Calcium-Silicate Hydrates Containing Aluminium: C-A-S-H II

2nd WORKSHOP Calcium-Silicate Hydrates Containing Aluminium: C-A-S-H II

Empa, Dübendorf, Switzerland April 23 – 24, 2018 Calcium-Silicate Hydrates Containing Aluminium: C-A-S-H II

Abstracts are ordered alphabetically according to the presentating author

Presentation author Title of the abstracts Page number

Andalibi, M. Reya On the mesoscale mechanism of synthetic calcium–silicate–hydrate precipitation: a population balance modeling approach ...... 1 Barzgar, Sonya Effect of aluminium on C-S-H structure, stability and solubility ...... 2 Bellmann, Frank Analysis of C-S-H growth rates in supersaturated conditions ...... 3 Bernard, Ellina Effect of magnesium on C-A-S-H ...... 4 Blanc, Philippe Thermodynamic properties of C-S-H, C-A-S-H and M-S-H phases: results from direct measurements and predictive modelling ...... 5 Churakov, Sergey Molecular level insight into ions uptale by C-(A)-S-H ...... 6 d'Espinose, Jean-Baptiste Molecular understanding of tricalcium silicate hydration in absence and in presence of aluminate ions ...... 7 Dufrêche, Jean-François Multi-scale modelling of silica interfaces: the role of the charges ...... 8 Fernandez-Martinez, Alejandro C-S-H nucleation pathways ...... 9 Geng, Guoqing Studying the intrinsic mechanical properties of calcium (alumino)silicate hydrate using synchrotron-radiation-based high-pressure x-ray diffraction ....10 Grangeon, Sylvain Structure and reactivity of nanocrystalline calcium silicate hydrates: the parallel with clay minerals ...... 11 Hay, Rotana Effect of aluminum inclusion on morphology and mechanical properties of calcium alumino silicate hydrate ...... 13 Heinz, Hendrik A C-S-H builder and interface modeling tools towards accurate reactive full electrolyte simulations of cement materials up to the micrometer scale ...... 14 John, Elisabeth The impact of the calcium to silicon ratio of C-S-H crystall seeds ...... 15 Kalinichev, Andrey Atomistic modeling of the structure and dynamics of water and ions in calcium silicate hydrate and calcium aluminate hydrate phases ...... 16 Kangni-Foli, Ekoe Carbonation of low calcium C-A-S-H ...... 17 Ke, Xinyuan Chloride sorption onto C-(N)-A-S-H gels as a function of Ca/Si and Al/Si ratios ...... 18 Kulik, Dmitrii C-A-S-H solid solutions: multisite vs quasichemical? ...... 19 Kunhi Mohamed, Aslam Atomic structure of Calcium Silicate Hydrate ...... 20 Labbez, Christophe From nucleation to particle assemblage: what can we learn from mesoscopic simulations ? ...... 21 Lefèvre, Grégory Probing the solid/solution interface by in situ real-time infrared spectroscopy ...... 22

Li, Jiaqi Synchrotron-based characterization of the chemistry and structure of calcium (alumino) silicate hydrate ...... 23 Liu, Sanheng Effect of Al on the dissolution rates of glass in hyperalkaline solutions ...... 24 Lothenbach, Barbara Calcium silicate hydrates with aluminium ...... 25 Mancini, Andrea Uptake of aluminium and iron by C-S-H ...... 26 Miron, Dan Extending GEMSFITS to THERMOFIT for using THERMOEXP experimental database for C-(A)-S-H ...... 27 Miron, Dan A PHREEQC version of CEMDATA'18 generated using ThermoMatch ...... 28 Naber, Christoph C-S-H precipitation kinetics in C3S paste and suspension ...... 29 Nguyen Tuan, Long Simulation of growing C-S-H phases using 3D sheet growth model ...... 30 Palacios, Marta New insights into the early reaction of alkaliactivated slag cements ...... 31 Patel, Ravi Ajitbhai Multiscale modeling of ion transport in cementitous system: Surface charge effects ...... 32 Plusquellec, Gilles Interaction between C-(A-)S-H and anions ...... 33 Prentice, Dale Thermodynamic modelling of synthetic C(-A)-S-H using the pitzer model – what is gained? ...... 34 Rößler, Christiane Characterisation of C-(A)-S-H phases using selected area diffraction in the TEM ...... 35 Siramont, Jirawan Synthetic Calcium Silicate Hydrate –formation kinetics – the key to understanding cement microstructures ...... 36 Steindl, Florian Roman Uptake of heavy metal ions during C-S-H precipitation ...... 37 Taube, Franziska Investigations on the intercalation of An(III)/Ln(III)-malate complexes in CSH phases ...... 38 Veryazov, Valera Multiscale (force field - semiempirical - density functional - ab initio) modelling of C-A-S-H ...... 39 Walker, Colin C-S-H gel solubility modeling at high temperatures ...... 40 Wolter, Jan-Martin Stability of U(VI) and CM(III) doped calcium silicate hydrate phases in high saline brines ...... 41 Yamada, Kazuo Alkali uptake evaluations of C-A-S-H with structure analysis by NMR and degradated OPC paste ...... 42 Yan, Yiru Structural and mechanical property determination of calcium silicate hydrate (C-S-H): a theoretical and experimental study ...... 43 Yang, Sheng-Yu The role of Aluminum in Calcium Silicate Hydrate Phases: A Multinuclear Solid-State NMR investigation ...... 44 Yin, Chennying Influence of calcium to silica ratio on H2 gas production in calcium silicate hydrate ...... 45

On the mesoscale mechanism of synthetic calcium–silicate–hydrate precipitation: a population balance modeling approach

Andalibi, M. R.1,2), Bowen, P.2) , Testino, A.1) 1) Energy and Environment Research Department, Paul Scherrer Institute (PSI), Villigen-PSI, Switzerland 2) Materials Institute, Powder Technology Laboratory (LTP), EPFL, Lausanne, Switzerland Corresponding author: Andalibi, M. R., email: [email protected]

Calcium–silicate–hydrate (C–S–H) is the most important product of cement hydration. Despite this importance, its formation mechanism is not well-understood. Here, we describe the novel application of a coupled thermodynamic-kinetic computational model based on a population balance equation in order to unravel the overall mechanism of synthetic C–S–H precipitation. The framework, embracing primary nucleation, true secondary nucleation, and molecular growth as the constituting sub-processes, is regressed to experimental Ca2+(aq) concentration vs. time data collected on a model synthetic C–S–H with Ca : Si = 2. Upon the critical appraisal of the model's adjustable parameters, which turn out to adopt rational values, simulations were performed to estimate various characteristics of the aforementioned model system (e.g., the kinetic speciation during the precipitation process, or the mechanisms and activation free energies of nucleation and growth phenomena). We mechanistically account for the evolution of the C–S–H mesostructure which is made up of defective crystallites around 3–6 nm thick, nematically packing together in two dimensions giving rise to foil- like polycrystalline particles around 100 nm in breadth, close to the experimentally observed values. The computational framework is generic and can be applied to other precipitation systems and cement hydration scenarios.

Figure 1: Schematic representation of C-S-H structure, from atomistic- to particle scale (~ 100 nm), and the proposed precipitation pathway (on the right); the differential equation at the bottom is the population balance equation which constitutes the core component of our computational framework.

Effect of aluminium on C-S-H structure, stability and solubility

Barzgar, S.1,2), Lothenbach, B.1), Ludwig, C. 2,3) 1) Empa, Laboratory of Concrete/ Construction Chemistry, 8610 Dübendorf, Switzerland 2) École Polytechnique Fédéral de Lausanne (EPFL), ENAC IIE GR-LUD, 1015 Lausanne, Switzerland 3) Paul Scherrer Institute (PSI), ENE LBK CPM, 5232 Villigen PSI, Switzerland

The production of cement is responsible for approximately 5% of man-made CO2 emissions. The replacement of Portland cement by supplementary cementitious materials (SCM) such as blast furnace slag, fly ash or calcined clay offers a very high potential to reduce these CO2 emissions. Due to the reaction of these SCM with Portland cement during hydration, additional calcium silicate hydrates (C-S-H) are formed [1]. The high Al2O3 and SiO2 content of the SCMs results in C-S-H compositions with more silicon and aluminium than in Portland cement which affects stability and durability of such cements as well as the sorption behavior of toxic heavy metals which might be present in SCM [2]. Therefore, it is crucial to determine the role of aluminium on C-S-H properties to predict the formed hydrate phase assemblages and their effects on durability. In this study, aluminium sorption isotherms including very low Al concentrations have been determined for C-S-H with different Ca/Si ratios. Sorption isotherms at different Ca/Si and Al/Si ratios enable us to determine the relative stability and capacity of different sorption sites in C-S-H. Moreover, the effect of alkali hydroxides on the aluminium uptake is investigated by ICP-MS and IC. IC and ICP results confirmed that the presence of sodium hydroxide leads to higher dissolved aluminium concentrations, which favors the aluminium uptake in C-S-H. Moreover, increasing the Al concentrations leads to more silicon concentrations in aqueous solution.

References: [1] Lothenbach, B., Nonat, A. Calcium silicate hydrates: Solid and liquid phase composition. Cement and Concrete Research, (2015) 57–70. [2] Chen, Q.Y., Tyrer, M., Hills, C.D., Yang, X.M., Carey, P. Immobilisation of heavy metal in cement-based solidification/stabilisation: A review. Waste Management 29 (2009) 390–403.

Analysis of C-S-H growth rates in supersaturated conditions

Bellmann, F.1), Scherer, G. W.2) 1) Bauhaus University Weimar, Coudraystrasse 11, 99423 Weimar, Germany 2) Princeton University, Eng. Quad. E-208, Princeton, NJ 08544, USA Corresponding author: Bellmann, F., email: [email protected]

The growth rate of calcium-silicate-hydrate (C-S-H) was analyzed by following the evolution of calcium and silicon concentrations in supersaturated solutions. In these experiments, the supersaturated solution was produced by mixing saturated calcium hydroxide solution and a solution obtained from the hydration of tricalcium silicate. A continuous decrease of the silicon concentration over time was observed during the experiments and the C-S-H formation rate was calculated from the amount of silicon that was precipitated between two consecutive analyses.

The data obtained in this study demonstrate that the growth rate of C-S-H depends mainly on the supersaturation with respect to this phase (Figure 1), the availability of calcite as a substrate for heterogeneous nucleation and the calcium concentration in solution. A mean value of approximately 11 nmol of C-S-H per m² per second was obtained for the interfacial growth rate of C-S-H in conditions that are relevant for the formation of tricalcium silicate.

The interfacial growth rates obtained in this study can be used for modelling C3S hydration kinetics.

Figure 1: C-S-H growth rate perpendicular to the substrate, G2, as a function of supersaturation

Effect of magnesium on C-A-S-H

Bernard, E.1), Lothenbach, B.2) 1) University of Bern, Institute of Geological Sciences, 3012 Bern, Switzerland 2) Empa, Laboratory for Concrete & Construction Chemistry, 8600 Dübendorf, Switzerland Corresponding author: Bernard, E., email: [email protected]

The pH of cementitious cement pastes containing Portland and silica rich SCMs such as fly ash or slags might be lower than those of typical Portland cement paste due to high silica contents. Such blends also contain less portlandite due to the pozzolanic reaction and a larger amount of aluminium which has been observed taken up in the calcium silicate hydrate (C-S-H) phases. The presence of magnesium and aluminium in the starting cementitious materials can lead to the precipitation of the magnesium and a part of the aluminium in hydrotalcite-like phases [1]. This study investigated the potential uptake of magnesium and aluminium in C-S-H phase in the context of low pH cements and also the stability of C-A-S-H and M-A-S-H versus hydrotalcite. The addition of reactive MgO to synthetic C-S-H (Ca/Si=0.8) suspensions increases the pH while C-S-H with lower polymerization degree and thus, a higher Ca/Si, is formed [2]. In the absence of aluminium, mainly brucite was observed by TGA plus a small amount of M-S-H by 29Si solid state NMR [2]. In this study, adding jointly MgO and metakaolin as aluminium source decreased the brucite content. TGA data showed the absence of hydrotalcite-like phase or amorphous aluminium hydroxide indicating an aluminium incorporation in the silicate phases. 29Si solid state NMR data confirmed the incorporation of aluminium in the C-A-S-H phases and the formation of more M-A-S-H than in a system without aluminium (Figure 1). The direct addition of hydrotalcite to C-S-H showed little reaction. Indeed, a large content of hydrotalcite was still observed by TGA and only few Q3 signals characteristic of M-(A-)S-H were detected by 29Si solid state NMR compared to the first experiment (Figure 1). Additionally, less aluminium seems to be incorporated in the C-A-S-H phases. The addition of hydrotalcite did also not change the pH of the solution of a C-A-S-H suspension (Al/Si=0.1) (~ 10.7); the stability of the C-S-H and hydrotalcite system seems to be mainly depenmiront of the pH of the solution.

Figure 1: 29Si SS MAS NMR spectra of C-S-H samples where MgO, MgO and metakaolin, and hydrotalcite have been added (1 year, at 20 °C), (C-S-H and C-A-S-H shown as references), aC-A-S-H samples synthesized in presence of 0.1M of NaOH (from the Ph.D. project of A. Di Giacomo, Empa).

References [1] A. Machner, M. Zajac, M.B. Haha, K.O. Kjellsen, M.R. Geiker, K. De Weerdt, Limitations of the hydrotalcite formation in Portland composite cement pastes containing dolomite and metakaolin, Cement and Concrete Research, (2018). [2] E. Bernard, B. Lothenbach, F. Le Goff, I. Pochard, A. Dauzères, Effect of magnesium on calcium silicate hydrate (C-S-H), Cement and Concrete Research, 97 (2017) 61-72.

Thermodynamic properties of C-S-H, C-A-S-H and M-S-H phases: results from direct measurements and predictive modelling

Roosz, C.1,2), Vieillard, P.3), Blanc, P.2), Gaboreau, S.2), Gailhanou, H.2), Braithwaite, D.4), Montouillout, V.5), Denoyel, R. 6), Henocq, P. 1), Madé1, B.

1) Andra, 1/7 rue Jean Monnet, Parc de la Croix Blanche, F-92298 Chatenay-Malabry Cedex, France 2) BRGM, 3 Av. Claude Guillemin, BP6009, F-45060 Orléans Cedex 2, France 3) CNRS/INSU,-FRE 3114 Hydrasa, 40 Av. Du Recteur Pineau, F-86022 Poitiers Cedex, France 4) IMAPEC, CEA-INAC/ UJF-Grenoble 1, 17 rue des martyrs, 38054 Grenoble, France 5) CNRS, CEMHTI, UPR 3079, 1D avenue de la Recherche Scientifique 45071, Orléans cedex 2, France 6) Aix Marseille University, CNRS, MADIREL UMR 7246, F-13397 Marseille Cedex, France Corresponding author: Blanc P.., email: [email protected]

How cement-based materials behave in storage operating conditions is important in the context of a radwaste deep geological repository. In particular, we need missing thermodynamic data for ternary and low-pH cementitious materials that have high aluminum oxide content. This work proposes the first calorimetric measurements performed on C-A-S-H and M-S-H minerals identified in the aforementioned cementitious matrix; pure C-S-H are also investigated as reference and to enhance the thermodynamic database. Coupled with solubility experiments and vapor adsorption isotherms, calorimetry measurements provide a complete set of thermodynamic parameters for C-S-H, C-A-S-H, and M-S-H. The measurements are completed by a model predicting thermodynamic properties, which is parameterized using results from this work. Predominance diagrams allow us to confirm the relation between M-S-H and low C/S phases. Calculations indicate that the competition with secondary Al-bearing phases (stratlingite, katoite, hydrogarnet) is held at the expense of C-A- S-H. Considering natural zeolites at equilibrium would make C-A-S-H disappear and further development could mitigate such disappearance if considering synthetic zeolites which form in cementitious chemical systems.

A) B)

Figure 1 – Predominance diagrams in the system SiO2-Al2O3-H2O without zeolites: A) C-S-H and C-A-S-H only ; B) including straetlingite, katoite and C3AH6

Molecular level insight into ions uptale by C-(A)-S-H

Churakov, S..1,2) Labbez, C.3) 1) Paul Scherrer Institute, OFLA 204a, 5232 Villigen-PSI, Switzerland 2) University of Bern, Blazer Str., 3012 Bern, Switzerland 3) UMR 6303 CNRS, Université de Bourgogne, F-21078 Dijon, France, Corresponding author: Churakov, S. V., email: [email protected]

Calcium-(Aluminium-)Silicate-Hydrates (C-(A-)S-H), controls to a large extent the uptake and transport properties of hardened cement paste. C-S-H has a foil like structure with very small coherent diffraction domains. The structure of each single foil is similar to the one of a naturally occurring mineral tobermorite with a non-stoichiometric calcium to silicon ratio (Ca/Si). The diffusive ion transport in cements with low w/c ratio is controlled by transport through so called “gel pores” confined by C-S-H particles. The composition of C-S-H varies with the activity of its constitutive species in solution, i.e. silicate, calcium and hydronium ions. In cements blended with supplementary cementitious materials such as slag, fly ash or metakaolin, the composition and structure of C-(A-)S-H is much less well-defined. The main open questions include the mechanism of the Al incorporation in the C-(A-)S-H and its consequences for the uptake of aqueous ions and their transport in C-(A- )S-H.

A deep mechanistic insight into the nature of ion-mineral interaction in C-(A)S-H can be obtained by multi-scale molecular simulations. Our coarse grain molecular thermodynamic model describes the pH dependent development of the surface charge due to deprotonation of the surface OH groups, adsorption of aqueous ions on the surface and exchange equilibrium between Al and Si bearing structural sites on C-(A-)S-H surface and the aqueous solution without use of adjustable parameters [1]. The surface sites densities are derived from structural properties of tobermorite. The protolysis constants of the surface groups and free energies are calculated from ab initio simulations [2,3]. The concentration dependent ion uptake is modelled by the Grand Canonical Monte Carlo Simulations [1]. Our coarse grain modelling approach also provide insight into spatial distribution of ion in gel pores of C-(A)S-H which is indispensable for pore scale modelling of the ion transport in cement.

References [1] Churakov, S. V., C. Labbez, L. Pegado and M. Sulpizi, Intrinsic acidity of surface sites in calcium-silicate- hydrates and its implication to their electrokinetic properties." Journal of Physical Chemistry C (2014), 118, 11752−11762. [2] Pegado, L., C. Labbez and S. V. Churakov, Mechanism of aluminium incorporation in C-S-H from ab initio calculations. Journal of Materials Chemistry A (2014), 2, 3477-3483. [3] Churakov, S. V. and C. Labbez Thermodynamics and Molecular Mechanism of Al Incorporation in Calcium Silicate Hydrates. Journal of Physical Chemistry C (2017), 121(8), 4412–4419.

Molecular understanding of tricalcium silicate hydration in absence and in presence of aluminate ions Pustovgar, E.1), Mishra, R.1), Palacios, M.1,2), Matschei, T.3), Flatt, R. J.1), d’Espinose de Lacaillerie, J.-B.4) 1) Institute for Building Materials, ETH Zürich, 8093 Zürich, Switzerland 2) Eduardo Torroja Institute for Construction Science, Serrano Galvache 4, 28033 Madrid, Spain 3) Growth & Innovation, Holcim Technology Ltd., 5113 Holderbank, Switzerland 4) Soft matter Science and Engineering, ESPCI Paris, 10 rue Vauquelin, 75005 Paris, France Corresponding author: d’Espinose de Lacaillerie, J.-B., email: [email protected]

The kinetics of hydration of tricalcium silicate is not yet fully understood. Different mechanisms have been proposed and NMR has proven to be very successful in elucidating the structure of hydrates when the reaction is stopped or completed. However, a precise picture of the dynamics of silicate hydrates formation in-situ is still lacking and impedes a definite choice between the proposed models. Here we show by a combination 29Si NMR, 27 29 Al NMR and calorimetry that the synthesis of carefully designed Si-enriched C3S allows following quantitatively the hydration process under conditions close to in situ ones.[1]

In water, we obtain the transient local molecular composition of the hydrates at different stage of hydration. In particular, during the deceleration period the hydrate precipitation rate decreases faster than the amount of hydroxylated C3S surface, suggesting that the C3S surface is partially covered by C-S-H and that the surface area available for silicate dissolution decreases. Furthermore, by 2D-NMR a distribution of silica chain length can be proposed. In the proposed scheme, although the average chain length is five, pentamers do not constitute the predominant occurrence.

Then, the effect of aluminate ions on the hydration of C3S is investigated. We show that a pH sensitive retardation of C3S hydration by aluminate ions occurs at early age of reaction, but that the amount of the hydrates formed increases later. These experimental results can be interpreted assuming that aluminates hinder C3S dissolution. This view is supported by molecular dynamics simulations establishing that aluminates can adsorb on hydroxylated C3S through ionic interactions between aluminate and calcium ions, as well as through hydrogen bonding with silicate surface groups. This interaction is pH dependent and, consequently, the retardation effect of aluminates varies during the advancement of hydration.[2]

References [1] E. Pustovgar, R.P. Sangodkar, A.S. Andreev, M. Palacios, B.F. Chmelka, R.J. Flatt, J.-B. d’Espinose de Lacaillerie, Understanding silicate hydration from quantitative analyses of hydrating tricalcium silicates, Nat. Commun. 7 (2016) 10952. doi:10.1038/ncomms10952. [2] E. Pustovgar, R.K. Mishra, M. Palacios, J.-B. d’Espinose de Lacaillerie, T. Matschei, A.S. Andreev, H. Heinz, R. Verel, R.J. Flatt, Influence of aluminates on the hydration kinetics of tricalcium silicate, Cem. Concr. Res. 100 (2017) 245–262. doi:10.1016/j.cemconres.2017.06.006.

Multi-scale modelling of silica interfaces: the role of the charges

Dufrêche, J.1), Siboulet, B.1), Hocine, S.1), Duvail, M.1), Turq, P. 1) , Hartcamp, R.2), Coasne, B.2)

1) Institute for Separation Chemistry of Marcoule (UMR 5257), CNRS/CEA/Universite Montpellier – ENSCM Centre de Marcoule, France 2) MultiScale Material Science for Energy and Environment, CNRS/MIT (UMI 3466), Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, USA

Corresponding author: Turq, P. Porous oxides such as silica glasses can exchange adsorbe numerous elements. They are especially useful for nuclear decontamination because they relatively well resist to radiolysis. Despite their significance, theirs interfacial and electrokinetic properties are not very well understood. There is no consensus for the description of the interface. The various length scales involved make this class of systems especially challenging for computational physics.

We present a multiscale approach in order to describe ion separation and specific effects in porous oxide. We will especially focus on electrokinetic phenomena. Extensive molecular simulations of aqueous electrolyte solutions confined in a charged amorphous silica slot have been performed. Generalyzing the McMillan-Mayer theory of electrolyte solution, the thermodynamical properties have been calculated in accordance with the experimental measurements.

For the dynamics, contrasting traditional models of the electric double layer, molecular dynamics simulations indicate that there is no stagnant layer, no Stern layer conduction, no outer Helmholtz layer. The description of the interface requires two points. First, a distinction has to be made between free and surface-bounded ions. The latter do not form a physical layer but rather a set of ion-surface contact pairs. Second, the mobility of the free ions relative to their bulk value is reduced. This effect proved to be from hydrodynamical origin needs to be included.

These two concepts, coupled to simple macroscopic equations, are sufficient to describe both equilibrium and electrokinetic (surface conductivity and electroosmotic flow) phenomena in agreement with the molecular simulations. The resulting macroscopic description is found to be valid up to very small porosities (typically 2 nm). Surface conduction is negative at high concentration, and the Bikerman formula is only valid at low concentration.

[1] S. Hocine, R. Hartkamp, B. Siboulet, M. Duvail, B. Coasne, P. Turq and J.-F. Dufrêche. J. Phys. Chem. C 120, 963 (2016). [2] R. Hartkamp, B. Siboulet, J.-F. Dufrêche and B. Coasne Phys. Chem. Chem. Phys. 17 24683 (2015) [3] B. Siboulet, B. Coasne, J.-F. Dufrêche. P. Turq J. Phys. Chem. B 115, 7881 (2011). [4] B. Siboulet, S. Hocine, R. Hartkamp M. Duvail, J.-F. Dufrêche. J. Phys. Chem. C 121 6756 (2017)

C-S-H nucleation pathways

Fernandez-Martinez, A.1), Van Driessche, A.1), Nicoleau, L.2), Labbez, C.3), Krautwurst, N.4), Kellermeier, M.5), Tremel W.4) 1) ISTerre, Univ. Grenoble Alpes & CNRS, 1381 rue de la Piscine, 38041 Grenoble, France. 2) BASF Construction Solutions GmbH, B08, Dr-Albert-Frank-Strasse 32, 83308 Trostberg, Germany. 3) Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR 6303 CNRS, Univ. de Bourgogne, F-21078 Dijon Cedex, France. 4) Institute of Inorganic Chemistry and Analytical Chemistry, Johannes Gutenberg-University Mainz. Duesbergweg 10 14, D-55099 Mainz, Germany. 5) Material Physics, BASF SE, Carl-Bosch-Straße 38, 67056 Ludwigshafen, Germany Corresponding author: Fernandez-Martinez, A., email: [email protected]

The nucleation of calcium silicate hydrates has been studied in the past using chemical methods, microscopy, tomography and diffraction-based techniques, at a variety of length scales. However, many details about the dynamics, and structural properties, of the early stages remain elusive due to the high reactivity of alite, which results in fast-coupled dissolution-precipitation processes and to the inherent poorly crystalline character of C-S- H.

Here, results from a new set of experiments based on the combination of state-of-the-art in situ X-ray scattering (at low and wide angles) and potentiometric titration will be presented. These results, combined with electron microscopy and modelling data, show that the homogeneous nucleation of C-S-H proceeds via multiple steps, including the initial formation of a disordered precipitate. Potential implications of this pathway will be discussed.

Studying the intrinsic mechanical properties of calcium (alumino)silicate hydrate using synchrotron-radiation-based high-pressure x-ray diffraction

Geng, Guoqing1), Vasin, Roman Nikolaevich 2), Dähn Rainer 1), Wieland, Erich 1), Wenk, Hans-Rudolf 3), Monteiro, Paulo J. M.4) 1) Laboratory for Waste Management, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland 2) Frank Laboratory of Neutron Physics, Joint Institute for Nuclear Research, Joliot-Curie 6, 141980 Dubna, Moscow Region, Russia 3) Department of Earth and Planetary Science, University of California, Berkeley, California 94720, United States 4) Department of Civil and Environmental Engineering, University of California, Berkeley, California 94720, United States Corresponding author: Geng, Guoqing, email: [email protected]

Calcium (alumino)silicate hydrate, or in cement chemistry notation C-(A-)S-H, is the principal binding phase in modern Portland cement (PC)-based concrete. Understanding its structure-property correlation is critical to modeling and optimizing the performance of PC concrete. Conventional lab methods often have limitations in addressing such problem, mainly due to the poorly-crystalline and multi-scale porous nature of C-(A-)S-H.

This abstract reports an innovative experimental approach to such problem. Using high-brilliance synchrotron- radiation-based x-ray as incident beam, the crystallographic information, e.g. lattice parameters and crystallite size, of nano-crystalline C-(A-)S-H can be reliably refined. Coupled with a high pressure diamond anvil cell, the anisotropic deformation of C-(A-)S-H nano grains can be quantified from XRD data at given pressure values. The results indicate that, the coherent domain of the lab-synthesized C-(A-)S-H is of platelet shape, i.e. shorter along the c-axis and longer along a- and b-axis. The mechanical property of these nano-platelets is transversely isotropic, with soft behavior along the normal of the layer structure (c-axis), but is direction-independent within the layer plane (a- and b-axis). The omission of the bridging site silicate does not affect the stiffness along the silicate chain direction (b-axis). The densification of the interlayer spacing, as well as the cross-liking of adjacent layers, is the major driven force to stiffen the C-(A-)S-H nano platelets. Rather than the Ca/Si, the basal spacing of C-(A-)S-H is a better index of its bulk modulus.

When the loosely packed nano C-(A-)S-H nano grains are under uniaxial compression and lateral confinement, the random orientation becomes strongly textured, i.e. most nano grains are orientated so that their c-axis are aligned with the direction of the compression. This in-turn changes the mechanical property of the whole C-(A- )S-H matrix. Discussions are given on C-(A-)S-H mechanical behavior evolution when subjected to external conditions such as indentation loading and drying environment.

Structure and reactivity of nanocrystalline calcium silicate hydrates: the parallel with clay minerals

Grangeon, S. 1), Claret, F. 1), Henocq, P. 2), Linard, Y.2). 1) BRGM, 3 avenue Claude Guillemin, 45060 Orléans Cedex 2 2) Andra, 1 – 7 rue Jean Monnet, 92298 Châtenay-Malabry, France Corresponding author. Email: [email protected]

Nanocrystalline calcium silicate hydrate (C-S-H) is the main hydration product of cements, in which they control the main chemical and mechanical properties, including ion and water sorption capacity [1]. This exceptional reactivity stems from the combination of nanocrystallinity and of the particular C-S-H structure. Indeed, like clay minerals, C-S-H has a lamellar structure, being built of layers of calcium in 7-fold coordination on which wollastonite-like Si ribbons are attached. Layers are, like clays, separated from each other by hydrated interlayer cations whose density depends on the layer charge. This charge is, like clays, generated by layer vacancies and isomorphic substitutions (e.g., Al3+ for Si4+ ). Finally, in a similar fashion as clays, C-S-H incorporates a variable amount of interlayer water, which depends on the nature and density of the layer charge and on the relative humidity. A sketch highlighting the similarities between clay minerals and C-S-H is presented in Figure .

Figure 1. Structural similarity between clays and C-S-H. In 2:1 clays (left), layers are built of cations in octahedral coordination (grey polyhedra), covered with cations in tetrahedral coordination (red tetrahedra). In C- S-H (right), layers are built of Ca in 7-fold coordination (grey polyhedra) covered by wollastonite-like Si chains (red polyhedra). In both cases, layers are separated from each other by hydrated interlayer cations (spheres).

Understanding the reactivity of C-S-H towards cations and water requires a sound description of the nature and density of the layer charge. This knowledge is also fundamental to our capacity to predict and model cement pore water. However, as for many clay minerals, C-S-H suffers from turbostratism that is the systematic presence of random translations and (or) rotations between adjacent layers that remain parallel. Consequently, many structure refinement methods, in particular Rietveld refinement of X-ray diffraction pattern, are hardly applicable.

Following the methodology developed for the analysis of clay minerals, we used a specific trial-and-error modelling of X-ray diffraction patterns to decipher the structure of C-S-H, allowing to unambiguously show that C-S-H is nanocrystalline turbostratic tobermorite [2]. This structure determination was confirmed using laboratory methods (e.g. transmission electron microscopy, Fourier-transformed infrared spectrometry), as well as synchrotron X-ray absorption spectrometry (XANES and EXAFS) [3, 4]. In particular, we could show that, at low Ca/Si ratio (~0.66), the structure contains about no layer vacancy, and that layer-to-layer distance is either 11 Å, if the adjacent layers are connected through the apical oxygen coordinating bridging Si atoms, or 14 Å if layers are not connected. The increase in Ca/Si ratio, up to ~1.25, is achieved through the increase in the number of vacancies in the Si bridging site (Figure ), which creates a layer charge deficit compensated for by interlayer Ca2+. This structural evolution is accompanied by a decrease in the layer-to-layer distance. Further increase in Ca/Si ratio cannot be achieved by such a mechanism, as 29Si MAS-NMR could show that the density of Si vacancies remains constant at Ca/Si > 1.25. Using differential pair-distribution function analysis of synchrotron X-ray high-energy scattering data, we could show that freeze-dried samples of high Ca/Si ratio contain nanosized (< 1 nm) Ca(OH)2 clusters [5], which contributes to increasing the macroscopic Ca/Si ratio while remaining X-ray silent when using conventional laboratory X-ray sources as a results of its minute size and minor abundance in the sample. A scheme of the prosed C-S-H structure evolution as a function of Ca/Si ratio is proposed in Figure .

Figure 2. Scheme of proposed C-S-H layer structure evolution as a function of Ca/Si ratio. The evolution proceeds mainly via omission of bridging Si tetrahedra in the wollastonite-like chains, plus incorporation of a discrete Ca(OH)2 phase at the highest ratios. Interlayer cations and water are omitted for the sake of clarity. Colours as in Figure .

This presentation will review these recent developments we made concerning C-S-H structure determination, and will replace it in light of existing structure models. In particular, we will show that our data are fully consistent with previous studies [e.g. 6, 7, 8] and support the Richardson and Groves’ model [9].

References [1] Richardson, Cem. Concr. Res., 38 (2008) 137-158. [2] Grangeon et al., Acta Crystallogr B, 69 (2013) 465-473. [3] Grangeon et al., Cem. Concr. Res., 52 (2013) 31-37. [4] Grangeon et al., J. Appl. Crystallogr., 49 (2016). [5] Grangeon et al., J. Appl. Crystallogr., 50 (2017). [6] Cong and Kirkpatrick, Adv. Cem. Based Mater., 3 (1996) 144-156. [7] Garbev et al., J Am Ceram Soc, 91 (2008) 3015-3023. [8] Nonat, Cem. Concr. Res., 34 (2004) 1521-1528. [9] Richardson, Acta Crystogr. B, 70 (2014) 903-923.

Effect of aluminum inclusion on morphology and mechanical properties of calcium alumino silicate hydrate

Hay, R.1), Li, J.2), Monteiro, P. J. M. 2), Celik, K.1) 1) New York University Abu Dhabi, Abu Dhabi, United Arab Emirates 2) University of California at Berkeley, 94720 Berkeley, United States Corresponding author: Celik, K., email: [email protected]

Calcium aluminate silicate hydrate (C-A-S-H) is one of the main binding phases in aluminium-rich blended cement. Existing literature shows that Al-incorporated crosslinking site increases the bulk modulus of calcium alumino silicate hydrate (C-A-S-H). However, structural, physical and mechanical properties of C-A-S-H are still a subject of research. Here, calcium (alumino)silicate hydrate (C-(A-)S-H) with bulk molar ratio Ca/(Al+Si) = 1 and Al/(Al+Si) = 0-0.33 were synthesized and equilibrated at 80°C for 4 months by mixing stoichiometric amounts of SiO2, CaO and CaO·Al2O3 at a water-to-solid mass ratio of 45 in a N2-filled glove box. X-ray diffraction (XRD), transmission electron microscopy (TEM) and thermogravimetry were used as methods of analysis. XRD results showed a comparable interlayer spacing of the C-A-S-H samples. Thermogravimetric analysis revealed that minor or trace of third aluminate hydrate and/or katoite present in the C-A-S-H products at Al/(Al+Si)=0.05- 0.2. TEM images showed that all C-(A-)S-H samples at these conditions are foil-like although the inclusion of aluminium led to a coarser morphology (Fig. 1), indicating a more polymerization and cross-linking of C-A-S-H. 27Al and 29Si nuclear magnetic resonance spectroscopy will be performed to study Al incorporation on the main dreierketten C-S-H chain. Nano-indentation work will be conducted to validate the effect of Al-incorporation on the mechanical properties of C-A-S-H with limited variation in the interlayer spacing.

Al/(Al+Si) = 0 Al/(Al+Si) = 0.1 Al/(Al+Si) = 0.33

Figure 1: Bright-field TEM images showing morphology of C-(A-)S-H at Al/(Al+Si) = 0, 0.1 and 0.33

A C-S-H builder and interface modeling tools towards accurate reactive full electrolyte simulations of cement materials up to the micrometer scale

Heinz, Hendrik1) 1) Department of Chemical and Biological Engineering, University of Colorado-Boulder, Boulder, CO 80309, USA Corresponding author: Heinz, Hendrik, e-mail: [email protected]

Abstract We will share a GUI-based automated C-S-H model builder across a range of C/S and C/H ratios, as well as new model developments for aluminate and oxide additives in cement that can be used with the Interface force field and common molecular dynamics and multiscale simulation programs (NAMD, GROMACS, LAMMPS, Materials Studio). We discuss the validation of models and first applications to elucidate the role of PCE polymers and carbohydrates in cement hydration in comparison with measurements. The full inclusion of ionic strength, pH, and metadynamics for sampling is described. The quality of structures, surface, and hydration energies as well as mechanical properties is of equal or higher reliability as DFT methods at far lower computational cost. Chemically detailed simulations are feasible up to the 1000 nm scale and can be combined with coarse-grain and finite element models based on the nanoscale data and experiments to extend into micrometer and macroscopic scale simulations of much improved quality.

The impact of the calcium to silicon ratio of C-S-H crystal seeds

John, E.1), Stephan, D.1) 1) TU Berlin, Gustav Meyer Allee 25, 13355 Berlin, Germany Corresponding author: John, E., email: [email protected]

The beneficial role of nanoparticle additions into cement is an intensively studied field in civil engineering and material science. Besides pozzolans like silica or quartz, the performance of calcium silicate hydrate (C-S-H) nanoparticles recently came to focus, due to their outstanding ability to accelerate the cement hydration process. The influence of the calcium to silicon ratio (Ca/Si) of the used C-S-H seeds however remains unclear. Thomas et. Al. could not find any significant performance differences of seeds, prepared by co-precipitation with a Ca/Si ratio from 0.8 to 1.5, in C3S. Land et. Al. observed that seeds, prepared mechanochemically led to a higher cement hydration maximum when richer in calcium [1, 2]. On the contrary, Alizadeh et. Al. found that the hydration of C3S was enhanced more, when C-S-H seeds with a lower Ca/Si were applied [3].

From literature, it is known that the Ca/Si ratio affects the compressive strength, density and surface area of the obtained C-S-H phases as well as the silicate chain length reflected in the Q1/Q2 ratio [4-6].

The current work aims to unite the current data by investigating and explaining the impact of the calcium to silicon ratio of C-S-H seeds onto the acceleration of cement and C3S hydration systematically. Therefore, different synthesis methods were applied to prepare C-S-H with a variety of properties, analysed by XRD, TGA, SEM and EDX. The hydration was followed using isothermal calorimetry. The impact of the relationship of different properties will be discussed in detail.

References: [1] Thomas, J. J.; Jennings, H. M.; Chen, J. J., The Journal of Physical Chemistry C., 2009, 113 (11), 4327–4334 [2] Land, G.; Stephan, D., ICCC, 2015 [3] Alizadeh, R.; Raki, L.; Makar, J. M.; Beaudoin, J. J.; Moudrakovski, I., Journal of Materials Chemistry., 2009, 19 (42), 7937-7946 [4] Thomas, J. J.; Jennings, H. M.; Allen, A. J., The Journal of Physical Chemistry C, 2010, 114 (17), 7594– 7601 [5] A. Nonat, The structure and stoichiometry of C-S-H, Cement and Concrete Research 34 (2004) 1521–1528 [6] W. Kunther, S. Ferreiro, J. Skibsted, Influence of the Ca/Si Ratio on the Compressive Strength of Cementitious Calcium-Silicate-Hydrate Binders, J. Mater. Chem. A (2017)

Atomistic modeling of the structure and dynamics of water and ions in Calcium Silicate Hydrate and calcium Aluminate hydrate phases

Kalinichev, A. G.1), 2) 1) Laboratoire SUBATECH (UMR 6457 - Institut Mines-Télécom Atlantique, Université de Nantes, CNRS/IN2P3), 44307 Nantes, France 2) International Laboratory for Supercomputer Atomistic Modelling and Multi-Scale Analysis, National Research University Higher School of Economics, 101000 Moscow, Russia Corresponding author: Kalinichev, A. G., email: [email protected]

For better understanding of cement hydration behaviour, it is essential to develop quantitative molecular scale picture of the structure and dynamics of the aqueous phase at the solid surfaces of cementitious materials and in their nanopores. Atomistic computer simulation is one of the rapidly developing and highly promising approaches capable to significantly complement (or even substitute) experimental investigations of such systems [1]. At cement interfaces, H2O molecules and hydrated ions simultaneously participate in several dynamic processes characterized by different, but equally important time- and length- scales. Most of these processes can be simultaneously studied by molecular dynamics computer simulations in a single MD run. On a relatively long time scale (~100-1000 ps and higher), it is possible to quantify the diffusional processes related to reformation of the entire interfacial hydrogen bonding network, surface adsorption of H2O molecules, and ions. However, the interfacial dynamics on the intermediate time scale (~1-10 ps) is dominated by the molecular librational and re- orientational motions. The librations (hindered rotations) of surface -OH- groups also occur at this time scale and control the reformation and breaking of individual H-bonds, but the strength of these bonds is directly correlated with the frequencies of intra-molecular -OH vibrations on the shortest, sub-ps time scale.

This talk will present an overview of applications of the ClayFF approach [2,3] to the atomistic modeling of cementitious systems and their aqueous interfaces, focusing on the hydrous calcium silicates and calcium aluminates (CSH, AFm, and AFt phases) and on the most recent improvements of the original ClayFF parameterization [4,5]. Good agreement with very diverse sources of experimental data (X-ray, NMR, IR, inelastic and quasielastic neutron scattering) covering multiple time- and length- scales relevant to many C(A)SH systems provides strong confidence that such models and approaches can be further extended to even broader ranges of compositions and processes.

References [1] R.K. Mishra, A.K. Mohamed, D. Geissbühler, H. Manzano, T. Jamil, R. Shahsavari, A.G. Kalinichev, S. Galmarini, L. Tao, H. Heinz, R. Pellenq, A.C.T. van Duin, S.C. Parker, R.J. Flatt, P. Bowen, Cemff: A force field database for cementitious materials including validations, applications and opportunities, Cem Concr Res, 102 (2017) 68-89. [2] R.T. Cygan, J.-J. Liang, A.G. Kalinichev, Molecular models of hydroxide, oxyhydroxide, and clay phases and the development of a general force field. J Phys Chem B 108 (2004) 1255-1266. [3] A.G. Kalinichev, J. Wang, R.J. Kirkpatrick, Molecular dynamics modeling of the structure, dynamics and energetics of mineral-water interfaces: Application to cement materials, Cem Concr Res, 37 (2007) 337- 347. [4] I. Androniuk, C. Landesman, P. Henocq, A.G. Kalinichev, Adsorption of gluconate and uranyl on C-S-H phases: Combination of wet chemistry experiments and molecular dynamics simulations for the binary systems. Phys Chem Earth, Parts A/B/C 99 (2017) 194-203. [5] M. Pouvreau, J.A. Greathouse, R.T. Cygan, A.G. Kalinichev, Structure of hydrated gibbsite and brucite edge surfaces: DFT results and further development of the ClayFF classical force field with metal-O-H angle bending terms, J Phys Chem C 121 (2017) 14757-71.

Carbonation of Low calcium C-A-S-H

Kangni-Foli, E.1) 2) 4), L'Hôpital, E.1), Dauzères, A.1), Le Bescop, P.2), Poyet, S.2), Charpentier, T.3), d’Espinose de Lacaillerie, J.-B. 4) 1) IRSN, Inst Radiat Protect & Nucl Safety, PRP DGE SRT LETIS, BP 17, F-92262 Fontenay Aux Roses, France 2) Den-Service d’Etude du Comportement des Radionucléides (SECR), CEA, Université de Paris-Saclay, F- 91191, France 3) NIMBE, CEA, CNRS, Université Paris-Saclay, CEA Saclay, F-91191 Gif-sur-Yvette Cedex, France 4) Soft matter Science and Engineering, ESPCI Paris, 10 rue Vauquelin, 75005 Paris, France Corresponding author: KANGNI-FOLI, E., email: [email protected] The use of low alkalinity cementitious materials - also known as low-pH materials - is considered in the French deep geological repository design for intermediate level long lived waste storage. The waste would be disposed in large underground cells that would be closed using bentonite (swelling clay) and low alkalinity concrete plugs. The main advantage of low alkalinity concrete is that the pH of the pore solution is low and would not impact the swelling ability of bentonite. Low alkalinity binders are obtained using large additions of siliceous bearing supplementary cementing materials such as slag or fly ash; the most effective being silica fume. Recent studies question the durability of such materials. Auroy et al. [1] found that atmospheric carbonation led to significant cracking and increase in transport properties and Serdar et al. [2] showed that the rate of carbonation-induced corrosion of steel was increased with regards to OPC. It is the objective of this study to gather information on the influence of atmospheric carbonation on the major component of low alkalinity cementitious materials: C-A-S- H.

In this study we have synthetized low C-(A-)S-H (Ca/Si ratio between 0.80 and 1.40) for various Al contents (Al/Si) ratio up to 0.10. Synthesis products are characterized by Nuclear Magnetic Resonance (NMR), ThermoGravimetric Analysis (TGA) and X-Ray Diffraction (XRD). This part of our work aims at better understanding the carbonation mechanisms of low-calcium C-(A-)S-H through the description of its structural evolution upon carbonation. As a first step, we focus on the evolution of the crystallo-chemistry going from the pristine to the fully carbonated state. Of particular interest is the representativeness of accelerated carbonation (at 3% CO2) compared to atmospheric carbonation. The impact of Al content on low Ca C-(A-)S-H carbonation process in terms of structure evolution, polymorphic abundance and reaction products will be assessed.

References [1] Auroy et al. (2015) Impact of carbonation on unsaturated water transport properties of cement-based materials, Cement and Concrete Research 74, 44-58. [2] Serdar et al. (2017) Carbonation of low-alkalinity mortars: influence on corrosion of steel and on mortar microstructure, Cement and Concrete Research 101, 33-45.

Chloride sorption onto C-(N)-A-S-H gels as a function of Ca/Si and Al/Si ratios

Ke, X., Bernal, S. A., Provis, J. L. Department of Materials Science and Engineering, The University of Sheffield, Sheffield S1 3JD, United Kingdom Corresponding author: Ke, X., email: [email protected]

The ingress of chloride from the environment, and its movement towards the interface between steel rebars and the cementitious matrix, is one of the main factors causing corrosion in reinforced concretes. The ionic interactions between free chloride and the cementitious phases present in the binder play an important role in the movement of free chloride within the concrete [1, 2].

This study aimed to investigate both the sorption and desorption process of chloride onto the sodium-containing calcium aluminosilicate hydrate (C-(N)-A-S-H) type gels. The reversibility of chloride interactions with C-(N)- A-S-H type gels under different conditions is an important factor to consider in a practical sense, as it reflects the stability of bound chloride with respect to the changing environment (e.g. carbonation, wet-dry cycles, etc.), therefore the outcomes of this study provide valuable insights for the prediction of concrete durability when exposed to complex service conditions. Here synthetic C-(N)-A-S-H type gels with various compositions (Ca/Si = 1.0 and 1.4, Al/Si = 0 and 0.1) were prepared. The chloride was introduced into the synthetic C-(N)-A-S-H type gel using two different methods, before and after the gel synthesis. Desorption experiments for both chloride-bearing gels were carried out in different background solutions, including Milli-Q water and, a simplified simulated cementitious pore solutions.

References [1] Ke, X., S.A. Bernal, and J.L. Provis, Uptake of chloride and carbonate by Mg-Al and Ca-Al layered double hydroxides in simulated pore solutions of alkali-activated slag cement. Cem. Concr. Res., 100 (2017) 1-13 [2] Ke, X., S.A. Bernal, and J.L. Provis. Chloride binding capacity of synthetic C-(A)-S-H type gels in alkali- activated slag simulated pore solutions. in 1st International Conference on Construction Materials for Sustainable Future. 2017. Zadar, Croatia.

C-A-S-H solid solutions: multisite VS quasichemical? Kulik, D. A.1), Miron, G.D.1), Lothenbach, B.2) 1) Laboratory for Waste Management, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland 2) Laboratory Concrete & Construction Chemistry, Empa, 8600 Dübendorf, Switzerland Corresponding author: Kulik, D. A., email: [email protected]

We still need to improve solid solution models of C-S-H in order to fix limitations/drawbacks of earlier models. These include (too) simple (ideal) mixing; lack of rigor (e.g. semi-empirical “downscaling” of end members [1]); difficulties upon extension for other cations than Ca+2 resulting in a non-incremental, trial-and error parameter fitting. The structural model of C-S-H requires simultaneous substitutions in 2 or 3 structural site types, which cannot be rigorously described with classical thermodynamic models for simple single-site mixing. This is why a multi-site (sublattice) model is needed to ensure structural consistency (reproducing e.g. mean chain length) and extensibility to more end members (e.g. K, Na, Al, Sr, Zn, U) relevant for cement chemistry and for understanding waste-cement interactions.

Our past attempts to fit a three-sublattice model of C-(N,K,A)-S-H to the available solubility data have elucidated the following problems: (i) account for vacancies on BT (bridging tetrahedral) sites results in too strong competition on IC (interlayer cationic) sites; (ii) cations should be allowed to enter spaces left after removed BTs (cf. [1]); (iii) there are doubts that the CU (calcium units) sites really exist; if not then much more Ca can be accommodated in IC+BT spaces [2]. In this context, do we need to re-consider the sublattice substitutions scheme? Can the independently evaluated atomistic defects be used for constructing a better model?

Galmarini [3] and Mohammed et al. [4] invented a “building block” (DNA-like) description to generate atomistic structural models for the 14 Å tobermorite structure C-S-H, and suggested the possibility to enhance future models by allowing the replacement of BTs by calcium ions and of silicate dimers - by hydroxyl groups. This corroborates our conclusions and opens up two alternative approaches to improve structurally consistent thermodynamic solid solution models of C-S-H.

We use these recent findings plus an upgraded GEMSFITS code with the ThermoHub database [5] to re- parameterise a re-formulated C-(A)-S-H solid solution model with only two sublattices (BT and IL) allowing substitution of BT silicon with Ca+2 and other cations. As a promising alternative, we formulate a new C-(A)-S- H model from the “building blocks” identified in [4] using a “quasichemical” approach to describe the configurational entropy plus the residual non-ideal interactions.

The goal is to compare both models for their performance in fitting exercises, to see which of the two solution models fits better against available data, and which one is easier to be extended for minor cations, e.g. Na, K and Al. The currently achieved progress will be reported.

This work is conducted as part of the SNF CASH-2 project (200021_169014/1).

References [1] D.A. Kulik, Improving the structural consistency of C-S-H solid solution thermodynamic models. Cem Concr Res, 41 (2011), 477-495. [2] A. Kumar, Synthetic calcium silicate hydrates. PhD. Diss. 7658 (2017), EPFL. [3] S. Galmarini, Atomistic simulation of cementitious systems. PhD. Diss. 5754 (2013), EP. [4] A.K. Mohammed, S.C. Parker, P.Bowen, S.Galmarini, An atomistic building block description of C-S-H - Towards a general C-S-H model. Cem Concr Res, in revision. [5] G.D. Miron, D.A. Kulik, B. Lothenbach (this workshop).

Atomic structure of Calcium Silicate Hydrate Aslam Kunhi Mohamed 1), Sandra Galmarini 2), Steve Parker 3)., Karen Scrivener 1), Paul Bowen1) 1) EPFL Switzerland 2) EMPA Dubendorf Switzerland, 3) University of Bath, UK [email protected]

Understanding the mechanisms of cement hydration has been quite difficult due to the complexity of the system and its continued reaction over time making it hard to observe experimentally. Although atomistic simulations might be useful to study how the presence of different species affect the nucleation and growth of hydrates, the atomic structure of calcium silicate hydrate (C-S-H), the main hydrate phase, is not clearly known or agreed upon and remains an open question. Proposed structures of C-S-H have been mainly based on tobermorite and a model structure is created by introducing defects in the original non-defective tobermorite structure. In general, a defective structure matching the Ca/Si ratio needed has been simply created by depolymerizing Si chains and/or by addition of calcium in the interlayer space.

In a recent work, Kunhi Mohamed et. al [1] proposed a new methodology to simplify the generation of C-S-H structure. Small chemical entities, derived from the breaking down of a 14 Å tobermorite unit cell, are arranged in three-dimensional space to create a variety of defective tobermorite unit cell structures which form the building blocks of C-S-H. The structural and energetic stability of these building blocks are assessed using molecular dynamics simulations and DFT calculations. This methodology has now been expanded to study the interactions of different stable building blocks and arrange them suitably to create stable C-S-H structures that are consistent with available experimental results.

[1] Kunhi Mohamed, A., Parker, S. C., Bowen, P., Galmarini, S. C., An atomistic building block description of C-S-H - Towards a realistic C-S-H model. Cem. Concr. Res. 2018, Just accepted.

From nucleation to particle assemblage: what can we learn from mesoscopic simulations?

Labbez, C., Turesson, M., Gauffinet, S., Nonat, A. Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR 6303 CNRS, Univ. Bourgogne Franche-Comté, F- 21078 Dijon Cedex, France. Corresponding author: Labbez, C., email: [email protected]

In the recent years lot of efforts have been devoted to molecular simulations of model cementitious systems. They have been used for example to predict the mechanical properties of pure C-S-H crystals, the interaction free energy between C-S-H particles, the deprotonation Gibbs free energy of the C-S-H surface silanols, the interaction free energy of small chemical species at the C-S-H/solution interface and the morphology of portlandite crystals in various equilibrium conditions. Although very successful, they are very limited in time and length scales. This makes difficult, if not impossible, to address the nucleation, growth and assemblage of C-S-H particles, at least using molecular simulations alone. In this contribution, we will see how mesoscopic simulations, which make used of the simple but sound physical primitive model, can be used to tackle and better understand these processes. In particular, we will see how they contributed to the recent finding of hybrid C-S-H mesocrystals, to shed light on the multi step nucleation of C-S-H and to rationalize the growth modes of C-S-H on C3S.

Probing the solid/solution interface by in Situ real-time infrared spectroscopy

Lefèvre, G. PSL Research University, Chimie ParisTech — CNRS, Institut de Recherche de Chimie, 11 Rue Pierre et Marie Curie, 75005, Paris, France Corresponding author: [email protected]

Infrared spectroscopy is a widespread tool to analyze solids in the mid-IR range. For example, silicate, carbonate, or hydroxide groups can be identified in dry samples, to complement X-Ray driffraction analysis. The same approach for solid/solution interfaces is constrained by the problem of the strong absorption of infrared beam by water. Accessories based on Attenuated Total Reflectance effect allow to record spectra of solid/solution interfaces, due to a small probed depth of the water. The protocol developed in the 90’s [1] consists in making a coating by drying a colloidal suspension onto the ATR crystal, and clamping a circulation cell (Fig. 1). Thus, the spectra of adsorbed species can be recorded in situ, and in real time.

Figure 1: Schematic of single internal reflection ATR crystal with an oxide layer in contact with an aqueous solution in a circulation cell, and example of its use to follow ion exchange in layered double hydroxide [3].

Numerous systems have been studied with this technique, to characterize mechanisms of growth [2], adsorption [1] or ion exchange [3]. I will describe the key points to use this method, and some examples of systems studied in our group, mainly to explore ion exchange phenomena in layered double hydroxides. The use of complementary methods as DFT calculations or zeta potential measurements to obtain an accurate description of the interfacial reactions will be illustrated too.

References [1] S.J. Hug, In situ Fourier transform infrared measurements of sulfate adsorption on hematite in aqueous solutions, J. Colloid Interface Sci. 188 (1997) 415-422. [2] C. Ebbert, G. Grundmeier, N. Buitkamp, A. Kröger, F. Messerschmidt, P. Thissen, Toward a microscopic understanding of the calcium–silicate–hydrates/water interface, Appl. Surf. Sci. 290 (2014) 207-214. [3] A. Davantès, G. Lefèvre, In situ real time infrared spectroscopy of sorption of (poly)molybdate ions into layered double hydroxides, J. Phys. Chem. A 117 (2013) 12922-12929.

Synchrotron-based characterization of the chemistry and structure of calcium (alumino) silicate hydrate

Li, J.1), Geng, G.2), Myers, R. J.3), Monteiro, P. J. M.1) 1) University of California at Berkeley, 94720 Berkeley, United States 2) Paul Scherrer Institute, PSI Aarebrücke, 5232 Villigen, Switzerland 3) University of Edinburgh, Old College, South Bridge, EH8 9YL Edinburgh, United Kingdom Corresponding author: Li, J., email: [email protected]

There is a lack of consensus on the coordination environment of Ca, and the nature and location of Al species in calcium alumino silicate hydrate (C-A-S-H). Calcium (alumino)silicate hydrate (C-(A-)S-H) with Ca/Si = 0.6- 1.6, Al/Si = 0-0.1, and equilibrated at 7-80°C are studied by nanoscale soft X-ray absorption spectroscopy at the Ca L2,3-, Al K- and Si K-edges and ptychographic imaging. CaO7 complexes in C-(A-)S-H and zeolitic Ca in the interlayer are highly distorted from an ideal octahedral coordination irrespective of Ca/Si, Al incorporation, or temperature. Third aluminate hydrate (TAH) is not Ca-bearing or its Ca is structurally similar to C-(A-)S-H, and does not resemble the Ca in AFm-phases. The chemical environment of Al in TAH is similar to Al(OH)3. TAH presents as a separate phase adjacent to C-A-S-H foils, the interlayer contains five-fold coordinated Al at 7- 50°C. Al K- and Si K-edges of C-(A-)S-H increase with higher polymerized aluminosilicate chains, implying Al- uptake occurring in bridging and/or cross-linked sites. C-(A-)S-H exhibits foil morphology, which is composed of nano-sized platelets with comparable thickness regardless of Ca/Si or Al inclusion at 7-50°C. Coarse foils are predominant in C-(A-)S-H at 80°C and are coarser with Al incorporation.

Figure 1: Ptychographic images of C-A-S-H synthesized at 20 °C and equilibrated for 182 days: A) Al/Si*=0.05, Ca/Si*=1; B) Al/Si*=0.05, Ca/Si*=1.6. Katoite, and Al(OH)3 is identified and indicated by yellow arrows in A and B, respectively. Ca/Si*= bulk Ca/Si. Al/Si*=bulk Al/Si

Effect of Al on the dissolution rates of glass in Hyperalkaline solutions

Liu, Sanheng.1), Ferrand, Karine.1), Jacques, Diederik .1), Lemmens, Karel 1) 1) Belgium Nuclear Research Center (SCKCEN), Boeretang 200, 2400, Mol, Belgium Corresponding author: Liu, Sanheng., email: [email protected]

Dissolution of aluminosilicate glass in hyperalkaline solutions (pH = 13.5 and 30 oC) showed that in the beginning Al dissolved as easily as B and Na from the glass. The concentration of Ca stayed lower than 2 mg/l suggesting that most Ca from the glass retained at the glass surface. Similarly, Si also partially retained at the glass surface. The rate of glass dissolution decreased progressively very soon after the start of the experiments. After certain time, Al peaked at a concentration about 70 and 40 mg/l in solution with Ca+Na+K and K, respectively. Geochemical modelling results showed that at this point the solution was still under-saturated with respect to gibbsite and Al-hydroxide. Afterwards Al concentration started to decrease slowly indicating Al incorporation into the glass surface. SEM revealed that the altered glass surface was enriched in K, Ca, Al and Si. No crystalline phases were detected by XRD analyse, even though the solution analyse showed that the solution became over-saturated with respect to phillipsite_K, tobermorite and jennite according to the Thermochimie database (according to the CEMDATA, the solution was still under-saturated with respect to CSHjen, CSHTob1 and CSHTob2). The solution was also far from saturation with respect to stratlingite according to both databases. Thermodynamic modelling assuming that Al concentration in the solution is controlled by precipitation of phillipiste_K can capture the trend of Al, i.e., increases first and then decreases again, but it results in a faster decrease of Al concentration. The slower decrease of Al concentration seems to be due to Al back diffusion into the glass surface instead of co-precipitating with Si as phillipsite_K out of the - solution, since at such high pH both Al and Si are present as negative charged species in the solution, Al(OH)4 2- and H2SiO4 respectively, and they repulse each other. The effect of incorporating Al to the glass surface causes a further decrease the dissolution rate. However, similar experiments at 70 oC showed that at some point when Al concentration in the solution became low enough a fast dissolution of the glass resumed [1].

References [1] S. Liu, K. Ferrand, K. Lemmens, Transport- and surface reaction-controlled SON68 glass dissolution at 30 °C and 70 °C and pH = 13.7, Appl Geochem, 61 (2015) 302-311.

Calcium silicate hydrates with aluminium

Lothenbach, B.1,2), L’Hôpital, E.,1) Di Giacomo, A.1) Barzgar, S. 1) 1) Laboratory Concrete & Construction Chemistry, Empa, 8600 Dübendorf, Switzerland 2) Department of Structural Engineering, NTNU, 7491 Trondheim, Norway Corresponding author: Lothenbach, B., email: [email protected]

Calcium silicate hydrates (C-S-H) are the main solid phase present in hydrated cements. The composition of C- S-H may change depending on the composition of the solution in which it is equilibrated. The silica chain length in C-S-H increases with the silicon concentrations and the calcium content in the interlayer space with the calcium concentrations.

Aluminium in C-S-H can be present in different sites depending on the Ca/Si ratio. At low Ca/Si ratios, aluminium substitutes silicon in the bridging position. At higher Ca/Si ratios the fraction of aluminium in bridging position decreases. At all Ca/Si ratios approximately 10% of the aluminium is present as penta- coordinated, AlV, which has been postulated to represent aluminium present in the interlayer. At Ca/Si >1 hexa- coordinated aluminium, AlVI, becomes important, which is attributed to the presence of an amorphous aluminate hydrate associated with C-S-H. Aluminium uptake in C-S-H increases strongly at higher aluminium concentrations in the solution.

Figure 1: 29Si NMR of C-A-S-H (a: 91 days, b: 1.5 years) with different Al/Si ratios [1] and Al uptake in C-S-H as a function of Al/Si and Ca/Si ratio [2].

Sodium and potassium are taken up preferentially at low calcium concentrations and thus by low Ca/Si C-S-H as present in e.g. in Portland cements blended with fly ash or silica fume. At a constant Ca/Si ratio, more alkalis are bound at higher pH than at lower pH values, as high pH values lower the calcium concentrations and increase the negative charge, while the presence of aluminium has no effect on alkali uptake. The presence of sodium and potassium does not significantly change the structure of C-S-H although at high concentrations a shortening of the silica chain length has been observed.

References [1] E. L'Hôpital, B. Lothenbach, G. Le Saout, D.A. Kulik, K. Scrivener, Incorporation of aluminium in calcium-silicate hydrate, Cem Concr Res, 75 (2015) 91-103. [2] E. L’Hôpital, B. Lothenbach, D.A. Kulik, K. Scrivener, Influence of calcium to silica ratio on aluminium uptake in calcium silicate hydrate. Cem Concr Res, 85 (2016) 111-121.

Uptake of aluminium and iron by C-S-H

Mancini, A.1,3), Wieland, E1) , Lothenbach, B.2) , Barzgar, S.2) Wehrli, B.3) 1) Paul Scherrer Institute, Laboratory for Waste Management, 5232 Villigen, Switzerland 2) Empa, Laboratory for Concrete & Construction Chemistry, Überlandstrasse 129, 8600 Dübendorf, Switzerland 3) ETH Zurich, Institute of Biogeochemistry and Pollutant Dynamics, Universitätstrasse 16, 8092 Zurich, Switzerland Corresponding author: Mancini, A., email: [email protected]

Calcium silicate hydrate (C-S-H) is the main hydration product in Portland cements and contributes in a significant way to the chemical and physical properties. Furthermore, C-S-H phases have a long-term stability and a high immobilization potential for cations and therefore play a fundamental role in controlling the radionuclide release from a cementitious near field. Aluminium has the affinity to be incorporated in C-S-H forming C-A-S-H phases, but it is not clear whether its uptake can be affected by other trivalent cations, for example Fe(III). Iron can be present in important quantities in different cementitious environment. It can be, for example, liberated by corrosion of reinforcement steel or be added as supplementary material, e.g. ground granulated blast furnace slag (GGBFS). The content of Fe in GGBFS is estimated to be 2-5 wt% in terms of [1] Fe2O3 .

Figure 1: Fe(III) and Al(III) uptake by a C-S-H phase with a C:S ratio of 0.8.

Kinetic and isotherm sorption experiment were carried out using 55Fe as radiotracer in order to determine Fe(III) uptake by cement phases. On the other hand, for aluminium, C-A-S-H phases were synthesized at different Ca/Al ratios and the aluminium uptake was estimated from mass balance calculations. Preliminary data for Fe and Al uptake in a C-S-H phase with Ca/Si ratio of 0.8 show a similar trend for Fe and Al at pH > 11, i.e. linear increase in uptake with increasing concentration. However, Al uptake seems to be higher compared to Fe at similar pH as evidenced from the Al data at pH 12.5 and those for Fe at pH 11.4. This could be attributed to different coordination environments of Al and Fe in C-S-H.

References [1] S.A. Bernal, V. Rose, J.L. Provis (2014). Mater. Chem. Phys. 146, 1.

Extending GEMSFITS to THERMOFIT for using THERMOEXP experimental database for C-(A)-S-H

Miron, G. D.1), Kulik, D. A.1), Lothenbach B.2) 1) Paul Scherrer Institut, Laboratory for Waste Management, 5232 Villigen PSI, Switzerland 2) Empa, Laboratory for Concrete & Construction Chemistry, CH-8600 Dübendorf, Switzerland Corresponding author: Miron, G. D., email: [email protected]

ThermoFit is a parameter optimization tool that builds upon the existing GEMSFITS [1]. It provides means for fitting input parameters of GEM (Gibbs energy minimization) geochemical-thermodynamic models, in an internally consistent way, against various types of experimental or geochemical data. Multiple parameters can be fitted simultaneously against selected experimental properties, along with several options for data weighting, parameter bounds and constraints, and statistical methods.

To be able to accurately model hydrated cement systems, a better representation of the aluminium, alkali and anions uptake in C-S-H phases is needed in chemical models. A large number of different experimental data such as wet chemistry (elemental concentrations in solutions, pH, sorption isotherms), solid phase (composition, water content, and structural data such as mean chain length), and error estimates for measured values are required for developing an internally consistent dataset of thermodynamic parameters (e.g. Gibbs energies, surface-, size-, or non-ideal terms) for the CASHNK aqueous-solid solution system [Kulik, Miron, Lothenbach, this workshop].

Extension of GEMSFITS into ThermoFit becomes possible with the current development of a ThermoExp experimental database, a large collection of various types of experimental data which are used by ThermoFit for simultaneous multi-parameter – multi-property fitting. ThermoExp uses a flexible data format and does not need a priori knowledge of data structure; many types of experimental data can be inserted, describing various experimental properties using hierarchically structured data. On top of the individual data records (e.g. experiment, recipe, data source) stored in the database, a property graph, which maintains links between different data objects (vertices) is present, making it easy to follow links between the data of various kinds. All the data can be connected through graph links of type “cites” to the appropriate bibliography references (“datasources”), thus providing a complete traceability.

The user can create own files (e.g. spreadsheets) containing descriptions of experiments, and upload them to the database without any dedicated format changes. This is done by using an import/export file which connects the input data types to the ThermoExp internal data records. The files should contain all the necessary information to re-create the experiments in GEMS [2] (e.g. temperature, pressure, amount of starting materials, etc.) together with the measured properties. Such experimental database can be used in different parameter optimization tasks, to re-fit existing models or parameterize new ones. On-going applications within the CASH-2 project will be discussed.

References [1] G. D. Miron, D. A. Kulik, S.V. Dmytrieva, T. Wagner, GEMSFITS: Code package for optimization of geochemical model parameters and inverse modeling, Appl Geochem, 55 (2015) 28-45. [2] D.A. Kulik, T. Wagner, S.V. Dmytrieva, G. Kosakowski, F. Hingerl, K.V. Chudnenko, U. Berner, GEM- Selektor geochemical modeling package: revised algorithm and GEMS3K numerical kernel for coupled simulation codes: Comp Geosci, 17 (2013) 1-24.

A PHREEQC version of CEMDATA'18 generated using ThermoMatch

Corresponding author: Miron, G. D., email: [email protected]

A new Cemdata’18 thermodynamic database (TDB) in GEM (Gibbs Energy Minimization) format has been developed [1], which includes standard-state thermodynamic properties of cement-relevant substances and some aqueous complexes determined from various experimental data published in recent years. This update is especially useful for modelling hydrated Portland-, Ca-aluminate-, Ca-sulfoaluminate- and blended cements, as well as alkali-activated materials. The internally consistent Cemdata’18 TDB should be used on top of the GEMS version of the PSI-Nagra TDB [2] supplied with the GEM-Selektor package (http://gems.web.psi.ch). Other popular chemical speciation codes such as PHREEQC [3], use LMA (Law of Mass Action) method that requires thermodynamic data in the form of log K values for both aqueous complexes and solids using formation reactions based on “master” species (usually aqueous ions and water). Earlier versions of Cemdata TDB have already been recasted into PHREEQC format [3]. In order to attract more PHREEQC users to model cementitious systems, the latest Cemdata'18 GEM-type TDB [1] with the related subset of general thermodynamic data from PSI-Nagra TDB [2] was exported into the PHREEQC “dat” format file, also available at https://www.empa.ch/cemdata. For the first time, this operation was automated using ThermoMatch, our new innovative tool for managing TDBs both for substances (GEM format) and reactions (LMA format). Cemdata'18 for GEMS was imported into ThermoMatch, which was can automatically generate sets of reactions from a complete list substances and a selection of master species, and can export the thus created ReactionSet into PHREEQC *.dat file. Supplementary data for aqueous, gaseous and solid species covering the list of elements present in Cemdata’18, were selected in ThermoMatch from the PSI/Nagra 12/07 [2] database. For generating the PHREEQC style reactions for product species, the following aqueous “master” species were selected, based on their generic predominance: Ca+2, Mg+2, Sr+2, Na+, K+, H+, CO3-2, SO4-2, Cl-, NO3-, AlO2-, FeO2-, SiO20, H2O0. For all the “product” species in the Cemdata18 database, the formation reactions were automatically generated, and their properties at 25 °C and 1 bar were calculated. To allow calculations of log10Ko as a function of temperature up to at least 100 °C, parameters for the log10Ko = f(T) polynomials were derived from the 3-term extrapolation method that assumes the ΔrSo ≠ 0, ΔrCpo ≠ 0 independent of temperature. Innovative tools such as ThermoMatch can quickly provide much more performance and flexibility to the cement chemistry community in supplying relevant TDBs of improved quality and consistency in various formats, to embrace as broad circles of specialists as possible. References [1] Lothenbach B., Kulik D.A., Matschei T., Balonis M., Baquerizo L., Dilnesa B.Z., Miron G.D., Myers R. (2017): Cemdata18: A chemical thermodynamic database for hydrated Portland cements and alkali-activated materials. Cem. Con. Res. (2017), in review. [2] T. Thoenen, W. Hummel, U. Berner, E. Curti, The PSI/Nagra Chemical Thermodynamic Data Base 12/07, PSI report 14-04, Villigen PSI, Switzerland (2014). [3] D.J. Parkhurst, C.A.J. Appelo, User's Guide to PHREEQC (version 2) - A computer program for speciation, batch reaction, one dimensional transport, and inverse geochemical calculations, USGS Report, Denver, Colorado (1999). [4] D. Jacques, L. Wang, E. Martens, D. Mallants, Thermodynamic database (CEMDATA07) for concrete in PHREEQC format, SCK-CEN Report (2013). [5] Miron G.D., Kulik D.A., Dmytrieva S.V., Thoenen T., The ThermoMatch User’s Guide, https://bitbucket.org/gems4/thermomatch (in prep., 2018).

C-S-H precipitation kinetics in C3S paste and suspension

Naber, C.1), Bellmann, F.2), Goetz-Neunhoeffer, F.1), Neubauer, J.1) 1) Friedrich-Alexander-University Erlangen-Nürnberg, GeoZentrum Nordbayern, Mineralogy, Schlossgarten 5a, 91056 Erlangen, DE 2) Bauhaus University Weimar, Coudraystraße 11A, 99423 Weimar, DE Corresponding author: Naber, C., email: [email protected]

Calcium-silicate-hydrate (C-S-H) precipitation kinetics are being discussed as the rate limiting reaction step during tricalcium silicate (C3S) hydration. The assessment of the C-S-H precipitation kinetics during hydration is not an easy task. The experimentally observable C3S dissolution and C-S-H precipitation rate at low w/c-ratios will always be nearly equal to one another due to the chemical coupling of both processes and the very small absolute ion content of the pore solution ion reservoir. However, knowledge of the theoretical C-S-H precipitation rate at conditions present during the course of C3S hydration might reveal, whether or not it is the rate limiting kinetic step for the overall hydration rate.

We determined the theoretical C-S-H precipitation rate during the hydration of C3S by combining data from different measurement techniques. For various w/c-ratios, consistent datasets were generated for the first 24 hours of C3S hydration. These datasets include pore solution compositions measured with ICP-OES, phase content developments from XRD and TGA and specific surface area measurements (BET) for a large number of time steps. From the pore solution concentrations, the aquatic species distributions, ion activities and C-S-H saturations states are calculated.

Interfacial C-S-H precipitation rates in solution as a function of the saturation state have recently been reported by Bellmann & Scherer [1]. We combine this rate function and the aforementioned datasets to calculate reaction rates from theoretical C-S-H precipitation rates. In a further step these rates are compared to measured overall C3S hydration rates and to reaction rates from theoretical C3S dissolution rates (Fig. 1). It could be shown, that the calculated C-S-H precipitation rates are in excellent agreement with the measured rates, at least up to the maximum of the C3S hydration. Not only are the features of the induction period and acceleration period reproduced, but also the absolute reaction rate values are matched closely. This hints towards the much debated concept, that the C-S-H precipitation is the rate limiting kinetic step during the induction- and acceleration period of the C3S hydration.

References [1] F. Bellmann, G.W. Scherer, Analysis of C-S-H growth rates in supersaturated conditions, Cem Concr Res (2017), article in press

Figure 1: Comparison of measured C3S hydration rate (XRD) with calculated theoretical reaction rates from C-S-H precipitation and C3S dissolution. 29

Simulation of growing C-S-H phases using 3D sheet growth model

Nguyen-Tuan, L.1), Etzold, M.A.2), Rößler, C.1), and Ludwig, H.M. 1)1) F. A. Finger Institute of Building Material Engineering, Bauhaus-University Weimar, Coudraystraße 11a, 99423 Weimar, Germany. 2) Department of Applied Mathematics and Theoretical Physics, University of Cambridge, Wilberforce Rd, Cambridge CB3 0WB, United Kingdom. Corresponding author: Nguyen-Tuan, L., Email: [email protected]

Despite more than 150 years of use, the physico-chemical processes of the hardening of hydrated cement remain the topic of intensive discussion. Calcium-silicate-hydrate (C-S-H) is of particular importance due to its critical role for strength and durability. Our incomplete knowledge of the 3D meso-structure of C-S-H has been identified as a bottleneck to innovation in the field.

Etzold et al. proposed a sheet model that can create a variety of structures which are in general agreement with those found in cement’s meso-structure [1]. This first model requires an unrealistically high nucleation site density on the cement grains. The disorder in this first model originates mostly in the random orientation of the initial seeds. Etzold et al. therefore investigated sheet growth creating disorder in the growth process itself. They found that a “bifurcation mode” creates highly disordered structures from a single nucleation site [2], so a large number of randomly oriented seeds is not required to introduce disorder. In this paper, we use this new growth mode in a numerical model to describe the formation of C– S–H in hydrating cement paste. The resulting model leads to highly disordered porous structures which fill space rapidly. As time evolves, the open structures densify whilst additional interlayer water is created, thus, leading to a densification as described by Muller et al. [3].

We also introduce strategies how the modelled structures can be compared quantitatively to experimental data, particularly SEM images and NMR.

References

[1] Etzold, M. A.; McDonald, P. J. and Routh, A. F. ‘Growth of sheets in 3D confinements - a model for the C– S–H meso-structure’, Cement and Concrete Research, 2014, Vol. 63: 137-142

[2] Etzold, M. A.; McDonald, P. J.; Faux, D. A. and Routh, A. F. ‘Filling of three-dimensional space by two- dimensional sheet growth’ Phys. Rev. E, American Physical Society, 2015, 92, 042106

[3] Muller, A. C. A; Scrivener, K. L. A; Gajewicz M. and McDonald P. J. ‘Densification of C–S–H measured by 1H–NMR relaxometry’. J. Phys. Chem. C, 2013 117(1):403–412.

New insights into the early reaction of alkali activated slag cements

Puertas, F.1), Gismera, S. 1), Alonso, M. M. 1), Blanco-Varela, M.T. 1), Hoyos-Montilla, A.2), Lanzón, M. 3), Moreno, R. 4), Lothenbach, B. 5), Palacios, M.1) 1) Eduardo Torroja Institute for Construction Sciences (IETcc-CSIC), C/ Serrano Galvache 4, 28033 Madrid, Spain 2) Universidad Nacional de Medellín, 223 Campus Robledo, 050034 Medellín, Colombia 3) Universidad Politécnica de Cartagena, Paseo Alfonso XIII 50, 30203 Cartagena, Spain 4) Ceramic and Glass Institute (ICV-CSIC), C/ Kelsen 5, Campus Cantoblanco 28049, Madrid, Spain 5) EMPA, Überland Str. 129, 8600 Dübendorf, Switzerland Corresponding author: Palacios, M., email: [email protected]

In the recent years, there is an increasing interest to reduce the environmental impact of concrete. One of the most studied strategies is the use of Portland clinker-free cement, such as alkali-activated materials, which are obtained by mixing aluminosilicates with a highly alkaline solution. Alkali-activated slag (AAS) cements are reported to have excellent engineering properties and durability, mainly when waterglass (WG) is used as activating solution.

Previous studies have addressed the influence of the composition of the activating solution and slag, on the microstructure of AAS cements, and consequently on the composition and structure of the main hydration product (C-A-S-H) [1], [2]. Most of these studies have been performed on already hardened samples, however, very few have been focused on the first minutes of slag reaction that can affect properties such as rheology of AAS mortars and concrete.

In this study, a deep characterization of the early hydration products of WG-AAS pastes has been performed. WG-AAS pastes have been characterized from 0 to 20 minutes of hydration by different techniques such as laser diffraction, XRD, FTIR, TG-MS, SEM-EDX, nitrogen adsorption or ICP-OES. Moreover, the early hydration has been modelled using the Gibbs free energy minimization program, GEMS. Results have confirmed the formation of an initial C-A-SH, with an increasing Al content with curing time.

References

[1] F. Puertas, M. Palacios, H. Manzano, J.S. Dolado, A. Rico, J. Rodríguez, A model for the C-A-S-H gel formed in alkali-activated slag cements, J. Eur. Ceram. Soc. 31 (2011) 2043–2056. [2] M. Ben Haha, G. Le Saout, F. Winnefeld, B. Lothenbach, Influence of activator type on hydration kinetics, hydrate assemblage and microstructural development of alkali activated blast-furnace slags, Cem. Concr. Res. 41 (2011) 301–310.

Multiscale modelling of ion transport in cementitious system: Surface charge effects

Yang, Y.1), Patel, R.2), Kosakowski, G.2), Churakov, S.2,3), Prasianakis, N.2) Wang, M.1) 1) CNMM, Tsinghua University, China 2) Laboratory for Waste Management, Paul Scherrer Institute, Switzerland 3) Institute of Geological Sciences, University of Bern, Switzerland Corresponding author: Patel, Ravi A., email: [email protected]

Ion transport is of key importance in assessing the impact of several degradation mechanisms such as sulphate attack, carbonation, leaching, chloride transport in marine structures and contaminant transport in hazardous waste disposalsystems. The diffusivity which characterizes the ability of ions to move in cementitious materials is also a key durability indicator. The ion transport (diffusivity) in cement is controlled by the transport through capillary pores and pores in C-S-H, the so called “gel pores”. The transport through gel pores becomes dominant in moderate to low w/c ratios as demonstrated by Patel et al. [1]. It has been also shown experimentally that for moderate to low w/c ratios anion diffusivity is higher than cation diffusivity [2]. This difference can be attributed to the negative surface charge of C-S-H. Therefore, in order to understand the transport mechanisms and to quantify the diffusivity of cementitious materials it becomes essential to take into account the effect of C- S-H surface charge.

In this study a pore-scale model is developed to predict diffusivity of cement paste from virtual microstructures using the lattice Boltzmann method. The effective diffusivity of C-S-H is predicted from the idealized C-S-H pore structure using modified Poisson-Nernst-Planck equations which takes into account steric and ion-ion correlation effects. To this end the concentration profiles predicted by Poisson-Nernst-Planck equations were corrected based on multiscale atomistic simulations [3]. The upscaled C-S-H diffusivity is used as an input at cement paste-scale where Fick’s law is used to predict effective diffusivity. Finally the model predictions are compared with experimental data available in literature.

References [1] Patel R, Perko J, Jacques D, De Schutter G, Ye G, Van Breugel K. Can a reliable prediction of cement paste transport properties be made using microstructure models? In: Azenha M, Gabrijel I, Schlicke D, Kanstad T, Jensen OM, editors. Service Life of Cement-Based Materials and Structures. Paris, France: RILEM; 2016. p. 203–10. [2] Chatterji, S., Transportation of ions through cement based materials. Part 3 experimental evidence for the basic equations and some important deductions. Cement and Concrete Research, 1994. 24(7): p. 1229-1236. [3] Churakov, S. V., Labbez, C., Pegado, L., & Sulpizi, M. (2014). Intrinsic Acidity of Surface Sites in Calcium Silicate Hydrates and Its Implication to Their Electrokinetic Properties. The Journal of Physical Chemistry C, 118(22), 11752–162

Interaction between C-(A-)S-H and anions

Plusquellec, G.1), Nonat, A.2) 1) RISE, Brinellgatan 4, 50115 Borås, Sweden 2) ICB, UMR 6303 CNRS - Université de Bourgogne, 9 avenue Alain Savary, 21078 Dijon Cedex, France Corresponding author: Plusquellec, G., email: [email protected]

The main phase of hydrated Portland cement is calcium silicate hydrate (C-S-H), which confers to the material its mechanical properties. The interactions between C-S-H and anions are an important issue mainly from the point of view of durability of cementitious materials. For example, chlorides intervene in the corrosion of the reinforcement bars of concrete, and the uptake/release of sulphate by C-S-H may cause expansion or cracking of the material (by the formation of delayed ettringite). The aim of this study is to investigate the interaction between anions and C-S-H and to model those interactions.

The interaction between C-(A-)S-H and anions (chloride, bromide and nitrate) has been studied using an in-situ analysis device which allowed us to analyse directly a C-S-H suspension without any exterior interference, e.g. filtration. This in-situ device can measure the conductivity, pH and ions activity (Ca2+, Cl- or Br-) of a C-(A-)S-H suspension while concentrated salt solution is added. C-(A-)S-H with various Ca/Si ratio have been used. The presence of C-S-H particles does not influence the activity of chloride in solution (see Figure 1), indicating that there is no adsorption of Cl- ions by C-S-H. Zeta potential measurements of C-S-H suspensions in equilibrium with various salt has also been carried out and confirmed the previous observation. Both in-situ analysis and zeta potential showed an adsorption of calcium by C-(A-)S-H.

The thermodynamic modelling was made by using the software PhreeqC in the Pitzer approximation, based on a C-S-H surface reaction model developed by Haas and Nonat [1]. This modelling allows us to calculate the ions sorption by C-S-H and then the conductivity and the pH of suspension which can be directly compared to the obtained experimental data.

250 Reference solution (without C-S-H particles) C-S-H suspension (Ca/Si = 1.0) 200

150

- Cl specific Cl - electrode 100

U (mV) U 50 0 50 100 150 200 [CaCl ] (mmol/L) 2 added Figure 1:Evolution of the response of the chloride specific electrode in function of the concentration of CaCl2 during the analysis of the reference solution (without C-S-H particles) and a C-S-H suspension with a Ca/Si ratio of 1.0. References [1] J. Haas, A. Nonat, From C–S–H to C–A–S–H: Experimental study and thermodynamic modelling, Cem Concr Res, 68 (2015) 124-138.

Thermodynamic modelling of synthetic C(-A)-S-H using the pitzer model – what is gained?

1)Prentice, D. P., 1)Bernal, S. A., 2)Bankhead, M., 2)Hayes, M. and 1)Provis, J. L. 1) Department of Materials Science and Engineering, University of Sheffield, Sir Robert Hadfield Building, Mappin Street, Sheffield, S1 3JD 2) National Nuclear Laboratory, UK Corresponding author: Prentice, D. P., email: [email protected]

In recent decades, thermodynamic modelling of cement phases has become more accurate and approachable as a result of the availability of software such as GEMSelektor [1]. Solid-solution modelling of the C(-A)-S-H phase, coupled with the Truesdell-Jones aqueous solution model, has become commonplace for predicting cement hydrate phase assemblages. The Truesdell-Jones model has an optimal ionic strength (IS) range between 0.5 and 1.0 M which is in the region observed in many hydrated cements. However, the complexity of cement systems based on changing precursor materials, or utilization of cement to produce wasteforms, may require a more sophisticated aqueous solution model. The Pitzer model [2] has the capacity to predict properties of aqueous solutions with IS from 0.5 to 6 M. The Pitzer model is significantly more complex and requires multiple input parameters, which makes it challenging to use. Here, a database lookup programme paired with GEMS was used to provide these parameters in a streamlined process. An adapted C(-A)-S-H model was used to compare the two aqueous solution modelling approaches by focusing on C-S-H and C-A-S-H formation in alkaline solution using data from the literature, e.g. [3]–[5]. Both models gave similar results in this system, but the available data concerning synthetic formation of these phases have an IS range of 0.1 to 0.5 M which does not allow for full utilization of the potential of the Pitzer model at higher IS.

References [1] D. A. Kulik, T. Wagner, S. V. Dmytrieva, G. Kosakowski, F. F. Hingerl, K. V. Chudnenko, and U. R. Berner, “GEM-Selektor geochemical modeling package: revised algorithm and GEMS3K numerical kernel for coupled simulation codes,” Comput. Geosci., vol. 17, pp. 1–24, 2013. [2] K. S. Pitzer, “Ion interaction approach: theory and data correlation,” in Activity Coefficients in Electrolyte Solutions, K.S. Pitzer (ed.), CRC Press, Boca Raton, 1991. [3] S. Y. Hong and F. P. Glasser, “Alkali sorption by C-S-H and C-A-S-H gels: Part II. Role of alumina,” Cem. Concr. Res., vol. 32, no. 7, pp. 1101–1111, 2002. [4] E. L’Hôpital, B. Lothenbach, K. Scrivener, and D. A. Kulik, “Alkali uptake in calcium alumina silicate hydrate (C-A-S-H),” Cem. Concr. Res., vol. 85, pp. 122–136, 2016. [5] S.-Y. Hong and F. P. Glasser, “Alkali binding in cement pastes,” Cem. Concr. Res., vol. 29, no. 12, pp. 1893–1903, 1999.

Characterisation of C-(A)-S-H phases using Slected area dirffraction in the TEM

Rößler, C.1), Sowoidnich, T.1), Ludwig, H.-M. 1) Bauhaus-Universität Weimar, Coudraystr.11, 99421 Weimar, Germany Corresponding author: Rößler, C., email: [email protected] Characterisation of structural properties of C-S-H phases is a key aspect for understanding properties of cement based materials. Nanoscale nature and water contained in structure of C-S-H pose a challenge for structural characterisation. As a result, C-S-H phases are often described as amorphous or gel-like. Previous studies showed that TEM diffraction experiments at low-electron dose and cryo-conditions clearly reveal a crystalline nature of single fibrous C-S-H and foil-like C-A-S-H phases obtained from cements hydrated for 28d [1, 2]. A similar approach was followed in the present study to characterise C-(A)-S-H phases synthesised from pure phases or in hydrated pastes with and without addition of biopolymer (BP).

A: B: C: Fibrous C-S-H from CEM I and SAED of fibrous C-S-H from SAED of foil-like C-A-S-H C3S hydration. CEM I and C3S+BP hydration. from hydration of CEM III/B.

D: E: F: Foil-like C-S-H from synthesis SAED of C-S-H from synthesis SAED of foil like C-A-S-H (Ca/Si 1.5). (Ca/Si 1.5) with and without BP. from synthesis (Ca/Si 1.5, Al/Si 0.05). Figure 1A-F: TEM images (A, D) and selected area diffraction pattern (SAED) obtained for C-(A)-S-H phases synthesised under variable conditions. Results in Figures 1 A-F clearly show that all investigated C-(A)-S-H phases possess a crystalline structure. Thereby SAED of C-S-H from synthesis have spottiest pattern and C-A-S-H from synthesis weak ring pattern. These results show that crystallinity/crystallite size of C-(A)-S-H decreases according to following order of displayed images: E>B>C>F. In addition, the presence of BP does not change the SAED of fibrous C-S-H but only the diffraction pattern of synthesised C-S-H (BP leading to slightly increased d-spacings). These results demonstrate the crystalline nature of all investigated C-(A)-S-H phases and point out that synthesised C-S-H are structurally not identical to fibrous C-S-H that were grown in CEM I and C3S pastes. References [1] C. Rößler, J. Stark, F. Steiniger, W. Tichelaar, Limited-Dose Electron Microscopy Reveals the Crystallinity of Fibrous C--S--H Phases, J. Am. Ceram. Soc., 89 (2006) 627-632. [2] C. Rößler, F. Steiniger, H.M. Ludwig, Characterization of C–S–H and C–A–S–H phases by electron microscopy imaging, diffraction, and energy dispersive X‐ray spectroscopy, J. Am. Ceram. Soc., 100 (2017) 1733-1742.

Synthetic Calcium Silicate Hydrate –formation kinetics – the key to understanding cement microstructures? Siramanont, J., Bowen, P. Powder Technology Laboratory (LTP), Materials Science and Engineering, Ecole Polytechnique Fédérale de Lausanne (EPFL,) Switzerland. Corresponding author: Bowen, P., email: [email protected]

The cement industry is the single biggest industrial production contributor to CO2 emissions, which is expected to double in the next 30 years. The reduction in this major CO2 emission source is one of the most urgent tasks facing the scientific and industrial community. The main route to do this economically seems to be by replacing the cement used in concrete with substitutes, such as calcined clays and industrial waste products (blast furnace slag, fly ash, silica fume) The main limitation with this approach is a reduction in reactivity compared to current cements which leads to low early age strength and higher costs. Such a formidable task cannot be done without a deep understanding about the hydration of cementitious materials, a process responsible for strength development in Portland cement-based concrete. Calcium-silicate-hydrate (C-S-H) is the most important phase formed during the hydration of cementitious materials, being responsible for many of their engineering properties. Therefore, acquiring a comprehensive knowledge about the formation pathway of C-S-H appears to be a key step toward a more sustainable cement industry. We have developed a synthetic approach by precipitation which allows any stoichiometry from 1 to 2 to be formed as a single phase uniform C-S-H particles [1]. The set-up also allows us to collect kinetic data (e.g. Ca2+ activity as a function of time) which has been used to develop a coupled thermodynamic-kinetic computational model based on a population balance equation [2]. From the model we can identify the kinetic speciation during the precipitation process and the mechanisms and activation free energies of nucleation and growth phenomena. We can perform the precipitation in the presence of organic additives or heterogeneous substrates such as quartz or limestone (CaCO3). We have successfully collected data both morphological and kinetic in the presence of either small organic molecules (D-gluconate) or quartz and limetstone. Both the organic molecules and the heterogeneous (non-reactive) substrates have an influence on the nucleation and growth. This depends on the formation of complexes with the D-gluconate and on the nucleation density for the heterogeneous solids. Discussion of the interpretation of the kinetic data using our population balance method will be presented.

Figure 1 Nanaofoils of synthetic C-S-H on Quartz (Q2) and limestone (L2) respectively.

[1] A. Kumar, B. J. Walder, A. K. Mohamed, A. Hofstetter, B. Srinivasan, A.J. Rossini, K. Scrivener, L. Emsley and P. Bowen ”The Atomic-Level Structure of Cementitious Calcium Silicate Hydrate”, J.Phys.Chem.C 121(32) 17188–17196 (2017). DOI: 10.1021/acs.jpcc.7b02439 [2] M.R.Andalibi, A.Kumar, B. Srinivasan, P. Bowen, K. Scrivener, C. Ludwig, A. Testino, “On the Mesoscale Mechanism of Calcium-Silicate-Hydrate Precipitation: A Population Balance Modeling Appraoch”, J.Mater.Chem.A – in press, 6(2) p.xxx-xxx, 2018

Uptake of heavy metal ions during C-S-H precipitation

Steindl, F. R. 1), Baldermann, A. 1), Galan, I. 1), Sakoparnig, M. 2), Mittermayr, F. 2) 1) Graz University of Technology, Institute of Applied Geosciences, Rechbauerstraße 12, 8010 Graz, Austria 2) Graz University of Technology, Institute of Technology and Testing of Building Materials, Inffeldgasse 24, 8010 Graz, Austria Corresponding author: Steindl, F. R., email: [email protected]

The sorption and leaching of heavy metal ions from cementitious materials is a long-running research field in concrete and cement science. Several cementitious phases such as C-A-S-H or AFt phases have been shown to take up large amounts of hazardous metal ions, both through chemisorption and physical absorption. However, experimental and thermodynamic data on the uptake of Co2+, Zn2+ and Cr3+/6+ during C-A-S-H formation is limited due to the complexity of C-A-S-H phase precipitation under various physicochemical conditions. Several synthesis methods for C-A-S-H, such as hydration of C3S, synthesis from a mixture of H2O, CaO and SiO2 and precipitation from a metasilicate solution using a highly soluble Ca2+ salt have been used in the literature. We present a study, conducted in the course of the Austrian FFG research project “Advanced and sustainable sprayed concrete (ASSpC)”, on the heavy metal removal potential by C-A-S-H using a co-precipitation approach. In detail, CaCl2, AlCl3.6H2O and heavy metal (either Co, Zn or Cr) chlorides were added to a CO2-free -1 94.3 mM Na2SiO3 solution (20 g l Na2SiO3.5H2O). The Al/Si molar ratio was fixed to 0.05 and the Ca/(Al+Si) molar ratio was varied between 0.6, 1.0 and 1.6. The molar ratio between the heavy metals and Si was varied between 0.02, 0.2 and 2. The precipitation reaction was monitored by pH and conductivity measurements and regular sampling of the liquid phase. The precipitation of C-A-S-H containing metal ions was an extremely fast process, as it can be seen by the almost complete removal of Si, Al and heavy metals in almost all experiments within minutes. The metal ion uptake increases in the order of Cr < Co < Zn. The removal efficiencies were > 99.8% for molar ratios of 0.02 and 0.2 and 31.5 to 76.5% for molar ratios of 2. The ongoing analyses of the solids by transmission electron microscopy in combination with hydrochemical modelling of fluid-solid equilibria will help to gain new insights into the crystal chemistry and solubility of Me-bearing C-A-S-H phases.

Investigations on the intercalation of An(III)/Ln(III)-malate complexes in C-S-H phases

Taube, F.1), Acker, M.1), Rossberg, A.2) Stumpf, Th. 2) 1) TU Dresden, Central Radionuclide Laboratory, 01062 Dresden, Germany 2) Helmholtz-Zentrum Dresden-Rossendorf, Institute of Resource Ecology, 01328 Dresden, Germany Corresponding author: Taube, F., email: [email protected]

Concrete widely serves as an engineering barrier and for waste conditioning in nuclear waste repositories. Organic additives are commonly used for tuning the physico-chemical and mechanical properties of fresh concrete. A variety of organic additives, like modified lignosulphonates, sulphonated melamine and naphthalene, polyacrylates and poly(hydroxo)carboxylates contain strong complexing functional groups, which can interact with radionuclides (RN) forming stable complexes. In the worst-case scenario of water intrusion into the waste repository, the concrete may degrade, so that the soluble organic additives will be leached out and will complex the RN. These complexes may contact the concrete. If the RN-complexes with organic ligands (additives) show higher formation constants than the surface complexation with CSH phases, they could intercalate in the cement interlayers. Thus, the mobility of the RN-complexes would be reduced. Consequently, for a long-term risk assessment in nuclear waste repositories, the interaction of the RNs, here Am(III) and its analogue Eu(III), with cement additives must be known. As cement additive malic acid (α-hydroxydicarboxylic acid) was chosen. Extended X-ray absorption fine structure (EXAFS) spectroscopy was used for the determination of the structural properties (coordination numbers and atomic distances) of Am(III)-malate complexes at highly alkaline pH values and of the ternary Am(III)/malate/CSH system. In the high pH range the EXAFS spectra of the Am- malate complexes show an Am-Am interaction at 3.5 - 4 Å. This observation would be in line with the formation of colloidal or precipitated Am(III)-malate complexes. Complementary, time-resolved laser-induced fluorescence measurements with Eu(III) and Eu(III)-malate complexes within CSH phases were conducted. In agreement with literature, at least two species were found irrespective of the presence of malate [1]. One species shows a long lifetime of 1500µs and the other species having a shorter lifetime of 300-700µs. This can be interpreted as one intercalated Eu3+ species and one superficial complex or precipitate.

2 Figure 1: Left: k -weighted Am LIII-edge EXAFS spectra and Fourier transform (FT) of an Am-malate-species at pH 11. Right: TRLFS spectra of aqueous CSH-phases with 1500 ppm Eu(III) (C/S=0.4), corresponding decay curve (inset).

References [1] I. Pointeau, B. Piriou, M. Fedoroff, F. Fromage, Sorption mechanisms of Eu3+ on CSH phases of hydrated cements, J. of Colloid and Interface Science 236 (2001) 252-259.

Multiscale (force field - semiempirical - density functional - ab initio) modelling of C-A-S-H

Veryazov, V. 1), Kovačević, G.2) 1) Lund University, Kemicentrum, POB 124, Lund 22100, Sweden 2) Ruđer Bošković Institute, Zagreb, Croatia Corresponding author: Veryazov V., email: [email protected]

Theoretical modeling of non-isovalent substitutions in concrete materials is a challenging problem. Atomistic, force-field based, simulations are widely used in prediction of the structure of materials and interfaces with solutions. However, the results obtained with force field potentials are very sensitive to the protocol used in the simulation: parametrization, assigned point charges, connectivity, etc. Quantum chemical methods are more expensive, but they are more reliable and can provide insight into the electronic structure of the material.

The cost of precise quantum chemical calculations, especially for the geometry optimization, impose us to apply multiscale procedure: starting from preliminary force-field simulations to semiempirical and density functional calculations. The final refinement of the electronic structure can be performed by ab initio calculations of small clusters embedded into an electrostatic field.

Earlier we applied a multiscale approach to study the structure of C-S-H with high concentration of Ca [1-2].

The current study has focus on Al additives in C-S-H, its positions in the crystal structure and the role in the improvements of macroscopic properties of concrete.

References [1] G. Kovačević, B. Persson, L. Nicoleau, A. Nonat, V. Veryazov, “Atomistic modeling of crystal structure of Ca1.67SiHx ”, Cement Concrete Research 67, 197-203 (2015) [2] G. Kovačević, L. Nicoleau, A. Nonat, V. Veryazov, “Revised atomistic models for the crystal structure of C- S-H with high C/S ratio”, Z. Phys. Chem. 230 (9), 1411-1424 (2016)

C-S-H gel solubility modelling at high temperatures

Walker, C. S.1), Anraku, S.1) Mitsui, S. 1) Oda, C. 1) Mihara, M. 1) Honda, A. 1) 1) Japan Atomic Energy Agency (JAEA), 4-33 Muramatsu, Tokai, 319-1194, Japan.

Calcium silicate hydrate (C-S-H) gel has long been recognized as the main component of hydrated ordinary Portland cement (OPC) based materials. C-S-H gel has a variable composition in terms of molar Ca/Si ratios ≈ (0.6 ± 0.1) to (1.7 ± 0.1), co-exists with amorphous silica (S(am)) for lower and portlandite (CH) for higher Ca/Si ratios, only dissolves congruently at Ca/Si ratios = (0.85 ± 0.05), and has an upper stability limit of t ≈ 90oC. JAEA have developed a C-S-H gel solubility model that is able to provide an account of all these features, which helps to improve predictions of the long-term chemical degradation of hydrated OPC based materials.

C-S-H gel solubility data have been compiled from 35 publications in the literature and supplemented in the current study with solubility experiments conducted at 25, 50 and 80oC, to give a master C-S-H gel solubility dataset containing over 1200 experiments. An increase in temperature causes a decrease in pH for all Ca/Si ratios, a decrease in Ca concentrations for Ca/Si ratios > 1.0, but has little effect on Si concentrations (Figure 1). Using selection criteria, the master C-S-H gel solubility dataset was reduced to 371 solutions that were considered representative of C-S-H gel at thermodynamic equilibrium.

These C-S-H gel solubility data were then used to derive a discrete solid phase (DSP) type model adapted from Walker et al. [1]. The DSP C-S-H gel model is based on two binary non-ideal solid solutions in aqueous solution (SSAS) described by solidus and solutus equations on Lippmann phase diagrams [2]. The first SSAS used end- members of S(am) and C0.8333SH1.3333 and the second CS1.2H1.56 and CH. Entropic terms [3] were included in Margules equations that were used to account for non-ideality of the two SSAS and delimit the compositional range of C-S-H gel for Ca/Si ratios = 0.55-1.65 at t = 0.01-90oC. Satisfactory predictions of the C-S-H gel solubility data could be made using this approach (Figure 1).

Figure 1: C-S-H gel solubility data shown as a function of molar Ca/Si ratio and temperature: [a] pH, [b] Ca, [c] Si. Symbols = experimental data, lines = model predictions at 25, 50 and 85oC. This study was partly funded by the Ministry of Economy, Trade and Industry of Japan through “The project for validating assessment methodology in geological disposal system” in JFY 2015.

References [1] C.S. Walker, S. Sutou, C. Oda, M. Mihara, A. Honda, Calcium silicate hydrate (C-S-H) gel solubility data and a discrete solid phase model at 25oC based on two binary non-ideal solid solutions, Cem Concr Res, 79 (2016) 1-30. [2] P.D. Glynn, E.J. Reardon, Solid-solution aqueous-solution equilibria: Thermodynamic theory and representation, Am J Sci, 290 (1990) 164-201. [3] J. Ganguly, W. Cheng, M. Tirone, Thermodynamics of aluminosilicate garnet solid solution: new experimental data, an optimized model, and thermodynamic applications, Contrib Mineral Petrol, 126 (1996) 137-151.

Stability of U(VI) and CM(III) doped calcium silicate hydrate phases in high saline brines

Wolter, J-M., Schmeide, K., Stumpf, T. Helmholtz-Zentrum Dresden - Rossendorf, Institute of Resource Ecology, Bautzner Landstraße 400, 01328, Germany Corresponding author: Wolter, J-M., email: [email protected]

U(VI) and Cm(III) doped calcium silicate hydrate (CSH) phases with different C/S ratios (1.0-2.0) were synthesized by direct actinide incorporation and characterized by time-resolved laser-induced fluorescence spectroscopy (TRLFS), infrared (IR) spectroscopy, powder X-ray diffraction (XRD), thermogravimetric analysis (TGA), differential scanning calorimetry (DSC) and scanning electron microscopy (SEM). The time-dependent release of Ca, Si, U and Cm from CSH phases into brines that contained either 2.5 M NaCl, 2.5 M NaCl/0.02 M Na2SO4, 2.5 M NaCl/0.02 M NaHCO3 or 0.02 M NaHCO3 for U(VI) doped CSH phases or 2.5 M NaCl/0.02 M NaHCO3 or 0.02 M NaHCO3 for Cm(III) doped CSH phases was monitored in batch leaching experiments for 30 or 60 days, respectively. Subsequently, leaching induced changes of the CSH structure and of the U(VI) or Cm(III) coordination environment were investigated with TRLFS, IR spectroscopy and XRD. Wavelength [nm] 455 476 500 526 556 588 625 667

0-1 0-2 60 d 14 d 0-3 Cm(III)/CSH main band 621.4 nm before leaching

Cm(III)/vaterite after leaching in 612.8 nm NaCl

after leaching in

NaCl/Na2SO4 Intensity [a. u.] [a. Intensity Intensity [a. u.] [a. Intensity Cm(III)/calcite after leaching in 608.0 nm

NaCl/0.02 M NaHCO3

after leaching in

NaCl/0.075 M NaHCO3 Cm(III)/CSH hot band 605.2 nm

22,000 21,000 20,000 19,000 18,000 17,000 16,000 15,000 600 605 610 615 620 625 630 Wavenumber [cm-1] Wavelength [nm]

Figure 1: TRLFS spectra of U(VI) doped CSH phases with a C/S ratio of 1.6 before and after leaching (l.) and site-selective TRLFS spectra of a Cm(III) doped CSH phase with a C/S ratio of 1.0 after leaching in 2.5 M NaCl/0.02 M NaHCO3 (r.). Site-selective TRLFS studies of the Cm(III)/CSH binding at 4 K revealed a fluent transition between two sorption sites causing a luminescence line-narrowing effect. The leached CSH phases showed pronounced differences in terms of decomposition behavior and actinide release depending on their C/S ratio and type of incorporated actinide. CSH phases with a lower C/S ratio were influenced strongly by NaHCO3 and showed a mobilization of U(VI) as Ca2UO2(CO3)3(aq). In contrast, Cm(III) was not leached out but as detected by site- selective TRLFS it is incorporated into calcite and vaterite (Fig. 1 (r.), 608 and 612.8 nm) formed during leaching in NaHCO3. The comparison between leaching experiments performed in 0.02 M NaHCO3 and 2.5 M NaCl/0.02 M NaHCO3 revealed that the presence of 2.5 M NaCl increases the U(VI) mobilization as Ca2UO2(CO3)3(aq) complex while no influence on the Cm(III) release was detectable.

Alkali uptake evaluations of C-A-S-H with structure analyse by NMR and of degradated OPC paste

Yamada, K.1), Haga, K.2), Watanabe, S.3), Harasawa, S.4) 1) National Institute for Environmental Studies, Fukushima Branch, Fukasaku 10-2, Miharu, Tamura, 963-7700 Fukushima, Japan 2) Taiheiyo Consultant, Co.Ltd., Tokyo Office, Higashi-Nihonbashi2-27-8, Chuo-ku, 103-0004 Tokyo, Japan 3) Tokyo Metropolitan Industrial Technology Research Institute, Aomi2-4-10, Koto-ku, 135-0064Tokyo, Japan 4) Taiheiyo Consultant, Co.Ltd., 2-2-16 Osaku, Sakura, 265-8655 Chiba, Japan Corresponding author: Yamada, K., email: [email protected]

Alkali uptake of C-A-S-H has been studies by many researchers. However, there still are some obscure points. First point is the change of crystalline structure during sorption test. Second point is the difference of crystalline structures between synthesized one and hydrated cement paste. Third point is the effect of silica from C-A-S-H by leaching of Ca and carbonation. Fourth point is the mechanism of different uptake amount depending on the kind of alkalis.

In this presentation, alkali uptake characteristics of C-A-S-H having various composition are studied. C-A-S-H having C/S = 0.8 – 1.2 and Al/Si = 0 -0.1 as starting composition using CaO, amorphous silica, γ-Al2O3 were synthesized at liquid/ solid ratio of 20, 20 ºC for 7 days and their silicate chain structures were evaluated by 29Si- and 27Al-NMR before and after immersion test in NaCl solution of 10mM with liquid/ solid ratio of 10 for 7 days. Of course, the structures of synthesized C-A-S-H were a little bit different from designed ratio. Major part of Al was in IV coordinate ad designed at low C/S ratio but less than half was in that form at higher C/S ratio near 1.1. After immersion test, Q1 decreased and Q2 increased, meaning the silicate chain length increased 13 % by reacting with Al remained in solution or other phases. Na uptake amount against Ca/(Al+Si) is shown in Fig. 4. Na uptake showed negative linear correlation with Ca(Al+Si). C-A-S-H having Ca/(Al+Si) ratio around 0.9 showed higher Na uptake but these C-A-S-H had tobermorite-like structure.

Obviously, more silanol groups at lower Ca/(Al+Si) ratio plays as adsorbing sites for alkalis. Therefore, leaching of Ca and carbonation generating free silanol more contribute to more uptake amount. Quantitative discussion will be made in the presentation. 6

2 C-S-H

1 C-A-S-H Na uptake (mmol)/100g (mmol)/100g solid of Nauptake 0 0.6 0.8 1.0 1.2 Ca/(Al+Si) of silicate chain from Si-NMR

Figure 1: Na uptake amount vs. Ca/(Al+Si) ratio determined by Si-NMR Acknowledgement: Part of this work was financially supported by the Japan Society for the Promotion of Science (JSPS, No.17H03292).

Structural and mechanical property determination of calcium silicate hydrate (C-S-H): a theoretical and experimental study

Yan, Y.1,2), Jiang, J.1,2), Hou, D.3) 1) School of Material Science and Engineering, Southeast University, Nanjing 211189, China 2) Jiangsu Key Laboratory of Construction Materials, Nanjing 211189, China 3) Qingdao University of Technology, Qingdao, China.

Calcium silicate hydrate (C-S-H), the main hydration product in Portland cement, influences the physical and mechanical properties and the durability of cementitious materials. Poorly crystalline C-S-H is synthesized and characterized by X-ray diffractometry (XRD), Fourier transform infrared spectroscopy (FT-IR), transmission electron microscope (TEM) and 29Si magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectroscopy. Density Functional Theory (DFT) and molecular dynamics modelling were applied to reveal the structural and mechanical properties of C-S-H. Comparison of experimental and theoretical results showed structural disorder in C-S-H and provided insight in the nanoscale characteristics of C-S-H.

Experiment-C-S-H (I) a Experiment-T11A b Simulation-C-S-H

Intensity(a.u.)

5 10 15 20 25 30 35 40 2 Figure 1 (a) The constructed C-S-H model. (b) X-ray data for synthesized C-S-H and simulated 11 Å tobermorite, C-S-H model.

The role of Aluminum in Calcium Silicate Hydrate Phases: A Multinuclear Solid-State NMR investigation

Yang, S.1), Lothenbach, B. 2), Skibsted, J. 1) 1) Department of Chemistry and Interdisciplinary Nanoscience Center (iNANO), Aarhus University, 8000 C Aarhus, Denmark 2) Laboratory for Concrete & Construction Chemistry, Empa, 8600 Dübendorf, Switzerland Corresponding author: Sheng-Yu Yang, e-mail: [email protected]

A significant reduction in CO2 emissions associated with cement production can be obtained by partly replacing Portland cement by SCMs, which in most cases are aluminosilicate-rich materials such as fly ashes, slags and calcined clays. During hydration of such Portland cement – SCM blends, this results in the formation of a calcium silicate hydrate (C-S-H) phase that incorporates a significant amount of aluminum. The present study aims to structurally characterize and quantify the uptake of Al in the C-S-H phase, which is a key-factor for our understanding of the fundamental chemistry of Portland cement blends with substantial SCM replacements.

Figure 1: 27Al MAS NMR spectra of a synthesized C-A-S-H sample with Ca/Si = 1.0 and Al/Si = 0.1, stored in a 0.1 M NaOH solution for three months. The spectra are acquired at three different magnetic fields (asterisks indicate spinning sidebands).

This presentation focusses on a series of C-A-S-H samples synthesized with Ca/Si ratios of 0.6 – 1.6, an Al/Si ratio 0.10, and concentrations of 0.1, 0.5, and 1.0 M NaOH. The samples have been characterized by 23Na, 27Al and 29Si MAS NMR with the main focus on the 27Al NMR experiments which have been performed at an ultra- high magnetic field of 22.3 T in addition to spectra at 14.1 T. Moreover, 27Al MQMAS NMR spectra have been acquired at both magnetic fields, which in particular improve the resolution of the resonances from tetrahedrally coordinated Al. For example, the 27Al MAS NMR spectra at 14.1 and 22.3 T in Figure 1 allow identification of three different Al(4) sites where the resonance at 61 ppm originates from an impurity of strätlingite in the sample. The 27Al and 29Si NMR spectra reveal that the use of 0.5 and 1.0 NaOH solutions in the synthesis gives nearly the same C-S-H compositions whereas 23Na NMR shows that Na+ ions are incorporated in the C-S-H structure.

Influence of calcium to silica ratio on H2 gas production in calcium silicate hydrate

Yin, C.1), Dannoux-Papin, A.1), Haas, J.1), Cau Dit Coumes, C.1) Renault, J-P.2) 1) CEA, DEN, MAR, DE2D, SEAD, Laboratoire d’études des Ciments et Bitumes pour le Conditionnement, 30207 Bagnols sur Cèze Cedex, France 2) CEA, DRF, IRAMIS, NIMBE, Laboratoire Interdisciplinaire sur l’Organisation Nanométrique et Supramoléculaire, 91191 Gif-sur-Yvette Cedex, France Corresponding author: YIN, C., email: [email protected]

Water radiolysis is one of the consequences of the interaction between a cementitious matrix and low or intermediate-level radioactive waste that are encapsulated within this matrix. Thus, for safety assessment, water radiolysis issued from pore solution and hydrates must be considered. This work is focused on the interactions at the interface between water and calcium silicate hydrate (C-S-H), a nanocrystallized product with a layered chemical structure. The first aim of this study is to understand the radiolytic mechanisms that promotes or inhibits the hydrogen production in C-S-H. Therefore, after characterization by several techniques (nitrogen gas adsorption-desorption, X-ray diffraction, thermogravimetric analysis, infrared and Raman spectroscopies, small- angle X-ray scattering), C-S-H with different C/S ratios were submitted to gamma irradiation to determine their H2 radiolytic yield, G(H2)i.The first results show that all the G(H2) values in C-S-H(cured at 60% RH) are approximately 2 to 6 times higher than those measured for bulk water. Hydrogen production is thus enhanced in C-S-H, probably due to a confinement effect in porosity1 (pore average diameter from 143 to 210 Å obtained using a Barrett-Joyener-Halenda model) and / or in the interlayer space2 (9 to 14 Å). Moreover, the radiolytic hydrogen production decreases with increasing C/S ratio from 0.8 to 1.4 (see Figure 1). The G(H2) value obtained at C/S ratio 1.6 may be overestimated, which might be due to the presence of traces of portlandite in the sample. A change in the average electronic density of the system in the water radiolysis requires to be

investigated in C-S-H, as well as that of bridging and non-bridging oxygens.

) -1

2,5

mol.J

-7

10 ( 2,0

1,5

1,0

0,5 H2 radiolytic yield of free water

0,44

normalized to water content water to normalized

) 2

H 0,0 (

G 0,8 1,0 1,2 1,4 1,6 C/S ratio Figure 1. The normalized H2 radiolytic yields as a function of the C/S ratio. The normalized G(H2), is equal to the G(H2) divided by its water mass fraction. The value obtained in liquid bulk water is given as a comparison (4.4 × 10-8 mol J-1) to point out the specific behavior of confined water. For the future work, experiments with Electron paramagnetic resonance spectroscopy will be conducted at low temperature in order to determine defects centers and to quantify their concentration issued from the radiolysis of C-S-H. References 1 Le Caër, S. et al. Radiolysis of confined water: Hydrogen production at a high dose rate. ChemPhysChem 6, 2585-2596 (2005). 2 Laine, M. et al. Reaction mechanisms in swelling clays under ionizing radiation: influence of the water amount and of the nature of the clay mineral. RSC Advances 7, 526-534, doi:10.1039/C6RA24861F (2017).