Robert Flatt | Swiss Federal Institute of Technology (ETH) (original) (raw)

Papers by Robert Flatt

Materials and Corrosion-Werkstoffe Und Korrosion, 2012

ABSTRACT This paper sums up the International RILEM Workshop on Present and Future Durability Cha... more ABSTRACT This paper sums up the International RILEM Workshop on Present and Future Durability Challenges for Reinforced Concrete Structures, held at ETH Zurich in Switzerland on 17–18 April 2012. Major focus is put on the discussions. During the latter it was revealed that one of the key durability challenges lies in predicting the performance of new materials, where the increasing diversity of cement and concrete plays a major role. As most current engineering models are not capable of predicting actual field performance adequately, a knowledge-based approach to durability will become more important than ever. Only a scientific approach on a multi-scale and interdisciplinary level will allow predicting the performance of new materials (where no long-term experience is available for fitting purposes). This will facilitate the use of more performance-based durability design that is urgently needed to promote innovative, long-lasting solutions.

The Journal of Physical Chemistry C, Apr 19, 2013

ABSTRACT

Cement and Concrete Research, 2008

The thermo-mechanics of damage during delayed ettringite formation have been examined. A thermody... more The thermo-mechanics of damage during delayed ettringite formation have been examined. A thermodynamic approach is used to evaluate the supersaturation under which ettringite may form and the crystallization pressures that may result. From these stresses at the pore scale and with the amount of ettringite forming, an average hydrostatic tensile stress in the solid is calculated and compared to the tensile strength of tested samples.

Cement and Concrete Research, 2015

Tricalcium silicate is the most abundant phase of Portland cement clinker. The presence of common... more Tricalcium silicate is the most abundant phase of Portland cement clinker. The presence of common chemical impurities such as Mg, Al 2 O 3 and Fe 2 O 3 is primarily significant due to its major impact on surface reactivity as well as considerable effects on structural, interfacial and transport properties, in addition to organic adsorption. The present study provides the first force field model of alite based on stoichiometric analysis. Alite model is developed on the basis of INTERFACE force field parameters validated for pure C 3 S atomistic model. Modeling of defect sites have been performed by two types of substitution: iso-stoichiometric defects (Ca → Mg) and non-stoichiometric substitution, such as Al 2 O 3 or Fe 2 O 3 .

Studies on cement suspension rheology are often complicated because of the hydration reactions of... more Studies on cement suspension rheology are often complicated because of the hydration reactions of cement, which modify surface properties over time. Less complex and better defined systems, model powders, have been used instead of cement for studies of interactions with superplasticizers. The powders previously proposed have all at least one major drawback: they cannot be used at high pH or do not have surface properties at this pH close to that of Portland cement suspensions. Dead burnt MgO and Mg(OH) 2 are good candidates. Their isoelectric points are around pH 12.4 and 12 respectively, in the same range as that of Portland cement. Unfortunately, Mg(OH) 2 powders initially chosen, produced by precipitation, were found to be porous and were rejected. Dead burnt MgO powders react with water, but they can be considered as non-reactive between 1/2 hour to 2 hours of contact with a basic solution. Such powders allowed us to study adsorption of superplasticizers, evolution of zeta potential as a function of adsorbed superplasticizer and rheology of suspensions. It is also possible to add different salts to study their influence on the preceding properties. MgO powders can be easily sintered to get compacts or single crystals which can be used for atomic force microscopy measurements with aqueous solutions of superplasticizers.

Synopsis: Dispersion mechanisms of superplasticizers have received much attention over the past y... more Synopsis: Dispersion mechanisms of superplasticizers have received much attention over the past years. Recent developments have brought very efficient superplasticizers where the dominant stabilizing mechanism is thought to be via steric repulsion. These new superplasticizers contain an adsorbing backbone onto which non adsorbing side chains are grafted with the objective of getting them to stretch out into the solution from the cement particle surface and induce the steric repulsion upon approach of other particles. Another feature of these polymers is that they induce only very small zeta potentials. Calculations of interaction energies indicate that these polymers act predominantly through steric repulsion. However, the same calculations could lead to the conclusion that all polymers can only act through steric repulsion. The calculation of the steric and electrostatic contributions are greatly dependent on the polymer adsorption conformation and the distribution of charge at the particle surface associated with these adsorbed polyelectrolytes. Many of the assumptions made in calculating interparticle forces are not necessarily good approximations for polyelectrolytes. This paper discusses the limits of the approximations currently used in such calculations and presents a more accurate model for the calculation of these forces. The main result, applicable for a wide range of superplasticizers, is that both electrostatic and steric repulsions should be taken into account, provided the electrostatic charge can be assumed to lie at the outer-bound of the adsorbed layer of superplasticizers. Such information is of primary importance for understanding and solving cement and superplasticizer incompatibilities, as well as for developing novel products.

In this paper, we present a new approach to study the reactivity of minerals in contact with wate... more In this paper, we present a new approach to study the reactivity of minerals in contact with water, in particular cementitious systems. A focused ion beam is used to carve micron-sized gaps into crystals of tricalcium silicate (alite), which is the main reactive phase in portland cement. By adding water into those gaps we obtain a "micro-reactor" to study cement hydration. Scanning electron microscopy is used to monitor the cement hydration. In particular, we can reproducibly study the growth of hydrates between surfaces mimicking particles in close contact. Questions such as the densification of hydrates in those gaps and its consequence on the development of mechanical strength can be addressed. Moreover the same methodology opens completely new perspectives to studying the role of chemical additives on cement hydration (accelerators, retarders, etc). While conceptually simple and presenting a great potential, this method is surrounded by numerous experimental difficulties. We will present our solutions to these; and illustrate the microscopic procedures most adequate to characterize the behavior in these "micro-reactors". We will also point to the remaining challenges, in particular in terms of advanced microscopic characterization.

Cement particles in suspension are agglomerated and form a three-dimension network. The water tra... more Cement particles in suspension are agglomerated and form a three-dimension network. The water trapped between particles has no more lubricating action. Superplasticizers, which disperse agglomerates, free the entrapped water and thus improve the fluidity. It is shown how theories of interparticle forces developed for ceramic materials can be adapted to the more complex case of cement suspensions. In particular, how this allows the (semi-)quantitative evaluation of the yield stress of cement suspensions. The particle size distribution and the activity of ions of the pore solution are taken into account. The total interaction force allowing agglomeration of cement particles take into consideration the sum of dispersion, electrostatic and steric forces. The model is dependent on input parameters, like polymer adsorption, the plane of origin for the calculation of the electrostatic force, the Hamaker constant of cementitious phases, which need to be treated in more detail.

Over recent years, polycarboxylate superplasticizers have found their way into grinding aids used... more Over recent years, polycarboxylate superplasticizers have found their way into grinding aids used in cement production to reduce the electrical energy consumption. The effectiveness of these large molecules challenges the pre-existing theories concerning the factors that govern the performance of grinding aids. This paper reports on molecular dynamics simulations to examine a physical property believed to control the effectiveness of grinding aids, namely their adsorption energy. The molecules selected are TIPA (Triisopropanol amine), TEA (Triethanol amine) and glycerine. The surfaces examined are dry and hydroxylated C 3 S surfaces, which are believed to be more representative of reality, since some humidity is always present during the grinding. Detailed results of this part of the work show that glycerine interacts relatively more with dry as well as hydroxylated surfaces of C 3 S both at 25°C, ambient temperature and 110°C, grinding temperature with respect to TIPA and TEA. These result help to better understand the specific interaction of these molecules with cement surfaces. In the second part of this work oligomers of some PCE superplasticizers are examined with similar numerical tools on dry and hydroxylated surfaces of C 3 S. Results for different types of these oligomers, together with the previous results, shed light onto the reasons why polycarboxylate superplasticizers have found to also be effective grinding aids in cement production.

Materials and Structures, 2015

Concrete compressive strength decreases 10 significantly with decreasing density and therefore, t... more Concrete compressive strength decreases 10 significantly with decreasing density and therefore, there 11 are few examples of structural grade concretes with 12 densities below 1,600 kg/m 3. Here we show the devel-13 opment of structural lightweight aggregate concrete in 14 the 1,200-1,600 kg/m 3 density range. Compressive 15 s t r e n g t h so fu pt o3 6M P aa r eo b t a i n e da t2 8d a y s .B y 16 using fibers, mixes with flexural strengths of up to 7 MPa 17 and high ductility in flexure are obtained at 28 days. 18 These results are significantly better than those in 19 existing literature at comparable densities. Compressive 20 strength of lightweight concrete depends on both paste 21 and aggregate properties, while the flexural strength 22 depends mostly on the volume fraction of fibers used. 23 Keywords Lightweight aggregate concrete Á Fibers Á 24 Compressive strength Á Flexural strength 25 1 Introduction 26 Lightweight aggregate concrete (LWAC) is useful for 27 applications where reducing the structural dead load of concrete elements is beneficial. These applications include flooring systems, high-rise buildings, largespan concrete structures, offshore oilrigs, and constructions in seismic regions, among others. Numerous studies on LWACs have been performed focusing on topics such as mix design, fresh concrete properties, compressive and flexural strengths, ductility, and durability. Typically, these studies are on either relatively heavy concretes (densities over 1,600 kg/m 3) or ultra-lightweight ones (densities below 1,200 kg/m 3). The former are structural grade lightweight concretes and can show very good properties with 28-day compressive strengths as high as 80 MPa [19, 23, 46]. Ultra-lightweight concretes, on the other side, typically have very low strengths. Such concretes typically show strengths below 15 MPa at 28 days and are usually used for non-structural applications such as insulation [11, 15, 38]. The compressive strength increases non-monotonically with density. In particular much higher strength values, reaching as high as 80 MPa are obtained above a density range of about 1,600-1,800 kg/m 3 [3, 5, 13, 19, 23, 47]. On the other hand, the strength values for densities below 1,600 kg/m 3 are much lower, probably due to lower density mixes using weak lightweight aggregates that cannot bear much load. Increasing the aggregate density increases their strength and stiffness. There is perhaps an optimum balance around a density range of 1,600-1,800 kg/m 3 when both aggregates and matrix have similar stiffness, which reduces stress concentrations [8, 9]. Alternatively, there may A1 Electronic supplementary material The online version of A2 this article (

Journal of the American Ceramic Society, 2015

A new approach to study the early hydration reaction of alite is presented. It relies on milling ... more A new approach to study the early hydration reaction of alite is presented. It relies on milling micron-sized gaps into alite grains using a Focused Ion Beam. This provides for a "microreactor" in which hydration in confined conditions can be followed after introducing aqueous solutions of any desired composition. Hydration is followed by stopping the reaction at desired ages using solvent exchange and imaging the gaps using a Scanning Electron Microscope. These microreactors offer several advantages over conventional methods to monitor alite hydration. In particular, we can study the time evolution of dissolution in a reproducible manner using a setup that mimics particles in close contact. This paper presents this new methodology and applies it to studying the effects of solution composition and selected chemical admixtures on the dissolution of alite. New insights into the role of these factors are obtained and discussed.

The Journal of Physical Chemistry C, 2013

ABSTRACT

Materials and Corrosion-Werkstoffe Und Korrosion, 2012

ABSTRACT This paper sums up the International RILEM Workshop on Present and Future Durability Cha... more ABSTRACT This paper sums up the International RILEM Workshop on Present and Future Durability Challenges for Reinforced Concrete Structures, held at ETH Zurich in Switzerland on 17–18 April 2012. Major focus is put on the discussions. During the latter it was revealed that one of the key durability challenges lies in predicting the performance of new materials, where the increasing diversity of cement and concrete plays a major role. As most current engineering models are not capable of predicting actual field performance adequately, a knowledge-based approach to durability will become more important than ever. Only a scientific approach on a multi-scale and interdisciplinary level will allow predicting the performance of new materials (where no long-term experience is available for fitting purposes). This will facilitate the use of more performance-based durability design that is urgently needed to promote innovative, long-lasting solutions.

The Journal of Physical Chemistry C, Apr 19, 2013

ABSTRACT

Cement and Concrete Research, 2008

The thermo-mechanics of damage during delayed ettringite formation have been examined. A thermody... more The thermo-mechanics of damage during delayed ettringite formation have been examined. A thermodynamic approach is used to evaluate the supersaturation under which ettringite may form and the crystallization pressures that may result. From these stresses at the pore scale and with the amount of ettringite forming, an average hydrostatic tensile stress in the solid is calculated and compared to the tensile strength of tested samples.

Cement and Concrete Research, 2015

Tricalcium silicate is the most abundant phase of Portland cement clinker. The presence of common... more Tricalcium silicate is the most abundant phase of Portland cement clinker. The presence of common chemical impurities such as Mg, Al 2 O 3 and Fe 2 O 3 is primarily significant due to its major impact on surface reactivity as well as considerable effects on structural, interfacial and transport properties, in addition to organic adsorption. The present study provides the first force field model of alite based on stoichiometric analysis. Alite model is developed on the basis of INTERFACE force field parameters validated for pure C 3 S atomistic model. Modeling of defect sites have been performed by two types of substitution: iso-stoichiometric defects (Ca → Mg) and non-stoichiometric substitution, such as Al 2 O 3 or Fe 2 O 3 .

Studies on cement suspension rheology are often complicated because of the hydration reactions of... more Studies on cement suspension rheology are often complicated because of the hydration reactions of cement, which modify surface properties over time. Less complex and better defined systems, model powders, have been used instead of cement for studies of interactions with superplasticizers. The powders previously proposed have all at least one major drawback: they cannot be used at high pH or do not have surface properties at this pH close to that of Portland cement suspensions. Dead burnt MgO and Mg(OH) 2 are good candidates. Their isoelectric points are around pH 12.4 and 12 respectively, in the same range as that of Portland cement. Unfortunately, Mg(OH) 2 powders initially chosen, produced by precipitation, were found to be porous and were rejected. Dead burnt MgO powders react with water, but they can be considered as non-reactive between 1/2 hour to 2 hours of contact with a basic solution. Such powders allowed us to study adsorption of superplasticizers, evolution of zeta potential as a function of adsorbed superplasticizer and rheology of suspensions. It is also possible to add different salts to study their influence on the preceding properties. MgO powders can be easily sintered to get compacts or single crystals which can be used for atomic force microscopy measurements with aqueous solutions of superplasticizers.

Synopsis: Dispersion mechanisms of superplasticizers have received much attention over the past y... more Synopsis: Dispersion mechanisms of superplasticizers have received much attention over the past years. Recent developments have brought very efficient superplasticizers where the dominant stabilizing mechanism is thought to be via steric repulsion. These new superplasticizers contain an adsorbing backbone onto which non adsorbing side chains are grafted with the objective of getting them to stretch out into the solution from the cement particle surface and induce the steric repulsion upon approach of other particles. Another feature of these polymers is that they induce only very small zeta potentials. Calculations of interaction energies indicate that these polymers act predominantly through steric repulsion. However, the same calculations could lead to the conclusion that all polymers can only act through steric repulsion. The calculation of the steric and electrostatic contributions are greatly dependent on the polymer adsorption conformation and the distribution of charge at the particle surface associated with these adsorbed polyelectrolytes. Many of the assumptions made in calculating interparticle forces are not necessarily good approximations for polyelectrolytes. This paper discusses the limits of the approximations currently used in such calculations and presents a more accurate model for the calculation of these forces. The main result, applicable for a wide range of superplasticizers, is that both electrostatic and steric repulsions should be taken into account, provided the electrostatic charge can be assumed to lie at the outer-bound of the adsorbed layer of superplasticizers. Such information is of primary importance for understanding and solving cement and superplasticizer incompatibilities, as well as for developing novel products.

In this paper, we present a new approach to study the reactivity of minerals in contact with wate... more In this paper, we present a new approach to study the reactivity of minerals in contact with water, in particular cementitious systems. A focused ion beam is used to carve micron-sized gaps into crystals of tricalcium silicate (alite), which is the main reactive phase in portland cement. By adding water into those gaps we obtain a "micro-reactor" to study cement hydration. Scanning electron microscopy is used to monitor the cement hydration. In particular, we can reproducibly study the growth of hydrates between surfaces mimicking particles in close contact. Questions such as the densification of hydrates in those gaps and its consequence on the development of mechanical strength can be addressed. Moreover the same methodology opens completely new perspectives to studying the role of chemical additives on cement hydration (accelerators, retarders, etc). While conceptually simple and presenting a great potential, this method is surrounded by numerous experimental difficulties. We will present our solutions to these; and illustrate the microscopic procedures most adequate to characterize the behavior in these "micro-reactors". We will also point to the remaining challenges, in particular in terms of advanced microscopic characterization.

Cement particles in suspension are agglomerated and form a three-dimension network. The water tra... more Cement particles in suspension are agglomerated and form a three-dimension network. The water trapped between particles has no more lubricating action. Superplasticizers, which disperse agglomerates, free the entrapped water and thus improve the fluidity. It is shown how theories of interparticle forces developed for ceramic materials can be adapted to the more complex case of cement suspensions. In particular, how this allows the (semi-)quantitative evaluation of the yield stress of cement suspensions. The particle size distribution and the activity of ions of the pore solution are taken into account. The total interaction force allowing agglomeration of cement particles take into consideration the sum of dispersion, electrostatic and steric forces. The model is dependent on input parameters, like polymer adsorption, the plane of origin for the calculation of the electrostatic force, the Hamaker constant of cementitious phases, which need to be treated in more detail.

Over recent years, polycarboxylate superplasticizers have found their way into grinding aids used... more Over recent years, polycarboxylate superplasticizers have found their way into grinding aids used in cement production to reduce the electrical energy consumption. The effectiveness of these large molecules challenges the pre-existing theories concerning the factors that govern the performance of grinding aids. This paper reports on molecular dynamics simulations to examine a physical property believed to control the effectiveness of grinding aids, namely their adsorption energy. The molecules selected are TIPA (Triisopropanol amine), TEA (Triethanol amine) and glycerine. The surfaces examined are dry and hydroxylated C 3 S surfaces, which are believed to be more representative of reality, since some humidity is always present during the grinding. Detailed results of this part of the work show that glycerine interacts relatively more with dry as well as hydroxylated surfaces of C 3 S both at 25°C, ambient temperature and 110°C, grinding temperature with respect to TIPA and TEA. These result help to better understand the specific interaction of these molecules with cement surfaces. In the second part of this work oligomers of some PCE superplasticizers are examined with similar numerical tools on dry and hydroxylated surfaces of C 3 S. Results for different types of these oligomers, together with the previous results, shed light onto the reasons why polycarboxylate superplasticizers have found to also be effective grinding aids in cement production.

Materials and Structures, 2015

Concrete compressive strength decreases 10 significantly with decreasing density and therefore, t... more Concrete compressive strength decreases 10 significantly with decreasing density and therefore, there 11 are few examples of structural grade concretes with 12 densities below 1,600 kg/m 3. Here we show the devel-13 opment of structural lightweight aggregate concrete in 14 the 1,200-1,600 kg/m 3 density range. Compressive 15 s t r e n g t h so fu pt o3 6M P aa r eo b t a i n e da t2 8d a y s .B y 16 using fibers, mixes with flexural strengths of up to 7 MPa 17 and high ductility in flexure are obtained at 28 days. 18 These results are significantly better than those in 19 existing literature at comparable densities. Compressive 20 strength of lightweight concrete depends on both paste 21 and aggregate properties, while the flexural strength 22 depends mostly on the volume fraction of fibers used. 23 Keywords Lightweight aggregate concrete Á Fibers Á 24 Compressive strength Á Flexural strength 25 1 Introduction 26 Lightweight aggregate concrete (LWAC) is useful for 27 applications where reducing the structural dead load of concrete elements is beneficial. These applications include flooring systems, high-rise buildings, largespan concrete structures, offshore oilrigs, and constructions in seismic regions, among others. Numerous studies on LWACs have been performed focusing on topics such as mix design, fresh concrete properties, compressive and flexural strengths, ductility, and durability. Typically, these studies are on either relatively heavy concretes (densities over 1,600 kg/m 3) or ultra-lightweight ones (densities below 1,200 kg/m 3). The former are structural grade lightweight concretes and can show very good properties with 28-day compressive strengths as high as 80 MPa [19, 23, 46]. Ultra-lightweight concretes, on the other side, typically have very low strengths. Such concretes typically show strengths below 15 MPa at 28 days and are usually used for non-structural applications such as insulation [11, 15, 38]. The compressive strength increases non-monotonically with density. In particular much higher strength values, reaching as high as 80 MPa are obtained above a density range of about 1,600-1,800 kg/m 3 [3, 5, 13, 19, 23, 47]. On the other hand, the strength values for densities below 1,600 kg/m 3 are much lower, probably due to lower density mixes using weak lightweight aggregates that cannot bear much load. Increasing the aggregate density increases their strength and stiffness. There is perhaps an optimum balance around a density range of 1,600-1,800 kg/m 3 when both aggregates and matrix have similar stiffness, which reduces stress concentrations [8, 9]. Alternatively, there may A1 Electronic supplementary material The online version of A2 this article (

Journal of the American Ceramic Society, 2015

A new approach to study the early hydration reaction of alite is presented. It relies on milling ... more A new approach to study the early hydration reaction of alite is presented. It relies on milling micron-sized gaps into alite grains using a Focused Ion Beam. This provides for a "microreactor" in which hydration in confined conditions can be followed after introducing aqueous solutions of any desired composition. Hydration is followed by stopping the reaction at desired ages using solvent exchange and imaging the gaps using a Scanning Electron Microscope. These microreactors offer several advantages over conventional methods to monitor alite hydration. In particular, we can study the time evolution of dissolution in a reproducible manner using a setup that mimics particles in close contact. This paper presents this new methodology and applies it to studying the effects of solution composition and selected chemical admixtures on the dissolution of alite. New insights into the role of these factors are obtained and discussed.

The Journal of Physical Chemistry C, 2013

ABSTRACT