar. Change in pore structure: Pourchez et al. [39] suggested that in the mortar modified by CE, air voids entrained into the mortar through mixing are connected by channels filled with pore water. The voids prevent water transport through them by means of slow vapour diffusion. Water adsorption: Bulichen et al. [46] first proposed water adsorption as one of the mechanisms for the water retention capability of the mortar. ESEM analyses showed that the CE chains adsorb water and swell under high humidity conditions; they also form a threedimensional network of gel. This adsorption of water on CE chains is in agreement with the study of Fringant et al. [59] for neutral polysaccharides; the hydrogen bonding between water molecules and (OH) groups of the polymer could be responsible for this effect. Besides (OH) groups, Feller et al [8] believed oxygen atoms also contribute to the water adsorption effect of CE. Formation of 3D hydrocolloidal polymer network: Each CE chain dissolved in aqueous solution possesses a space that is characterised by gyro radius [54]. Bulichen et al. [46] proposed that if the CE chain is of a sufficiently large size, it will be able to block the pores of the filter in the test method, thereby preventing water loss. The water retention of the mortar is dependent on the concentration and chemistry of CE. The increase in CE content generally increases the water retention capability of the mortar [46]. It was found that the increase in the CE Mw improves the water retention of the mortar (Fig. 33) [38, 39]. In contrast with the impact of the Mw of CE, the effect of other chemical parameters of CEs, such as type of substitution, DS, MS, and the distribution of substitutions, on water retention is still ambiguous. For the discussion provided in Water retention section, the complete dissolution of CE should be assumed. However, CE products are normally employed in powder form, and the rate of CE dissolution is critical to the water retention of the mortar [58]. The more the CE dissolved, the more the CE contributes to water retention of the mortar. Patural et al. [58] studied the effect of CE particle size on the water retention and found that the smaller the particles, the greater the water retention. Mechanical properties The mechanical properties which are critical to polymermodified mortars include compressive strength, tensile strength, and/or flexural strength. Compressive strength is a measure of the ability of the mortars to resist compressive stresses. In general, compressive strength is important to all mortars for applications, but more critical for floor applications. Tensile strength indicates the ability of the mortars to resist tensile stresses. In practice, the flexural strength is also commonly used to show the ability of the mortars to resist the tensile stress. The tensile strength or flexural strength is important to mortars in applications that are subjected to expansion or contraction between the layers within the mortar or between the mortars and the substrates or tiles. CE addition has both positive and negative effects on the mechanical properties of the mortars [6]. CE addition may enhance the mechanical properties of the mortars due to the more cohesive microstructure of the mortars. This improvement in the microstructure is from the reduction in the porosity at the ITZ due to the result of the dispersing effect of CE addition; from the reduction in the number of cracks owing to the integration of CE within the cement matrix; and from the bridges between CE film and the cement and aggregate matrix [6, 32]. This effect of CEs was observed by Khayat et al. [32] when he found that although the CE-modified mortar with certain composition has even slightly higher air content than the unmodified mortar, the former shows higher compressive and flexural strengths. Although the air content was not measured in studies by Sign et al. [24] and Coarna et al. [60], the improvement in the cohesion of the microstructure is also believed to result from the increase in the mechanical strengths of the certain CE-modified mortars compared to the unmodified one. In contrast, air entrainment causes reduction in the mechanical properties of the mortars [6, 32]. The higher the air content, the greater the reduction in the mechanical strengths. The retardation of CE may also contribute to the reduction in these properties of the mortars. The more the retarding effect of CEs, the lower the strength of the mortar [6]. The loss of strength from CE films under high moisture environments is also responsible for the degradation in the mechanical strengths of the mortars under these environments. The more the CE presents as a film within the mortar, the greater the loss of strengths of the mortar under high moisture environments. Due to simultaneous positive and negative effects of CEs addition, the influence of CE addition on the mechanical strengths depends on several factors, including mortar composition (i.e., water content, CE content, chemical nature of CEs), the mixing procedure, the age of the mortars, and the curing conditions [6]. In order to mitigate the negative impacts of CEs on the mechanical strengths, addition of additives (such as accelerators and defoamers) has been investigated [6, 24]. The limited number of studies that focus on the mechanical properties of CE-modified mortars leads to a minimal supporting data to the effects of these factors and additives. Effect of water content The increase in water content results in increased porosity of the mortars. The increase in porosity tends to reduce the compressive strength and the flexural strength. However, the increase in the water content may enhance the dispersing effect of CEs, thereby resulting in an improvement in the cohesiveness of the microstructure and enhancement of the mechanical strengths. If the increase in strength due to the presence of a more cohesive structure is higher than the reduction by the increase in porosity, the increase in water content improves the mechanical strengths, as in the case of MC-modified mortars (Table 3). In contrast, if the extent of strength increase due to the more cohesive structure is lower than the strength reduction due to the increase in the porosity, an increase in the water content can lower the mechanical strength, as seen in the case of HEC-modified mortars (Table 3). Effect of CE content By increasing the CE content, Singh et al. [24] found that there does exist a critical concentration of CE at which the measured mechanical strength of the mortar reaches a highest value (Fig. 34). At contents smaller than the critical value, the increase in CE dosage further enhances the mechanical strength. The increase in CE content normally leads to an increase in the retardation of cement hydration and the porosity, and this then tends to degrade the mechanical strength. However, the increase in CE content Table 3 Effect of w/c ratio on the compressive and flexural strength of mortars [6] Compressive strength at 28 days (MPa) Fig. 34 Generic figure showing the influence of CE concentration on compressive/tensile strength of the mortar, based on [24] in this period produces a better dispersing effect, enhancing the cohesion of the microstructure of the mortars and therefore improving the mechanical strength of the mortars. In this range of the CE content, the enhancement in the mechanical strength due to the better cohesion of the microstructure is greater than the degradation due to retardation of cement hydration and air entrainment. In contrast, at contents higher than the critical value, the increase in CE dosage degrades the mechanical properties of the mortar. This is because the enhancement in the mechanical strength due to the improved cohesion of the microstructure is lower than the degradation due to cement hydration and air entrainment. The value of the critical concentration of CE depends on the age of the sample, and the values are 0, 0.5, and 2.0 wt% for 7, 28, and 91 days, respectively (Figs. 35, 36) [24]. The change to higher CE concentration at longer )a 70 P (M60 h t ng 50 e r t es 40 v i ss 30 e r pm20 o C10 Unmodified Fig. 35 Influence of HEC content (HEC/cement by weight) on the compressive strength of the mortars at different ages, adapted from [24] Unmodified Fig. 36 Influence of HEC content (HEC/cement by weight) on the tensile strength of the mortars at different ages, adapted from [24] ageing times of the mortars suggests that this could be due to the retardation effect and CE film formation, which contributes to the strength of the mortars. Coarna et al. [60] studied the influence of the content of MHPC (0, 3, 5, and 7 wt%) on the flexural strength of the mortars and determined that a critical content of CE (3 wt%) exists that provides the highest value of flexural strength. Effect of curing conditions Under highly wet or moist curing conditions, the hydration of cement is promoted, while the CE film swells and loses its strength [6]. Overall, the wet/moisture curing produces a lower strength compared to the strength of the mortars under dry curing conditions [6, 60]. Coarna et al. [60] compared the flexural strength of MHPC-modified mortars between dry curing (6070 % relative humidity) and moisture curing (9098 % relative humidity) and found that the samples produced under dry curing conditions showed higher strengths than those produced under moisture curing. The study by Knapen [6] also shows the loss of compressive and flexural strengths of the mortars under moisture and water conditions when compared to the strength of the mortars before and after curing under these conditions. Knapen [6] also found that the mortars curing under wet conditions suffered losses in strength more dramatically than those cured under moisture conditions. The loss in strength of the mortars under moisture and water conditions can be recovered if the dry curing is followed as the film can form again and thus contribute to the strengths of the mortars. It should be noted that the migration of the CE film during the wet curing period may change the location of the CE film formed after the drying period following the wet period. Hence, the strength recovery of the mortar may be affected accordingly. Effect of additives in mitigating the degradation of mechanical strength by CEs Singh et al. [24] studied the effect of oxalic acid on the mechanical strengths of mortars modified by CEs during the first 91 days of curing and found that the strength of the CE-modified mortars was enhanced by oxalic acid addition. This was due to the reduction in the retardation effect of CEs owing to the addition of oxalic acid. Knapen [6] found that the use of defoamers helps to reduce the degradation in mechanical strengths of the CE-modified mortars by reducing air entrainment by CE addition. Water absorption and permeability Polymer-modified mortars in many applications, such as cement render, repair mortars, tile adhesives for outdoors, and for swimming pools, are subjected to contact with water. Under these circumstances, the mortars are required to have minimal water adsorption and permeability. This is because the water that gets adsorbed and is transported through the mortars may degrade the mortars and the construction structures covered by the mortars. The CE addition is found to inhibit the water adsorption of the mortars [24, 31, 39]. Ohama [31] and Singh et al. [24] explained this in terms of the swelling effect of CE film that seals the capillary pores. Pourchez et al. [39] correlated this inhibition to the effect of the void system. These voids which are connected through the capillary channels prevent fast flow of water and promote slow vapour flow. The increase in CE content further reduces the water permeability [24, 31]. Pourchez et al. [39] elucidated the importance of CE chemistry on the water permeability of the mortar. Mortar using HECs have higher water absorption than those employing MHPCs. Chemical resistance In some applications, the mortars are subjected to contact with chemicals. For example, tile adhesives that are used for swimming pools come into regular contact with sanitary chemicals. In these cases, chemical resistance is a desirable property of the mortars. The modification using CE was found to improve the chemical resistance of the mortar to H2SO4 and HCl [24, 32, 61]. Fu et al. [61] and Khayat et al. [32] believed that the reduction in porosity at the ITZ is responsible for that effect, while Singh et al. [24] assigned this improvement to the filling of CE film within the capillary pores. The higher the CE content, the greater the chemical resistance of the mortar. Volume change Volume changes (shrinkage or expansion) are critical to the mechanical properties of polymer-modified mortars. They are usually not desirable for hardened mortars as these may result in the formation of cracks between the mortar and tiles or substrate due to differences in contraction or expansion. CE addition increases the expansion under moisture or wet conditions and shrinkage under dry conditions [6]. This is due to the swelling of CE under moist and wet conditions and due to the water released during drying. This increase in the volume resulting from CE addition is a disadvantage as it could cause cracking within the mortar or at the interface between the mortar and the substrates or tiles. In order to minimise these negative effects of CE, the CE content should be at minimal levels. Summary: approaches for effective use of CEs and prospective future research The present work shows that CE addition has both positive and negative impacts on the properties of fresh and hardened mortars. The addition of CEs improves the rheology of the fresh mortars because, at a sufficiently low CE dosage, the yield stress and viscosity of the mortar are reduced and this enhances the flowability of the mortars [53]. This improvement in flowability is critical to mortars that require good flow, such as self-levelling mortars. The addition of suitable levels of CE helps to increase the yield stress of the fresh mortar and this allows the mortars to resist sagging [32, 53]. This sagging resistance is critical to render and repair mortars used for walls, for tile adhesives in the prevention of tile slippage, and for floor tiles in the prevention of early-stage slumping. The addition of CEs also increases the viscosity of the pore solution, increases the cohesiveness, and prevents phase separation of the mortar components [4]. The extension of the slow reaction period of cement hydration due to the retardation effect by CE modification increases the length of time during which the rheology of the mortars does not change significantly [29]. This has the advantage of allowing workers more time to fix the tiles. CE-modified mortars have greater water retention capacity and this allows the use of thin layers of mortars rather than thicker layers to compensate for the water lost due to evaporation and absorption by the substrate [1, 2]. The addition of an appropriate CE content helps to increase the mechanical strengths of mortars due to the enhancement in cohesion of the mortar microstructure (reduction in the porosity at the ITZ, presence of welldistributed cement particles, and reduction in the number of cracks) due to the dispersing effect of CEs and the integration between CE and the cement phases [6, 24, 32, 60]. The presence of a CE film within the microstructure of the mortars also reduces the water absorption, permeability, and subsequent risk of chemical attack [24, 31, 39]. Besides positive effects, CE addition may have a negative effect on fresh and hardened mortars. In order to mitigate this negative impact, appropriate CE choices and other additives can be used. To ensure the water retention capacity of the mortars, a certain dosage of CE is required and this may result in excessive values of the yield stress and viscosity of the mortars. In order to reduce this increase in the rheological parameters, CE grades of low Mw should be used. This CE choice allows for the use of higher CE contents to obtain the desired water capacity by exploiting the adsorption mechanism, while the yield stress and viscosity are still low enough such that they increase more slowly with concentration compared to CE grades of higher molecular weights [19, 54]. The retarding effect of CEs may cause undesired rate of strength development. The use of CE with the appropriate chemistry (type, DS, MS, and Mw) helps to minimise the retardation of CE because this depends strongly on the chemistry of CEs [5, 7, 38]. Moreover, the use of an appropriate type and dosage of accelerator allows for an increase in the rate of strength development [24]. CE addition also may cause an excessive amount of air entrainment, which reduces the strengths of mortars [6]. Defoamers can be used in order to reduce the air entrainment. The water sensitivity of the CE film causes degradation of the microstructure and strength of mortars under moist and wet conditions [6, 60]. The migration of CE under wet conditions also may negate advantages such as increased strength and reduced water absorption, permeability, and chemical attack following dry conditions. These potential disadvantages would arise owing to the consequent change in the position of the CE film. The use of the appropriate dosage and chemistry of CEs is critical to minimise the risk of formation of the CE film [6]. Although there is a significant number of studies that have focussed on the influence of CEs on cement mortars, the complexity of the physico-chemical nature of CEs is such that the impact of CEs is not fully understood. The adsorption of CEs on cement is considered to be critical for an understanding of their influence on cement hydration and to determine the amount of CE that can be used to avoid film formation and so mitigate the loss of strength under moist and wet conditions. Although several mechanisms of CE adsorption have been proposed (hydrogen bonds, complex formation, hydrophobic interaction, and electrostatic forces), there does not appear to be any published studies aimed at confirming these mechanisms. Although the effect of the Mw of CEs and the MS and DS of CEs with one type of ether (e.g., MC and HEC) has been investigated, the unambiguous influence of DS and MS of CEs with a mixture of ethers (e.g., MHEC and MHPC) on the adsorption on cement has not been reported. A significant understanding of the mechanism of the retardation effect of CE addition on the hydration of pure cement clinker phases (C3S, C2S, C3A, and C4AF) and Portland cement has been obtained. CE addition does not affect the initial reaction period of C3S hydration but it reduces the rate of the slow reaction period (i.e., extends the setting time) and reduces the rate of CSH and CH formation during the acceleration and deceleration periods [5, 21]. In the hydration of C3A without calcium sulphate, CE addition inhibits the transformation of metastable calcium hydroxyaluminates (C2AH8 and C4AH13) into the stable form C3AH6 [20, 28]. In the hydration of C3A in the presence of calcium sulphate, addition of CEs does not affect the first ettringite formation but it inhibits the formation of second ettringite [5, 20]. CE addition retards the hydration of C4AF and inhibits the transformation of metastable phases, such as C2(A,F)H8 and C4(A,F)H13 to C3(A,F)H6 [28]. In the hydration of Portland cement, CE addition does not affect the initial reaction of C3S and first ettringite formation but it increases the length of the slow reaction period and reduces the rates of CH and CSH formation, second ettringite formation, and subsequent CH and CS H formation [57, 29, 30]. However, there exists only a limited understanding of the effect of CEs on the transformation of ettringite to monosulphate as well as the formation of C4(A,F)H13. There are several mechanisms that have been proposed to explain the retardation effect of CEs, including the poisoning of the nucleation and growth of hydrates as well as reduction in dissolution and transportation of dissolved ions by forming an insoluble complex around the cement particles, thereby forming a protective membrane [6]. The concentration and chemical nature of CEs have been found to affect their retardation effect on the hydration of cement [57, 20, 24, 29, 38]. The effects of CE concentration, Mw, DS, and MS in general and that of CEs with single type of ether (such as MCs, HECs) are clear. In contrast, the effects of DS and MS of CEs containing two types of ethers (i.e., MHEC and MHPC) on the retardation of cement hydration remain ambiguous. The mechanism of air entrainment and stabilisation by CE addition is well established. The air is first entrained into the mortars as air bubbles due to the reduction in surface energy; with increasing time, these air bubbles tend to coalesce into bigger ones [6, 39]. The size and distribution of these air bubbles after mixing and after hardening depend strongly on the CE content and the chemical nature of the CEs. However, the influence of DS of CEs in general and the effect of DS and MS of CEs along containing two ethers on air entrainment and stabilisation do not appear to have been reported. CE addition increases the viscosity of the pore solution of the mortars but it may increase or decrease the yield stress and viscosity of the mortars depending on the CE content [32, 53]. CE-modified mortars show shear thinning behaviour and associated hysteresis [53]. Although the CE chemistry is believed to affect the rheology of the mortars, the influence of the type, DS, and MS have not been identified clearly. The increase in water retention capacity of mortars modified by CEs is attributed to four possible mechanisms: increase in viscosity, change of porous structure, adsorption, and formation of hydrocolloidals [31, 39, 46]. However, there does not appear to be any published study identifying which mechanism is dominant or how the CE chemistry affects the relative dominance of these mechanisms. Increasing the CE concentration or Mw is known to increase the water retention capacity of the mortars [39, 46] but the effects of the type, DS, and MS of CEs on water retention are not known. Acknowledgements The authors acknowledge the financial and technical support by Bostik Australia Pty. Ltd. Company toward this project. The authors also would like to thank the University of New South Wales for the tuition waiver scholarship to support for this work.


This is a preview of a remote PDF: http://link.springer.com/content/pdf/10.1007%2Fs10853-013-7732-8.pdf

D. D. Nguyen, L. P. Devlin, P. Koshy, C. C. Sorrell. Impact of water-soluble cellulose ethers on polymer-modified mortars, Journal of Materials Science, 2014, 923-951, DOI: 10.1007/s10853-013-7732-8