HAL Id: hal-03234247
https://hal.archives-ouvertes.fr/hal-03234247
Submitted on 25 May 2021HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Phase assemblage of cement pastes with SCM at
different ages
Mickael Saillio, Véronique Baroghel Bouny, Matthieu Bertin, Sylvain Pradelle,
Julien Vincent
To cite this version:
Phase assemblage of cement pastes with SCM at different ages.
Mickael Saillioa*, Véronique Baroghel-Bounyb, Matthieu Bertinb, Sylvain Pradelleb, Julien
Vincent a.
aParis-Est University, MAST, CPDM, IFSTTAR F-77447 Marne-La-Vallée, France bParis-Est University, MAST, FM2D, IFSTTAR F-77447 Marne-La-Vallée, France
Abstract
For economical and ecological reasons, supplementary cementitious materials (SCM) are increasingly used in concrete. Although having a lot of advantages, concretes with SCM have different cement matrices and do not have the same evolution as a function of time as CEM I concretes. In this paper, the microstructure of various cement pastes with/without SCM (slag GGBS, fly ash FA and metakaolin MK), has been investigated as a function of the curing time.
The microstructure was characterized not only by usual techniques such as XRD and TGA-DTA, but also by 29Si
and 27Al NMR spectroscopy. The use of a combination of techniques for microstructural characterization
allowed to quantify the proportion of each cementitious phase. This quantification showed the evolution of the cementitious matrices as a function of the binder content and of the curing time. For example, the alumina content in the C-S-H and the average length of C-S-H chains are higher for cement pastes with SCM than for OPC ones. Hydrated alumina phases proportions (AFt, AFm and TAH) are also higher in SCM cement paste. For MK and FA cement pastes, in addition to pozzolanic reactions, it seems that a small part of calcium from clinker hydration is used to form these aluminate phases.
*Corresponding author. Tel.: +33 1 81 66 82 39 E-mail address: [email protected]
Keywords: SCM, Curing time, hydration, TGA/DTA, 29Si and 27Al NMR spectroscopy.
1 Introduction
In non-aggressive environment (e.g. without CO2), the hydrated Portland cement consists of hydrated
phases (C-S-H, Portlandite, AFt, AFm,…) in equilibrium with a pore solution whose pH is around 13.5 [1-5]. In order to reduce carbon footprint and to improve some properties of the materials, clinker is partially replaced by supplementary cementitious materials (SCM). These SCMs can be silica fume, fly ash (FA), slag (ground granulated blast furnace slag, GGBS) or some other products, which can be found locally such as metakaolin (MK), glass powder... These SCMs affect the properties of
cementitious materials. Binders, for example, which contain GGBS are used in marine environment since the chloride diffusion coefficient of GGBS-mixtures is significantly reduced [6-9]. However, these binders form less portlandite than Ordinary Portland Cement (OPC) binders [10-12] as well for binders with fly ash or metakaolin [13-15]. Consequently, corrosion initiated by carbonation can be an issue for these materials using SCM with respect to OPC materials [16-20]. In these SCM materials, pozzolanic reactions consume a part of portlandite to form C(-A)-S-H [14,15,21]. Furthermore, evolution of the microstructure as a function of the age is not the same in presence or not of SCM [15][22-27] and in particular the pore structure [13,15,28]. There are numerous experimental studies about binders with SCM on one or two phases but few comparing all the phase assemblage of various cementitious matrices. In general, phase assemblage is more studied experimentally on reduced number of phases (mainly portlandite or C(-A)-S-H or sometimes on AFm/AFt phases) [4,14,22,29]. In addition, there are also studies about phase assemblage obtained by numerical modelling [12][30-32] taking into the thermodynamic and/or kinetics of the hydrated products.
The aim of this research is to investigate the effect of the binder type on various cementitious matrices. An experimental campaign has been carried out on various cementitious materials with or without SCM. The SCMs considered here are GGBS, FA and MK. Microstructure is investigated by differential thermogravimetry (TGA-DTA), X-ray diffraction (XRD) and 29Si and 27Al NMR MAS
© 2019. This manuscript version is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
Version of Record: https://www.sciencedirect.com/science/article/pii/S0950061819317313
spectroscopy. Quantification of various mineral phases (hydrated or not) has been obtained by combination of these techniques.
2 Experimental
2.1 materials studied
Various cement pastes were designed with the same clinker. The main constituents of the clinker are given in tables 1 and 2. The water to binder ratio (w/b) is equal to 0.50 for all the mixtures. The studied binders are CEM I (OPC with 97% clinker), CEM III/A (with 62% GGBS) denoted CEM III GGBS(62%), CEM III/C (with 82% GGBS) denoted CEM III GGBS(82%), CEM I + 20% FA denoted CEM I FA(20%), CEM I + 30% FA denoted CEM I FA(30%), CEM I + 40% FA denoted CEM I FA(40%), CEM I + 10% MK denoted CEM I MK(10%) and CEM I + 25% MK denoted CEM I MK(25%).
Table 1: Chemical compositions of the cement and SCM tested (% in mass) obtained by ICP-AES (Inductively Coupled Plasma Atomic Emission Spectroscopy).
CaO SiO2 Fe2O3 Al2O3 TiO2 MgO Na2O K2O MnO SO3
CEM I 62.53 19.54 2.90 4.98 0.30 0.84 0.30 0.82 0.09 2.97 CEM III/A 49.77 29.86 1.28 8.10 0.46 4.61 0.40 0.56 0.16 2.29 CEM III/C 45.70 32.00 1.00 9.90 0.50 5.80 0.63 0.54 0.20 2.00 MK 2.00 66.29 4.29 21.30 1.12 0.25 0.84 0.49 0.00 0.08 FA 4.30 51.59 6.58 23.78 1.03 0.49 1.09 3.05 0.11 3.05 Table 2: Mineralogical composition of the CEM I from BOGUE calculation (% in mass). BOGUE calculation determines mineralogical composition (C3S, C2S, C3A and C4AF) from elementary analysis obtained by
ICP-AES.
C3S C2S C3A C4AF
CEM I 51.23 28.32 9.90 8.81
Three main water-curing times (28 days, 90 days and 1 year) were chosen, in order to take into account the evolution of the microstructure as a function of the age. The bulk porosity of the cement pastes (measured by two different techniques) is also presented in table 3. In addition, some
additionnal characterizations were made after 7 days of water-curing in order to investigate the early age behaviour.
After water-curing, samples are crushed in small pieces of 1-2g. A short drying was performed on the samples (at 40°C during 24h) before the analyses described in section 2.2.
Table 3: Porosity of cement pastes obtained by porosity assessed by water and porosity assessed by mercury intrusion (in brackets) [28]. (-) values not obtained.
Bulk porosity (%) 28 days 91 days 1 year
2.2. Microstructural characterization
2.2.1. TGA- DTA
Thermogravimetric analyses (TGA) and Differential thermal analyses (DTA) [33,34] were performed with a simultaneous thermal analyzer by heating from 25°C to 1250°C with 10°C/min steps. These techniques are used here to quantify the water amount lost from calcium hydroxide (portlandite) and C-S-H + Ettringite (noted C-S-H+E), as well as the CO2 content lost from calcium carbonates of the
samples. Degree of reaction is also obtained by measuring bound water. The device used in the thermo-gravimetric studies is a NETZSCH STA 409.
2.2.2. XRD
X-ray diffraction [32] allows one to identify the crystallized phases (e.g. portlandite, calcium carbonates, monocarboaluminate, C2S, C3S …) and, in this study, to estimate their proportions and
evolution with time. The XRD analyses were performed using the PHILIPS PW 3830 diffractometer with the Kα radiation of cobalt (40 KV and 30 mA). The scan step size of the diffractometer was 0.02°/s in the range of 2ϴ from 4 and 76 °.
2.2.3 NMR
NMR spectroscopy [35] gives access, at a local scale, to the immediate environment of a nucleus. 27Al
and 29Si nuclei have been observed here by the MAS (Magical Angle Spining) technique [19,35].
Geometrical configurations of a nucleus are a function of the chemical shift. NMR spectra have been obtained here with a Bruker Avance 500MHz apparatus (11.74 T magnetic field).
Geometrical configurations of aluminum can be observed by 27Al NMR [19,36,37]. Aluminum in
tetrahedral configuration Al(IV) is generally attributed to aluminum substituted for silicon in C-S-H chains and residual anhydrous cement. Aluminum in octahedral Al(VI) configuration is divided into three components : AFt, AFm and a constituent which is, according to Andersen et al. [38,39], an amorphous/disordered aluminum hydroxide or a calcium aluminate hydrate (TAH) not observed by XRD [39]. The representative peak of AFm phases integrates multiple peaks (such as Friedel’s salt, monosulfoaluminate, monocarboaluminate phases, Künzel’salt…). However, XRD can provide complementary information in order to identify what are the phases involved and their relative
proportion. Finally, aluminum in pentahedral configuration Al(V) represents aluminum substituted for calcium in the C-S-H interlayers or present in non-hydrated phases generally in small amounts. The 29Si MAS NMR spectra allow identification of the different types of silicon tetrahedral present in
the samples [34,35,40,41]. In NMR spectroscopy, silicon tetrahedral are designated as Qn , where Q
represents the silicon tetrahedron bonded to four oxygen atoms and n denotes the connectivity, i.e. the number of other Q units attached to the SiO4 tetrahedron under study. In cement pastes, Qo species
characterize the anhydrous silicates, while Q1 and Q2 represent the C-S-H chains. Q3 and Q4
correspond to polymerized chains of C-S-H or silica gel, wich can be observed, for example, in carbonated samples. In the presence of Al, others species can be observed such as Qn (mAl), where m
represents the connectivity of silicon tetrahedron attached to the aluminate tetrahedron. From
quantitative 29Si NMR spectra, the average length of C-S-H chains can be estimated as 2 + 2 × Q2/Q1,
as well as the hydration rate obtained by 1-Q0 [35].
In this study, characterizations were made on crushed specimens of the cement pastes. Due to the high number of samples, a selection of NMR experiments on sample has been made.
2.2.4 Calculation of phase assemblage
Consequently, other tests were performed for TGA or XRD. These final results are presented in the article. However, if aberrant values depended on NMR, it was not possible to test again for the same reason said earlier (availability of NRM spectroscopy). There are multiple reason for these aberrant values. First, NMR and TGA/XRD experiments are not performed on the same location and for example, equilibrium of aluminate phases are sensitive to the exposure conditions (drying, humidity, …). It is possible that some samples were altered during transport. In addition, the representativeness of the sample can also be the cause of the problem, given the small quantities tested in the various technics.
3 Results and discussion
3.1 SCM effect on phase assemblage for cement paste in the long term
The proportions (after 365 days of water curing) of each cementitious phase obtained from various techniques (see annexes) are presented in figure 1 for different cement pastes. Inaddition, XRD data is also presented (see figure 2).
The amount of portlandite is reduced in the case of cement paste with SCM in particular for CEM III GGBS(82%) (see figure 1), as also confirmed by XRD results. Portlandite is produced only by the clinker, the decrease in the proportion of portlandite when the proportion of clinker decreases is thus quite expected and as already observed in numerous studies [24,42,43,44]. In addition, the pozzolanic reaction and slag activation are known for consuming a part of portlandite [24,44]. In order to separate the effect of dilution from the consumption of portlandite due to the pozzolanic reaction and slag activation, it is necessary to calculate the portlandite amount reported to the clinker proportion of the binder (Figure 1B). Based on these results, the pozzolanic reaction with FA and MK consumes obviously more portlandite than the activation of slag. In comparison to portlandite amount of CEM I cement paste (Figure 1A), the lower portlandite amount of CEM III GGBS(62%) is only due to the dilution effect. However, GGBS seem to consume a part of portlandite in CEM III GGBS(82%). It may due to the fact we used a commercial CEM III/C. If the GGBS content of this cement is not exactly 82% as mentioned on the technical sheet, it can induce also variations of clinker content and consequently variations of portlandite after hydration of the CEM III/C cement. However, in comparison, previous observations by TGA [11] made on cement paste with slag show the same portlandite amount (between 3 and 10% function of GGBS content of the cement) as in this study. In addition, another study [14,24] on cement paste with 30% fly ash showed a similar amount of portlandite as seen here as well a study with metakaolin [45].
The amount of C-S-H is increased in the case of cement paste with SCM, except CEM I FA(30%). In general, C-S-H amount of SCM materials are less important than expected. This may be due to the different states of SCM hydration (see evolution with time in section 3.2). This is consistent with the anhydrous phase proportions: they are higher for CEM I FA(30%), CEM III GGBS(62%) and CEM III GGBS(82%) than for CEM I after 365 days of water curing (see figure 1). According to 29Si NMR
CaCO3 (calcite according XRD results in figure 2) is also present in all the samples. This means that,
despite the precautions taken, a part of the samples has been carbonated (probably during the crushing process before analysis). In addition, thermal analysis of anhydrous CEM I, CEM III/A, CEM III/C, FA and MK revealed a small quantity of CaCO3 in these cements. Finally, CEM I with FA have the
highest amount of CaCO3 (and it increases with FA substitution) since anhydrous FA, before reaction,
already contains calcite (7.4%).
In this part, global variation between AFm phases and AFt phases (Ettringite) are discussed since it is difficult to separate by NMR 27Al the various forms of AFm phases (such as Monosulfoaluminate,
Monocarboaluminate, Hemicarboaluminate …). Cement pastes with SCM contain more AFm phases than OPC (except for CEM III GGBS(82%)). This result is explained by the high proportion of alumina in SCM binders and can be observed in other studies [13,27]. It also seems that the proportion of AFm phases decreases with the SCM ratio. As already observed on C-S-H, regarding amount of AFm + AFt reported to clinker (see figure 1B), SCM cement paste produce more of these aluminate phases than OPC. This result is particularly interesting because the production of these phases need calcium and there is almost no calcium available in anhydrous MK and FA (calcium is only calcium carbonate in these anhydrous SCMs). Consequently, a part of portlandite seems to be consumed to form these phases, in addition to form C-S-H during pozzolanic reaction. Concerning AFm/AFt ratio presented in figure 1C at 365 days (in green), it is higher for cement paste with SCM in comparison to CEM I cement paste. Since SO3 amount is higher in anhydrous CEM I than in CEM III/A, CEM III/C
and MK (see table 1), ettringite (AFt phases) is more stable than AFm for CEM I cement paste. Consequently, AFm/AFt ratio is lower for this cement paste than the one for the other cement pastes. SO3 amount is the same in anhydrous CEM I and in CEM I FA(30%) but AFm/AFt ratio is clearly
higher for this last. The form of SO3 is different in anhydrous CEM I (gypsum) than in FA (anhydrite)
and maybe this SO3 in FA is just partially dissolved during hydration and pozzolanic reaction. As
observed by XRD (see figure 2), hemicarboaluminate (Hc) and monocarboaluminate (Mc) seem to be present in all cement pastes in addition to traces of monosulfoaluminate (Ms). Equilibrium between these AFm types depends on the cement binder (the quantity of total AFm phase which can be formed) and the presence of CaCO3 [3,13,27]. For CEM I, CaCO3 amount is higher and AFm amount is lower
than for other cement pastes. Consequently, Mc is the preferential form. For other binders, AFm amount is higher and CaCO3 amount is lower (except for binders with FA superior to 30%).
Figure 1: Proportions of the main phases in various cement pastes after 365-day water curing, obtained by combination of 29Si and 27Al NMR and TGA/DTG results. A) unit is % reported to the hardened cement paste B)
unit is % reported to clinker C) Ratio AFm/AFt. (*) aberrant value. 0 0,5 1 1,5 2 2,5 3 CEM I CEM I MK(10%) CEM I MK(25%) CEM I FA(30%) CEM III GGBS(62%) CEM III GGBS(82%) AFm/AFt
28 days 91 days 365 days
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%
C-S-H Portlandite AFt AFm C2S C3S CaCO3 FA or GGBS Other phases
% o f ce m e n t p a st e
CEM I CEM I MK(10%) CEM I MK(25%) CEM I FA(30%) CEM III GGBS(62%) CEM III GGBS(82%)
0,0 0,5 1,0 1,5 2,0 2,5
Portlandite C-S-H AFt AFm
Figure 2: XRD data of cement pastes after 365-day water curing. E – Ettringite, Ms – Monosulfoaluminate Hc – hemicarboaluminate, Mc – monocarboaluminate, P – Portlandite, C – Calcite Q – Quartz, AN – Alite and Belite.
0
100
200
300
400
500
600
700
800
900
1000
5
10
15
20
25
30
35
40
In
te
n
si
ty
a
.u
.
2thetha
CEM I
CEM I MK(10%)
CEM I MK(25%)
0
100
200
300
400
500
600
700
800
900
1000
5
10
15
20
25
30
35
40
In
te
n
si
ty
a
.u
.
2theta
CEM I
CEM III GGBS(62%)
CEM III GGBS(82%)
0
100
200
300
400
500
600
700
800
900
1000
5
10
15
20
25
30
35
40
In
te
n
si
ty
a
.u
.
2theta
CEM I
CEM FA(20%)
CEM I FA(30%)
CEM I FA(40%)
3.2 Age effect on cement pastes hydration
The proportions of the anhydrous phases decrease as a function of the water curing time consistently with the increase of the C-S-H proportions (see figure 3, 4 and 5), whatever the cement paste and confirmed by XRD, as expected. It seems there are some modifications of the aluminate phases equilibrium during hydration (see also figure 1C). However, some values appears aberrant (e.g. CEM I at 91 days) and therefore, these samples need to be tested again for confirmation (27Al NMR).
The AFm/AFt ratio decreases with time for CEM I (if the value at 91 days is retrieved) and increases for the other cement pastes. SO3 amount is lower in cement paste with SCM than in CEM I one
(except CEM I FA(30%) cement paste). During the pozzolanic or the slag activation, this SO3 amount
might be not sufficient and consequently the AFm formation is favored rather than the ettringite formation. For CEM I FA(30%), SO3 amount from FA might be not completely accessible. In
addition, XRD shows small evolutions of AFm phase equilibrium. It seems there is less Hc and more Mc at 365 days for CEM I, CEM I MK(10%) and the three cement pastes with FA than 7 days. Consequently, it seems that equilibrium of aluminate phases continues to evolve with time.
The amount of portlandite increases with water curing time for CEM I and CEM I MK(10%), stays constant for CEM III GGBS(62%) and CEM III GGBS(82%) after 28 days and decreases for CEM I MK(25%) and CEM I FA(30%). Consequently, portlandite is consumed by MK and FA during pozzolanic reaction and a small part seems to be consumed by slag activation. The amount of portlandite reported to clinker as a function of water curing time is presented in figure 6. For cement pastes with FA, portlandite amount increases until 28 days and then decreases. It can be deduced that pozzolanic reaction consumes more portlandite than it is produced by clinker after 28 days. For CEM I MK(25%), portlandite amount decreases after 7 days and is strongly lower than CEM I. At the
beginning, kinetic of pozzolanic reaction with MK seems to be quicker than with FA. A filler effect can also occurs as reported [47]. However, after 365 days, reactivity of MK and FA seems similar. Figure 6 show the portlandite amount of cement paste with MK or FA, as a function of cement substitution for each water curing time. A linear relation appears for 91 days and 365 days confirming that, in long term, MK and FA have a similar reactivity, at the opposite case of in early age.
In the case of this study, it is possible to access to the percentage of C2S, C3S, FA and GGBS which
are reacted (called here reactivity). This is particularly interesting since clinker anhydrous phases can be separated of FA or of GGBS in SCM cement paste. For MK, it was not possible to calculate these anhydrous SCM amount because by 29Si NMR, the peak of this phase was indistinguishable to the
other peaks of Si. Results are given in figure 7 (right side). Reactivity of anhydrous phases increases with water curing time as expected and confirmed by XRD data (see figure 3, 4 and 5 in right). Reactivity of C2S and C3S in SCM cement paste are lower than in OPC, except in CEM III
GGBS(62%) where the values are close. Perhaps, the formation of hydrates on the surface limits the dissolution of anhydrous phases since C-S-H amount reported to clinker is higher in SCM cement paste (see figure 7 left). There is still high amount of SCM remaining after 365 days of water curing. It is difficult to say if these remaining phases could be hydrated in the future or if they are inert due to the presence of non-reactive phases (quartz).
In addition, hydration rate directly obtained by 29Si NMR (1-Q0) and degree of reaction (called
also with the reactivity of anhydrous phases obtained by coupling all technics (see figure 7). This can be explained since 29Si NMR focus only on Si and TGA only on water of hydrated phases. For this
last, it assumes a particular phase assemblage of cementitious matrix and knowing stoichiometry of C-S-H (C/S and H/S). However, it is difficult to obtain this stoichiometry value in particular for SCM cement pastes. Numerous previous studies tried to assess the hydration rate or degree of reaction or hydration [48-54]. Results often depends on technics used to determine this (NMR, TGA, SEM,…) and cannot be compared. In addition, according [55], uncertainties can be evaluated to 5%.
Figure 3: Proportions of the main phases in various cement pastes obtained by combination of 29Si and 27Al
NMR and TGA/DTG results (in left) and XRD data of cement pastes (in right) as a function of water curing time. E – Ettringite, A – Monosulfoaluminate, Hc – hemicarboaluminate, Mc – Monocarboaluminate, P – Portlandite, C – Calcite Q – Quartz, AN – Alite and Belite.
0 100 200 300 400 500 600 700 800 5 10 15 20 25 30 35 40 In te n si ty u .a . 2theta CEM I
7 days 28 days 91 days 365 days CEM I anh
0 100 200 300 400 500 600 700 800 5 10 15 20 25 30 35 40 In te n si ty u .a . 2theta CEM I MK(10%) 0 100 200 300 400 500 600 700 800 5 10 15 20 25 30 35 40 In te n si ty u .a . 2theta CEM I MK(25%) 0% 5% 10% 15% 20% 25% 30% 35% 40%
C-S-H Portlandite AFt AFm C2S C3S CaCO3 Other phases (TAH, Iron…)
CEM I
7 days 28 days 91 days 365 days
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%
C-S-H Portlandite AFt AFm C2S C3S CaCO3 Other phases (TAH, Iron…) CEM I MK(10%) 0% 5% 10% 15% 20% 25% 30% 35% 40%
Figure 4: Proportions of the main phases in various cement pastes obtained by combination of 29Si and 27Al
NMR and TGA/DTG results (in left) and XRD data of cement pastes (in right) as a function of water curing time. E – Ettringite, A – Monosulfoaluminate, Hc – hemicarboaluminate, Mc – Monocarboaluminate, P – Portlandite, C – Calcite Q – Quartz, AN – Alite and Belite.
0 100 200 300 400 500 600 700 800 5 10 15 20 25 30 35 40 In te n si ty u .a . 2theta CEM I FA(30%)
7 days 28 days 91 days 365 days CEM I anh
0 100 200 300 400 500 600 700 800 5 10 15 20 25 30 35 40 In te n si ty u .a . 2theta CEM III GGBS(62%)
7 days 28 days 91 days 365 days CEM III anh
Ms 0 100 200 300 400 500 600 700 800 5 10 15 20 25 30 35 40 In te n si ty u .a . 2theta CEM III GGBS(82%)
7 days 28 days 91 days 365 days CEM III anh
0% 5% 10% 15% 20% 25% 30% 35% 40%
C-S-H Portlandite AFt AFm C2S C3S CaCO3 FA anh Other phases (TAH,…)
CEM I FA(30%)
7 days 28 days 91 days 365 days
0% 5% 10% 15% 20% 25% 30% 35% 40%
C-S-H Portlandite AFt AFm C2S C3S CaCO3 GGBS anh Other phases (TAH, …) CEM III GGBS(62%) 0% 5% 10% 15% 20% 25% 30% 35% 40%
Figure 5: Proportions of portlandite and calcium carbonate in various cement pastes obtained by TGA/DTG results (in left) and XRD data of cement pastes (in right) as a function of water curing time. E – Ettringite, A – Monosulfoaluminate, Hc – hemicarboaluminate ,Mc – Monocarboaluminate, P – Portlandite, C – Calcite Q – Quartz, AN – Alite and Belite.
Figure 6: Proportions of portlandite as function of cement paste. Unit is reported to clinker. 0 100 200 300 400 500 600 700 800 5 10 15 20 25 30 35 40 In te n si ty u .a . 2theta CEM I FA(20%)
7 days 28 days 91 days 365 days CEM I anh
E Ms Hc Mc E P Q C P AN P 0 100 200 300 400 500 600 700 800 5 10 15 20 25 30 35 40 In te n si ty u .a . 2theta CEM I FA(40%)
7 days 28 days 91 days 365 days CEM I anh
E Ms Hc Mc E P Q P C AN P 0% 5% 10% 15% 20% 25% 30% 35% 40% Portlandite CaCO3 CEM I FA(20%)
7 days 28 days 91 days 365 days
0% 5% 10% 15% 20% 25% 30% 35% 40% Portlandite CaCO3 CEM I FA(40%) 0 0,1 0,2 0,3 CEM I CEM I MK(10%) CEM I MK(25%) CEM I FA(20%) CEM I FA(30%) CEM I FA(40%) CEM III GGBS(62%) CEM III GGBS(82%) g o f p o rt la n d it e p e r g o f cl in ke r
7 days 28 days 91 days 365 days
y = -0,07x + 0,20 R² = 0,67 y = -0,10x + 0,22 R² = 0,41 y = -0,28x + 0,25 R² = 0,94 y = -0,36x + 0,24 R² = 0,98 0 0,1 0,2 0,3 0,4 0% 10% 20% 30% 40% g o f p o rt la n d it e p e r g o f cl in ke r % of cement substitution by FA or MK
Figure 7: Proportion of each phase reported to clinker (in left) and reactivity of anhydrous phases (in right). (*) aberrant values. 0 0,5 1 1,5 2 2,5
Portlandite C-S-H AFt AFm
g p e r g o f cl in ke r CEM I
7 days 28 days 91 days 365 days
0% 20% 40% 60% 80% 100% C2S C3S R e a ct iv it y CEM I 28 days 91 days 365 days
0 0,5 1 1,5 2 2,5
Portlandite C-S-H AFt AFm
g p e r g o f cl in k e r CEM I MK(10%) 0% 20% 40% 60% 80% 100% C2S C3S R e a ct iv it y CEM I MK(10%) 0,0 0,5 1,0 1,5 2,0 2,5
Portlandite C-S-H AFt AFm
g p e r g o f cl in ke r CEM I MK(25%) 0% 20% 40% 60% 80% 100% C2S C3S R e a ct iv it y CEM I MK(25%) 0,0 0,5 1,0 1,5 2,0 2,5
Portlandite C-S-H AFt AFm
g p e r g o f cl in ke r CEM I FA(30%) 0% 20% 40% 60% 80% 100% C2S C3S FA R e a ct iv it y CEM I FA(30%) 0 0,5 1 1,5 2 2,5
Portlandite C-S-H AFt AFm
g p e r g o f cl in k e r CEM III GGBS(62%) 0% 20% 40% 60% 80% 100% C2S C3S GGBS R e a ct iv it y CEM III GGBS(62%) 0 0,5 1 1,5 2 2,5
Portlandite C-S-H AFt AFm
Figure 8: Hydration rate obtained by 29Si NMR (in left) and degree of reaction directly obtained by TGA (in
right).
3.3 C-S-H characteristics depending of the binder and of the age
Since amount of C-S-H and reactivity of clinker and SCM are known, it is possible to distinguish origin of C-S-H produced from clinker or from SCM (here FA or GGBS). Results of these are presented in figure 9 as a function of water curing time. Hydration kinetics can be assessed but, in the present study, only three times have been tested at the best (from 28 days to 365 days). C-S-H from SCM has major contribution for CEM III GGBS(62%) and CEM III GGBS(82%) after 28 days. FA reacts obviously slower than GGBS. In the long term, the proportion of C-S-H from SCM (e.g. 38% of all C-S-H for CEM I FA(30%)) is slightly higher than cement substitution (e.g. FA 30% of the binder for CEM I FA(30%)). The same tendency is observed for binder with GGBS in the long term.
Almost all the C/S ratios obtained (see figure 10) are similar to previous studies. The average value is between 0.6 and 2.0 according to [14]. In addition, the C/S of CEM III GGBS(62%) should be lower than OPC one, whatever the age of the material [11,22], and this is the case for CEM III GGBS(82%) after 91 days but not at 28 days. The C/S of cement pastes with FA and MK are lower than OPC one and seem to decrease with cement substitution by SCM. In fact, during pozzolanic reactions, SCM needs portlandite and uses calcium from this phase to form C-S-H. However, this calcium amount incorporated in C-S-H is lower than calcium amount directly incorporated in C-S-H by clinker hydration. For example, calculation gives C/S=0.8 for C-S-H from FA pozzolanic reaction in CEM I FA(30%) cement paste at one year, compare to C/S=1.7 if CEM I cement paste is taken as reference. There is alumina in C-S-H and it is substituted to calcium or silicium according NMR results (see section 2.2.3). Therefore, it is possible to include them in calculation of C/S, which becomes:
+ ( )
+ ( )
However, deconvolution of 27Al NMR spectra shows that the peak of Al(V) is almost negligible
comparing the peak of Al(IV) whatever the materials. Consequently, ( )
( ) ratios are lower than C/S
ones and become much lower than value of 0.6 for cement pastes with MK or FA. Below this value, C-S-H is considered more like a silica gel with calcium which is in contradiction with 29Si NMR
results showing that they are actually C-S-H and not silica gel.
relationship, developed originally by Richardson [2], to define the maximum allowable alumina substitution:
(1) = . ( )
.
In the present study, calculations are made for all cement pastes and not just for cement paste for GGBS. This maximum value (A/S) is respected for CEM I, CEM I MK(10%) and CEM I MK(25%). This value is largely exceeded for the other cement pastes. The formula depends on C/S and maybe for CEM III GGBS(62%), this ratio was overvalued. However, in the best scenario, if the minimal value of C/S is taken (0.6 as said before) for calculation, the A/S ratios of CEM I FA(30%) and of CEM III GGBS(82%) are still too high. In addition, these results may due to 27Al NMR experiments. During
deconvolution of spectrum, it is difficult to clearly separate SCM peak from all the other peaks and consequently alumina in C-S-H could be overestimated explaining the contradictory results.
The H/S ratio in C-S-H is presented figure 11. It seems to increase between 28 and 91 days and almost all value are included in the interval between values generally obtained in the literature (1 to 4
indicated in orange lines on the figure 11), excepted for CEM III GGBS(82%), for CEM I MK(25%) at 28 days and for CEM I FA(30%) at 28 days. However, in general, the values are low (inferior to 2.5). This may be due to the method used here. In fact, samples were analyzed after a short drying step (at 40°C during 24h). In addition, these samples were dried at 40°C until constant mass in order to obtain the water still contained in the porosity, without removing the water chemically bound in hydrated phases (S-H and ettringite). This water was removed from the TGA results concerning C-S-H+ettringite peak (within ambient temperature and 250°C) and represents 4-9% in cement paste mass. It seems that this drying step takes some water from C-S-H in addition to water from capillary pores as reported in [56-57]. In addition to this difficulty, porosity assessed by water evolves with water curing time as presented in table 3 as well as microstructure. In fact, drying kinetics are not the same considering the binder with or without SCM [56]. Consequently, H/S in C-S-H are lower than expected and C-S-H proportions, as presented in figure 1, are maybe slightly underestimated. In conclusion, as already reported in literature [53-54], separation of free water in pores from water chemically bound in C-S-H is quite difficult. For comparison, the W/(clinker+SCM) ratios are also presented in figure 11. These data were obtained by TGA and this value includes evaporable and chemically bond water. These ratios are lower than the theoretical value of 0.5. However, this value is only reached if there is no evaporation. The evolutions of W/(clinker+SCM) ratios are rather similar to H/S ones.
Finally, with 29Si NMR results, average length of C-S-H chains can be assessed (see figure 12). These
lengths seem to increase with cement substitution and water curing time. Another recent study show similar results [58]. In addition, the study reported in [11] shows that this average length increases with increasing GGBS proportion. In this case, cement paste are 20 years old. This could be an explanation of the difference in results as average length of C-S-H increases over time. In addition, detailed 29Si NMR results (not presented here) show a higher proportion of Q3 and Q4 and maybe the
presence of Q4(1Al) for cement paste with SCM, indicating a higher polymerization between C-S-H
chains. This is consistent with previously presented 27Al NMR results indicating the presence of higher
Figure 9: Amount of C-S-H according to origin (clinker or SCM) as a function of water curing time. Proportion of C-S-H are indicated on figure. For example, in CEM I, the C-S-H come from 100% of the clinker.
y = 0,0117ln(x) + 0,1149 R² = 0,708 y = 0,0489ln(x) + 0,0528 R² = 0,9192 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0 100 200 300 400 g o f C -S -H p ro d u ce d p e r g o f b in d e r Time (day)
C-S-H produced for CEM III GGBS(62%)
from clinker from GGBS
58% 62% 65% 42% 38% 35% 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0 100 200 300 400 g o f C -S -H p ro d u ce d p e r g o f b in d e r Time (day)
C-S-H produced for CEM III GGBS(82%)
from clinker from GGBS
90% 87% 10% 13% y = 0,0098ln(x) + 0,2145 R² = 0,9019 y = 0,0523ln(x) - 0,1461 R² = 1 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0 100 200 300 400 g o f C -S -H p ro d u ce d p e r g o f b in d e r Time (day)
C-S-H produced for CEM I FA(30%)
from clinker from FA
10% 38% 25% 90% 75% 62% y = 0,057ln(x) + 0,2234 R² = 0,4971 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0 100 200 300 400 g o f C -S -H p ro d u ce d p e r g o f b in d e r Time (day)
C-S-H produced for CEM I
from clinker
100%
Figure 10: Evolution of C/S and A/S in C-S-H as a function of the water curing time for various cement pastes obtained by combining 29Si NMR and TGA/DTA (see appendices). Orange line: interval between values
generally reported in the literature. Purple arrow : maximum theoretical values according Chen [32]. (*) aberrant value.
Figure 11: Evolution of H/S in C-S-H as a function of the water curing time for various cement pastes obtained by combining 29Si NMR and TGA/DTA (see appendices) and W/(clinker+SCM) directly obtained by TGA (W:
chemically bound water + evaporable water). Orange line: interval between values generally obtained in the literature. (*) aberrant value.
Figure 12: Average length of C-S-H chains as a function of the water curing time for various cement pastes obtained by 29Si NMR (see section 2.3.3).
0 0,5 1 1,5 2 2,5 CEM I CEM I MK(10%) CEM I MK(25%) CEM I FA(30%) CEM III GGBS(62%) CEM III GGBS(82%) C/S in C-S-H
28 days 91 days 365 days
0,0 0,1 0,2 0,3 CEM I CEM I MK(10%) CEM I MK(25%) CEM I FA(30%) CEM III GGBS(62%) CEM III GGBS(82%) A/S in C-S-H
28 days 91 days 365 days
* * * 0,0 0,1 0,2 0,3 0,4 CEM I CEM I MK(10%) CEM I MK(25%) CEM I FA(20%) CEM I FA(30%) CEM I FA(40%) CEM III GGBS(62%) CEM III GGBS(82%) W/(clinker + SCM)
7 days 28 days 91 days 365 days
0 1 2 3 4 CEM I CEM I MK(10%) CEM I MK(25%) CEM I FA(30%) CEM III GGBS(62%) CEM III GGBS(82%) H/S in C-S-H
28 days 91 days 365 days
0 1 2 3 4 5 6 7 8 9 CEM I CEM I MK(10%) CEM I MK(25%) CEM I FA(30%) CEM III GGBS(62%) CEM III GGBS(82%) A v e ra g e le n g th o f C -S -H c h a in s
4 Conclusion
In this paper, quantification of each mineral phase has been obtained by coupling of various
techniques (XRD, DTG/TGA, 29Si and 27Al NMR MAS spectroscopy and elementary analysis by
ICP-AES). This quantification shows the effect of SCM and age of various pastes on the proportion of each phase. After one year, there is higher proportion of C-S-H in SCM cement pastes (except in CEM I FA (30%)) and there is still large amount of anhydrous phases in CEM III GGBS(62%), CEM III
GGBS(82%) and CEM I FA(30%) samples that are quantified. However, it is impossible to say from the investigations carried out here if they could be reactive or are totally inert without performing additional tests at longer age (higher than 365 days).
The pozzolanic reactions consume more portlandite that the activation of slag. The C/S of the C-S-H is higher for OPC pastes than for cement pastes with fly ash or metakaolin. It seems particularly low for MK cement paste. This decrease of portlandite amount and a lower C/S have consequences on the buffering capacity of these cementitious materials. Consequently, in CO2 environment, these materials
are less resistant to the carbonation. On the other hand, the alumina content in the C-S-H and the average length of C-S-H chains are higher for pastes with SCM than for OPC ones. Hydrated alumina phases proportions (AFt, AFm and TAH) are higher in SCM cement paste. For MK and FA cement pastes, in addition to pozzolanic reactions, it seems that a small part of calcium from clinker hydration is used to form these hydrated aluminate phases. This could also explain why the C/S is so low for these materials. Equilibrium of aluminate phases is also quite different. There is more ettringite in CEM I cement paste and more AFm phase in SCM cement paste, mainly due to the initial amount of SO3 (higher in anhydrous CEM I compare to the other binders). This is also probably due to the form
of SO3 in binder (gypsum in CEM I, anhydrite in FA…) and consequently its capacity to react in
solution. For CEM I, CEM I with MK or CEM I with FA, AFm phases evolve from
hemicarboaluminate to monocarboaluminate in the long term. These AFm phases are probably due to the presence of calcium carbonate. For CEM III with GGBS, the crystalline form of AFm phases is not clear in the long term. It appears more like solid solution of various AFm phases. These differences in aluminate phase equilibrium and the differences of C-S-H have consequences on the chloride binding capacity of the materials. Consequently, there is an effect on the chloride ingress between CEM I and CEM I with SCM (in addition to the effect of the porous network).
This study will be complemented by tests at younger age (inferior to 28 days) in order to better understand hydration and pozzolanic kinetics. In addition, these results can be used to compare the phases assemblage obtained by Rietveld analysis and by modelling, in particular for hydration of cement paste with SCM. Finally, the method proposed here can be also applied in the study of cementitious materials in aggressive environment (presence of CO2 or chlorides) in order to study the
evolution of the phase assemblage.
Acknowledgements
The authors are grateful to J.B.D’Espinose de Lacaillerie from ESPCI (France) for NMR experiments, to G.Platret, B.Duchesne from IFSTTAR (France) for XRD and TGA/DTG investigations and to P.Gegout and F.Barberon from Bouygues-TP (France) for materials and for the partial funding of this study.
References
[1] W. Taylor, Cement chemistry, academic press, London, 1997.
[2] I.G. Richardson, The nature of C–S–H in hardened cements, Cement and Concrete Research 29 (1999) 1131-1147.
[4] T. Matschei, B. Lothenbach, F.P. Glasser, Thermodynamic properties of Portland cement hydrates in the system CaO–Al2O3–SiO2–CaSO4–CaCO3–H2O, Cement and Concrete Research 37 (2007) 1379-1410.
[5] K. Scrivener, A. Nonat, Hydration of cementitious materials, present and future, Cement and Concrete research 41:7 (2011) 651-665.
[6] P.S. Mangat, J.M. Khatib, B.T. Molloy, Microstructure, chloride diffusion and reinforcement corrosion in blended cement paste and concrete, Cement and Concrete Composites 16 (1994) 73-81.
[7] C. Arya, Y. Xu, Effect of cement type on chloride binding and corrosion of steel in concrete, Cement and Concrete Research 25 (1995) 893-902.
[8] V. Bouteiller, C. Cremona, V. Baroghel-Bouny, A. Maloula, Corrosion initiation of reinforced
concretes based on Portland or GGBS cements: Chloride contents and electrochemical characterizations versus time, Cement and Concrete Research 42 (2012) 1456-1467.
[9] S. Pradelle, M. Thiery, V. Baroghel-Bouny, Sensitivity analysis of chloride ingress models: Case of concretes immersed in seawater, Construction and Building Materials 136 (2017) 44-56.
[10] J. Duchesne, M.A. Berubé, The effectiveness of supplementary cementing materials in suppressing expansion due to ASR: Another look at the reaction mechanisms part 1: Concrete expansion and portlandite depletion, Cement and Concrete Research 24 (1994) 73-82.
[11] R. Taylor, I. Richardson, R. Brydson, Composition and microstructure of 20-year-old ordinary Portland cement-ground granulated blast-furnace slag blends containing 0 to 100% slag, Cement and Concrete Research 40 (2010) 971-983.
[12] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials, Cement and Concrete Research 41 (2011) 1244-1256.
[13] J.I. Escalante-Garcia, J.H. Sharp, The chemical composition and microstructure of hydration products in blended cements, Cement and Concrete Composites 26 (2004) 967-976.
[14] A. Girao, I. Richardson, R. Taylor, R. Brydson, Composition, morphology and nanostructure of C-S-H in 70% white Portland cement-30% fly ash blends hydrated at 55 °C, Cement and Concrete Research 40 (2010) 1350-1359.
[15] M. Saillio, V. Baroghel-Bouny, S. Pradelle, Various durability aspects of calcined Kaolin-Blended Portland cement pastes and concretes, Proceedings of Calcined clays for sustainable concrete, Lausanne: RILEM Bookseries 10 (2015) 491-499.
[16] J. Cowie, F.P Glasser, The reaction between cement and natural waters containing dissolved carbon dioxide, Advances in Cement Research 4:15 (1992) 119-134.
[17] V.T. Ngala, C.L. Page, Effects of carbonation on pore structure and diffusional properties of hydrated cement pastes, Cement and concrete research 27 (1997) 995-1007.
[18] P.H.R. Borges, J.O. Costa, N.B. Milestone, C.J. Lynsdale, Carbonation of CH and C-S-H in composite cement pastes containing high amounts of BFS, Cement and Concrete Research 40 (2010) 284-292. [19] M. Saillio, V. Baroghel-Bouny, Chloride binding in sound and carbonated cementitious materials with
various types of binder, Construction and Building Materials 68 (2014) 82-91.
[20] A. Leeman, P. Nygaard, J. Kaufman, R. Loser, Relation between carbonation resistance, mix design and exposure of mortar and concrete, Cement and Concrete Composites 62 (2015) 33-43.
[21] C. Love, I. Richardson, A. Brough, Composition and structure of CSH in white portland cement -20% metakaolin pastes hydrated at 25°C, Cement and Concrete Research 37 (2007) 109-117.
[22] J.J. Thomas, A.J. Allen, H. Jennings, Density and water content of nanoscale solid C–S–H formed in alkali-activated slag (AAS) paste and implications for chemical shrinkage, Cement and Concrete Research 42 (2012) 377-383
[23] I.G. Richardson, G.W. Groves, The structure of the calcium silicate phases present in hardened pastes of white Portland cement/blast-furnace slag blends, Journal of Materials Science 32 (1997) 4793-4802. [24] S. Hanehara, F. Tomosawa, M. Kobayakawa, K.R. Hwang, Effects of water/powder ratio, mixing ratio
of fly ash, and curing temperature on pozzolanic reaction of fly ash in cement paste, Cement and Concrete Research 31 (2001) 31-19.
[25] D. Dyer, R.K. Dhir, Hydration reactions of cement combinations containing vitrified incinerator fly ash, Cement and Concrete Research 34 (2004) 849-856.
[26] K. Luke, E. Lachowski, Internal composition of 20-year-old fly ash and slag-blended ordinary Portland cement pastes, Journal of the American Ceramic Society 91 (2008) 4084-4092.
[27] P.T. Durdziński, M. Ben Haha, M. Zajac, K. Scrivener, Phase assemblage of composite cements, Cement and Concrete Research 99 (2017) 172-182.
[28] E. Berodier, K. Scrivener, Evolution of pore structure in blended systems, Cement and Concrete Research 73 (2015) 25-35.
[30] J.J. Thomas, J.J. Biernacki, J.W. Bullard, S. Bishnoi, J.S. Dolado, G.W. Scherer, A. Luttge, Modeling and simulation of cement hydration kinetics and microstructure development, Cement and Concrete Research 411 (2011) 1 257-1278.
[31] X.Y. Wang, H.S. Lee, Modeling the hydration of concrete incorporating fly ash or slag, Cement and Concrete Research 40 (2010) 984-996.
[32] W. Chen, H.J.H. Brouwers, The hydration of slag, part 1:reaction models for alkali-activated slag, Journal of Materials Science 42 (2007) 428-443.
[33] G. Villain, M Thiery, G. Platret, Measurement of carbonation profiles in concrete: Thermogravimetry, chemical analysis and gammadensimetry, Cement and Concrete Research 37:8 (2007) 1182-1192. [34] K. Scrivener, R. Snellings, B. Lothenbach, A pratical guide to microstructural analysis of cementitious
materials, CRC Press, Boca Raton. 2016.
[35] G. Engelhardt, D. Michel, High resolution 29Si NMR of silicates and Zeolites, Wiley, New York, 1987.
[36] M.R. Jones, D.E. Macphee, J.A. Chudek, G. Hunter, R. Lannegrand, R. Talero, S.N. Scrimgeour, 27Al
MAS NMR of AFm and AFt phases and the formation of Friedel's salt, Cement and Concrete Research 33 (2003) 177-182.
[37] G.K. Sun, J.F. Young, R.J. Kirkpatrick, The role of Al in C-S-H: NMR, XRD, and compositional results for precipitated samples, Cement and Concrete Research 36 (2006) 18-29.
[38] M.D. Andersen, H.J. Jakobsen, J. Skibsted, A new aluminium-hydrate species in hydrated Portland cements characterized by 27Al and 29Si MAS NMR spectroscopy, Cement and Concrete Research 36 (2006) 3-17.
[39] J. Skibsted, E. Henderson, H.J. Jakobsen, Characterization of calcium aluminate phases in cements by
27Al MAS NMR spectroscopy, Inorganic Chemistry 32 (1993) 1013-1027.
[40] T. Sevelsted, J. Skibsted, Carbonation of C-S-H and C-A-S-H samples studied by 13C, 27Al and 29Si
MAS NMR spectroscopy, Cement and Concrete Research 71 (2015) 56-65.
[41] X. Cong, R.J. Kirkpatrick, 29Si MAS NMR study of the structure of calcium silicate hydrate, Advances in Cement Based Materials 3 (1996) 144-156.
[42] R.B. Freeman, R.L. Carrasquillo, Influence of the method of fly ash incorporation on the sulfate resistance of fly ash concrete, Cement and Concrete Composites 13 (1991) 209-217.
[43] E. Badogiannis, The effect of metakaolin on concrete properties, Proceedings of international Congress : Challenges of Concrete Construction, Dundee, 2002, 81-89.
[44] M. Ben Haha, B. Lothenbach, G. Le Saout, F. Winnefeld, Influence of slag chemistry on the hydration of alkali-activated blast-furnace slag – Part II: effect of Al2O3, Cement and Concrete Research 42 (2012) 74-83.
[45] B.B. Sabir, S. Wild, J. Bai, Metakaolin and calcined clays as pozzolans for concrete: a review, Cement and Concrete Composites, 23 (2001) 441-454.
[46] E. Sakai, S. Miyahara, S. Ohsawa, S.H. Lee, M. Daimon, Hydration of fly ash cement, Cement and Concrete Research 35 (2005) 1135-1140.
[47] W.A. Gutteridge, J.A. Dalziel, Filler cement: the effect of the secondary component on the hydration of Portland cement: part I. A fine non-hydraulic filler, Cement and Concrete Research 20 (1990) 778-782. [48] M. Ben Haha, K. De Weerdt, B. Lothenbach, Quantification of the degree of reaction of fly ash,
Cement and Concrete Research 40 (2010) 1620-1629.
[49] V. Kocaba, E. Gallucci, K. Scrivener, Methods for determination of degree of reaction of slag in blended cement pastes, Cement and Concrete Research 42 (2012) 511-525.
[50] S. Lumley, R.S. Gollop, G.K. Moir, H.F.W. Taylor, Degrees of reaction of the slag in some blends with Portland cement, Cement and Concrete Research 26 (1996) 139-151.
[51] X. Feng, E.J. Garboczi, DP.Bentz, P.E. Stutzman, T.O. Mason, Estimation of the degree of hydration of blended cement pastes by a scanning electron microscope point-counting procedure, Cement and Concrete Research 34 (2004) 1787-1793.
[52] X. Wu X, D.M. Roy, C.A. Langton, Early stage hydration of slag cement, Cement and Concrete Research 13 (1983) 277-286.
[53] I. Pane, W. Hansen, Investigation of blended cement hydration by isothermal calorimetry and thermal analysis, Cement and Concrete Research 35 (2005) 1155-1164.
[54] J.I. Bhatty, Hydration versus strength in a portland cement developed from domestic mineral wastes — a comparative study, Thermochimica Acta 106 (1986) 93-103.
[55] P. Durdzinski, M. Ben Haha, N. De Belie, E. Gruyaert, B. Lothenbach, E. Menendez-Mendez, J.L. Provis, A. Scholer, C. Stabler, Z. Tan, Y. Villagran Zaccardi, A. Vollpracht, F. Winnefeld, M. Zajac, K. Scrivener, Outcomes of the RILEM roundrobin on degree of reaction of slag and fly ash in blended cements, Materials and Structures 50 (2017) 0-15.
[57] V. Baroghel-Bouny, Microstructural and hydric characterization of cement paste and ordinary and high-performance concrete in French, PhD thesis, ENPC, 1994.
Appendices
Results from chemical analysis (see table 1), TGA/DTA, XRD and NMR are combined to obtain each phase proportion except portlandite amount which is directly obtained from TGA/DTG results.
1 Quantification of portlandite (1) % !"#$%&!'=%()*+,,./,*0)123*+4
567 × M !"#$%&!'
Where % !"#$%&!'denotes the mass proportion of portlandite, %:;<=>, ?@AB CDEFBG is the percentage of water lost by portlandite obtained by TGA, MH6Iis mass molar of water and M !"#$%&!'the mass molar of portlandite
2 Quantification of aluminate phases In 100 g of cement paste, the total alumina quantity (nK"6IL in mol) is:
(2) nK"6IL =
MN67LO×P.,QQRR°S
4MN67L × 100
Where PK"6ILV denotes the mass proportion of alumina obtained by chemical analyses in anhydrous binder,
?P, ° is the remaining mass proportion of binder after 1100°C thermal treatment of the hydrated cement paste
obtained by TGA, and MK"6ILis the molar mass of alumina.
It is assumed that AFt phases are only ettringite. According to 27Al NMR results, the mass proportion
of AFt phases (%KW!) is:
(3) %KW!= nK"6IL× PKW!,X4Y× MZ
Where PKW!,X4Y is the molar proportion of AFt phases obtained by 27Al NMR and MZ is the molar mass of ettringite (1255.7 g/mol).
It is assumed that AFm phases are only monosulfoaluminate phase. According to 27Al NMR results,
the mass proportion of AFm phases (% [\) is:
(4) %KW]= nK"6IL× PKW],X4Y× MKW]
Where P[\,^_` is the molar proportion of AFm phases obtained by 27Al NMR and a[\ is the average molar mass of AFm phases (622 g/mol).
In addition, the AFm/AFt ratio can be obtained from these equations.
3 Quantification of C-S-H
Each molar quantity of SiO2, CaO, H2O and Al2O3 contained in C-S-H are determined for 100 g of
cement paste.
3.1 Molar quantity of SiO2 contained in C-S-H
(5) n bI6 =
de76O×P.,QQRR°S
4de76 × 100
Where Pf&I6V denotes the mass proportion of alumina obtained by chemical analyses in anhydrous binder,
?P, ° is the remaining mass proportion of binder after 1100°C thermal treatment of the hydrated cement paste
obtained by TGA, and Mf&I6is the molar mass of alumina. According to 29Si NMR results, the molar quantity of SiO
2 (nf&I6, g) contained in C-S-H for 100 g of
cement paste is:
(6) nf&I6,hfH= nbI6× Pf& hfH,X4Y
Where Pf& hfH,X4Y is the molar proportion of C-S-H obtained by 29Si NMR.
3.2 Molar quantity of watercontained in C-S-H
The molar quantity of water contained in C-S-H (nH6I,hfH) for 100 g of cement paste is determined by
TGA/DTA. However, the quantity of water of etttringite needs to be quantified before and it is assumed that this water lost is only due to C-S-H, ettringite and remaining water in porosity.
(7) nH6I,hfH= (PhfH'!Z,ijK× 100 −%Mlm × L6 × n567no − %p>== H6I)/MH6I
Where PhfH'!Z,ijK is the mass proportion of water lost from CSH+E obtained by TGA/DTG, %KW! × 4 × 4567
o
denotes the quantity of water contained in ettringite, %p>== H6Iis the quantity of water inside the porous network
obtained by drying until constant mass at 40°C, and MH6I is the molar mass of water (in %mass). 3.3 Molar quantity of CaO contained in C-S-H
The molar quantity of CaO contained in C-S-H is the most difficult value to asses. The quantity of calcium contained in the other phases is first evaluated and then the molar quantity of CaO contained in C-S-H is deduced from the whole calcium quantity for 100 g of cement paste.
• Calcium molar quantity of portlandite (nh#,r !"#$%&!')
(8) nh#,r !"#$%&!'=%4stumNvwxemy
ztumNvwxemy× 100
Where % !"#$%&!' is the mass proportion of portlandite obtained by equation (1) and Mr !"#$%&!' is the molar mass of porlandite.
• Calcium molar quantity of each aluminate phase AFt (nh#,KW!) and AFm (nh#,KW])
There are 6 (respectively 3) mol of calcium by mol of ettringite (respectively AFm). (9) nh#,KW!= nK"6IL× PKW!,X4Y× 6
(10) nh#,KW]= nK"6IL× PKW],X4Y× 3
The calcium quantities contained in C3A, C4AF and TAH are assumed as negligible with respect to the
calcium quantities contained in the other phases.
• Calcium molar quantities of anhydrous phases C3S and C2S
According to 29Si NMR results, the calcium molar quantities of C
(11) D ;, L = nf&I6× Pb L ,^_`× 3 (12) D ;, 6 = nf&I6× Pb 6 ,^_`× 2
Where nf&I6is the total Si quantity in 100g of cement paste, Pb L ,^_` is the molar proportion of C3S phases
obtained by 29Si NMR and where P
b 6 ,^_` is the molar proportion of C2S phases obtained by
29Si NMR.
• Calcium molar quantity of CaCO3
The calculated molar quantity of CaCO3 is low since the materials are non-carbonated samples.
However, a part of material can be carbonated during crushed step and drying despite the taken precaution and in this study CEM I, CEM III/A, CEM III/C, FA and MK contain also a small part of CaCO3 observed by XRD (calcite) and quantified by TGA.
(13) nh#,h#hIL =
~v~7L,•€M
4~v~7L × 100
Where Ph#hIL,ijK is the mass proportion of CaCO3 obtained by TGA/DTG and Mh#hILis the molar mass of
CaCO3.
Consequently, the molar quantity of CaO (nh#I,hfH) contained in C-S-H for 100 g of cement paste is:
(14) D ;•, g= (nh#I− ∑ Ca, other phases)
Where nh#I is the molar quantity of CaO contained in 100 g of cement paste and ∑ Ca, other phases denotes the sum of calcium contained in other phases than C-S-H.
3.4 Molar quantity of alumina in C-S-H (nK"6IL,hfH)
The quantity of alumina in C-S-H is very low for OPC sample and is high for cement paste with SCM. (15) nK"6IL,hfH= nK"6IL× PK"(Œ• •),X4Y
Where PK"(Œ• •),X4Y is the molar proportion of alumina present in C-S-H obtained by 27Al NMR.
According to equations 6,7,14 and 15, the mass proportion of C-S-H (%],hfH) for 100 g of cement
paste is:
(16) %],hfH= D ;•, g× Mh#I+ nf&I6,hfH× Mf&I6+ nH6I,hfH× MH6I+ nK"6IL,hfH× MK"6IL In addition, the C/S, H/S and A/S ratios of the C-S-H can be obtained from these equations.
4 Quantification of anhydrous phases (C2S or C3S)
The mass proportions of C2S or C3S % 6 L are given by the following equations:
(17) % 6 = nf&I6× Pb 6 ,^_`× M 6 (18) % L = nf&I6× Pb L ,^_`× M L
Where Pb 6 ,^_` or Pb L ,^_` are the molar proportions of anhydrous phases obtained by
29Si NMR and
M 6 L is respectively the molar mass of or .
The amount of the other phases (TAH, quartz, iron…) represents the uncertainty of the calculations. The higher the percentage of these phases, the greater the uncertainty.
(19) % !Ž' rŽ#•'•= 100 − ∑ other phases
Here, it is difficult to evaluate general uncertainties for all phases, especially when results depend of NMR spectroscopy. In fact, to our knowledge, there is no study about measurement uncertainties obtained by this technics.
6 Adaptation of calculations due to presence of SCM in binders In the presence of SCM (such as metakaolin or fly ash), the binder contents for the different elements (SiO2, Al2O3, CaO ...) take into account the contribution of the SCM. For example, the mass
proportion of Al2O3 for CEM I MK(25%) anhydrous binder is equal to:
(20) PK"6IL,V= PK"6IL,•"&$‘' × 0,75 + PK"6IL,4“× 0,25
Finally, the proportions of GGBS (%””• ) and FA (%[ ), which are not reacted, are calculated in SCM
cement paste:
(21) %””• = nf&I6× Pb ””• ,^_`× M””• (22) %[ = nf&I6× Pb [ ,^_`× M[
Where Pb ””• ,^_` and Pb [ ,^_` are the molar proportions of anhydrous phases obtained by 29Si NMR and M””• and M[ are respectively the molar mass of GGBS and FA. For calculation of M””• and M[ , we consider that these contain only SiO2 and Al2O3.
