Combining Plasticizers Retarders and Accelerators

Katholieke Universiteit Leuven Faculteit Ingenieurswetenschappen Departement Burgerlijke Bouwkunde Norwegian University of Science and Technology Faculty of Natural Sciences and Technology Department of Materials Science and Engineering

Combining Plasticizers/Retarders And Accelerators

E2006 Promotor: prof. dr. H. Justnes prof. dr. ir. D. Van Gemert

Klaartje De Weerdt Dirk Reynders

Katholieke Universiteit Leuven Faculteit Ingenieurswetenschappen Academiejaar: 2005-2006 Departement: Burgerlijke Bouwkunde Adres en telefoon: Kasteelpark Arenberg 40 3001 Heverlee 016/32 16 54 Naam en voornaam studenten: De Weerdt Klaartje Reynders Dirk Titel eindwerk: Combineren van plastificeerders/vertragers en versnellers Korte inhoud eindwerk: De combinatie van plastificeerders/vertragers en versnellers werd bestudeerd met drie mogelijke toepassingen in het achterhoofd: 1) het tegengaan van het vertragend effect van plastificeerders zonder de reologie sterk te wijzigen, 2) de activatie van vertraagd beton op de werf na veilig transport in warme streken of steden met onvoorspelbaar verkeer en 3) het oververtragen van overschotten aan vers beton gevolgd door activatie na n of meerdere dagen. De experimenten werden grotendeels uitgevoerd op cementpasta. Een Paar-Physica MCR 300 rheometer werd gebruikt ter bepaling van de reologie en een TAM Air isotherme calorimeter ter bepaling van de hydratiecurves. Er werd vastgesteld voor toepassing 1) dat calciumnitraat het vertragend effect van natrium en calcium lignosulfonaat sterk terugschroeft en in het geval van polyacrylaat zelfs volledig wegneemt terwijl de combinaties werken als plastificeerders, voor toepassing 2) dat de combinatie natriumgluconaat/calciumnitraat een mogelijk werkend systeem is en voor toepassing 3) dat de combinatie citroenzuur/calciumnitraat het hergebruik van overschotten aan vers beton op een later tijdstip mogelijk maakt.

Promotor: prof. dr. ir. D. Van Gemert prof. dr. H. Justnes Assessoren: prof. dr. ir. L. Vandewalle ir. G. Heirman

Katholieke Universiteit Leuven Faculteit Ingenieurswetenschappen Year: 2005-2006 Department: Burgerlijke Bouwkunde Address en tel.: Kasteelpark Arenberg 40 3001 Heverlee 016/32 16 54 Name and surname students: De Weerdt Klaartje Reynders Dirk Title of thesis: Combining plasticizers/retarders and accelerators Summary of thesis: The combination of plasticizers/retarders with accelerators has been studied in view of three potential concrete applications: 1) counteracting retardation of plasticizers without negatively affecting rheology too much, 2) activating retarded concrete at site after safe transport in hot climate or cities with unpredictable traffic and 3) over-retarding residual fresh concrete one day and activating it next day or after several days. The experimental work is largely carried out on cement paste using a Paar-Physica MCR 300 rheometer to determine flow curves and gel strength and a TAM Air isothermal calorimeter for determination of heat of hydration curves. It has been found for application 1) that calcium nitrate strongly reduces retardation of sodium and calcium lignosulphonates and even cancels retardation of polyacrylates, whereas the blend also has plasticizing effects, for 2) that sodium gluconate/calcium nitrate is a potentially effective system and for 3) that citric acid/calcium nitrate may facilitate later use of residual fresh concrete.

Promotor: prof. dr. ir. D. Van Gemert prof. dr. H. Justnes Assessors: prof. dr. ir. L. Vandewalle ir. G. Heirman

Table of Contents1 2 Introduction Background on cement, cement hydration, rheology and admixtures 2.1 2.2 2.3 2.4 2.5 3 Cement .................................................................................................................. Cement hydration .................................................................................................. Rheology ............................................................................................................... Plasticizers/retarders ............................................................................................. Calcium nitrate ...................................................................................................... 1 4 4 5 9 13 22 24

Materials and apparatus

3.1 Materials................................................................................................................ 24 3.2 Apparatus .............................................................................................................. 27 4 Counteracting plasticizer retardation 4.1 4.2 4.3 4.4 5 Introduction ........................................................................................................... Calorimetric and rheological measurements......................................................... Mortar measurements............................................................................................ General conclusion................................................................................................ 34 34 35 75 80 81 81 81 95 98 101 110 111 111 111 114 116 120 126 127

Long transport of fresh concrete 5.1 5.2 5.3 5.4 5.5 5.6 Introduction ........................................................................................................... Sodium lignosulphonate........................................................................................ Citric acid .............................................................................................................. Lead nitrate............................................................................................................ Sodium gluconate.................................................................................................. General conclusion................................................................................................

6

Reutilizing residual fresh concrete 6.1 6.2 6.3 6.4 6.5 6.6 Introduction ........................................................................................................... Phase I Screening of retarders............................................................................ Phase II Determination of required retarder dosage .......................................... Phase III Activation using calcium nitrate ......................................................... Phase IV Strength measurements....................................................................... General conclusion................................................................................................

7

Conclusions

Chapter 1 IntroductionThis thesis continues a long tradition of Erasmus exchanges between the Katholieke Universiteit Leuven (Belgium) and the Norges Teknisk-Naturvitenskapelige Universitet i Trondheim (Norway). For many years students have been studying advanced aspects of cementitious materials. Thys, A. and Vanparijs, F. ([1]) studied the longterm performance of concrete with calcium nitrate, Ardoullie, B. and Hendrix, E. ([2]) focused on the chemical shrinkage of cementitious pastes and mortars, Clemmens, F. and Depuydt, P. ([3]) investigated early hydration of Portland cements, the thesis of Van Dooren, M. ([4]) concerned the factors influencing the workability of fresh concrete, and Brouwers, K. ([5]) studied a number of cold weather accelerators. In this thesis the combination of plasticizers/retarders and accelerators has been investigated in view of three different potential concrete applications. The first application, which made up the major part of this study, focused on the fact that plasticizers that are used to increase flow for cementitious materials at equal water-to-cement ratio also to a variable extent retard setting as a side effect. The objective was to find an accelerator that at least partially would counteract this retardation without negatively affecting the rheology too much. Whereas earlier studies on this topic focused on plastic viscosity at high shear rate (i.e. relevant for mixing) and relatively low dosages of plasticizer, the study reported here focused on the lower shear rate range (i.e. relevant for pouring concrete) and higher dosages of plasticizer. The results of this study are presented in Chapter 4. These results are valuable elements in evaluating the combined use of plasticizers and accelerators, as it was e.g. applied during construction of Statoils Troll platform (Figure 1.1), a huge gas platform located 80 km north-west of Bergen (Norway) that reaches 303 m below the surface of the sea. During the construction of its 350 m tall base an accelerator has been used to speed

1

Chapter 1: Introduction

2

up the slip forming process of the plasticized concrete as construction works were behind schedule. The second application concerns long transport of fresh concrete. The preliminary study was largely carried out on paste. It was investigated if a concrete mix from a ready mix plant after being deliberately over-retarded for long transport in for instance hot climate or cities with unpredictable traffic (e.g. traffic jam) could be activated by adding an accelerator in the revolving drum close to the construction site before pumping the concrete in place. Results are discussed in Chapter 5. The third potential application, presented in Chapter 6, concerns the search for a system to preserve residual fresh concrete for a few days (e.g. over a weekend) followed by activation before use. However, it might also be used as an overnight concept. Whereas recently a freezing preservation technique has been proposed as method for reutilizing left-over concrete, this study concentrated on a technique consisting of over-retardation of residual fresh concrete followed by later activation using an accelerator.

Figure 1.1 Troll gas platform (1996)

Chapter 1: Introduction

3

The necessary background on cement, cement hydration, rheology and admixtures is given in Chapter 2. Chapter 3 introduces and describes the materials and the apparatus that have been used throughout this work.

Chapter 2 Background on cement, cement hydration, rheology and admixtures2.1 CementCement chemists use in general a short hand notation, C = CaO, S = SiO2, A =Al2O3, F = Fe2O3 and S = SO3, for the main elements in the chemical analyses of cement, inaddition to H = H2O to describe hydration processes. The elements are determined by X-ray fluorescence or analytical chemistry and given as the corresponding oxides. Assuming that the only minerals in the cement are alite (C3S), belite (C2S), aluminate phase (C3A), ferrite phase (C4AF) and anhydrite ( C S ) the content of these minerals may be calculated through mass balances. The first four minerals are formed during equilibrium conditions in the burning of the cement clinker, while the latter mineral (or gypsum, C S H 2 ) is added to the mill when clinker is ground to cement. In specification sheets, the content of other oxides is also given: N (Na2O), K (K2O) and M (MgO). Free lime is the content of free CaO due to insufficient burning or due to the decomposition of C3S into C2S and free lime if the cooling rate is too low. The specific surface area (m2/kg) of cement is commonly determined directly by an air permeability method called the Blaine method. In addition to the specific area, the particle size is of importance for the hydration rate of cement, since the hydration takes place at the interface between the cement grain and the water phase. However, it is important to realise that the surface of a cement grain is inhomogeneous. The distribution of C3S/C2S- and C3A/C4AF-domains are determined by the milling process and the difference in resistance against fracture. Since cement grains are composite grains with possibly all 4 major phases in one grain, efforts to simulate

4

Chapter 2: Background

5

cement by adding corresponding amounts of individual minerals will therefore fail. (Justnes, H., [6], p.10)

2.2. Cement hydrationIn the discussion of rheology of cement paste and the interaction with plasticizing admixtures and retarders, it is of importance to know something about the hydration until setting. It is sometimes believed that no hydration takes place in the so-called dormant period between water addition and initial setting, while actually a substantial growth of hydration products takes place on the surface of the cement grains. (Justnes, H., [6], p.10)

2.2.1 The interstitial phases C3A/C4AFIn the absence of calcium sulphates the first hydration product of C3A which appears to grow at the C3A surface is gel-like. Later this material transforms into hexagonal crystals corresponding to the phases C2AH8 and C4AH19. The formation of the hexagonal phases slows down further hydration of C3A as they function as a hydration barrier. Finally the hexagonal phases convert to the thermodynamically stable cubic phase C3AH6 disrupting the diffusion barrier, after which the hydration proceeds with a fairly high speed. The overall hydration process may thus be written as 2 C 3 A + 27 H C 2 AH 8 + C 4 AH19 2 C 3 AH 6 + 15 H (hexagonal phases) (cubic phase)

In the presence of calcium sulphate (as in a Portland cement) the amount of hydration of C3A in the initial state of hydration is distinctly reduced when compared to that consumed in the absence of C S . Needle-shaped crystals of ettringite are formed as the main hydration product:

C3 A + 3 CSH 2 + 26 H C6 AS3 H32 Minor amounts of the monosulphate C 4 A S H 12 or even C 4 AH19 may also be formed if an imbalance exists between the reactivity of C3A and the dissolution rate of2 calcium sulphate, resulting in an insufficient supply of SO 4- - ions.

Then ettringite formation is accompanied by a significant liberation of heat. After a rapid initial reaction, the hydration rate is slowed down significantly. The length of this dormant period may vary and increases with increasing amounts of calcium sulphate in the original paste.


6

A faster hydration, associated with a second heat release maximum, gets under way after all the available amount of calcium sulphate has been consumed. Under these conditions the ettringite, formed initially, reacts with additional amounts of tricalcium aluminate, resulting in the formation of calcium aluminate monosulphate hydrate (monosulphate): C 6 A S3 H 32 + 2 C 3 A + 4 H 3 C 4 A S H 12 As ettringite is gradually consumed, hexagonal calcium aluminate hydrate ( C 4 AH19 ) also starts to form. It may be present in the form of a solid solution with C 4 A S H 12 or as separate crystals. The origin of the dormant period, characterised by a distinctly reduced hydration rate, is not obvious and several theories have been forwarded to explain it. The theory most widely accepted assumes the build-up of a layer of ettringite at the surface of C3A that acts as a barrier responsible for slowing down the hydration. Ettringite is formed in a through-solution reaction and precipitates at the surface of C3A due to its limited solubility in the presence of sulphates. The validity of this theory has been questioned arguing that the deposited ettringite crystals are not dense enough to account for the retardation of hydration. The four proceeding alternative theories have been proposed: i) The impervious layer consists of water-deficient hexagonal hydrate2 stabilised by incorporation of SO 4- . It is formed on the surface of C3A and

ii)

becomes covered by ettringite. C3A dissolves incongruently in the liquid phase, leaving an aluminate rich layer on the surface. Ca2+ - ions are adsorbed on it, thus reducing the number of active dissolution sites and thereby the rate of C3A dissolution. A subsequent adsorption of sulphate ions results in a further reduction of the dissolution rate.

iii)

2 SO 4- - ions are adsorbed on the surface of C3A forming a barrier. Contrary

to this theory it has been found that C3A is not slowed down if the calcium iv) sulphate is replaced by sodium sulphate. Formation of an amorphous layer at the C3A surface that acts as an osmotic membrane and slows down the hydration of C3A.

The termination of the dormant period appears to be due to a breakdown of the protective layer, as the added calcium sulphate becomes consumed and ettringite is converted to monosulphate. In this through-solution reaction both C3A and ettringite dissolve and monosulphate is precipitated from the liquid phase in the matrix.


7

The composition of the calcium aluminoferrite phase (ferrite phase), usually written as C4AF, may vary between about C4A1.4F0.6 and C4A0.6F1.4. Under comparable conditions the hydration products formed in the hydration of the ferrite phase are in many aspects similar to those formed by the hydration of C3A although the rates differ and the aluminium in the products is partially substituted by ferric ions. The reactivity of the ferrite may vary over a wide range, but seems to increase with increasing A/F ratio.

2.2.2 The main mineral alite C3SThe hydration of alite can be divided into 4 periods: a) Pre-induction period: Immediately after contact with water, an intense, but short-lived hydration of C3S gets under way. An intense liberation of heat may be observed in this stage of hydration. The duration of this period is typically no more than a few minutes. b) Induction (dormant) period: The pre-induction period is followed by a period in which the rate of reaction slows down significantly. At the same time the liberation of heat is significantly reduced. This period lasts typically a few hours. c) Acceleration (post-induction) period: After several hours the rate of hydration accelerates suddenly and reaches a maximum within about 5 to 10 hours. The beginning of the acceleration period coincides roughly with the beginning of the second main heat evolution peak. The Ca(OH)2 concentration in the liquid phase attains a maximum at this time and begins to decline. Crystalline calcium hydroxide (portlandite) starts to precipitate. The initial set as determined by Vicat-needle is often just after the start of this period and the final setting time just before the ending of it. d) Deceleration period: After reaching a maximum the rate of hydration starts to slow down gradually, however, a measurable reaction may still persist even after months of curing. The reason for this is that the hydration reaction becomes diffusion controlled due to hydration products growing around the unhydrated cement core in increasingly thickness.

Chapter 2: Background The overall alite hydration reaction may ideally be written as 2 C 3S + 7 H C 3S 2 H 4 + 3 CH

8

The calcium hydroxide, CH, is crystalline, while the calcium silicate hydrate is amorphous with a variable composition and therefore often simply denoted CSH-gel.

2.2.3 Hydration and setting of ordinary Portland cementThe overall hydration of ordinary Portland cement is basically a combination of the description of the interstitial phase with gypsum and alite as discussed in the preceding sections. Which of the two dominates the setting is still a matter of discussion and probably depends on the cement composition The hydration of Portland cement can be associated with the liberation of hydration heat. Figure 2.1 shows the heat evolution curve for a typical Portland cement.

Rate of Heat Evolution

Dissolution Ettringite and CSH gel Formation

Formation of Monosulfate

Rapid Formation of CSH and CH

Induction Period Increase in Ca2+ and OH- Concentration

DiffusionControlled Reactions Final Set

Initial Set

Min

Hours

Days Time of Hydration

Figure 2.1 Hydration heat evolution of an ordinary Portland cement. (Justnes, H., [6], p. 10)

In cements containing at least a fraction of the K+ in the form of potassium sulphate, the hydration process may be marked by a distinct initial endothermic peak immediately after mixing which is due to the dissolution of this cement constituent in the mixing water. A rather intense liberation of heat with a maximum within a few


9

minutes is due to the initial rapid hydration of C3S and C3A. Hydration of calcium sulphate hemihydrate to dehydrate may also contribute to this exothermic peak. After a distinct minimum, due to the existence of a dormant period in which the overall rate of hydration is slowed down, a second, mean exothermic peak, with a maximum after a few hours, becomes apparent. It is mainly due to the hydration of C3S and the formation of the CSH phase and portlandite. After that, the rate of heat release slows down gradually and reaches very low values within a few days. In most but not all cements, a shoulder or small peak may be observed at the descending branch of the main peak, which is probably due to renewed ettringite formation, there may even be a second shoulder which is attributed to ettringite-monosulphate conversion. (Hewlett, P., [7], p. 270-271)

2.3 Rheology2.3.1 General viscosityIn his Principa published in 1687, Isaac Newton formulated the following hypothesis about steady simple shearing flow: The resistance which arises from the lack of slipperiness of the parts of the liquid, other things being equal, is proportional to the velocity with which the parts of the liquid are separated from each other. This is shown in Figure 2.2.

Figure 2.2 Steady simple shearing flow. (Justnes, H., [6], p. 3)This lack of slipperiness is what we now call viscosity. It is synonymous with internal friction and is a measure of resistance to flow. The force per unit area required to produce the motion F/A is denoted shear stress ( ) and is proportional to the velocity gradient U/d (or shear rate, ). The constant of proportionality, , is called the shear viscosity (also called apparent viscosity):

=


10

The simplest rheological behaviour for liquids is the Newtonian viscous flow and Hookes law for solid materials. Ideal viscous (or Newtonian) flow behaviour is described using Newtons law

= Examples of ideal viscous materials are low molecular liquids such as water, solvents, mineral oils, etc. and they are often called Newtonian liquids. Hookes law states that the shear force acting on a solid is proportional to the resulting deformation

= G where G is the rigidity modulus. Many materials especially those of colloidal nature show a mechanic behaviour in between these to border lines (Hookes an Newtons laws), i.e. they have both plastic and elastic properties and are called viscoelastic. Samples with a yield point only begin to flow when the external forces acting on the material are larger than the internal structural forces. Below the yield point, the material shows elastic behaviour, i.e. it behaves like a rigid solid that under load displays only a very small degree of deformation that does not remain after removing the load. To describe the rheology of samples showing a yield point the Bingham model is often used. The Bingham model was extended by Herschel/Bulkley to include samples with apparent yield point due to shear thinning or thickening:

= 0 + p pp = 1 for samples with Bingham behaviour (true yield point) p < 1 for samples exhibiting shear thinning (apparent yield point) p > 1 for samples with shear thickening behaviour

Shear thinning is a reduction of viscosity with increasing shear rate in steady flow. Samples with shear thinning behaviour can be macromolecule solutions or melts where the individual molecules are entangled. Under high shear load the macromolecules will stretch out and may be disentangled, causing a reduction of the viscosity. Furthermore, in dispersions or suspensions shearing can cause particles to orient in the flow direction, agglomerates to disintegrate or particles to change their


11

form. During this process the interaction forces between the particles usually decrease and this also lowers the flow resistance. Shear thickening is an increase of viscosity with increasing shear rate. Shear thickening flow behaviour occurs in concentrated chemically unlinked polymers due to mechanical entanglements between the mostly branched molecule chains. The higher the shear load the more the molecule chains prevent each other from moving. If, during the shear process with highly concentrated suspensions, the particles touch each other more and more the consequences are similar: the resistance to flow increases. Cement paste has shear thinning properties due to both agglomerates of cement grains and growth of needle-shaped ettringite in the fresh state. An extreme case of particles that will change shape under shear load easily are entrained air bubbles. There is often more air in concrete than in cement paste, and this may make it difficult to correlate the concrete rheological properties with those of the same paste using the particle-matrix model. Note that concrete with 5 volume percentage air corresponds to 15 20 volume percentage air in the matrix, something that clearly will affect the matrix rheology.

2.3.2 Flow resistanceNumerous rheological models have been proposed to describe cementitious materials. The Bingham model has become very popular due to its simplicity and ability to describe cementitious flow. The model describes the shear stress ( ) as a function of yield stress ( 0 ), plastic viscosity ( p ) and shear rate ( ) as = 0 + p

The concept of yield stress is sometimes a very good approximation for practical purposes. It is however clear that the Bingham model often only applies for limited parts of the flow curve if the tested material has shear thinning or shear thickening flow behaviour. The Bingham model is dependent on the shear rate range for shear thickening materials. The shear thickening behaviour results furthermore in negative yield stress values at the high shear rate, which has no physical meaning (see Figure 2.3). There is a similar strong effect of the shear rate range on the flow parameters of a shear thinning paste.


12

p

0

Figure 2.3 Shear thickening behaviour resulting in negative yield stress values when using the Bingham model. The Hershel/Buckley equation = 0 + p p can be used to fit flow curves of pastes

showing shear thinning or shear thickening behaviour. However, it may be difficult to compare viscosities ( p ) for different mixes with different

p-factors. Negative yield stress values ( 0 ) with no physical meaning can sometimes also be obtained using the Hershel/Buckley equation. Therefore the area under the flow curve (Vikan, H. and Justnes, H., [8]) was chosen as a measure of flow resistance (Figure 2.4). This parameter, from here on referred to as flow resistance, shall be used throughout to work to describe the flow curve. The flow resistance will always be a positive value and not depend on curve shape.

flow resistance

Figure 2.4 Flow resistance.


13

Furthermore, the choice between two parameters for correlation, as for the Bingham model, can be omitted. It can be shown (Vikan, H. and Justnes, H., [8]) that the area under the flow curve represents something more physical than an apparent yield stress from Bingham modeling. In a parallel plate set-up with shear area, A [m2], and gap h [m] between the plates:

=

F A v h

[N/m2 or Pa] [m/s.m or s-1]

=

where F [N] is the force used to rotate the upper plate and v [m/s] the velocity.

F v F v F v Area under the curve = = = = A h V A h where V [m3] is the volume of the sample. The unit of the area under the curve is then [N.m/m3.s or J/m3.s or W/m3]. It is in other words the power required to make a unit volume of the paste flow with the prescribed rate in the selected range. The power,

P [W], required to mix concrete for a certain time interval is actually sometimes measured by simply monitoring voltage (U [V]) and current (I [A]) driving theelectrical motor of the mixer, since P = U.I.

2.4 Plasticizers/retarders2.4.1. IntroductionWater-reducing admixtures or plasticizers are all hydrophilic surfactants which, when dissolved in water, deflocculate and disperse particles of cement. By preventing the formation of conglomerates of cement particles in suspension, less water is required to produce a paste of a given consistency or concrete of particular workability. Maintaining low water contents whilst achieving an acceptable level of workability results in higher strengths for given cement content as well as lower permeability and reduced shrinkage. An important consequence of the reduction in the permeability is a major enhancement of its durability. The permeability of concrete to gases (oxygen, CO2), and water (carrying chlorides, sulfates, acids and carbonates) is of major importance with respect to its durability. Retarding admixtures, which extend the hydration induction period and thereby lengthening the setting times, are often treated together with plasticizing admixtures as the main components used for retarding mixtures are also present in water-reducing


14

admixtures. As a result, many retarders tend to reduce mixing water and many water reducers tend to retard the setting of concrete. A much greater reduction in the volume of mixing water can be achieved using socalled superplasticizers or high-range water-reducing admixtures in case of concretes of normal workability. Normal water reducers are capable of reducing water requirement by about 10-15%. Further reductions can be obtained at higher dosages but this may result in undesirable effect on setting, air content, bleeding, segregation and hardening characteristics of concrete. Superplasticizers are capable of reducing water contents by about 30%. (Ramachandran, V.S., [9], p. 211) Much of the following is based on Rheology of Cement based Binders State-of-theArt by H. Justnes ([6]).

2.4.2. Common plasticizer typesThere are four generations of plasticizers/water reducers in terms of time of discovery/use: 1. Salts of hydrocarboxylic acids with strong retarding effects 2. Calcium or sodium lignosulphonate (denoted CLS or NLS) as by-products from pulping industry with medium retarding properties. 3. Synthetic compounds like naphtalene-sulphonate-formaldehyde condensates (SNF) and sulphonated melamine-formaldehyde condensates (SMF) with small retarding properties. 4. Synthetic polyacrylates with grafted polyether side chains (PA) with small retarding properties. The first generation plasticizers, the salts of organic hydroxycarboxylic acids, are mostly used for their dominating retarding behavior. As the name implies, the hydrocarboxylic acids have several hydroxyl (OH) groups and either one or two terminal carboxylic acids (COOH) groups attached to a relatively short carbon chain. Figure 2.5 illustrates some typical hydroxycarboxylic acids which can be used as water reducing or retarding admixtures. Gluconic acid is perhaps the most widely used admixture. Citric, tartaric, mucic, malic, salicylic, heptonic, saccharic and tannic acid can also be used for the same purpose. Usually they are synthetized chemically


15

Figure 2.5 Typical hydrocarboxylic acids used in water reducing admixtures. (Ramachandran, V.S., [9], p.126)and have a very high degree of purity as they are used as raw materials by pharmaceutical and food industries. Some aliphatic hydrocarboxylic acids, however, can also be produced from fermentation or oxidation of carbohydrates and for this reason are also called sugar acids. Hydrocarboxylic acids can be used alone as retarders or water-reducing and retarding admixtures. For use as normal and accelerating water reducers they must be mixed with an accelerator. (Ramachandran, V.S., [9], p. 125) The second generation plasticizers, the lignosulphonates, are still the most widely used raw material in the production of water reducing admixtures. Lignosulphonates are sulphonated macromolecules from partial decomposition of lignin by calcium hydrogen sulphite. Under sulphite pulping, lignin is sulphonated and rendered water soluble. The spent sulphite liquor contains sulphonated lignin fragments of different molecular sizes and sugar monomers after removing the pulp. It can be further purified by fermentation to remove hexoses and by ultrafiltration to enrich larger molecular fractions. In addition to chemical modification of functional groups for special applications, simple treatment by sodium sulphate will ion exchange calcium


16

through formation of gypsum that is removed. A fragment of a lignosulphonate is illustrated in Figure 2.6. Fractionation to enrich larger molecular fractions increases the effectiveness of lignosulphonate as a dispersant for cement in water and reduces the retarding effect. Sodium lignosulphonates retard in general less than calcium lignosulphonates.

Figure 2.6 Fragment of lignosulphonate. (Justnes, H., [6], p. 30)Due to the size of the molecule, it cannot be ruled out that lignosulphonates disperse cement both through electrostatic repulsion and steric hindrance. The average molecular weight of common lignosulphonates used as plasticizers for cement may be about 5,000-10,000. It is assumed that the structure of lignosulphonates in solution consists of a mainly hydrophobic hydrocarbon core with sulphonic groups positioned at the surface. The bulk of the model is assumed to be made up of cross linked, polyaromatic chains which are randomly coiled. The negatively charged groups are positioned mainly on the surface or near the surface of the particle, and a double layer


17

of counter ions is present in the solvent. The lignosulphonate molecules behave as expanding polyelectrolytes as they expand at low and contract at high salt concentrations. The third generation plasticizers, the synthesized polymers with sulphonated groups, are not covered here as they were not used in this work. The fourth generation of plasticizers is based on a polyacrylate (PA) backbone that is obtained by free radical polymerization of different vinyl monomers. This backbone may vary widely in composition depending on the choice of monomers as shown in Figure 2.7. The next step is to graft on side chains of polyether (polyethylene oxide). Variations in the nature and relative proportions of the different monomers in the copolymer yield a group of products having broad ranges of physico-chemical and functional properties. Since some of the polyacrylates seem to enhance the segregation tendencies, they are often combined with viscosifiers to counteract this effect.

Figure 2.7 Illustration of a generic group of polyacrylate copolymers where R1 equals H or CH3, R2 is a poly-ether side chain (e.g., polyethylene oxide) and X is a polar (e.g., CN) or ionic (e.g., SO3) group. (Ramachandran, V.S. et al., [10], p.52) 2.4.3. Mechanisms of dispersionThere are generally two main mechanisms which explain how plasticizers disperse particles in a suspension: electrostatic repulsion and steric hindrance. These two mechanisms are sketched Figure 2.8 and Figure 2.9 respectively. Since its ionic lattice is cut, any fractured mineral particle will have domains of positive and negative charged sites. Negatively charged polymers (common feature of most plasticizers) will absorb to the positive charged sites and render the total particle surface negatively charged. As negatively charged particles approach each other there will be an electrostatic repulsion preventing them from getting close and attach to form


18

Figure 2.8 Sketch of how negative charged polymers may adsorb to both positively and negatively charged domains of particles. The resulting overall negative charge of the particles will prevent them to form agglomerates by electrostatic repulsion and they will stay dispersed. The electrostatic repulsion effect increases with increasing charge density of the adsorbed molecule. (Ramachandran, V.S. et al, [10], p.200)

Figure 2.9 Sketch of branched macromolecules adsorbing on the surface of grains that will create steric hindrance for them to get close enough to form agglomerates. The size effect of steric hindrance increases with increasing molecular weight (or actual size) of the adsorbed molecule. (Ramachandran, V.S. et al, [10], p.201)agglomerates. The latest generation of grafted polymers may also have some negative charges on their backbone that can co-ordinate on the positive sites but it should be noted that the ester group of acrylates may co-ordinate strongly to calcium anyway without any charge. The grafted polyether chains perpendicular to the backbone may stretch out and hinder the particles to get close enough to form agglomerates. This so-called steric hindrance is based on the size of the adsorbed molecules perpendicular to the particle surface. This is shown in Figure 2.10.


19

Figure 2.10 Idealized model on how a grafted polymer will lead to steric hindrance by adsorbing the polymer backbone to the surface and stretching the grafted side chains into the water phase. (Justnes, H., [6], p. 26)

The model of the grafted polymer dispersing according to steric hindrance in Figure 2.10 may be a simplification. It would then be necessary for all the intermolecular bonds (van der Waals type hydrogen bonds) to break and unwind the polyether chains to let them stretch out into the water phase (even though the hydrophilic nature of polyethers may aid in stabilizing such configuration). Alternatively, the molecules may stay unwound as polymeric balls or micelles that equally well will lead to steric hindrance (see Figure 2.11). While the first three generations of plasticizers are said to rely on electrostatic repulsion as mechanism for their dispersion of cement agglomerates, the fourth generation is the first to be designed to function through steric hindrance.

Macromolecular micelles

Cement surface Figure 2.11 Model of how macromolecules with strong intramolecular forces still may disperse through steric hindrance as polymer balls or micelles (after Justnes, H., [6], p. 26)Another effect that will prevent agglomerates formation is called depletion as sketched in Figure 2.12. The mechanism of this is that surplus polymer will not be adsorbed and will stay in the water phase between the particles and for this reason prevents them from getting close enough to form agglomerates.


20

cement particle s

cement particle s

polymer Figure 2.12 Surplus polymer in the water phase (not adsorbed) may prevent the cement particles to get close enough to form agglomerates. This depletion effect will not disperse by itself, but rather help stabilize dispersions by preventing flocculation. (after Justnes, H., [6], p. 27)

Rheology may also be improved by a tribology effect as sketched in Figure 2.13. Tribology is the science of friction, abrasion and lubrication. Low molecular weight compounds may reduce the friction between particles and also reduce the surface tension of the water face.

cement particle s

cement particle s

Low molecular weight compound Figure 2.13 Low molecular compounds in the water phase may improve rheology of particle suspensions by lubrication and by lowering the surface tension of the water phase, which may be denoted as a tribology effect. (after Justnes, H., [6], p. 27)Initial rheology of cement paste is also governed by early hydration, unlike inert particles suspensions (e.g. limestone). Thus, there are other mechanisms of how plasticizers may improve rheology of cement pastes. One is adsorption to active sites


21

and retardation of the formation of hydration products (see Figure 2.14), another is changing the morphology of the hydration products formed by reducing growth (see Figure 2.15) or by intercalation in the hydration products (see Figure 2.16).

Figure 2.14 Rheology in cement pastes may improve due to less hydration caused by adsorbed polymers co-ordinating to active sites (). The effect increases with decreasing size of the molecules. LMW = low molecular weight and HMW = high molecular weight. (Ramachandran, V.S. et al, [10], p.201)

Figure 2.15 Schematic illustration of hydration nucleation and growth inhibition by adsorbed molecules. Selective adsorption on crystal planes can give morphology changes. (Ramachandran, V.S. et al, [10], p.208)


22

Figure 2.16 Intercalation of plasticizer in hydration product with structural alteration (e.g. lignosulphonates with hydration products of C3A). (Ramachandran, V.S. et al, [10], p.209)

2.5 Calcium nitrateThis section is based on the paper Setting Accelerator Calcium Nitrate,

Fundamentals, Performance and Applications by Justnes, H. and Nygaard, E. ([11]). In the past a growing concern about the chloride-induced corrosion of reinforcing barsembedded in Portland cement concrete has led to the development of a number of chloride-free set accelerating admixtures to replace the widely used calcium chloride accelerator. In 1981, calcium nitrate, Ca(NO3)2, was proposed as a basic component of a set accelerating admixture. Calcium nitrate, denoted as CN, works as a pure set accelerator (see Figure 2.17), and not as a strength development accelerator. The pure set accelerating effect is beneficial in preventing any increase in maximum temperature in massive constructions due to the heat of hydration. In spite of this, an increase in long term compressive strength is often observed, probably due to binder morphology changes.

Hardening

Setting

Reference

Figure 2.17 Difference between set and hardening accelerators.


23

The effectiveness of CN as a setting accelerator for cement is dependent on the cement type. The set accelerating efficiency appeared to be correlated with the belite, C2S, content, while no correlation between set accelerating efficiency and C3A has been found. In order to find the reason for the linear correlation between accelerator efficiency and belite content, and possibly the mechanism of CN as set accelerator for cement, Justnes and Nygaard undertook a thorough analysis of the water in cement pastes from mixing to paste setting for two different cement types (HS65 and P30). For both cement pastes the most noticeable change when 1.55 % CN by weight of the cement was added, was that the calcium concentration increased and the sulphate concentration decreased. Thus, the mechanism for accelerated setting is twofold: i) ii) an increased calcium concentration leads to a faster super-saturation of the fluid with respect to calcium hydroxide, Ca(OH)2, while a lower sulphate concentration will lead to slower/less formation of ettringite which will shorten the onset of aluminate, C3A, hydration.

The difference between the two cements was that P30 contained much more of the mineral aphthitalite, K3Na(SO4)2, which leads to a high initial sulphate concentration in the fluid. When CN was added, much of the calcium precipitated as sparingly soluble gypsum. Even when 1.55 % CN was added to the P30 paste, the sulphate concentration in the fluid was higher than in the water of HS65 paste without CN. At the same time, the calcium concentration in the fluid of P30 with CN was only slightly higher than for HS65 without CN. The Ca2+ concentration in the water of HS65 paste, on the other hand, was increased with about 4 times when 1.55 % CN was added. Thus, the reason why CN did not accelerate the setting of P30 was that it contained a very soluble alkali sulphate originating from the clinker process. The correlation between belite content and set accelerating efficiency is understandable since belite can incorporate a portion of the total alkalies in its structure and consequently prevent them from taking part in the early fluid chemistry since belite is a slow reacting mineral. Hence, for a series of cements, with about equal total alkali content and increasing belite content, it is expected that the set accelerating efficiency of CN will increase. On the other hand, in an investigation of calcium acetate, chloride and nitrate on belite hydration, it has been found that after 1 day, the chemically bound water was 6 times larger when 2 % CN was mixed in the water, while 2 % calcium acetate and 2 % calcium chloride only increased the 1 day chemically bound water by 30 % compared with the reference. Therefore, a special influence of CN on -C2S can not be excluded.

Chapter 3 Materials and apparatusThe purpose of this chapter is to introduce and describe the materials and the apparatus that have been used frequently throughout this work.

3.1 Materials3.1.1. Cements Two Portland cements have been used in this thesis. Their physical characteristics are given in Table 3.1, chemical analysis according to producer and minerals by Bogue estimation is given in Table 3.2 and the mineralogy of the cements determined by multicomponent Rietveld analyses of XRD profiles, specific surface determined by the Blaine method and content of easily soluble alkalis determined by plasmaemissionspectrometry are given in Table 3.3. Table 3.1 Physical characteristics of Portland cements according to EN 196 Cement type Fineness: Grains + 90 m Grains + 64 m Grains 24 m Grains 30 m Blaine (m2/kg) Water demand Le Chatelier Initial set time c (MPa) at 1 day 2 days 7 days 28 days CEM I 52.5 R - LA 1.7% 4.1% 66.3% 75.6% 359 26.7% 0.5 mm 145 min. 17.1 27.5 42.5 58.6 CEM I 42.5 RR* 0.1% 0.5% 89.2% 94.8% 546 32.0% 0 mm 115 min. 32.7 39.9 49.3 58.9

24

Chapter 3: Materials and apparatus

25

Table 3.2 Chemical analysis (%) of the Portland cements according to producer and minerals (%) by Bogue estimation. Cement type Chemical analyses CaO SiO2 Al2O3 Fe2O3 SO3 MgO Free CaO K2O Na2O Equiv. Na2O Cr6+ (ppm) Carbon Chloride LOI Fly Ash Minerals by Bogue C3 S C2 S C3 A C4AF CS CEM I 52.5 R - LA CEM I 42.5 RR*

63.71 20.92 4.21 3.49 2.67 1.87 0.84 0.46 0.19 0.49 0.30 0.17 0.02 1.72 -

61.98 20.15 4.99 3.36 3.55 2.36 1.23 1.08 0.42 1.13 0.00 0.04 0.03 1.34 -

50.4 22.0 5.3 10.6 5.8

50.7 19.5 7.5 10.2 7.7

(* The RR term refers to the Norwegian standard NS 3086 (2003) where RR means extra demands to 1 and 2 day strength compared to R. 42.5 RR should then have characteristic 1 day strength 20.0 MPa and 2 day strength 30.0 MPa.) It can be seen that the CEM I 42.5 RR cement had a higher alkali and C3A content and a higher specific surface than the CEM I 52.5 R LA cement and, as a consequence of the latter two, had a higher water demand. CEM I 42.5 RR cement pastes were therefore prepared with a w/c ratio of 0.50, whereas CEM I 52.5 R LA cement pastes were prepared with a w/c ratio of 0.40 throughout this work.


26

Table 3.3 Mineral composition (%) and alkali content of Portland cements obtained by QXRD and plasmaemissionspectrometry Cement type Alite Belite Ferrite Cubic aluminate Orthorombic aluminate Lime Periclase Gypsum Hemihydrate Anhydrite Calcite Portlandite Quartz Arcanite Mullite Amporhous Blaine K (%) Na (%) Naeqv (%) CEM I 52.5 R - LA 65.0 12.9 9.6 0.5 3.0 0.6 0.3 1.4 1.5 0.4 4.0 0.3 0.4 0.0 364 0.32 0.74 0.26 CEM I 42.5 RR 64.7 14.8 7.5 5.9 1.1 1.0 1.6 0.0 1.8 0.6 0.5 0.3 0.0 0.3 546 0.92 0.22 0.76

3.1.2. Plasticizers/retarders Borregaard Lignotech, Sarpsborg, Norway delivered two lignosulphonate powders denoted as Ultrazine Na and Ultrazine Ca. Ultrazine Ca (CLS) was sugar reduced and large molecular size enriched by ultra filtration of the basic calcium lignosulphonate obtained in the sulfite process on spruce. In Ultrazine Na (NLS) the calcium in Ultrazine Ca has been ion exchanged with sodium. Solutions with 30% dry matter were prepared before use. A polyether grafted polyacrylate water solution containing 18% solids and a viscosifying agent has also been used as a plasticizer. The molecular weight of the polyacrylate was 220,000. A number of substances were used as retarders. They were all of analytical laboratory grade: - citric acid (C6H8O7 H2O ) - sodium salt of gluconic acid (C6H11NaO7) - sodium salt of tartaric acid (Na2C4H4O6 2H2O, right-turning form) - lead nitrate (Pb(NO3)2) - zinc acetate (Zn(CH3OO)2 2H2O) - sucrose (C12H22O11)


27

The trisodiumphosphate (Na3PO4 12H2O) used in this work was from technical quality. Household sugar was also used as a retarder.

3.1.3. Accelerator Technical calcium nitrate (CN) was used as an accelerator. Its formula may be written as xNH4NO3 yCa(NO3)2 zH2O, and named xyz CN according to short hand practice. The CN used in the present work had x = 0.092, y = 0.500 and z = 0.826, or in other words 19.00%Ca2+, 1.57% NH + , 64.68% NO3 and 14.10% H2O. The CN was delivered in the form of 4

granules by Yara, Porsgrunn, Norway. Calcium nitrate was also used in the form of a 50% aqueous solution of pure calcium nitrate Ca(NO3)2, also obtained from Yara. The fluid is colourless, viscous and can easily be blended into the mixing water.

3.2. Apparatus3.2.1. MixerThe cement pastes were blended in a high shear mixer by Braun (MR5550CA) and by Tefal (Rondo 500) as illustrated in Figure 3.1. The mixers had a rotational speed of approximately 800 rpm. It will be notified which of the blenders has been used in each chapter. The blending was performed by adding cement to the water and mixing for minute, resting for 5 minutes and blending again for 1 minute.

Figure 3.1 High shear blenders from Braun (left) and Tefal (right)


28

3.2.2. RheometerRheological measurements have been performed with a MCR 300 rheometer produced by Paar Physica (Figure 3.2). A parallel-plate measuring system was used as illustrated in Figure 3.3. This measuring system consisted of two plates. The surfaces of both the bob and the motionless plate were flat, but the upper plate had a serrated surface of 150 m depth to avoid slippage.

Figure 3.2 MCR 300 rheometer by Paar Physica

Figure 3.3 The parallel plate measuring system (Mezger T., [12], p. 177)

The geometry of the upper plate is determined by the plate radius R being 2.5 cm. The distance H between the two parallel plates must be much smaller than the radius R and has been recommended to be at least 10 times larger than the largest of the particles of the sample (Mezger T., [12], p. 177-179). The average particle size of unhydrated cement being


29

approximately 10 m (Taylor, [13]), the gap between the plates was set to 1 mm for all measurements. The temperature controlled bottom plate was set to 20 C. The parallel plate measuring system makes it possible to measure dispersions containing relatively large particles as well as samples with three-dimensional structures. The measuring system has however also a number of disadvantages. There is no constant shear gradient in the measurement gap because the shear rate (or shear deformation) increases in value from zero at the center of the plate to the maximum at the edge. Furthermore, several unwanted phenomena can occur at the edge of the plate: inhomogeneities, emptying of the gap, flowingoff and spreading of the sample, evaporation of water, or skin formation (Mezger T., [12], p. 180-181). To reduce evaporation both upper and lower plates were covered with a plastic ring and a metallic lid while a water trap attached to the upper plate was filled with water to ensure saturated water pressure. The following measuring sequence was used to determine the flow resistance (area under the (down) flow curve in the range from 2 to 50 1/s), the gel strength after 10 seconds of resting and the gel strength after 10 minutes of resting: 1. 1 minute with constant shear rate ( ) of 100 1/s to stir up the paste 2. 1 minute resting 3. Stress ( ) shear rate ( ) curve with linear sweep of from 2 up to 200 1/s in 30 points lasting 6 s each (up curve) 4. Stress ( ) shear rate ( ) curve with linear sweep of from 200 down to 2 1/s in 30 points lasting 6 s each (down curve) 5. 10 s resting 6. Shear rate ( ) stress ( ) curve with logarithmic sweep of from 1 to 100 Pa in 30 points lasting 6 s each to measure the gel strength after 10 s rest 7. 10 minutes resting 8. Shear rate ( ) stress ( ) curve with logarithmic sweep of from 1 to 400 Pa in 70 points lasting 6 s each to measure the gel strength after 10 minutes rest The recording of the shear rate ( ) stress ( ) curves was stopped whenever the shear rate ( ) exceeded 300 1/s to prevent the sample from being lost from the measurement gap. A flow chart of the mixing and measurement sequence is shown in Figure 3.4.


30

Shear ratemixing minute mixing 1 minute

gel strength up curve 1 minute at 100 1/s down curve

gel strength

transfer to rheometer

5 minutes rest

8 minutes

1 minute rest

10 seconds rest

10 minutes rest

Time

Figure 3.4 Flow chart of the mixing and measurement sequenceThe reproducibility of the rheological measurements was investigated for two different cement pastes. The cement pastes were made with distilled water. The plasticizer was added to the water. Cement paste 1 was prepared with CEM I 52.5 R LA cement and 0.30% sodium lignosulphonate by weight and a w/c ratio of 0.40. Paste 2 was prepared with CEM I 42.5 RR cement and 0.50% sodium lignosulphonate by weight and a w/c ratio of 0.50. Total paste volume was approximately 250 ml. Each of the two cement pastes was prepared 5 times. The rheological data has been transformed into flow resistance (area under the flow curve in the range from 2 to 50 1/s), gel strength after 10 seconds of rest and gel strength after 10 minutes of rest. The results are shown in Table 3.3 for cement paste 1 and Table 3.5 for paste 2. The data show that the reproducibility of the flow resistance is reasonable. Measurements of the gel strength show higher deviations, especially for the 10 minute gel strength of the CEM I 52.5 R LA cement pastes which had a standard deviation of 27%.


31

Table 3.4 Reproducibility of rheological measurements for cement paste 1 (w/c=0.40 CEM I 52.5 R LA 0.30% Ultrazine Na) PASTE 1 Flow resistance [Pa/s] 391 383 394 419 384 394 15 4% Gel strength [Pa] 10 sec. 10 min. 2.4 14.2 2.4 13.0 2.8 9.2 2.8 10.0 2.8 7.1 2.7 10.7 0.2 3 9% 27%

Average Standard deviation % standard dev.

Table 3.5 Reproducibility of rheological measurements for cement paste 2 (w/c=0.50 CEM I 42.5 RR 0.50% Ultrazine Na) PASTE 2 Flow resistance [Pa/s] 2119 2375 2455 2343 2392 2337 128 5% Gel strength [Pa] 10 sec. 10 min. 22.2 36.8 22.2 36.8 26.1 40.1 22.2 40.1 22.2 36.8 2.7 38.1 1.7 2 7% 5%

Average Standard deviation % standard dev.

3.2.3. CalorimeterAn eight-channel TAM Air Isothermal Calorimeter from Thermometric AB, Sweden was used for the heat of hydration measurements (Figure 3.5). The calorimeter was calibrated at 20 C. The hydration heat was measured by weighing 6 to 7 grams of cement paste into a glass ampoule after which the ampoule was sealed and loaded into the calorimeter. The ampoules were wiped with a paper tissue to make sure that they were perfectly clean and dry when they were inserted into the calorimeter. When studying the heat of hydration measurements it should be kept in mind that when an ampoule is loaded into the calorimeter the temperature of the calorimeter will be disturbed. If the temperature of the ampoule is 2 degrees higher than the thermostat temperature, an exothermic heat flow, showing an exponential decay, of roughly 400 mW is observed. This phenomenon explains the exponential decay in specific heat which is observed in the first hour after mixing.


32

Figure 3.5 TAM Air Isothermal Calorimeter

3.2.4. Adsorption of plasticizersTo measure the consumed amount of lignosulfonate on the cement a UV Spectrophotometer from Thermo Spectronic was used as illustrated in Figure 3.6. The adsorption measurements in this work utilized a wavelength of 285 nm. Pore solutions were extracted from the cement pastes by filtering the pastes through 0.45 m filter paper on a Bchner funnel using low vacuum 15 minutes after water addition. They were then diluted 25, 50 or 100 times with a solution of artificial pore water (NaOH and KOH with a K/Na molar ratio equal to 2 and pH = 13.2). The amount of plasticizer in the water phase was read from calibration curves which had been made with a dilution series of each of the two lignosulfonates being used in this work. The difference between the added and the measured content of plasticizer gave the bound portion.

Figure 3.6 UV Spectrophotometer from Thermo Spectronic


33

The consumption of polyacrylate on cement was determined by measuring Total Organic Carbon (TOC) left in the pore water with a Shimadzu TOC Analyzer 5000A. The Shimadzu TOC 5000A works by converting organic matter to carbon dioxide by combustion with a catalyst that promotes the redox reaction with oxygen. The reaction takes place at a temperature of 680 C. The amount of carbon dioxide formed is measured to determine the carbon content. The amount of plasticizer bound to the cement is given by the difference between the added and the measured content of organic carbon.

Chapter 4 Counteracting plasticizer retardation

4.1 IntroductionPlasticizers are used to increase flow for cementitious materials at equal water-to-cement ratio, but will also to a variable extent retard cement setting as a side effect. The objective was to find an accelerator that at least partially would counteract this retardation without negatively affecting the rheology too much. Earlier papers (Justnes, H., Petersen, B.G., [14] and [15]) focusing on this topic studied rheological properties at high shear rate (i.e. relevant for mixing) for relatively low dosages of plasticizer, whereas the study reported in this chapter focused on the lower shear rate range (i.e. relevant for pouring concrete) and higher dosages of plasticizer. Three different plasticizers were tested in the present study, but the accelerator was chosen to be calcium nitrate. The experimental work is largely carried out on cement paste using a Physica MCR 300 rheometer to determine flow curves and gel strength and an isothermal calorimeter for determination of heat of hydration curves. Two promising admixture blends were also tried out in mortar.

34

Chapter 4: Counteracting plasticizer retardation

35

4.2 Calorimetric and rheological measurements4.2.1. Experimental The investigated cement pastes were made with distilled water. Plasticizer and accelerator were added to the water before mixing, except for one series of pastes marked with DA (delayed addition), where the plasticizer was added 5 minutes after the start of initial blending in a 30% aqueous solution. Both a CEM I 52.5 R LA and a CEM I 42.5 RR Portland cement were used. Three different plasticizers were studied: a sodium lignosulphonate (NLS), a calcium lignosulphonate (CLS) and a polyether grafted polyacrylate (PA). The setting accelerator calcium nitrate (CN), available in a 50% aqueous solution, was used to counteract the retardation. A more detailed description of both plasticizers and accelerator can be found in Chapter 3. Table 4.1 provides an overview of the experimental program.

Table 4.1 Experimental program Plasticizer Accelerator Reference (0%) 0.15% NLS* 0.15% NLS DA* 0.30% NLS 0.50% NLS 0.00% CN 0.30% CLS 0.25% CN 0.50% CLS 0.50% CN 0.75% CN 0.10% PA 1.00% CN CEM I 42.5 RR Reference (0%) (w/c = 0.50) 0.50% NLS 1.00% NLS 0.50% CLS 1.00% CLS 0.10% PA (* The 1.00% CN dosage was not studied for these series.) Cement type CEM I 52.5 R LA (w/c = 0.40)

In Chapter 3 it was pointed out that the CEM I 42.5 RR cement had a higher alkali and C3A content and a higher specific surface than the CEM I 52.5 R LA cement and, as a consequence of the latter two, had a higher water demand. CEM I 42.5 RR cement pastes were therefore prepared with a w/c ratio of 0.50, whereas CEM I 52.5 R LA cement pastes were prepared with a w/c ratio of 0.40 throughout this work. Total paste volume was approximately 250 ml. The blending was performed in a high shear mixer of Braun (see 3.2.1) by adding the cement to the water containing plasticizer and/or accelerator and mixing for minute, resting for 5


36

minutes and blending again for 1 minute. The cement pastes containing 0.15% sodium lignosulphonate were mixed with a high shear mixer by Tefal using the same blending sequence. The heat of hydration versus time curves were measured by accurately weighing 6 to 7 grams of cement paste into a glass ampoule after which the ampoule was sealed and loaded into the calorimeter. The rheological properties were studied by performing the measurement sequence discussed in section 3.2.2 on the cement pastes 15 minutes after the start of the blending: To measure the consumed (adsorbed and intercalated) amount of plasticizer by cement, pore solutions were extracted from the cement pastes by filtering the pastes through 0.45 m filter paper on a Bchner funnel using low vacuum 15 minutes after water addition. The consumed amount of lignosulphonate was determined using a UV Spectrophotometer from Thermo Spectronic. The adsorption measurements in this work utilized a wavelength of 285 nm. The pore solutions were diluted 25, 50 or 100 times with a solution of artificial pore water (NaOH and KOH with a K/Na molar ratio equal to 2 and pH = 13.5). The amount of plasticizer in the water phase was read from calibration curves which had been made with a dilution series of each of the two lignosulphonates being used in this work. The calibration curves for NLS and CLS are given in Figure 4.1 and Figure 4.2 respectively. The difference between the added and the measured content of plasticizer gave the consumed amount. The consumption of polyacrylate by cement was determined by measuring Total Organic Carbon (TOC) left in the pore water with a Shimadzu TOC Analyzer 5000A.


37

Calibration curve, NLS0.9 0.8 0.7 Absorbance 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 % Added y = 137.5844x R2 = 0.9992

Figure 4.1 Calibration curve for adsorbance of sodium lignosulphonate (NLS).

Calibration curve, CLS0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.001

Absorbance

y = 137.2718x R2 = 0.9997

0.002

0.003

0.004

0.005

0.006

0.007

% Added

Figure 4.2 Calibration curve for adsorbance of calcium lignosulphonate (CLS). Prior to discussing the results, we shall provide an overview of the way in which the read outs from the rheometer were converted into flow resistance (area under the flow curve in the range from 2 to 50 1/s, see also Chapter 2), gel strength after 10 seconds of rest and gel strength after 10 minutes of rest. The measurements on the cement paste made with CEM I 52.5 R LA cement without any admixtures shall be used to illustrate this:


38

1. The flow resistance is defined as the area under the down flow curve in the range from 2 to 50 1/s. The down curve for the paste made with CEM I 52.5 R LA cement is shown in Figure 4.3. Table 4.2 shows the read outs from the rheometer. The area under the curve was determined by calculating the average of the shear stresses for every two consecutive measuring points in the range from 2 to 50 1/s and multiplying this by the difference in shear rate for these points. In this case a value of 2283 Pa/s was found for the flow resistance.

Table 4.2 Rheometer read outs for the down curve. Meas. Pt. 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 Shear Rate [1/s] 200 193 186 180 173 166 159 152 145 139 132 125 118 111 104 97.6 90.8 83.9 77.1 70.3 63.4 56.6 49.8 43.0 36.1 29.3 22.5 15.7 8.83 2.01 Shear Stress [Pa] 98.2 96.9 95.7 94.6 93.4 92.9 91.1 89.9 88.8 87.5 86.4 85.1 83.8 82.5 81.1 79.5 77.7 75.7 73.6 71.4 69.4 67.3 64.7 61.4 58.0 53.3 47.3 40.1 30.8 22.2


39

Down Curve120 100 Shear Stress [Pa] 80 60 40 20 0 0 50 100 Shear Rate [1/s] 150 200

Figure 4.3 Down curve.

2. The 10 sec. gel strength can be derived from the shear rate ( ) stress ( ) curve with logarithmic sweep of from 1 to 100 Pa in 30 points lasting 6 s each. The curve is plotted in Figure 4.4. The rheometer read outs are given in Table 4.3. The 10 sec. gel strength was calculated by taking the average of the shear stresses of measuring points 19 and 20 (Table 4.3) as the breakthrough happened somewhere in between. That way a value of 19 Pa was found for the 10 sec. gel strength.

10 sec. gel strength180 160 140 Shear Rate [1/s] 120 100 80 60 40 20 0 0 20 40 60 80 100 Shear Stress [Pa]

gel strength

Figure 4.4 Shear rate stress curve to determine the 10 sec. gel strength.

Chapter 4: Counteracting plasticizer retardation Table 4.3 Rheometer read outs to determine the 10 sec gel strength. Meas. Pt. 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 Shear Rate [1/s] 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.44 9.32 12.9 16.0 21.3 26.1 35.0 48.7 73.6 110 155 Shear Stress [Pa] 1.00 1.17 1.37 1.61 1.89 2.21 2.59 3.04 3.56 4.18 4.89 5.74 6.72 7.88 9.24 10.8 12.7 14.9 17.4 20.4 24.0 28.1 32.9 38.6 45.2 53.0 62.1 72.8 85.3 100

40

3. The calculation of the 10 min. gel strength is completely similar to that of the 10 sec. gel strength and shall therefore not be treated.

Chapter 4: Counteracting plasticizer retardation 4.2.2. Results and discussion for reference pastes

41

Figure 4.5 shows the flow resistances for both CEM I 52.5 R LA and CEM I 42.5 RR reference cement pastes. The flow resistance of the CEM I 42.5 RR cement paste (w/c = 0.50) is higher than the CEM I 52.5 R LA paste (w/c = 0.40) in spite of the higher water-to-cement ratio. This is due to the higher specific surface and the content of cubic C3A. Addition of calcium nitrate appeared to have no effect on the flow resistance of these pastes.

reference3500 CEM I 52.5 R LA CEM I 42.5 RR

Flow resistance (Pa/s)

3000 2500 2000 1500 1000 500 0 0.00

0.25

0.50

0.75

1.00

Calcium nitrate (%)

Figure 4.5 Flow resistance for CEM I 52.5 R LA (w/c = 0.40) and CEM I 42.5 RR reference cement pastes (w/c = 0.50) for different dosages of calcium nitrate.

The gel strengths after 10 seconds of rest are depicted in Figure 4.6. In case of CEM I 52.5 R LA cement paste, an increasing 10 seconds gel strength was observed for increasing calcium nitrate dosages up to 0.50%. Figure 4.7 shows the gel strengths after 10 minutes of rest. For both cement types an increasing (albeit less pronounced in case of CEM I 42.5 RR cement) gel strength can be seen for increasing calcium nitrate dosages.


42

reference35

CEM I 52.5 R LA

10 sec. gel strength (Pa)

30 25 20 15 10 5 0 0.00

CEM I 42.5 RR

0.25

0.50

0.75

1.00

Calcium nitrate (%)

Figure 4.6 Gel strength after 10 seconds of rest for CEM I 52.5 R LA (w/c = 0.40) and CEM I 42.5 RR reference cement pastes (w/c = 0.50) for different dosages of calcium nitrate.

reference300

CEM I 52.5 R LA CEM I 42.5 RR

10 min. gel strength (Pa)

250 200 150 100 50 0 0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

Figure 4.7 Gel strength after 10 minutes of rest for CEM I 52.5 R LA (w/c = 0.40) and CEM I 42.5 RR reference cement pastes (w/c = 0.50) for different dosages of calcium nitrate.


43

The heat of hydration curves are shown in Figure 4.8 and Figure 4.9. It can be seen that calcium nitrate speeded up hydration with approximately two hours for both cement types. The peak in the hydration curve for the pastes without calcium nitrate was seen at about 9 hours after water addition.CEM I 52.5 R LA - w/c = 0.40 - reference2.5

1.00 % CN 2 0.75 % CN

Rate of hydration heat (mW/g)

0.50 % CN 1.5

0.25 % CN 1 0.00 % CN

0.5

0 1 3 5 7 9 11 13 15 17 19 21 23 25

Time (hours)

Figure 4.8 Heat of hydration curves for CEM I 52.5 R LA cement pastes (w/c = 0.40) for different dosages of calcium nitrate.CEM I 42.5 RR - w/c = 0.50 - reference4

3.5

3

1.00 % CN 0.75 % CN

Rate of hydration heat (mW/g)

2.5

2

1.5 0.00 % CN 1 0.25 % CN 0.50 % CN 0.5

0 1 3 5 7 9 11 13 15 17 19

Time (hours)

Figure 4.9 Heat of hydration curves for CEM I 42.5 RR cement pastes (w/c = 0.50) for different dosages of calcium nitrate.

Chapter 4: Counteracting plasticizer retardation 4.2.3. Results and discussion for sodium lignosulphonate

44

Flow resistances, gel strengths after 10 seconds and 10 minutes of rest measured on CEM I 52.5 R LA cement pastes (w/c = 0.40) are listed in Table 4.4, Table 4.5 and Table 4.6, respectively. Those measured on CEM I 42.5 RR cement pastes (w/c = 0.50) are listed in Table 4.7, Table 4.8 and Table 4.9. Table 4.4 Flow resistance (Pa/s) for CEM I 52.5 R LA cement paste (w/c=0.40). Flow resistance [Pa/s] Reference 0.15% NLS 0.15% NLS DA 0.30% NLS 0.50% NLS Calcium nitrate [%] 0.25 0.50 0.75 2253 2515 2418 1973 1815 2060 618 727 839 651 819 1030 287 528 671

0.00 2283 1552 683 353 147

1.00 2372

1201 881

Table 4.5 Gel strength after 10 seconds of rest (Pa) for CEM I 52.5 R LA cement paste (w/c=0.40). 10 sec. gel strength [Pa] Reference 0.15% NLS 0.15% NLS DA 0.30% NLS 0.50% NLS 0.00 18.9 22.2 5.3 2.4

Documents

Combining Plasticizers Retarders and Accelerators