Two-Step Nucleation Process of Calcium Silicate Hydrate, the

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Two-Step Nucleation Process of Calcium Silicate Hydrate, the Nano-Brick of Cement Nina Krautwurst, Luc Nicoleau, Michael Dietzsch, Ingo Lieberwirth, Christophe Labbez, Alejandro Fernandez-Martinez, Alexander van Driessche, Bastian Barton, Sebastian Leukel, and Wolfgang Tremel Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.7b04245 • Publication Date (Web): 17 Apr 2018 Downloaded from http://pubs.acs.org on April 17, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Two-Step Nucleation Process of Calcium Silicate Hydrate, the Nano-Brick of Cement Nina Krautwurst1, Luc Nicoleau2, Michael Dietzsch2, Ingo Lieberwirth3, Christophe Labbez4, Alejandro Fernandez-Martinez5, Alexander van Driessche5, Bastian Barton1, Sebastian Leukel1and Wolfgang Tremel1 1

Institute of Inorganic Chemistry and Analytical Chemistry, Johannes Gutenberg-University Mainz, Duesbergweg 10-14, D-55099 Mainz, Germany. 2 BASF Construction Solutions GmbH, B08, Dr-Albert-Frank-Strasse 32, 83308 Trostberg, Germany. 3 Max Planck-Institut für Polymerforschung, Ackermannweg 8-10, D-55128 Mainz, Germany. 4 Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR 6303 CNRS, Univ. Bourgogne Franche-Compté, F-21078 Dijon Cedex, France. 5 ISTerre - CNRS & Univ. Grenoble Alpes, 1381 Rue de la Piscine, 38041 Grenoble, France.

Despite a millennial history and the ubiquitous presence of cement in everyday life the molecular processes underlying its hydration behavior, like the formation of calcium-silicate-hydrate (C-S-H), the binding phase of concrete, are mostly unexplored. Using time-resolved potentiometry and turbidimetry combined with dynamic light scattering, small angle X-ray scattering and cryo-TEM we demonstrate C-S-H formation to proceed via a complex two-step pathway. In the first step, amorphous and dispersed spheroids are formed, whose composition is depleted in calcium compared to C-S-H and charge compensated with sodium. In the second step these amorphous spheroids crystallize to tobermorite-type C-S-H. The crystallization is accompanied by a sodium/calcium cation exchange and aggregation. Understanding the formation of C-S-H via amorphous liquid precursors may allow for a better understanding of the topography of the nucleation in cement paste and thus the percolation of hydration products leading to the mechanical setting as well as the retarding effect of known chemical species like aluminum ions and polycarboxylate ethers.

INTRODUCTION

particular calcium carbonate. Calcium carbonate polymorphs have been reported to nucleate from aggregation and coalescence of pre-nucleation clusters formed through a liquid-liquid phase separation process.3, 13 These processes are accompanied by dehydration/solidification that eventually leads to mineralization.14 Similar mechanisms were proposed for other mineral systems.5 Systems of industrial relevance have been investigated less, the only exception being calcium sulfate.14 Many studies on calcium silicate hydrate (C-S-H) rely on atomistic simulations,15 but aside from a study by Nonat and coworkers16 that interprets conductivity results in solution within the framework of the classical nucleation theory (CNT) the early stages of C-S-H formation have not yet been unraveled. C-S-H belongs to the family of inosilicates, whose structure is derived from that of natural 14 Å-tobermorite. It contains calcium sheets flanked on each side by linear “Dreierketten” silicate chains. The composition of C-S-H, i.e. the calcium to silicon ratio (Ca/Si) and the amount of water molecules, varies depending on the concentration of calcium hydroxide in solution (Figure 1).17 In regular cement suspensions, the solution is typically close to the solubility limit of Portlandite (Ca(OH)2 = 23mM at 23°C) or slightly supersaturated (up to 33 mM). Under those conditions, the infinite silicate chains of the tobermorite structure are reduced to silicate rows containing silicate dimers, occasionally bridged by a third silicate tetrahedron. Although the overall Ca/Si ratio in C-S-H may vary from 0.66 to 2, its range is restricted to values between 1.4 and 1.7 in cement pastes.19

Cement is the principal component of concrete, which is ubiquitous in the construction industry. With an annual consumption of 4 billion tons, it is the most used industrial material and therefore tightly linked to global economy.1 Modern cement (i.e. Portland cement) is a reactive mixture of calcium silicates, aluminates and sulfates present as several anhydrous polymorphs stabilized by various alkali-metal ions. These phases react with water and form a complex suspension of hydrated amorphous and crystalline phases.2 During the reaction with water (hydration), a percolating network of hydration products develops, which determines the setting and hardening behavior of the material. Despite many decades of intensive research on cement hydration, insights on the early stages of this process remain elusive. Detailed understanding of the initial formation of calcium silicate hydrate (C-S-H), the most cohesive phase, is paramount for the development of new additives to improve the mechanical performance of concrete and to reduce the cement portion in the next generation of concrete. The past decade has witnessed major advances in our comprehension of the early stages of crystallization.3-6 Many, so-called “non-classical” nucleation pathways, including the formation of precursors, have been proposed to account for the formation of the final crystalline phase. These multi-step nucleation processes comprise aggregation-based mechanisms of aqueous species ranging from multi-ion complexes to fully formed nanocrystals. The majority of these studies were concerned with proteins7 or minerals of geological and biological interest,8-12 in ACS Paragon Plus Environment

Chemistry of Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 11

titration device (Titrando 905) that controls two dosing units (Dosino 800). The system is supplemented by a pH module (Metrohm, pH module 867). The pH in the sample was monitored in real time with a glass iUnitrode electrode (Metrohm, No. 6.0278.300), while the turbidity of the samples can be monitored with an optical sensor (Spectrosense, Metrohm No. 6.1109.110). All experiments were carried out in a lab-built PTFE titration vessel (Figure S1). The vessel was closed and protected from the atmosphere in order to minimize potential artifacts that may arise from diffusion of atmospheric CO2 or evaporation of the solution. The room temperature was controlled at T = 293.15 K. A solution supersaturated with respect to calcium silicate hydrate and undersaturated with respect to portlandite was used as starting solution. The chemicals used and the ion concentrations of the different supersaturated solutions are shown in Table S1 and S2. The desired pH of 12.86 was reached by addition of sodium hydroxide (solution at 1 M). In order to ensure the same ionic strength, sodium chloride was added to the solution in order to reach an ionic strength I = 0.245 M. To prevent carbonation of the supersaturated solution, all solutions were freshly prepared with nitrogen purged deionized water. In addition, all experiments were performed in a closed Teflon vessel by passing a dry nitrogen flux. The freshly prepared solutions were filtered with a Millex-GV, 0.22 µm, PVDF, 33 mm to prevent the intrusion of dust. The sodium metasilicate solution was added with a dosing speed of 1.2 mL/min while the solution was stirred (stirring speed: 960 rpm). Characterization. Transmission electron microscopy (TEM) was carried out on a Tecnai 12 and a Tecnai F20 transmission electron microscope equipped with field emission gun and working at 120 kV and 220 kV, respectively. TEM snapshots were prepared by taking a drop of the crystallization solution during the in situ measurement and placing it on copper grids coated with amorphous carbon. Scanning electron microscopy (STEM) images were obtained with a FISCHONE high angular annular dark field (HAADF) detector and elemental analysis was performed by energy-dispersive X-ray (EDX) spectroscopy and quantified by EDAX Genesis (Version 6.4.1.). The spacing of the atomic planes was obtained by a profile plot (Image J: Image J 1.50a, National Institutes of Health, USA, Java 1.6.0_46). Frozen-hydrated (cryo) samples were prepared by drop-casting a C-S-H suspension on a TEM grid covered by holey carbon film (Quantifoil 2/2), followed by blotting and plungefreezing in liquid ethane (Gatan Cryoplunge 3). For the subsequent examination this specimen was transferred to a TEM (FEI Tecnai F20) keeping cryogenic conditions using a cryo TEM holder (Gatan 926). The cryo-TEM measurements were performed at 200 kV with a Tecnai 20 (FEI, Hillsboro, OR, USA) equipped with a TWIN-lens and a 2kx2k CCD by Gatan (Pleasanton, CA, USA). For electron diffraction measurements under cryo-TEM conditions the operation mode of the TEM was adjusted to the selected area diffraction mode (SAED). By inserting a small 10 µm C2 condenser lens a fine, parallel beam was formed, which could be focused to a diameter of several nm. Since the vitrified water as well as the sample are extremely beam sensitive, the total beam current was reduced to approx. 5 pA. The electron diffraction was recorded on a 2k CCD camera at binning 4 using an exposure time of 500 ms. Inductively-coupled plasma equipped with optical emission spectroscopy (Spectro Ciros Visions ICP-OES) was used to determine the silicon and calcium content in the solution. At designated time intervals, 0.2 mL of the suspension were

Figure 1. Evolution of the calcium to silicon ratio (x=Ca/Si) as a function of the concentration in calcium hydroxide in solution (data extracted from Haas et al.18). Schematic structures representative of C-S-H a, b and g are shown. For the sake of clarity, only one sheet is shown without the water molecules. Two ways were identified to increase the Ca/Si: the release of the bridging Si tetrahedra and/or the compensation and even the overcompensation of silanolate charges by Ca2+ ions. The Na+ ions are not accounted in CS-H but they have been represented here to obtain a neutral structure. For more detail, the reader is referred to ref. 18.

Disregarding the concomitant dissolution of anhydrous cement phases which are kinetically coupled to the precipitation of hydrates during the hydration of cement in paste,20 the precipitation reaction of C-S-H can be formulated as follows: xCa2++2xOH-+H4SiO4→(CaO)x-(SiO2)-(H2O)1.8 (Eq.1) with x as Ca to Si molar ratio. H4SiO4 is chosen as silicate species for mass balancing even though the solution does contain a mix of H4SiO4, H3SiO4- and H2SiO42-. In ordinary Portland cement, the first marked physical transformation, i.e. the setting, is mostly due to the precipitation of C-S-H around the cement grains forming and strengthening the percolating network.21-24 Experimental and modelling studies have shown that the crystal growth of C-S-H is strongly limited,25 which indicates that the precipitation is driven by nucleation rather than by crystal growth. Thus, the nucleation of C-S-H is a crucial step in the hydration kinetics. Moreover, the heterogeneous precipitation nucleation of C-S-H in dissolution etch-pits can locally annihilate the dissolution of anhydrous silicate phases.26 Aluminum ions27 and organic admixtures28 retard C-S-H nucleation whereas controlled seeding represents an attractive way to accelerate the evolution of the mechanical properties of cement for industrial applications.29 Here, we demonstrate that the homogeneous nucleation of C-S-H is a two-step mechanism proceeding via amorphous spheroidal intermediates. C-S-H nucleation is heterogeneous under regular operating conditions, even though homogeneous nucleation has been identified in the presence of polycarboxylate ether polymers.30 In order to unravel the underlying elementary steps, the complexity of the system has been reduced by using idealized representative homogeneous conditions, i.e. the formation occurs in solution free of particles at typical pH and ionic conditions of cement pastes. EXPERIMENTAL SECTION Synthesis. Potentiometric and turbidimetric measurements were performed using a computer-controlled system from Metrohm (Filderstadt, Germany), operated with the customdesigned software Tiamo (v2.4). The setup consists of a

2

ACS Paragon Plus Environment

Page 3 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials between 0.31 nm and 0.35 nm demonstrate a distribution of the d spacing, indicating the presence of strong structural variations between regions of short-range ordering. These results are in line with the results of previous X-ray diffraction (XRD) studies on C-S-H that revealed the presence of a broad diffuse peak in the range of 0.25-0.31 nm.2 A comparison of Figure 2b and c reveals that the periodicity of short-range ordering in the diffuse rings varies from region to region.

collected and centrifuged for 5 minutes at 5000 rpm to separate the solution from larger particles for analysis. The supernatant was acidified with nitric acid (8 M to stop the reaction) and diluted with deionized water. In order to achieve the highest precision, the calibration was performed with a matrix-matching, i.e, the calcium and silicon standards were prepared with the same concentration in NaCl as in samples. All time-resolved small angle X-ray scattering (SAXS) measurements were carried out at beamline I02 at the ESRF. The setup uses the monochromatic, highly collimated, and intense beam in a pinhole configuration with sample-to-detector distance variable from 0.6 m to 30 m. Experiments were performed using a monochromatic X-ray beam at 12.4 keV and two-dimensional (2D) scattered intensities were collected at small-angles with a Pilatus 300K (2D large area pixel-array detector49). Transmission was measured by means of a photodiode installed in the beam-stop of the SAXS detector. For each experiment, a series of backgrounds and reference samples were measured including the empty capillary cell, cell filled with water and cell filled with the initial, unmixed CaCl2 solution at the used concentration and temperatures. In all SAXS measurements, the acquisition time per frame varied between experiments (from 1 to 30 s per frame). This time frame was based on previously off-line tested reaction times for the various conditions. A Kapton capillary (ID = 0.1461 cm) was used for the measurements through which solution was allowed to flow. This setup maintained a constant X-ray beam over the course of a single experiment, allowing SAXS patterns taken at different times to be quantitatively compared. Solution flowed continuously through the capillary over the course of the experiment using peristaltic pumps at the inlet and outlet with a flow rate of ∼100 mL/h. Time-dependent turbidity was measured with an optical sensor which consists of a light source, two glass fiber (light) guides, a concave mirror, and an amplifier.

Figure 2. Characterization of the final reaction product. (a) TEM micrograph of the collected solid phase. (b, c) SAED pattern of the final product. (d) STEM EDX spectrum and (e) area from which the EDX pattern was recorded.

RESULTS AND DISCUSSION

Chemical preparation. The homogeneous nucleation of C-SH was studied from solutions supersaturated with respect to C-S-H and undersaturated (or slightly supersaturated) with respect to Portlandite (Ca(OH)2). Supersaturation was reached after slow titration of sodium silicate to a lime solution with a custom-designed set-up (Figure S1). Representative conditions of cement solution were approached, i.e. a high pH (pH = 12.6), a calcium concentration close to the solubility of Portlandite (Ca = 15 mM), a low Si concentration (Si ≤ 200 µM) and a high ionic strength (I = 0.245 M). A solution containing 100 µmol/L of Si and 80 mmol/L of NaOH was characterized, which is reported here as reference solution. To prevent carbonation, all solutions were freshly prepared with nitrogen purged deionized water. All experiments were performed in a closed Teflon vessel under nitrogen. No solubilized CO2 was detected by titration (CO2 < 1 µM Figure S2). Structural characterization by electron microscopy. The final product, obtained after 1000 minutes, was characterized by transmission electron microscopy (TEM). The TEM micrograph (Figure 2a) showed a product with foil-like morphology, as previously reported for C-S-H.31, 32 It was analyzed by selected area electron diffraction (SAED, Figure 2b and c). The rotational average from the diffraction pattern has maxima at 5.3 nm-1 and 3.2 nm-1. A periodicity of ~5.3 nm-1 corresponds to the main 110 reflection associated with the Ca-O layer repeat C-S-H along the silicate chain of C-S-H. The broad rings

Figure 3. Evolution of turbidity and solution concentrations. (a, b, c) Results related to the reference experiment containing 100 µmol/L of Si and 80 mmol/L of NaOH. (a) Time evolution of the turbidity (black) and the activity of hydroxide (blue). (b) Si and Ca concentrations (mean values averaged over three experiments) analyzed by ICP-OES. (c) Evolution of the supersaturation (β) with respect to C-S-H and Ca(OH)2 solubilities, calculated from the solution concentrations. Evolution of the turbidity for various

3



conditions: at different silicate concentrations (d), pH values (e) and stirring speeds (f) as a function of time. The standard deviation of the turbidity potential is below 5 mV.

Page 4 of 11

An increased stirring speed (Figure 3f) retarded the Uturb drop in stage II and III. The stirring speed, e.g. the rate of mass transport (diffusion, convection, shear force, etc.) affects the formation/consumption dynamics of the resulting material.. Particle size and morphology. Size and morphology of the objects formed in the supersaturated reference solution (Figure 4a) were determined by dynamic and static light scattering (DLS, SLS). Measurements were carried out on samples transferred into dedicated cuvettes. Objects with a hydrodynamic radius (RH,app) of ~60 nm ± 3 nm were detected already during the very first stage of the reaction. This indicates that they were formed right after silicate addition (i.e. during the 5 minutes that were needed for sample transfer for DLS measurement). The particle size was independent of the incidence angle, which shows that the particle size distribution is fairly narrow.35 RH,app increases linearly from 60 to 90 nm during the next 60 min (Figure 4a). After the first 60 minutes, no further increase of RH,app was observed during stage I. Figure 4b displays the evolution of the gyration radius Rg calculated from SLS data as a function of time. The Guinier plots show a linear correlation yielding Rg = 70 ± 3 nm (Figure S5). A change of the silicate concentration from 50 to 200 µM had no effect on Rg in stage II (Figure S5) according to the results of the SLS experiments. Therefore, the Uturb drop for increasing silicate concentration in stage II must be due to a higher number of scattering particles (Figure 3d, vide supra). This suggest that particle formation is fast compared to their growth. The higher silicate concentration in stage I and II leads to a higher number of spheroids, while the first Uturb drop may be associated with the time needed to reach the detection limit. The structure parameter ρ (Rg/RH,app) of 0.77 indicates that the objects are fairly spherical.35 No morphological changes were detected by SLS during the next four hours (Figure 4b).

Compositional analysis. The composition was analyzed by scanning transmission electron microscopy (STEM) combined with energy dispersive X-ray spectroscopy (EDX, Figure 2d and e, vide supra). The product contains calcium and silicon with a molar ratio of 1.41 ± 0.02 together with minor amounts of sodium (2.2 ± 0.6 at. %) and chloride (1.8 ± 0.9 at. %) ions. The Ca/Si ratio and the z-potential value of +14.50 ± 0.5 mV (determined by electrophoresis) match the values expected from the synthesis of C-S-H for such calcium hydroxide concentrations.33,34 This is in accordance with the precipitation of β-C-S-H, according to a thermodynamic model devised by Haas and Nonat,18 where the negative surface charges are overcompensated by Ca2+ ions. The physico-chemical evolution of the suspension during C-S-H nucleation was monitored by turbidimetry and pH measurements (Figure 3a). The Si and Ca concentration was determined by analyzing the supernatants by inductively coupled plasma optical emission spectrometry (ICPOES) (Figure 3b). The degree of supersaturation (β), was calculated from the ratio of the ion activity product to the solubility product with respect to β-C-S-H and Portlandite (Figure 3c). C-S-H nucleation The nucleation process can be divided into three main stages according to the characteristic evolution of the turbidity (Figure 3a) as clear start solutions (Uturb = 800 mV) turned more turbid in two steps (two drops of Uturb in stages I and III). The drops in Uturb and the hydroxide ion activity run parallel in stage I. The pH shows only a significant change in stage I, but no change in stage III (Figure 3a). The silicate concentration drops from 0.10 to 0.04 mM during stage I and II (Figure 3b). The calcium concentration remained constant because it is two orders of magnitude higher than the other concentrations at the start. The evolution of silicate concentration, pH and turbidity are correlated during stages I and II. Supersaturation. After adding the sodium silicate solution, the system is supersaturated with respect to β-C-S-H (Figure 3c). The silicate concentration decreases rapidly during stages I and II and more slowly afterwards. The solubility equilibrium was almost reached at the end of stage II within the accuracy of the measurement (~4 µmol/L in Si). The supersaturation index (log b) with respect to Portlandite is close to zero. Indeed, no Portlandite formation was detected by IR spectroscopy (Figure S4). As no abrupt change in chemical composition was observed in stage III, secondary precipitation of other solids like calcium carbonate can be excluded as well. The silicate concentration (Si = 50-200 µmol/L) was varied in order to change the supersaturation with respect to β-C-S-H in the starting solution. The supersaturation affects the drop of Uturb in stage I: The higher the supersaturation, the earlier and the sharper the drop. The turbidity in stage II and III is independent of the silicate concentration. This indicates that the second drop of Uturb is not linked directly to the supersaturation (Figure 3d). An increase in pH accelerates the first and the second Uturb drops (Figure 3e). When the pH was increased to 13.0, the final C-S-H product was formed after only 600 minutes. The higher pH increased the supersaturation with respect to β-C-SH and the deprotonation of the silanol groups, i.e. the charge density of the silicate chains in C-S-H.

Figure 4. Light and synchrotron X-ray scattering analysis of the supersaturated reference solution. (a) Hydrodynamic radius determined by DLS at three different angles during the first 5 to 100 minutes. (b) Evolution of the gyration radius calculated from the Guinier plot (ln Iq over q2) obtained by SLS during stages I and II (400 minutes). (c) Time resolved in situ SAXS pattern during the first 15 min. The formation of small primary entities/scatterers is shown through the increase in intensity with the fits of scattering curves (red). (d) Radius evolution resulting from the fits of scattering curves (size distribution represented by the error bars).

Particle size and morphology during the early stages analyzed by SAXS. The DLS measurements required a transfer of the supersaturated solution into a cuvette and thus did not allow

4


Page 5 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


any stirring. In order to circumvent this problem and to monitor the size evolution of the objects from the very first minute, the early stages of the reaction were probed using synchrotronbased small angle X-ray scattering (SAXS). The same titration set-up was used with a circulation loop powered by a peristaltic pump and connected to a capillary permitting in situ analysis. Figure 4c reveals that (i) the first objects form already during the first two minutes and (ii) continue to grow linearly (Figure 4d). The scattering curves were fitted assuming the presence of spherical particles with a log-normal size distribution. Figure 4c shows the radius evolution of these spherical objects during the first 15 minutes. The radius of the first particles determined by SAXS (~22 nm) is significantly lower than that from DLS (~60 nm), although their growth over time is similar (0.55 and 0.44 nm/min). This difference in particle size might be attributed to the effect of stirring.

compatible with a Ca/Si ratio of 1.13 ± 0.33. In addition, the spherical particles contained sodium (19.7 ± 0.2 at.-%) and chloride (5.3 ± 0.8 at.-%), yielding a Ca/Na ratio of 2. A few particles showed incipient crystalline features (circled particle in Figure 5a), which could readily be identified after 600 min. Irregular fringes inside the spheroids are compatible with medium-range order (Figure 5d). Figure 5 shows that the spheroids transformed into an ordered structure with layers spaced by 1.36 ± 0.3 nm. The particle morphology and the change of the chemical composition of the spheroids are in line with the formation of an amorphous C-S-H intermediate during stage I. It transforms into a product with medium range order that matches the 14 Å tobermorite-type structural periodicity.33 The presence of sodium and chloride ions in the droplets contrasts with the composition of the final reaction product. This indicates that sodium and chloride are expelled during stage II while the Ca/Si ratio increases from 1.13 to 1.41. This is compatible with a Na/Ca cation exchange because the silicate condensation is completed in the middle of stage II.

Figure 5. C-S-H crystallization from the reference supersaturated solution observed by Cryo-TEM, NAED and STEM EDX. (a) Cryo-TEM image of the suspended particles 300 min after addition of the silicate solution. (b) SAED pattern of the spheroid after 300 min (c) STEM EDX spectrum of the spheroid after 300 min. (d) Cryo-TEM image of the suspension and (e) high resolution image after 600 min. (f) Profile plot along the black dotted line shown in high resolution image (e) of the particles formed after 600 min in solution.

Figure 6. Modeling results. (a) Simulations of the average number of Ca2+, Na+ and anions (OH- and Cl-) per silicate dimer (Si2O76-) as a function of the volume of the cubic simulation box. The simulations are performed in the semi-grand canonical ensemble with one silicate dimer per box. The volume of the box is therefore proportional to the density in silicate dimers in the spheroids. The ions excepted the silicate dimer are allowed to move in and out of the simulation box according to their set chemical potential, i.e. they are in equilibrium with the bulk solution. A Ca2+/Na+ ratio of 2 corresponds to a volume of ~16³ Å³ per dimer (vs), see dashed line. In C-S-H, this volume (vC-S-H) corresponds to ~6³ A³, and the surface charge density is ~-4 e/nm² in those conditions of pH as well as of Na+ and Ca2+ concentrations.36 Thus, the spheroids are approximatively vs/vC-S-H = (16/6)³ ~19 times less dense than C-S-H and their surface is (vs/vC-S-H)2/3 = (16/6)² ~ 7 times less charged. (b) Interaction surface free energy (W(r)) curves as a function of the distance

Snapshots of the reaction by cryo-TEM. In the liquid phase cryo-TEM provides access to metastable structures in liquids even prior to particle formation as snapshots of the sample at the instant of cryogenic vitrification. Cryo-TEM performed 300 min (during stage II) after the addition of the silicate solution (Figure 5a and Figure S7) confirmed the presence of spherical particles at this stage with a radius of about 32 ± 9 nm, close to that obtained by SAXS. These objects in stage II are predominantly non-crystalline as no diffraction rings in the SAED pattern were observed (Figure 5b). Ca and Si contents of 40.2 ± 0.5 and 35.2 ± 0.8 at.-% determined by STEM EDX (Figure 5c) are

5



between two negatively charged infinite plates (see Figure S8) for different charge densities. A charge of -0.6 e/nm² corresponds to the surface charge density of the spheroids as calculated in (a). A charge of -4 e/nm² matches the surface charge density of C-S-H. The interaction is purely repulsive at low charge density and becomes attractive below -1 e/nm².

Page 6 of 11

The amorphous and crystalline C-S-H phases differ in their density and cation composition. This is revealed by the different material contrast of the TEM images of the spheroids and the crystallizing C-S-H phase and by the different Ca/Na ratios as explained by the charge modeling (Figure 6a). The low silicate density (i.e. low charge) inside the spheroids ensures their colloidal stability and explains their existence as dispersed particles on long-time scales. A low density contributes to faster relaxation, rotational freedom of the silicate polyhedra within the spheroids and to the distortion of bond lengths and angles36 in order to accommodate the motifs into the crystalline C-S-H structure. Second chemical step. Well-defined diffraction rings in cryoTEM SAED indicate the appearance of crystalline order related to C-S-H during stages II and III. In contrast to the first step, the initial silicate concentration does not affect the time of the second Uturb drop. The solution is already close to the solubility equilibrium of C-S-H before the ordering process and the second step is unlikely to be driven directly by the deviation from the C-S-H solubility equilibrium. As the crystallization starts in stage II and particles are dispersed in stage II but aggregated in stage III, the second increase in turbidity is likely the result of the concomitant transformation of the spheroids into β-C-S-H and the aggregation of resulting β-C-S-H crystallites. Based on electrostatic considerations, the Na+ and Ca2+ ions are bound to the silicate tetrahedra. The formation of an ordered arrangement leads to an alignment of the silicate layers. The concomitant release of water molecules (and the associated increase of charge) leads to a densification inside the spheroids, which in turn favors Ca2+ adsorption at the expense of Na+ ions.38 The crystallization is therefore accompanied by an exchange of sodium against Ca2+ ions. The simulations show that this progressive densification is linked directly to an increase of the surface charge density and leads to an aggregation of the spheroids (Figure 6b). The onset of aggregation appears to be the consequence of crystalline ordering. As further support, higher pH accelerates the second Uturb drop (Figure 3e, vide supra). Higher pH leads to a deprotonation of the silanol groups, i.e. to a higher charge density of the silicate chains in C-S-H. The formation of amorphous intermediates is often encountered in sol-gel processing of ceramics. Typically, the initial formation of poorly crystalline intermediates by condensation of dissolved monomer (or oligomer) precursors is more exothermic than the formation of a crystalline phase,39,40 which is accompanied by a release of water molecules. It is only weakly exothermic or even endothermic. As a result, the process becomes purely entropy-driven and eventually leads to denser compounds. The higher density of the charged species goes along with stronger attractive interactions between the crystallizing objects and therefore faster aggregation. The proposed two-step mechanism of the homogeneous nucleation of C-S-H is illustrated in Figure 7a. Two-step nucleation vs. one-step model from classical nucleation theory. Our results describe the formation of crystalline C-S-H particles via a two-step mechanism. It fundamentally differs from the classical nucleation approach used by Garrault-Gauffinet to explain the homogeneous nucleation of C-S-H

Modelling the structure of the C-S-H intermediate. The experimental Ca/Si ratio (1.41) and the z-potential (+14.5 mV) correspond to b-C-S-H at equilibrium according to the model of Nonat and Haas18 that predicts a theoretical Ca/Si ratio of 1.45 and a surface potential of +8 mV. The final β-C-S-H product should contain 93% of silicate dimers. A charge balance calculation shows that the speciation in the spheroids, based on the composition obtained by STEM/EDX, is very similar to that of b-C-S-H. The apparent negative charge deficit inside the spheroids may be explained by the presence of 72-100% of silicate dimers (Table S8). Assuming identical ion activity inside and outside the spheroids for Ca2+ and Na+, a theoretical estimate of the silicate density inside the spheroids can be obtained from a modelling of the ionic Na/Ca ratio. This is performed here at the level of the primitive model which has been shown to predict with good accuracy adsorption isotherms of simple ions on C-S-H and electrokinetic properties of C-S-H.36 The modeling results indicate that the density with respect to silicate inside the spheroids is ~ 19 times lower than in the final C-S-H (Figure 6a). The significantly lower silicate density inside the spheroids compared to C-S-H indicates that the transformation of the calcium silicate spheroids into final b-C-S-H is accompanied by a densification, i.e. water molecules are released. This densification leads to an increase of the surface charge density (s), which in turn drives the aggregation of the system (Figure 6b). Despite strong attractive forces between highly charged (s = -4 e/nm²) C-S-H platelets resulting from strong ion-ion correlation forces,37 the low charge density in the spheroids (s = -0.6 e/nm²) prevents aggregation because repulsive forces predominate. Chemical reaction model The combined experimental and theoretical results from scattering, STEM EDX, cryo-TEM and modelling allow sketching a physico-chemical model for the nucleation of C-S-H. The two turbidity drops appear to be the consequence of two well-defined chemical steps. First chemical step. The first turbidity change corresponds to the formation of amorphous calcium silicate spheroids with diameters of ~50 ± 10 nm. Their composition is compatible with C-S-H chemistry, but they are depleted in calcium compared to the final equilibrium state. The amorphous spheroids form rapidly after addition of silicate to a calcium hydroxide solution, but they do not form in the absence of calcium (Figure S6). Ca2+ is known to catalyze the oligomerization of silicate ions.38 Still, as the pH in this study is high and the silicate concentration low, oligomerization is clearly disfavored. In addition, a decrease of the starting pH shifts the first Uturb drop to later times (Figure 3e, vide supra). Thus, spheroid formation is not directly associated with the formation of Si-O-Si bonds, and increasing pH and Si concentration will only result in higher supersaturation. Furthermore, the solution (based on Si speciation) approaches the solubility limit of C-S-H already during stage II, i.e., the chemical potentials of the spheroids and C-S-H are similar. Thus, it is reasonable to assume that the spheroid formation is triggered by similar molecular calciumsilicate interactions as in crystalline C-S-H.

from supersaturated solutions which was monitored by conductometry.16 An apparent induction time, tind, could be assigned to the

first Uturb drop in our study. A deviation from equilibrium can be calculated from the initial conditions by taking β-C-S-H as equilibrium phase into account. Based on classical nucleation theory tind is linked to the supersaturation β by

6


Page 7 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ln(tind) = (fV2 Ω3)/(kT)3ln-2(β)

Chemistry of Materials (Eq.2)

Discussion There are several questions concerning the two-step mechanism for C-S-H formation: (a) Why are the amorphous precursors formed? Interfacial surface energy may explain the occurrence of the amorphous precursor.43 For sufficiently small particle sizes, the lower surface energy makes the more disordered phase more stable. All crystals must go from small to large and the lower interfacial energy coupled with the solubility ensures that the barrier must be smaller regardless of the concentration. Therefore, the system will naturally go through the more stable intermediate precursor phase. When the supersaturation is high, many - stable and metastable - phases are accessible to the system as the free energy barrier can be easily overcome for each of them. Since the amorphous phase is more soluble than the crystalline phase, the impact of supersaturation on the barrier is more significant for the crystalline phase. For high supersaturations, the barrier will be small for both, regardless of differences in interfacial energy. At some point, the factor that determines which phase nucleates first will no longer be the barrier, it will be the kinetic prefactor, which contains a Boltzmann factor, whose argument is the effective activation barrier for the reaction divided by kT.41 As the activation barrier for the formation of the amorphous phase is likely to be smaller than that for the crystalline phase, one can assume that the amorphous precursor will form for kinetic reasons. Despite of being representative of real conditions, it is difficult to know whether the supersaturations in our study are “high or rather low”, i.e. whether the nucleation follows a path dictated by thermodynamics or rather by kinetics. More generally, supersaturations are difficult to compare here as they are calculated as a function of the number of atoms/molecules in the chemical formula. (b) Are the observed spheroids solid-amorphous or liquid-like? Figure 7a shows in a schematic manner the evolution of the turbidity with time, i.e. it reflects the experimental state of affairs. The curve shape indicates that the amorphous precursor is formed via spinodal phase separation (while a small barrier at the initial portion of the curve would indicate that the amorphous precursor nucleates). This observation is in line with the notion that the interconnected structures and aggregates of particles in sol−gel processes (formation of silica, metal oxides or polymers) are related to spinodal processes.44-47 In those, the reaction starts with a homogeneous solution. Subsequently, a phase separation is induced by a chemical reaction that leads to a separation into two phases, at least one of them being solidrich (polymer or ceramic material). This observation indicates that the observed C-S-H spheroids are solid-amorphous rather than liquid-like. (c) Do the speroids play an active role in crystal nucleation or are they passive bystanders? The amorphous spheroids are very likely the locus of the crystallization because (i) the start of the ordering is observed within the spheroids (cryo-TEM snapshot) and (ii) no other objects are formed in solution due to quite low supersaturations in bulk and the particle size remains constant (in situ scattering studies). As an evidence of the chemical transformation, the crystallization within the spheroids is associated to a Na+-Ca2+ ion exchange between the bulk and the spheroids. (d) Are the spheroids the only objects resulting from a non-classical nucleation process or do they act as substrate for heterogeneous nucleation? The combined cryo-TEM and scattering (SLS, DLS, SAXS) results suggest that the spheroids do not serve as substrate for a heterogeneous nucleation. C-S-H

where V, Ω and f are the molar volume, the surface energy, and the shape factor, and T, β and k are the temperature (298 K), the supersaturation, and the kinetic constant.41 Figure 7b highlights that the evolution of ln tind is proportional to ln-2 β. It indicates that the formation of the spheroids could be described within the framework of classical nucleation theory from a thermodynamic point of view. The change in slope observed by GarraultGauffinet (Figure 7b) may be caused from the change between amorphous intermediate-driven and crystalline phase-driven formations. In that case, the amorphous intermediate can only be observed when the transition kinetics is slow, i.e., for low to moderate supersaturations.

Figure 7. Two-step pathway theory versus classical nucleation theory (CNT). (a) Evolution of the turbidity (red curve) with the different species emerging along the reaction the path: (I) Supersaturated solution, (II) liquid amorphous intermediate, (III) crystalline domains and (IV) final b-C-S-H platelets. (V) Model of C−S−H structure represented by 14 Å tobermorite. Reprinted with permission from ref. 42. Copyright 2017 American Chemical Society. The first chemical step, defined by the first Uturb. drop in stage I, is the formation of liquid amorphous intermediates rich in silicate dimers and calcium. The intermediate has a composition in silicate and calcium similar to that of β-C-S-H with additional sodium ions and water molecules. The second step encompasses a Ca/Na exchange (stage II to stage III) resulting in the formation of β-C-S-H crystallites which aggregate and give rise to the second Uturb. drop. (b) Typical plot in the framework of the CNT used for the calculation of the energy penalty (interfacial energy) to form a solid phase from supersaturated solution.

Figure 7b gives a clue for the chemical model. The formation of the nuclei involves at least an intermediate. GarraultGauffinet’s data highlight two domains of supersaturation. In both of them, a linear dependence with ln-2b was observed. Without experimental evidence, Garrault-Gauffinet proposed a switch between heterogeneous (low b) and homogeneous (high b) nucleation processes. Regardless of any consideration on the geometry and nature of the forming objects, the plot indicates that the energy penalty to form the amorphous spheroids as intermediate is significantly lower.

7



nucleation is, however, certainly heterogeneous under regular operating conditions, i.e. in concrete systems. Therefore, the two-step pathway seen here, although not incompatible, might be evaded in real cements because heterogeneous nucleation can favor another kinetic path as well as other polymorphs. A two-step mechanism, where crystal nuclei grow inside previously formed dense liquid clusters of mesoscopic size has been reported first for protein crystals.48,49 This non-classical nucleation pathway has been observed later also for inorganic50 organic,51 and colloidal52 systems. Direct observations of crystal nuclei forming within dense liquid clusters have been confined mostly to colloids.52 For proteins, clusters of protein molecules with mesoscopic sizes have been suggested as precursors. They were assumed to are liquid-like and serve as nucleation precursors,53,54 but eventually turned out to be amorphous solid particles that act as heterogeneous nucleation sites.55 The fact that crystalline phases do not form inside them but rather on the surface of the precursor particles appears to be a key difference between C-S-H (and other polycondensating materials) and protein systems. (e) Is the aggregation of spheroids crucial? We demonstrated the key role of aggregation during the transformation of spheroids into a crystalline product. It has been shown elsewhere that the aggregation lowers the overall energy of the system 56 and also allows a rapid propagation of the chemical transformation due to the percolating character of the aggregates.57 It is evident that aggregation is favored at high spheroid concentrations, i.e. at high supersaturations. The energy transition could be simply due to statically low aggregation at low supersaturations which in turn would enable the observation of these transient amorphous spheroids.

Page 8 of 11

ASSOCIATED CONTENT The Supporting Information is available free of charge on the ACS Publications webseite at DOI: 10.1021/acs.chemmater.xxxxxx. Titration setup with additional information (Figure S1), potentiometric determination of carbonate (Figure S2), time-dependent turbidity of colloidal silica (Figure S3), IR spectrum of C-S-H reaction product (Figure S4), Guinier plots showing the size evolution of particles with time (Figure S5), time-de2+ pendent turbidity for free Ca in solution (Figure S6), cryoTEM and SAED (Figure S7), model for determining the interactions between two spheroids (Figure S8), chemicals (Table S1), supersaturated solutions (Table S2), particle morphology and structure factor (Table S3), composition and concentration of solutions (Table S4), particle surface compositions (Table S5), solution equilibria (Table S6), chemicals for turbidimetric measurements (Table S7), Ca and Si speciation (Table S8), SAXS data evaluation (Table S9), Guinier plot (Table S10). Details concerning turbidimetric measurements, prevention of carbonation, silicate speciation, calculation of ion activities, and models used for simulations. AUTHOR INFORMATION Corresponding Author: [email protected] The authors declare no competing financial interest. AUTHOR CONTRIBUTIONS

N. K. carried out the experiments, L. N. initiated the study, L. N., M. D. and W. T. conceived the experiments, I. L., C. L., A. F.-M., A. D., B. B. and S. L. contributed to methods, N. K. wrote the manuscript with contributions of L.N., F. N., A. F.M., A. D, and W. T. All authors have given approval to the final version of the manuscript.

SUMMARY

The nucleation of very few minerals has been investigated in detail and rationalized based on complex models. This study shows that C-S-H nuclei are formed in two distinct steps that can be differentiated by their crystalline order, their calcium contents and their density. However, it is unknown whether the amorphous spheroids are a requirement for C-S-H formation in real cement pastes. As a matter of fact, their amount and dispersion state would be crucial for building up the C-S-H network structure. Along this path we demonstrated how advection affects spheroid formation and may be correlated to the effect of mixing during the early hydration of cement.58 The stabilization of spheroids by polymers may be at the origin of the frequent and adverse retardation observed with the use of the polycarboxylate ether superplasticizers. As a support, Valentini et al.59 proposed that C-S-H nucleation switches from heterogeneous to homogeneous in the presence of superplasticizers, and Nicoleau et al.28 used similar polymers to stabilize single C-S-H particles. Aluminum ions were reported to distort the local ordering in C-S-H. 42,59 This may be an explanation for a longer or more difficult crystallization of the spheroids into C-S-H in the presence of aluminum and the known poisoning effect of aluminum during the early stage of cement hydration. A fundamental understanding of the mechanism of C-S-H formation at a molecular level is essential for a rational design of novel, cement-based construction materials. The implication of such fundamental understanding are creating novel avenue for “designing” the hydration process and the materials properties. Our study indicates how the evolution of the chemical composition may become a tool for scientifically guided design of cements and hydrated nanocomposites.

ACKNOWLEDGEMENTS N. K. thanks Dr. Karl Fischer (University Mainz) for help with SLS and DLS data processing. This work was supported by Johannes Gutenberg Universität and BASF Construction Solutions (FS Nr. 50015). REFERENCES (1) Worrell, E.; Price, L.; Martin, N.; Hendriks, C.; Meida, L. O. Carbon dioxide emissions from the global cement industry 1. Annu. Rev. Energy Environ. 2001, 26, 303–329. (2) Taylor, H. F. W.; Cement Chemistry, Thomas Telford Publishing, London 1997. (3) Gebauer, D.; Kellermeier, M.; Gale, J. D.; Bergström, L.; Cölfen, H. Pre-nucleation clusters as solute precursors in crystallisation. Chem. Soc. Rev. 2014, 43, 2348−2371. (4) De Yoreo, J. J.; Gilbert, P. U. P. A.; Sommerdijk, N. A. J. M.; Lee Penn, R.; Whitelam, S.; Joester, D.; Zhang, H.; Rimer, J. D.; Navrotsky, A.; Banfield, J. F.; Wallace, A. F.; Marc Michel, F.; Meldrum, F. C.; Cölfen, H.; Dove, P. M. Crystallization by particle attachment in synthetic, biogenic, and geologic environments. Science 2015, 349, 6247/1-9. (5) Rieger, J.; Kellermeier, M.; Nicoleau, L. Formation of nanoparticles and nanostructures-an industrial perspective on CaCO3, cement, and polymers. Angew. Chem. Int. Ed. 2014, 53, 12380-12396. (6) Thanh, T.-K; Maclean, N.; Mahiddine, S. Mechanisms of Nucleation and Growth of Nanoparticles in Solution. Chem. Rev. 2014, 114, 7610−7630. (7) Sleutel, M.; Van Driessche, A. E. S. The role of Clusters in nonclassical nucleation and growth of protein crystals. Proc. Natl Acad. Sci. USA 2014, 111, E546- E553.

8


Page 9 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(8) De Yoreo, J. J. The two-step mechanism of nucleation of crystals in solution. Nat. Mater. 2013, 12, 284−285. (9) Sun, W.; Booth, S; Myerson, A.; Hughes, C.; Pan, H.; Coquerel, G.; Brandel, C.; Meekes, H.; Mazzotti, M.; Fabian, L.; Black, S.; Vekilov, P; Back, K.; Toroz, D.; Lovelock, J.; Sefcik, J.; Rasmuson, A.; Breynaert, E.; Sear, R.; Hammond, R.; Ward, M.; Threlfall, T.; De Yoreo, J.; Davey, R.; Ristic, R.; Lewtas, K.; Roberts, K.; Hare, A.; Gich, M.; Cölfen, H.; Likhatskiy, M.; ter Horst, J. Time and space resolved methods: general discussion. Farad. Discuss. 2015, 179, 247267. (10) Sun, S. T.; Zhang, P.; Gebauer, D.; Cölfen, H. Distinct shortrange order is inherent to amorphous calcium carbonate clusters smaller than 2 nm. Angew. Chem., Int. Ed 2016, 55, 12206-12209. (11) Rao, A.; Vásquez-Quitral, P.; Fernández, M. S.; Berg, J. K.; Sánchez, M.; Drechsler, M., Neira-Carrillo, A., Arias, J. L., Gebauer; D., Cölfen, H., pH-Dependent Schemes of Calcium Carbonate Formation in the Presence of Alginates. Cryst. Growth Des. 2016, 16, 1349-1359. (12) Faatz, M.; Grohn, F.; Wegner, G. Amorphous calcium carbonate: Synthesis and potential intermediate in biomineralization. Adv. Mater 2014, 16, 153-159. (13) A. F. Wallace, L.O. Hedges, A. Fernandez-Martinez, P. Raiteri, J.D. Gale, G.A. Waychunas, S. Whitelam, J.F. Banfield, J. J. De Yoreo, Microscopic evidence for liquid-liquid separation in supersaturated CaCO3 solutions. Science 2013, 341, 885-889. (14) Van Driessche1, E. S.; Benning, L. G.; Rodriguez-Blanco, J. D.; Ossorio, M.; Bots, P.; García-Ruiz, J. M. The role and implications of bassanite as a stable precursor phase to gypsum precipitation. Science 2012, 336, 69-72. (15) Abdolhosseini Qomi, M. J.; Krakowiak, K. J.; Bauchy, M.; Stewart, K. L.; Shahsavari, R.; Jagannathan, D.; Brommer, D. B.; Baronnet, A.; Buehler; M. J.; Yip, S.; Ulm, F.-J; Van Vliet, K. J.; Pellenq, R. J-.M. Combinatorial molecular optimization of cement hydrates. Nat. Commun. 2014, 5, 4960/1-10. (16) Garrault-Gauﬃnet, S.; Nonat, A. Experimental investigation of calcium silicate hydrate (C-S-H) nucleation. J. Cryst. Growth 1999, 200, 565-574. (17) Damidot, D.; Rettel, A.; Capmas, A. Action of admixtures on Fondu cement: Part 1. Lithium and sodium salts compared. Adv. Cem. Res. 1996, 8, 83. (18) Haas, J.; Nonat, A. From C-S-H to C-A-S-H: Experimental study and thermodynamic modelling. Cem. Concr. Res. 2015, 68, 124-138. (19) Renaudin, G.; Russias, J.; Leroux, F.; Frizon, F.; CauditCoumes, C. Structural characterization of C–S–H and C–A–S–H samples-Part I: Long-range order investigated by Rietveld analyses. J. Solid State Chem. 2009, 182, 3312-3319. (20) Nicoleau, L.; Nonat, A. A new view on the kinetics of tricalcium silicate hydration. Cem. Conc. Res. 2016, 86, 1-11. (21) Ioannidou, K.; Kanduč, M.; Li, L.; Frenkel, D.;. Dobnikar, J, Del Gado, E. The crucial effect of early-stage gelation on the mechanical properties of cement hydrates. Nat. Comm. 2016, 7, 12106/1-9. (22) Masoero, E.; Del Gado, E.; Pellenq, R. J.; Ulm, F. -J.; Yip, S. Nanostructure and Nanomechanics of Cement: Polydisperse Colloidal Packing. Phys. Rev. Lett. 2012, 109, 155503/1-4. (23) Roussel, N.; Ovarlez, G.; Garrault, S.; Brumaud, C. The origins of thixotropy of fresh cement pastes. Cem. Concr. Res. 2012, 42, 148157. (24) Nonat, A. The structure and stoichiometry of C-S-H. Cem. Concr. Res. 2004, 34, 1521-1528. (25) Labbez, C.,; Jönsson, B.; Woodward, C.; Nonat, A., Delhorme, M. The growth of charged platelets. Phys. Chem. Chem. Phys. 2014, 16, 23800-23808. (26) Hua, Q.; Aboustaita, M.; Kima, T.; Tyler Leya, M; Hananb, J. C.; Bullardc, J.; Winarskid, R.; Rosed, V. Direct three-dimensional observation of the microstructure and chemistry of C3S hydration. Cem. Concr. Res. 2016, 88, 157-169. (27) Begarin, F.; Garrault, S.;. Nonat, A, Nicoleau, L. Hydration of alite containing aluminium. Adv. Appl. Cer. 2011, 110, 127-130. (28) Marchon, D.; Sulser, U.; Eberhardt, A.; Flatt, R. J. Molecular design of comb-shaped polycarboxylate dispersants for environmentally friendly concrete. Soft Matter 2013, 9, 10719-10728.

(29) Nicoleau, L.; Gädt, T.; Chitu, L.; Maier, G.; Paris, O. Oriented aggregation of calcium silicate hydrate platelets by the use of comblike copolymers. Soft Matter 2013, 9, 4864-74. (30) Sowoidnich, T.; Rachowski, T.; Rößler, C.; Völkel, A.; Ludwig, H.-M. Calcium complexation and cluster formation as principal modes of action of polymers used as superplasticizer in cement systems. Cem. Concr. Res. 2015, 73, 42-50. (31) Rodriguez, T. Composition, silicate anion structure and morphology of calcium silicate hydrates (C-S-H) synthesised by silicalime reaction and by controlled hydration of tricalcium silicate (C3S). Adv. Appl. Cer. 2015, 114, 362-371. (32). Richardson, I. G Tobermorite/jennite and tobermorite calcium hydroxide-based model for the structure of CSH: Applicability to hardened pastes of tricalcium silicate, β-dicalcium silicate, Portland cement and bends of Portland cement with blast furnace slag, metakaolin or silica fume. Cem. Concr. Res. 2004, 34, 1733-1777. (33) Lothenbach, B.; Nonat, A. Calcium silicate hydrates: Solid and liquid phase composition. Cem. Concr. Res. 2015, 8, 57–70. (34) Nachbaur, L.; Nkinamubanzi, P.-C.; Nonat, A.; Mutin, J.-C. Electrokinetic properties which conrol the coagulation of silicate cement suspernsions during early age hydration. Cem. Concr. Res. 1998, 202, 261-268. (35) Schaertl, W. Light Scattering from Polymer Solutions and Nanoparticle Dispersions, Springer, Heidelberg 2007. (36) Labbez, C.; Jönsson, B.; Pochard, I.; Nonat, A.; Cabane, B. Surface Charge Density and Electrokinetic Potential of Highly Charged Minerals: Experiments and Monte Carlo Simulations on Calcium Silicate Hydrate. J. Phys. Chem. B 2006, 110, 9219–9230. (37) Jönsson, B; Nonat, A.; Labbez, C.; Cabane, B.; Wennerström, H. Controlling the Cohesion of Cement Paste. Langmuir 2005, 21, 9211–9221. (38) Gaboriaud, F.; Nonat, A., Chaumont, D.; Craievich, A.; Hanquet, B. Si-29 NMR and Small-Angle X-Ray Scattering Studies of the Effect of Alkaline Ions (Li+, Na+, and K+) in Silico-Alkaline Sols. J. Phys. Chem. 1999, 103, 2091–2099. (39). Maniar, P. D; Navrotsky, A.; Rabinovich, E. M.; Ying, J. Y.; Benziger, J. B. Energetics and structure of sol-gel silicas. J. NonCryst. Sol. 1990, 124, 101-111. (40) Yang, S., In Situ Calorimetric, Structural, and Compositional Study of Zeolite Synthesis in the System 5.15 Na2O− 1.00 Al2O3− 3.28 SiO2−165 H2O. J. Phys. Chem. B. 2000, 104, 6071-6080. (41) Kashchiev, D. Nucleation: basic theory with applications, Butterworth-Heinemann, 2000. (42) Özçelik, V. O.; White, C. E. Nanoscale Charge-Balancing Mechanism in Alkali-substituted Calcium−Silicate−Hydrate Gels. J. Phys. Chem. Lett. 2016, 7, 5266−5272. (43) Navrotsky, A. Energetic clues to pathways to biomineralization: Precursors, clusters, and nanoparticles. Proc. Natl. Acad. Sci U.S.A. 2004, 101, 12096-2101. (44) Nakanishi, K. Pore Structure Control of Silica Gels Based on Phase Separation. J. Porous Mater. 1997, 4, 67−112. (45) Nakanishi, K.; Tanaka, N. Sol-Gel with Phase Separation. Hierarchically Porous Materials Optimized for High-Performance Liquid Chromatography Separations. Acc. Chem. Res. 2007, 40, 863− 873. (46) Nakanishi, K.; Amatani, T.; Yano, S.; Kodaira, T. Multiscale Templating of Siloxane Gels via Polymerization-Induced Phase Separation. Chem. Mater. 2008, 20, 1108−1115. (47) Wen, J.; Wilkes, G. L. Organic/Inorganic Hybrid Network Materials by the Sol-Gel Approach. Chem. Mater. 1996, 8, 1667−1681. (48) ten Wolde, P. R.; Frenkel, D. Enhancement of protein crystal nucleation by critical density fluctuations. Science 1997, 277, 1975– 1978. (49) Talanquer, V.; Oxtoby, D. W. Crystal nucleation in the presence of a metastable critical point. J. Chem. Phys. 1998, 109, 223–227. (50) Nielsen, M. H.; Aloni, S.; De Yoreo, J. J. In situ TEM imaging of CaCO3 nucleation reveals coexistence of direct and indirect pathways. Science 2014, 345, 1158–1162. (51) Harano, K.; Homma, T.; Niimi, Y.; Suenaga, M.; Leibler, L.; Nakamura, E. Heterogeneous nucleation of organic crystals mediated by single-molecule templates. Nat. Mater. 2012, 11, 877–881.

9



(52) Savage, J. R.; Dinsmore, A.D. Experimental evidence for twostep nucleation in colloidal crystallization. Phys. Rev. Lett. 2012, 102, 198302. (53) Galkin, O.; Vekilov, P. G. Control of protein crystal nucleation around the metastable liquid–liquid phase boundary. Proc. Natl. Acad, Sci. U. S. A. 2000, 97, 6277–6281. (54) Vekilov, P. G. Crystal nucleation: Nucleus in a droplet. Nat. Mater. 2012, 11, 838–840. (55) Yamazaki, T.; Kimura, Y.; Vekilov, P. G. Furukawa, E.; Shirai, M.; Matsumoto, H.; van Driessche, A. E. S.; Tsukamoto, K. Two types of amorphous protein particles facilitate crystal nucleation. Proc. Natl. Acad, Sci. U. S. A. 2017, 114, 2154–2159. (56) Delhorme, M.; Labbez, C.; Turesson, M.; Lesniewska, E.; Woodward, C. E.; Jönsson, B. Aggregation of Calcium Silicate Hydrate Nanoplatelets. Langmuir 2016, 32, 2058-2066. (57) Baronov, A.; Bufkin, K.; Shaw, D. W.; Johnson, B. L; Patrick, D. L. A simple model of burst nucleation. Phys.Chem.Chem.Phys. 2015, 17, 20846-20852. (58) Juilland, P. J; Kumar, A.; Gallucci, E.; Flatt, R. J.; Scrivener, K. L. Effect of mixing on the early hydration of alite and OPC systems. Cem. Concr. Res. 2002, 42, 1175-1188. (59) Valentini, L.; Favero, M.; Dalconi, M. C.; Russo, V.; Ferrari, G.; Artioli, G. Kinetic Model of Calcium-Silicate Hydrate Nucleation and Growth in the Presence of PCE Superplasticizers. Cryst. Growth Des. 2016, 16, 2404-2415. (60) White, C. E.; Daemen, L. L.; Hartl, M.; Page, K. Intrinsic differences in atomic ordering of calcium (alumino)silicate hydrates in conventional and alkali-activated cements. Cem. Concr. Res. 2015, 67, 66-73.

10


Page 10 of 11

Page 11 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TOC graphic

11


Two-Step Nucleation Process of Calcium Silicate Hydrate, the

Recommend Documents