Send Orders for Reprints to
[email protected] Current Inorganic Chemistry, 2015, 5, 00-00
1
Crystal Chemistry of Iron Containing Cementitious AFm Layered Hydrates Guillaume Renaudin1,2,*,#, Adel Mesbah1,3, Belay Zeleke Dilnesa4, Michel Francois5 and Barbara Lothenbach4 1
Clermont Université, ENSCCF, Institut de Chimie de Clermont-Ferrand, BP 10448, F-63000 Clermont-Ferrand, France; 2CNRS, UMR 6296, ICCF, F-63171 Aubière, France; 3ICSM, UMR5257 CEA/CNRS/UM2/ENSCM, Site de Marcoule, Bât. 426, BP 17171, F-30207 Bagnol/Cèze cedex, France; 4Empa, Laboratory for Concrete & Construction Chemistry, Überlandstrasse 129, CH-8600 Dübendorf, Switzerland; 5Université de Lorraine, Institut Jean Lamour – UMR 7198, BP 70239, 54506 Vandœuvre-lès-Nancy Cedex, France Abstract: The crystal structure of the three main Fe-containing AFm phases (Al2O3-Fe2O3-mono: family of lamellar calcium alumina-ferrite hydrates) encountered in cement hydration process are characterized and compared with their Al-analogs. This includes AFm phases containing sulfate (which is present in Portland cement to regulate the hydration kinetic), carbonate (which is present in Portland cements, or originates from atmospheric carbon oxide) and chloride (either from the water used or from the environment). The results show that Fe-AFm and Al-AFm compounds are not (or rarely) isostructural. Iron in AFm phases does not simply substitute aluminium. Fe-carbonate has a rhombohedral symmetry whereas Al-carbonate has a triclinic symmetry, with carbonate anions located in different crystallographic sites in both compounds. Fe-Friedel’s salt corresponds to a 3R polytype whereas Al-Friedel’s salt corresponds to a 6R polytype. Both compounds have a temperature dependent transition with two different HT- and two different LT-polymorphs descriptions (HT: high-temperature, LT: low-temperature). Only Fesulfate and Al-sulfate are isostructural. Despite this isostructural feature, only limited solid solutions have been observed between both sulfate end-members. In a general way, this system (when considering sulfate, carbonate and chloride with aluminium and iron) leads to extremely complicated subsystems with limited solid solutions. The crystallographic studies and comparisons developed here have been fully completed by thermodynamic characterisations in order to make possible thermodynamic modelling of the hydrates assemblage during the hydration process and the aging of Portland concrete.
Keywords: Cement hydrates, crystal structure, iron, solid solution, X-ray diffraction. 1. INTRODUCTION The hydrates formed during the hydration of Portland cement are mainly C-S-H (calcium silicate hydrate), portlandite (Ca(OH)2), ettringite (the AFt, Al2O3-Fe2O3-tri, family with composition Ca6Al2(OH)6·(SO4)3·26H2O) and lamellar AFm (the Al2O3-Fe2O3-mono family: lamellar calcium alumina-ferrite hydrates) phases. AFm hydrates are obtained from the C3A (3CaO·Al2O3) and C2(A,F) (‘4CaO·Al2O3·Fe2O3’) – the two calcium aluminate anhydrous compounds present in clinker – hydration process in the presence of sulfate, carbonate, chloride and hydroxide. The general formula of AFm phases is Ca2Al(OH)6·X·nH2O. X represents a single charged or half a double charged anion located in the interlayer region of these lamellar hydrates. The layered structure of AFm compounds is based on two layers; a positively charged rigid main layer [Ca2Al(OH)6] + and a negatively charged [X·nH2O]- interlayer corresponding to the hydrocalumite type structure (an ordered modification *Address correspondence to this author at the ENSCCF, Institut de Chimie de Clermont-Ferrand, Clermont-Ferrand, France; Tel: 04-73-40-73-36; Fax: 04-73-40-73-33; E-mail:
[email protected] # Present address: ICSM, UMR5257 CEA/CNRS/UM2/ENSCM, Site de Marcoule, Bât. 426, BP 17171, F-30207 Bagnol/Cèze cedex, France 1877-9441/15 $58.00+.00
from the large LDH family described by the hydrotalcite type structure in which bivalent and trivalent cations are disordered in a unique crystallographic site). Different anions and number of water molecules can be present in the interlayer region depending on the relative humidity and temperature. The most important anions in cement chemistry are OH- (pH around 12.5), SO42- (presence of gypsum to regulate the cement hydration kinetic), CO32- (from atmospheric contamination, cement addition or environment) and Cl- (from water or environment). The main layers are composed of Ca(OH)6 octahedra (portlandite type) with every third Ca2+ substituted by Al3+ and/or Fe3+ in an ordered way. Crystal chemistry, solubility products and thermodynamic data are well-known for the Al-containing AFm hydrates as comprehensive experimental studies were carried in the past [1-3]. Due to the low Fe-content in Portland cement (between 2 and 5 weight percent of iron oxide, mainly present in C2(A,F)) the first series of experimental data on Fecontaining AFm hydrates have been determined quite recently [4-8]. Thermodynamic studies on Fe-monosulfate [6], Fe-monocarbonate [7] and Fe-hydrogarnet [8] have been published last years. The fate of Fe during the hydration of cement depends on the cement. In the absence of other clinker phases, C2(A,F) reacts to form Fe-containing ettrin© 2015 Bentham Science Publishers
2 Current Inorganic Chemistry, 2015, Vol. 5, No. 2
gite, monosulfate or monocarbonate, depending on the presence of calcium, sulfate or carbonate. In many studies, the presence of amorphous iron hydroxide and the solid solution formation with their Al-containing analogous were assumed [9-14]. The almost equivalent ionic radii in octahedral environment of Al3+ (0.55 Å) and Fe3+ (0.65 Å) [15] and the well-known octahedral Al to Fe substitution in clays minerals have suggested the (Al,Fe)-AFm solids solutions formation. Whether and to what extent the Fe-containing phases and their solid solutions formed in PC (Portland cement) is presently poorly understood. Investigations on the fate of Fe in cementitious materials are complicated by the fact that identification of the Fe-containing minerals in hydrated cement using standard techniques, such as X-ray diffraction (XRD), thermogravimetric analysis (TGA) and Scanning Electron Microscopy (SEM), is difficult due to significant overlap of the signals from the Fe-containing phases with those of the respective Al analogs. Further, detection of amorphous Fe-containing phases in hydrated cement is difficult using standard techniques. Nowadays the development of new cement formulations [16,17] enhances the necessity to have a full and detailed understanding of each element that acts during cement hydration. The present study exposes recent structural investigations of the main Fe-based AFm phases. Crystallographic data on Al- and Fe-based compounds show that solids solutions are highly limited for carbonate, sulfate and also chloride compounds. Despite similar crystal chemistry for all AFm phases, the existence of different space groups clearly evidences the absence of complete solids solutions. This has already been described for the different anions observed in Al-based AFm hydrates: R3 for monosulfoaluminate [18], R3c for hemicarboaluminate [19], P1 or P1 for monocarboaluminate [20,21], C2/c for Friedel’s salt [22], P3c1 for nitrate [23]. The comparison of the crystal structures of Al-based AFm compounds with that of Fe-based AFm compounds can indicate whether the formation of solids solutions based on Al to Fe substitution are possible. In the present paper the three most important anions in cement chemistry have been considered; i.e. sulfate, carbonate and chloride. In the complex cement matrix, the amount of Fe incorporated in AFm phases will influence the AFt/AFm ratio and thus the hydrates volume and finally the hydrated cement properties. A complete understanding of the characteristics of the hydrates is important since the material properties of cement-based materials are related to the chemical environment and properties of the hydrated phases.
Renaudin et al.
with a liquid over solid ratio ~20. CaO was prepared by heating CaCO3 at 1000°C. Ca2Fe2O5 was prepared as previously described. More details are given in thermodynamic study of the compound [7]. -
PE-bottles at different temperatures (20, 50 and 80°C) were used to store the samples. Ageing effect (or reaction time) has been considered to improve the crystallinity of the powder. The solid and liquid phases were separated by vacuum filtration through 0.45 μm nylon filter after equilibration. All manipulations were done in a nitrogen filled glove box to minimize atmospheric CO2 contamination. 2.2. Powder X-ray Diffraction -
Fe-monocarbonate Ca2Fe(OH)6·(CO3)·nH2O and FeFriedel’s salt Ca2Fe(OH)6·Cl·nH2O: synchrotron powder diffraction data were collected at Swiss-Norwegian Beam Line (SNBL) at the European Synchrotron Radiation Facility (ESRF), Grenoble, France. The powder material was introduced into glass capillaries (0.5 mm diameter). Data collection was performed at 295 K at a wavelength of = 0.72085 Å using a MAR345 image plate detector with the highest resolution (34503450 pixels with a pixel size of 100 μm). The calculated absorption coefficient mμR (m = powder packing factor, μ = linear absorption coefficient, R = radius of the capillary) was estimated at 0.65. Three sample-to-detector distances were used (150, 250 and 350 mm) in order to combine the advantages of high resolution and extended 2 range. The detector parameters and the wavelength were calibrated with NIST LaB6. The exposure time was 60 s with a rotation of the capillary by 60°. The two-dimensional data were integrated with the Fit2D program which produced the correct intensity in relative scale. This 2D detector was used in order to perfectly define the background, to observe very weak diffraction peaks, and to improve the accuracy of the integrated intensities by achieving a better powder average. Uncertainties of the integrated intensities were calculated at each 2-point applying Poisson statistics to the intensity data, considering the geometry of the detector. The Instrument Resolution Function was determined from the LaB6 data.
-
Fe-monosulfate Ca2Fe(OH)6·(SO4)·nH2O: X-ray powder diffraction measurements were carried out using CuK radiation on a PANalytical X’Pert Pro MPD diffractometer in a –2 configuration at room temperature in the interval 3° < 2 < 120°, with a step size of 2 = 0.0167° and a counting time of 500 s for each data value. A total counting time of ~8 h was used for each sample.
-
Structure transition of Fe-Friedel salt’s Ca2Fe(OH)6·Cl·nH2O: low temperature measurements
2. EXPERIMENTAL METHOD 2.1. Syntheses -
-
Synthesis of Fe-monosulfate Ca2Fe(OH)6·(SO4)·nH2O: addition of the appropriate amount of Ca2Fe2O5 (C2F), gypsum and CaO to 50 mL of 0.4 M KOH solution with a liquid over solid ratio ~20. CaO was prepared by heating CaCO3 at 1000°C. Ca2Fe2O5 was prepared by mixing CaCO3 with Fe2O3 and burning at 1400°C and 1350°C for 24 h respectively. More details are given in thermodynamic study of the compound [6]. Synthesis of Fe-monocarbonate Ca2Fe(OH)6·(CO3)· nH2O: addition of the appropriate amount of Ca2Fe2O5 (C2F), calcite and CaO to 50 mL of 0.1 M KOH solution
Synthesis of Fe-Friedel’s salt Ca2Fe(OH)6·Cl·nH2O: addition of the appropriate amount of reactants at a liquid/solid ratio ~20. Three different synthesis protocols have been used: a) FeCl3.6H2O and CaO in 0.1M KOH (pH = 11.94), b) C2F, CaCl2.2H2O and CaO in distilled water (pH = 12.39) and c) C2F, CaCl2.2H2O, and CaO in 0.1 M KOH (pH = 12.84). More details are given in [4]. CaO was prepared by heating CaCO3 at 1000°C. Ca2Fe2O5 was prepared as previously described.
Crystal Chemistry of Iron Containing Cementitious AFm Layered Hydrates
were realized in order to investigate the structure transition of the Fe-Friedel’s salt compound. A series of diffraction patterns was recorded in the temperature range +25 °C to -50°C at interval of -5°C and in the range and 60°C – -130°C at interval of -10°C, using X-Pert Pro diffractometer (Panalytical) with Bragg-Brentano geometry, a Ni filtered CuK radiation, a Multistrip Detector Xcelerator (Panalytical) and a TTK 450 HT chamber (Anton Paar). The data are recorded between 10-50° (2) in step of 0.0167° (2) and a measuring time of 21 seconds per step. 3. RESULTS Crystal structure description of Al-based AFm hydrates have been extensively studied, and are today well described for sulfate, carbonate, chloride and nitrate anions [18-22]. Boron-containing AFm-phase has also been recently described [24]. Table 1 gathered crystallographic data of the mains Al-AFm phases. Solid solutions in these Al-AFm systems have also been extensively investigated. Contrary to former expectations of a unique crystallographic description with extensive anionic exchange, recent results indicated that solids solutions are limited in the sulfate-carbonate-chloride systems [25-27] due to the different anion-dependent space groups encountered. The crystal structure of monosulfoaluminate was first solved by Allmann in 1977 in the trigonal R3 space group [18]. This description, with the sulfate anion in the center of the interlayer region linked to main layer via hydrogen bonding, was enlarged for the whole AFm family during several decades. During the years 2000-2010, crystallographic studies have shown that each anion has its own environment in the interlayer region leading to distinct space groups (from triclinic to rhombohedral description) [20-23]. Ten years after, studies on bi-anionic compounds have allowed to obtain much more crystallographic data of AFm compounds [25,26]. These studies have shown restricted solids solutions in cement system between sulfate, carbonate and chloride [27]. For a long time, aluminum to iron substitution was simply assumed [28]. Recent studies allowed exploring the miscibility between Al3+ and Fe3+ in these systems [4]. The obtained results allow now investigating the formation of solids solution in the (Al,Fe)-AFm system taking into account crystallographic considerations. 3.1. The Ca2(Al,Fe)(OH)6·(CO3)·nH2O System: Monocarboaluminate Versus Monocarboferrate The monocarboferrate (Fe-Mc) sample was prepared at room temperature. The kinetic of reaction of Ca2Fe2O5 ferrite phase was slow and 3 years equilibration time was necessary to obtain Fe-Mc only with some traces of calcite (i.e. to observe the disappearance of Fe-hemicarbonate). Higher temperature (50°C and 80°C) increased the kinetic of formation but destabilized Fe-Mc to portlandite and hematite [7]. High quality diffraction data, using synchrotron measurement, was used to solve and refine the structure of Fe-Mc. The crystal data and multi-pattern refinement (using data from the different samples-to-detector distances) parameters have been previously published with the thermodynamic study [7]. Monocarboaluminate (Al-Mc) has a triclinic symmetry while
Current Inorganic Chemistry, 2015, Vol. 5, No. 2
3
Fe-Mc exhibits a trigonal symmetry, and 3CaO·Al2O3·CaCO3·11H2O contains less water than 3CaO·Fe2O3·CaCO3·12H2O. The main difference is in the carbonate anion environment: CO32- is directly linked to main layer in Al-Mc (connected to Ca2+), whereas CO32- is located at the center of the interlayer in Fe-Mc (with hydrogen bonding only). Fig. (1) illustrates these two distinct crystallographic situations. The general representation (Fig. 1, top) show the layered structure usually encountered for AFm phases with the main layer composed of sevenfold coordinated Ca2+ together with six-fold coordinated trivalent cations (Al3+ or Fe3+). The [Ca2Al(OH)6]+ and [Ca2Fe(OH)6] + main layers are similar. The Ca2+ cations are shifted out of the main layer central plane (defined by Al3+ cations) to a quite extent: ± 0.57 Å for Al-Mc (average value) [20] and ± 0.54 Å for Fe-Mc. The 0.03 Å difference can be attributed to the different carbonate connection mode or to the higher ionic radius of Fe3+ (0.65 Å) versus 0.55 Å for Al3+ [15], which induces a smaller distortion in the ordered rigid main layer due to smaller Ca2+/Fe3+ radius ratio. The interlayer region contains anionic species and water molecules. The slightly larger ionic radii of Fe3+ (compared to Al3+) enlarges the basal a lattice parameter (considering the hexagonal lattice), which is close to 5.92 Å for Fe-Mc (Table 2) compared to ~5.76 Å for the hexagonal lattice of Al-based AFm phases (Table 1). Carbonate anions are directly connected to the main layer in Al-Mc with an ionic bonding between one oxygen atom belonging to the carbonate and to a Ca2+ cation. The case of Fe-Mc is totally different with CO32- located at the center of the interlayer connected to main layer by a network of hydrogen bonds only. Details of the structure around carbonate anions (Fig. 1, bottom) illustrate the two different carbonate environments. For Fe-Mc, each carbonate anion is linked to six Ca-polyhedron via six hydrogen bonds between oxygen atoms belonging to carbonate and water molecule linked to Ca2+. Each oxygen atom from the carbonate group shares two hydrogen bonds with the two adjacent main layers. The structure of Al-Mc is different; the carbonate groups are also connected by three hydrogen bonds with one main layer, but with an ionic bond (see the red bond in Fig. 1, bottom left) and a hydrogen bond with the octahedrally coordinated Al3+ cation with the adjacent main layer. In addition, the carbonate groups share also hydrogen bonds with water molecules in the interlayer region. It is surprising to observe such a different carbonate environment when exchanging Al3+ cation by Fe3+. Previous study on the (Al,Fe)-Mc system has clearly evidenced the absence of iron to aluminum substitution in Al-Mc, as well as the absence of aluminum to iron substitution in Fe-Mc [7]. The Al- and Fe-end members have different symmetry with different environment for carbonate anion, such that not even a limited solid solution occurs in this system, contrary to previous reports [29]. As already reported for Al-AFm phases, Raman spectroscopy can easily distinguish between carbonate bonded to main layer via ionic bonding to Ca2+ (symmetry stretching mode at 1068 cm-1) and carbonate located at the center of interlayer bonded with hydrogen bonds only (symmetric stretching mode at 1086 cm-1) [26]. Raman spectra realized on Fe-Mc, in agreement with structural description, have shown the carbonate symmetric stretching at 1085 cm-1 corresponding to hydrogen bonding only [7].
4 Current Inorganic Chemistry, 2015, Vol. 5, No. 2
Table 1.
Renaudin et al.
Crystallographic data of Al-AFm phases. The Z number is calculated assuming the motif [Ca2Al(OH)6]·[X·nH2O]. Compound name Chemical composition
Symmetry, space group Z
Lattice parameters Unit cell volume
Year, Reference
Monosulfoaluminate 3CaO·Al2O3 ·CaSO 4·12H2O
Trigonal, R3 3
a = 5.7586 (3) Å, b = 26.7946 (12) Å, V = 769.51 (1) Å3.
1977, Allmann [18]
Triclinic, P1 2
a = 5.775 (1) Å, b = 8.469 (1) Å, c = 9.923 (3) Å, = 64.77 (2) °, = 82.75 (2) °, = 81.43 (2) °, V = 433.0 (2) Å3.
1998, François et al. [20]
Disordered monocarboaluminate D-3CaO·Al2O3 ·CaCO 3·11H2O
Triclinic, P1 2
a = 5.7422 (4) Å, b = 5.7444 (4) Å, c = 15.091 (3) Å, = 92.29 (1) °, = 87.45 (1) °, = 119.547 (7) °, V = 432.5 (41) Å3.
1999, Renaudin et al. [21]
AFm-(NO3-) 3CaO·Al2O3 ·Ca(NO3) 2·10H2O
Trigonal, P3c1 2
a = 5.7445 (8) Å, b = 17.235 (5) Å, V = 492.55 (16) Å3.
1999, Renaudin et al. [23]
a = 9.960 (4) Å, b = 5.7320 (2) Å, c = 16.268 (7) Å, = 104.471 (2) °, V = 432.5 (41) Å3.
2002, Rapin et al. [22]
Ordered monocarboaluminate O-3CaO·Al2O3 ·CaCO 3·11H2O
Friedel’s salt 3CaO·Al2O3 ·CaCl2·10H2O
Monoclinic, C2/c 2
AFm-(Cl-,CO3 2-) 3CaO·Al2O3 ·CaCO3 ·CaCl2·11H2O
Trigonal, R3c 6
a = 5.7557 (1) Å, b = 46.947 (1) Å, V = 1346.90 (5) Å3.
2011, Mesbah et al. [26]
Kuzel’s salt 3CaO·Al2O3 ·CaSO 4·CaCl2·11H2O
Trigonal, R3 6
a = 5.7508 (2) Å, b = 50.418 (3) Å, V = 1444.04 (11) Å3.
2011, Mesbah et al. [25]
Hemicarboaluminate 3CaO·Al2O3 ·CaCO3 ·Ca(OH)2 ·10H2O
Trigonal, R3c 6
a = 5.77570 (10) Å, b = 48.812 (2) Å, V = 1410.15 (8) Å3.
2012, Runcevski et al. [19]
3.2. The Ca2(Al,Fe)(OH)6·(SO4)·nH2O System: Monosulfoaluminate Versus Monosulfoferrate The kinetic of monosulfoferrate (Fe-Ms) precipitation increases with temperature. Both at 50°C and at 80°C Fe-Ms is formed within 7 days, but at 80°C Fe-Ms decomposes to portlandite and hematite after 28 days. Crystal structure investigations were performed on two well-crystalline samples: one synthesized at room temperature for two years and one synthesized at 50°C for one year [6]. The crystal structure of Fe-Ms is isostructural to monosulfoaluminate (AlMs); the first AFm crystal structure solved by Allmann [18] in 1977. The representation of its structure in Fig. 2a is valid for both Fe-Ms and Al-Ms compounds. The structure is trigonal, described in the R3 space group (Table 2), with two kinds of disorder around the anionic sulfate group: i) statistical disorder between one sulfate and three water
molecules and ii) orientation up/down disorder of the sulfate anion. In agreement with the larger ionic radii of Fe3+, the a basal lattice parameter increases from Al-Ms (5.76 Å) to FeMs (5.89 Å). In contrast the interlayer distance (or the hexagonal c lattice parameter) decreases from 8.932 Å for AlMs to 8.875 Å for Fe-Ms, although the same amount of water has been observed in the interlayer region: 3CaO·Al2O3·CaSO4·12H2O and 3CaO·Fe2O3·CaSO4·12H2O. The decrease of interlayer distance indicates stronger hydrogen bonding around SO42- in the case of Fe-Mc. The strengthening of hydrogen bond network around sulfate group (greys dotted links in Fig. 2a) in Fe-Ms is illustrated by Raman spectroscopy (Fig. 2b) by the shift of the [SO4] 1 mode toward higher Raman shift (992 cm-1 for Fe-Ms, compared to 982 cm-1 for Al-Ms) and a smaller Raman shift for the intense hydrogen bonds vibration (3640 cm-1 for Fe-Ms, compared to 3688 cm-1 for Al-Ms). Despite the same crystal
Crystal Chemistry of Iron Containing Cementitious AFm Layered Hydrates
Current Inorganic Chemistry, 2015, Vol. 5, No. 2
5
Fig. (1). General representations (Top) of the crystal structure of Al-Mc (left) and Fe-Mc (right), and details showing the carbonate anion environment (bottom).
Fig. (2). Representation of the crystal structure of Fe-Ms (a) and parts of the Raman spectra for Al-Ms and Fe-Ms (b). Hydrogen bonds are represented by grey dotted links.
structure of both end members, a miscibility gap has been evidenced in the range 0.45 < Al/(Al+Fe) < 0.95 [6], again contrary to previous findings indicating an ideal solid solution [30]. Aluminum to iron substitution is only observed in Fe-Ms with the following general formulae [Ca2(Fe1xAlx)(OH)6]·[(SO4)·3H2O] with 0 x < and only a very limited iron to aluminium substitution is possible for the AlMs end member. The origin of the existence of the miscibility gap is difficult to explain as both solids present the same crystallographic structure. Whatever, such observation clearly indicates that iron cannot simply be considered as an aluminium substitution in cement hydrates.
3.3. The Ca2(Al,Fe)(OH)6·Cl·nH2O System: Al-Friedel’s Salt Versus Fe-Friedel’s Salt Fe-Friedel’s salt started to form from C2F after 7 days equilibration time. Portlandite and Fe-hydroxide coprecipitate with Fe-Friedel’s salt as observed for the two previous Fe-AFm phases. The presence of portlandite and calcite is detected on powder diffraction patterns after 3 years of equilibration time. The sharp peaks of the FeFriedel’s salt indicate the presence of well crystalline solid. Fig. (3a) shows that the XRD patterns of Fe-Friedel’s salt synthesized at different pH values (pH = 11.94, 12.39 and
6 Current Inorganic Chemistry, 2015, Vol. 5, No. 2
Table 2.
Renaudin et al.
Crystallographic data of Fe-AFm phases. The Z number is calculated assuming the motif: [Ca2Fe(OH)6]·[X·nH2O].
Compound name Chemical composition
Symmetry, space group Z
Lattice parameters Unit cell volume
Year, Reference
Fe-sulfate 3CaO·Fe2O3 ·CaSO 4·12H2O
Trigonal, R3 3
a = 5.8864 (2) Å, c = 26.614 (2) Å, V = 798.62 (6) Å3.
2012, Dilnesa et al. [6]
Fe-carbonate 3CaO·Fe2O3 ·CaCO3 ·12H2O
Trigonal, R3c 6
a = 5.9196 (1) Å, c = 47.8796 (10) Å, V = 1453.01 (4) Å3.
2011, Dilnesa et al. [7]
Fe-Friedel’s salt 3CaO·Fe2O3 ·CaCl2·10H2O
Trigonal, R3 3
a = 5.8262 (5) Å, c = 23.417 (1) Å, V = 688.4 (1) Å3.
2002, Rousselot et al. [31]
Fe-Friedel’s salt 3CaO·Fe2O3 ·CaCl2·10H2O
Trigonal, R3 3
a = 5.9000 (3) Å, c = 23.740 (6) Å, V = 715.7 (3) Å3.
2001, Rapin thesis [32]
Fig. (3). a: X-ray powder pattern ( = 1.5418 Å) from Fe-Friedel’s salt synthesized at different pH values, b: Representation of the crystal structure of Fe-Friedel’s salt, and c: Rietveld plot ( = 0.697751 Å) together with Raman spectrum showing the carbonate contamination. Hydrogen bonds are represented by grey dotted links.
12.84) are very similar. At higher pH values, however, a slight peak shift of the main diffraction peak towards higher 2 values is observed, which could indicate the uptake of additional hydroxide – or carbonate contamination – in the interlayer of Fe-Friedel’s salt. The solid synthesized at pH = 12.4 equilibrated for 500 days was used for the crystallographic investigation. Multipattern Rietveld refinement was performed (see Rietveld plot in Fig. 3c) to solve the structure of Fe-Friedel’s salt. In agreement with previous studies, the crystal structure of Fe-Friedel’s salt was described in the rhombohedral R3 space group [31,32] with refined lattice parameters a = 5.8567 (2) Å and c = 23.314 (1) Å (V = 692.57 (5) Å3). A general representation of the structure is shown in Fig. (3b), the structure parameters are gathered in Table 3. The Fe-Friedel’s salt is composed of a positively charged main layer [Ca2Fe(OH)6]+ with a negatively charged interlayer [Cl·2H2O]- corresponding to the refined composi-
tion 3CaO·Fe2O3·CaCl2·10H2O (similarly to the well-known Al-Friedel’s salt). The large Biso value observed for the interlayer species (chloride and water molecules) can be explained either by dynamical disorder within the interlayer region or by carbonate contamination which imply carbonate to chloride substitution in the 3b site. The carbonate contamination has been evidenced by Raman spectroscopy (Fig. 3c top) with the [CO3] symmetric stretching observed at 1088 cm-1 corresponding to carbonate anion located at the center of the interlayer region [26]. The structure of Al-Friedel’s salt has been investigated by Terzis et al. [33] with the monoclinic C2/c symmetry, and revisited by Rapin et al. [22,34] with the same symmetry. This monoclinic description corresponds to the low temperature LT-polymorph of Friedel’s salt which transforms to the rhombohedral R3c HT-polymorph above 37 °C [22,34]. The temperature of transition is dependent on the carbonate
Crystal Chemistry of Iron Containing Cementitious AFm Layered Hydrates
Table 3.
Current Inorganic Chemistry, 2015, Vol. 5, No. 2
7
Refined structure parameters of Fe-Friedel’s salt (standards deviation are indicated in parentheses): a = 5.8567 (2) Å and c = 23.314 (1) Å (V = 692.57 (5) Å3).
Atom
Wyckoff
x
y
z
Biso (Å3 )
Occupancy
Ca
6c
2/3
1/3
0.0255 (2)
1.2 (1)
1
Fe
3a
0
0
0
1.2 (1)
1
O (OH)
18f
0.270 (1)
-0.053 (1)
0.0445 (3)
0.9 (2)
1
Cl
3b
0
0
1/2
7.0 (4)
1
O (H2O)
6c
2/3
1/3
0.1355 (5)
4.0 (4)
1
Fig. (4). In-situ low temperature PXRD measurement from +25°C down to -130°C on Fe-Friedel’s salt sample ( = 1.5418 Å).
contamination; the transition temperature decreases if carbonate substitutes chloride in the interlayer [35]. Small carbonate contamination during synthesis (from atmospheric CO2) is sufficient to observe at room temperature the HTrhombohedral polymorph. The co-precipitation method to prepare Al-Friedel’s salt resulted in a new R3 description (instead of R3c ) [31]. These two polytypes, labeled 3R and 6R for the R3 and the R3c descriptions, have also been observed for the bromide analog [36]. Whereas the coprecipitation method was not use here to prepare FeFriedel’s salt, its structure corresponds to the 3R HTpolytype. Low temperature in-situ XRD measurements have been performed (between +25°C and -130°C) in order to complete the comparison between Fe- and Al-Friedel’s salt. Fig. (4) shows the 24 temperature dependent powder patterns. At room temperature the sample is mainly composed of Fe-
Friedel’s salt (about 90 wt %) with impurities of calcite and portlandite. Below -45°C ice is visible in the powder pattern, which becomes more and more important down to -130°C. The ice formation is due to an anomalous leak in the low temperature chamber, nevertheless did not perturb our sample characterization. Careful analysis of powder patterns shown in Fig. (4) evidences small modifications during the temperature decrease; especially diffraction peaks at about 38.5° and 42.5° showing enlargement and shift when the temperature decreases below -15°C. The details shown in Fig. (5) illustrate the modification appearing at around 15°C: a shift towards lower angles together with a broadening of the diffraction peaks during cooling. The transition is not well described by the appearance of new superstructure diffraction peaks. Nevertheless the sharp shift toward low angles, indicating the increase of unit cell volume during cooling, can be explained only by a structural transition. A
8 Current Inorganic Chemistry, 2015, Vol. 5, No. 2
Renaudin et al.
Fig. (5). Details on the in-situ low temperature PXRD measurement from +25°C down to -130°C showing the evolution of the intense diffraction peaks (003), (006) and (104) of Fe-Friedel’s salt ( = 1.5418 Å). The bold line corresponds to the powder pattern recorded at -15°C, where the transition is observed during cooling.
Fig. (6). Temperature dependences of the (pseudo-)hexagonal a (top left) and c (top right) lattice parameters, the corresponding unit cell volume (bottom left) and the interlayer distance (bottom left).
Crystal Chemistry of Iron Containing Cementitious AFm Layered Hydrates
DSC measurement on the sample, not shown here, has shown an endothermic signal at around -15°C. Rietveld refinements of the lattice parameters, assuming a continuous hexagonal lattice, were performed to follow their temperature dependence (Fig. 6). The two curves illustrating the changes during heating and cooling show the reversible feature of the transition. The c lattice parameter clearly increases during cooling at around -15°C leading to an increase of ~0.06 Å of the interlayer distance around the temperature of transition. The evolution of the a lattice parameter is not so distinct, although a small increase is also observed during cooling with an incident around -20°C. These observations on the lattice show a structural transition at around -15°C, although the resolution of the powder patterns did not allow elucidating the symmetry of the low temperature Fe-Friedel’s salt polymorph. By comparison with observation made on the iodide AFm compound [Ca2Al(OH)6]·[I·2H2O] [36], we assumed for Fe-Friedel’s salt a triclinic LT-polymorph. Previous crystallographic studies performed on a series of halide Al-AFm (Cl-, Br- and I-) have shown that all R3c HT-polymorphs (6R) present a monoclinic C2/c LT-polymorph, whereas the R3 HTpolymorph (3R) of the iodide compound presents a triclinic LT-polymorph (structure not solved yet). High quality low temperature synchrotron measurements are needed to pursue the crystallographic study of this triclinic LT-polymorph. Present results on Fe-Friedel’s salt show that the temperature dependent phase transition, well known for Al-Friedel’s salt, remains when iron substitutes aluminum. This phase transition corresponds to an order-disorder transition of the hydrogen bond network around the halide anion [36]. Nevertheless the reason of the existence of two couples of polymorphs ( R3c that transforms in C2/c, and R3 that transforms in P 1 ) is not yet clear; these two couples of polymorphs seems to be synthesis dependent, halide anion dependent, and also trivalent cation (Al3+/Fe3+) dependent.
Current Inorganic Chemistry, 2015, Vol. 5, No. 2
mined the hydrogen bond network and the cohesion between the main layers; i.e. the interlayer region commands the crystal symmetry (from high rhombohedral symmetry to low triclinic symmetry). Present observations on Fe-AFm compounds show now that the nature of the trivalent cation, which is contained at the center of the rigid main layer, has also an impact on the crystal symmetry of the compound. This is certainly also due to interaction with the hydrogen bond network, but is less evident to understand as the size difference between Al3+ and Fe3+ cations is relatively small. The present paper shows the complexity of the crystallographic description of the AFm phases. All the phases discussed here could be present during the hydration of Portland cements as well as various solid solutions. This complexity has an impact not only on the hydration process, but also on the relative fraction of AFm and AFt phases and thus on the final properties of the concrete. This indicates the importance to well understand such structural differences to correctly predict the hydrates assemblage of concrete over time. Thermodynamic calculations of such complicated systems need reliable thermodynamic data of the different hydrates formed [4] based on good description and understanding of the different structures. CONFLICT OF INTEREST The authors confirm that this article content has no conflict of interest. ACKNOWLEDGEMENTS The authors thank SNBL-ESRF (Yaroslav Filinchuk) for the in-house beam time allocation used for the characterization of Fe-carbonate and Fe-Friedel’s salt (HT-polymorph). REFERENCES [1]
CONCLUSION The present paper illustrates the crystal chemical complexity of the AFm family (hydrocalumite type structure). Behind the apparent simplicity of the main structural features valuable for all compounds belonging to the AFm family numerous fine crystallographic descriptions are hidden. AFm compounds are based on a double layer structure: a main layer with the composition [Ca2M(OH)6]+, M3+ = Al3+ and Fe3+, and an interlayer with [X·nH2O]-, X- = SO42-, CO32and Cl- and n = 2, 3 and 4 for the compounds discussed here. The main layer is extremely rigid with the ordering of the bivalent and trivalent cations due to the large difference in ionic radius, in comparison with the LDH family (Layered Double Hydroxide with the hydrotalcite type structure for which bivalent and trivalent cations are disordered in a unique crystallographic site). On the other hand the interlayer region with anionic species and water molecules appears more flexible with different possible disorders (dynamic, statistic, orientation). Few decades ago, all AFm compounds were supposed to exhibit the same rhombohedral structure as Al-monosulfate, the first AFm crystal structure solved [18]. The work performed on Al-AFm compound has shown that the interlayer region is not at all that flexible. The interlayer description depends on the anion which deter-
9
[2] [3]
[4]
[5] [6]
[7] [8]
[9] [10]
Lothenbach, B., Matschei, T., Möschner, G., Glasser, F.P. Thermodynamic modelling of the effect of temperature on the hydration and porosity of Portland cement, Cem. Concr. Res., 2008, 38, 1-18. Lothenbach, B., Winnefeld, F. Thermodynamic modelling of the hydration of Portland Cement, Cem. Concr. Res., 2006, 36, 209226. Matschei, T., Lothenbach, B., Glasser, F.P. Thermodynamic properties of Portland cement hydrates in the system CaO-Al2O3-SiO 2CaSO4-CaCO 3-H2 O, Cem. Concr. Res., 2007, 37 1379-1410. Dilnesa, B.Z. Fe-containing hydrates and their fate during cement hydration: thermodynamique data and experimental study, PhD Thesis N° 5262, Ecole Polytechnique Fédérale de Lausanne, Switzerland, 2011. Dilnesa, B.Z., Wieland, E., Lothenbach, B., Dähn, R., Scrivener, K.L. Fe-containing phases in hydrated cements, Cem. Concr. Res., 2014, 58, 45-55. Dilnesa, B.Z., Lothenbach, B., Renaudin, G., Wichser, A., Wieland, E. Stability of monosulfate in the presence of iron, J. Am. Ceram. Soc., 2012, 95(10), 3305-3316. Dilnesa, B.Z., Lothenbach, B., Le Saout, G., Renaudin, G., Mesbah, A., Filinchuk, Y., Wichser, A., Wieland, E. Iron in carbonate containing AFm phases, Cem. Concr. Res., 2011, 41, 311-323. Dilnesa, B.Z., Lothenbach, B., Renaudin, G., Wichser, A., Kulik, D. Synthesis and characterization of hydrogarnet Ca3(AlxFe1x)2(SiO4) y(OH)4(3-y), Cem. Concr. Res., 2014, 59, 96-111. Rogers, D.E., Aldridge, L.P. Hydrates of calcium ferrites and calcium aluminoferrites, Cem. Concr. Res., 1977, 7, 399-410. Fortune, J.M., Coey, J.M.D. Hydration products of calcium aluminoferrite, Cem. Concr. Res., 1983, 13, 696-702.
10 Current Inorganic Chemistry, 2015, Vol. 5, No. 2 [11]
[12] [13] [14]
[15] [16] [17]
[18] [19]
[20]
[21] [22]
[23]
[24]
[25]
Renaudin et al.
Fukuhara, M., Goto, S., Asaga, K., Daimon, M., Kondo, R. Mechanisms and kinetics of C4AF hydration with gypsum Cem. Concr. Res., 1981, 11, 407-414. Collier, N., Milestone, N.B., Hill, J., Godfrey, I.H. The disposal of radioactive ferric floc Waste Manag., 2006, 26, 769-775. Gallucci, E., Mathur, P., Scrivener, K. Microstructural development of early age hydration shells around cement grains Cem. Concr. Res., 2010, 40, 4-13. Harchand, K.S., Kumar, R., Chandra, K., Vishwamittar, Mössbauer and X-ray investigations of some Portland cements, Cem. Concr. Res., 1984, 14, 170-176. Shannon, R. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides, Acta Cryst. A, 1976, A32(5), 751-767. Aïtcin, P.C. Cement of yesterday and today, Concrete of tomorrow, Cem. Conc. Res., 2000, 30, 1349-1359. Shi, C., Ferandez Jimenez, A., Palomo, A. New cements for th 21st century: the pursuit of an alternative to Portland cement, Cem. Conc. Res., 2011, 41, 750-763. Allmann, R. Refinement of the hybrid layer structure [Ca2Al(OH)6 ] +·[SO4·3H2O]-, Neues Jahrb. Mineral. Monatsch., 1977, 3, 136-144. Runcevski, T., Dinnebier, R.E., Magdysyuk, O.V., Pollmann, H. Crystal structures of calcium hemicarboaluminate and carbonated calcium hemicarboaluminate from synchrotron powder diffraction data, Acta Cryst. B., 2012, B68, 493-500. François, M., Renaudin, G., Evrard, O. A cimentitious compound with composition 3CaO·Al2O3 ·CaCO3 ·11H2O, Acta Cryst. C, 1998, C54(9), 1214-1217. Renaudin, G., François, M., Evrard, O. Order and disorder in the lamellar hydrated tetracalcium monocarboaluminate compound, Cem. Concr. Res., 1999, 29(1), 63-69. Rapin, J.P., Renaudin, G., Elkaim, E., François, M. Structural transition of Friedel’s Salt 3CaO·Al2O3·CaCl2·10H2O studied by synchrotron powder diffraction, Cem. Concr. Res., 2002, 32(4), 513519. Renaudin, G., François, M. The lamellar double hydroxide (LDH) compound with composition 3CaO·Al2O3·Ca(NO 3)2 ·10H2O, Acta Cryst. C, 1999, C55(6), 835-838. Champenois, J.B., Mesbah, A., Cau-Dit-Coumes, C., Renaudin, G., Leroux, F., Mercier, C., Revel, B., Damidot, D. Crystal structures of boro-AFm and boro-AFt phases, Cem. Concr. Res., 2012, 42, 1362-1370. Mesbah, A., François, M., Cau-dit-Coumes, C., Frizon, F., Filinchuk, Y., Leroux, F., Ravaux, J., Renaudin, G. Crystal structure
Received: September 24, 2014
[26] [27]
[28] [29]
[30]
[31]
[32]
[33] [34]
[35]
[36]
of Kuzel’s salt CaO·Al2O3 ·1/2CaSO4·1/2CaCl2·11H2O determined by synchrotron powder diffraction, Cem. Concr. Res., 2011, 41, 504-509. Mesbah, A., Cau-dit-Coumes, C., Frizon, F., Leroux, F., Ravaux, J., Renaudin, G. A new investigation of the Cl--CO32- substitution in AFm phases, J. Am. Ceram. Soc., 2011, 94, 1901-1910. Mesbah, A., Cau-Dit-Coumes, C., Renaudin, G., Frizon, F., Leroux, F. Uptake of chloride and carbonate by calcium monosulfoaluminate, Cem. Concr. Res., 2012, 42, 1157-1165. Taylor, H.F.W. Cement Chemistry, Academic Press Limited, Second printing, London 1992. Ecker, M., Pöllmann, H.J. Synthesis and characterisation of CO3containing Ca-Al-Fe-hydrates, Proceeding of the 10th International Cogress on the Chemistry of Cement, Goteborg, Sweden, 1997, 2ii032. Ecker, M. Diadochiebeziehungen in Calciumaluminatferraten und Deren Hydrationsprodukten, PhD Thesis, Hallesches Jahrbuch für Geowissenschaften, Rheihe B, Martin-Luther-Universität HalleWittenberg, 1998. Rousselot, I., Taviot-Gueho, C., Leroux, F., Leone, P., Palvadeau, P., Besse, J.P. Insights on the structural chemistry of hydrocalumite and hydrotalcite-like materials: Investigation of the series Ca2M3+(OH)6 Cl·2H2O (M3+: Al3+, Ga3+, Fe 3+, and Sc3+) by X-ray powder diffraction, J. Solid State Chem., 2002, 167, 137-144. Rapin, J.P. Synthèse et étude cristallochimique de quelques aluminates et ferrites calciques hydratés de formule [Ca2M(OH)6] +·[X,nH2O]- avec X = Cl, Br, I, ClO4 , CO 3SO 4, CrO4SO 4 et M = Al, Fe, PhD Thesis, Nancy Université, France, 2001. Terzis, A., Filippakis, S., Kuzel, H.J., Bruzlaff, H. The Crystal Structure of Ca2Al(OH)6Cl·2H2O, Zeitschr. Kristallogr., 1987, 181(1–4), 29-34. Renaudin, G., Kubel, F., Rivera, J.P., Francois, M. Structural Phase Transition and High Temperature Phase Structure of Friedels Salt, 3CaO·Al2O3 ·CaCl2·10H2O, Cem. Concr. Res., 1999, 29(12), 19371942. Mesbah, A., Rapin, J.P., François, M., Cau-dit-Coumes, C., Frizon, F., Leroux F., Renaudin, G. Crystal structures and phase transition of cementitious bi-anionic AFm-(Cl-,CO 32-) compounds, J. Amer. Ceram. Soc., 2011, 94(1), 261-268. Renaudin, G., Rapin, J.P., Elkaim, E., François, M. Polytypes and polymorphs in the related Friedel’s salt [Ca2Al(OH)6] +[X·2H2O]halide series, Cem. Concr. Res., 2004, 34(10), 1845-1852.
Revised: January 15, 2015
Accepted: January 16, 2015
