Combining Plasticizers/Retarders And Accelerators · high shear rate (i.e. relevant for mixing) and...

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



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 één 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



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 Contents

1 Introduction 1

2 Background on cement, cement hydration, rheology and admixtures 4

2.1 Cement .................................................................................................................. 4

2.2 Cement hydration .................................................................................................. 5

2.3 Rheology ............................................................................................................... 9

2.4 Plasticizers/retarders ............................................................................................. 13

2.5 Calcium nitrate ...................................................................................................... 22

3 Materials and apparatus 24

3.1 Materials................................................................................................................ 24

3.2 Apparatus .............................................................................................................. 27

4 Counteracting plasticizer retardation 34

4.1 Introduction ........................................................................................................... 34

4.2 Calorimetric and rheological measurements......................................................... 35

4.3 Mortar measurements............................................................................................ 75

4.4 General conclusion................................................................................................ 80

5 Long transport of fresh concrete 81

5.1 Introduction ........................................................................................................... 81

5.2 Sodium lignosulphonate........................................................................................ 81

5.3 Citric acid .............................................................................................................. 95

5.4 Lead nitrate............................................................................................................ 98

5.5 Sodium gluconate.................................................................................................. 101


6 Reutilizing residual fresh concrete 111

6.1 Introduction ........................................................................................................... 111

6.2 Phase I – Screening of retarders............................................................................ 111

6.3 Phase II – Determination of required retarder dosage .......................................... 114

6.4 Phase III – Activation using calcium nitrate ......................................................... 116

6.5 Phase IV – Strength measurements....................................................................... 120


7 Conclusions 127

1

Chapter 1

Introduction

This 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 Statoil’s 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

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.

4

Chapter 2

Background on cement, cement

hydration, rheology and admixtures

2.1 Cement

Cement 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, in

addition 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 ( SC ) 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, 2HSC ) 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

Chapter 2: Background 5

cement by adding corresponding amounts of individual minerals will therefore fail.

(Justnes, H., [6], p.10)

2.2. Cement hydration

In 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/C4AF

In 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

phase) (cubic phases) (hexagonal

H 15 AHC 2 AHC AHC H 27 A C 2 63194823 +→+→+

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 SC . Needle-shaped crystals of ettringite are formed as the

main hydration product:

3 2 6 3 32C A 3 CSH 26 H C AS H+ + →

Minor amounts of the monosulphate 124 HSAC or even 194AHC may also be formed

if an imbalance exists between the reactivity of C3A and the dissolution rate of

calcium sulphate, resulting in an insufficient supply of -24SO - 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.


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):

12433236 HSAC 3 H 4 A C 2 HSAC →++

As ettringite is gradually consumed, hexagonal calcium aluminate hydrate ( 194AHC )

also starts to form. It may be present in the form of a solid solution with 124 HSAC 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 hydrate

stabilised by incorporation of -24SO . It is formed on the surface of C3A and

becomes covered by ettringite.

ii) 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) -24SO - 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

sulphate is replaced by sodium sulphate.

iv) 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.


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 C3S

The 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.


The overall alite hydration reaction may ideally be written as

CH 3 HSC H 7 SC 2 4233 +→+

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 cement

The 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.

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

Dissolution Ettringite and CSH gel Formation

Induction Period Increase in Ca2+ and OH- Concentration

Initial Set

Final Set

Rapid Formationof CSH and CH

Diffusion-Controlled Reactions

Formation ofMonosulfate

Min Hours Days

Rate

of

Hea

t E

vo

luti

on

Time of Hydration


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 Rheology

2.3.1 General viscosity

In 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 (τ ) and is proportional to

the “velocity gradient” U/d (or “shear rate”, γɺ ). The constant of proportionality, η ,

is called the shear viscosity (also called “apparent” viscosity):

γ

τηɺ

=


The simplest rheological behaviour for liquids is the Newtonian viscous flow and

Hooke’s law for solid materials. Ideal viscous (or Newtonian) flow behaviour is

described using Newton’s law

γητ ɺ⋅=

Examples of ideal viscous materials are low molecular liquids such as water, solvents,

mineral oils, etc. and they are often called Newtonian liquids.

Hooke’s law states that the shear force acting on a solid is proportional to the resulting

deformation

γτ ⋅= G

where G is the “rigidity modulus”.

Many materials – especially those of colloidal nature – show a mechanic behaviour in

between these to border lines (Hooke’s an Newton’s 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:

p

p γµττ ɺ⋅+= 0

p = 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


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 resistance

Numerous 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 (τ ) as a function of

yield stress ( 0τ̂ ), plastic viscosity ( pµ ) and shear rate (γɺ ) as

0 pˆτ = τ + µ ⋅ γɺ

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.


Figure 2.3 Shear thickening behaviour resulting in negative yield stress values when

using the Bingham model.

The Hershel/Buckley equation p0 pˆτ = τ + µ ⋅ γɺ 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

µ ) for different mixes with different

p-factors. Negative yield stress values ( 0̂τ ) 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.

Figure 2.4 Flow resistance.

γɺ

τ

0τ̂

pµ

γɺ

τ

flow resistance


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:

A

F=τ [N/m2 or Pa]

h

v∆=∆γɺ [m/s.m or s-1]

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

Area under the curve V

vF

hA

vF

h

v

A

F ∆⋅=

⋅

∆⋅=

∆⋅

=∆⋅= γτ ɺ

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/m

3.s or W/m

3]. 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 the

electrical motor of the mixer, since P = U.I.

2.4 Plasticizers/retarders

2.4.1. Introduction

Water-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


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 so-

called 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-the-

Art’ by H. Justnes ([6]).

2.4.2. Common plasticizer types

There 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


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


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, poly-

aromatic 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


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 dispersion

There 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


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.


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.

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.

Macromolecular micelles

Cement surface


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.

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

cementparticles

cementparticles

Low molecular weight compound

cementparticles

cementparticles

polymer


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)


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 nitrate

This 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 bars

embedded 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.

Figure 2.17 Difference between set and hardening accelerators.

Reference

Setting

Hardening


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) an increased calcium concentration leads to a faster super-saturation of the

fluid with respect to calcium hydroxide, Ca(OH)2, while

ii) 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 β-C2S can not be excluded.

24

Chapter 3

Materials and apparatus

The purpose of this chapter is to introduce and describe the materials and the apparatus that

have been used frequently throughout this work.

3.1 Materials

3.1.1. 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 CEM I

52.5 R - LA

CEM I

42.5 RR*

Fineness:

Grains + 90 µm

Grains + 64 µm

Grains – 24 µm

Grains – 30 µm

Blaine (m2/kg)

1.7%

4.1%

66.3%

75.6%

359

0.1%

0.5%

89.2%

94.8%

546

Water demand 26.7% 32.0%

Le Chatelier 0.5 mm 0 mm

Initial set time 145 min. 115 min.

σσσσc (MPa) at 1 day

2 days

7 days

28 days

17.1

27.5

42.5

58.6

32.7

39.9

49.3

58.9

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

CEM I

52.5 R - LA

CEM I

42.5 RR*

Chemical

analyses

CaO

SiO2

Al2O3

Fe2O3

SO3

MgO

Free CaO

K2O

Na2O

Equiv. Na2O

Cr6+

(ppm)

Carbon

Chloride

LOI

Fly Ash

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

-

Minerals

by Bogue

C3S

C2S

C3A

C4AF

CS

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.


Table 3.3 Mineral composition (%) and alkali content of Portland cements obtained by

QXRD and plasmaemissionspectrometry

Cement

type

CEM I

52.5 R - LA

CEM I

42.5 RR

Alite

Belite

Ferrite

Cubic

aluminate

Orthorombic

aluminate

Lime

Periclase

Gypsum

Hemihydrate

Anhydrite

Calcite

Portlandite

Quartz

Arcanite

Mullite

Amporhous

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

-

-

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

-

-

Blaine 364 546

K (%)

Na (%)

Naeqv (%)

0.32

0.74

0.26

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)


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% +4NH , 64.68% -

3NO and 14.10% H2O. The CN was delivered in the form of

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. Apparatus

3.2.1. Mixer

The 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 ½ minute, resting for 5 minutes

and blending again for 1 minute.

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


3.2.2. Rheometer

Rheological 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 µm 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


approximately 10 µm (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, flowing-

off 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 (γɺ ) of 100 1/s to stir up the paste

2. 1 minute resting

3. Stress (τ ) – shear rate (γɺ ) curve with linear sweep of γɺ from 2 up to 200 1/s in 30

points lasting 6 s each (up curve)

4. Stress (τ ) – shear rate (γɺ ) curve with linear sweep of γɺ from 200 down to 2 1/s in 30

points lasting 6 s each (down curve)

5. 10 s resting

6. Shear rate (γɺ ) – stress (τ ) curve with logarithmic sweep of τ 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 (γɺ ) – stress (τ ) curve with logarithmic sweep of τ 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 (γɺ ) – stress (τ ) curves was stopped whenever the shear rate

(γɺ ) 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.


The 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%.

Shear rate

Time

mixing

½ minute

mixing

1 minute

5 minutes

rest

8 ½ minutes

1 minute

rest

10 seconds

rest

10 minutes

rest

1 minute

at 100 1/s

up

curve

down

curve

gel

strength

gel

strength

transfer to

rheometer

Figure 3.4 Flow chart of the mixing and measurement sequence


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)

Gel strength [Pa] Flow resistance

[Pa/s] 10 sec. 10 min.

391 2.4 14.2

383 2.4 13.0

394 2.8 9.2

419 2.8 10.0

PASTE 1

384 2.8 7.1

Average 394 2.7 10.7

Standard deviation 15 0.2 3

% standard dev. 4% 9% 27%

Table 3.5 Reproducibility of rheological measurements for cement paste 2

(w/c=0.50 – CEM I 42.5 RR – 0.50% Ultrazine Na)

Gel strength [Pa] Flow resistance

[Pa/s] 10 sec. 10 min.

2119 22.2 36.8

2375 22.2 36.8

2455 26.1 40.1

2343 22.2 40.1

PASTE 2

2392 22.2 36.8

Average 2337 2.7 38.1

Standard deviation 128 1.7 2

% standard dev. 5% 7% 5%

3.2.3. Calorimeter

An 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.


Figure 3.5 TAM Air Isothermal Calorimeter

3.2.4. Adsorption of plasticizers

To 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 µm filter paper on a Büchner 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


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.

34

Chapter 4

Counteracting plasticizer retardation

4.1 Introduction

Plasticizers 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.

Chapter 4: Counteracting plasticizer retardation 35

4.2 Calorimetric and rheological measurements

4.2.1. 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

Cement type Plasticizer Accelerator

CEM I 52.5 R LA Reference (0%)

(w/c = 0.40) 0.15% NLS*

0.15% NLS DA*

0.30% NLS

0.50% NLS

0.30% CLS

0.50% CLS

0.10% PA

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

0.00% CN

0.25% CN

0.50% CN

0.75% CN

1.00% CN

(* The 1.00% CN dosage was not studied for these series.)

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 ½ minute, resting for 5


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 µm filter

paper on a Büchner 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.


Calibration curve, NLS

y = 137.5844x

R2 = 0.9992

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007

% Added

Ab

so

rba

nc

e

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

Calibration curve, CLS

y = 137.2718x

R2 = 0.9997

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007

% Added

Ab

so

rba

nc

e

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:


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. Shear Rate [1/s] Shear Stress [Pa]

1 200 98.2

2 193 96.9

3 186 95.7

4 180 94.6

5 173 93.4

6 166 92.9

7 159 91.1

8 152 89.9

9 145 88.8

10 139 87.5

11 132 86.4

12 125 85.1

13 118 83.8

14 111 82.5

15 104 81.1

16 97.6 79.5

17 90.8 77.7

18 83.9 75.7

19 77.1 73.6

20 70.3 71.4

21 63.4 69.4

22 56.6 67.3

23 49.8 64.7

24 43.0 61.4

25 36.1 58.0

26 29.3 53.3

27 22.5 47.3

28 15.7 40.1

29 8.83 30.8

30 2.01 22.2


Down Curve

0

20

40

60

80

100

120

0 50 100 150 200

Shear Rate [1/s]

Sh

ea

r S

tre

ss

[P

a]

Figure 4.3 Down curve.

2. The 10 sec. gel strength can be derived from the shear rate (γɺ ) – stress (τ ) curve with

logarithmic sweep of τ 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 strength

0

20

40

60

80

100

120

140

160

180

0 20 40 60 80 100

Shear Stress [Pa]

Sh

ear

Rate

[1/s

]

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

gel strength


Table 4.3 Rheometer read outs to determine the 10 sec gel strength.

Meas. Pt. Shear Rate [1/s] Shear Stress [Pa]

1 0.00 1.00

2 0.00 1.17

3 0.00 1.37

4 0.00 1.61

5 0.00 1.89

6 0.00 2.21

7 0.00 2.59

8 0.00 3.04

9 0.00 3.56

10 0.00 4.18

11 0.00 4.89

12 0.00 5.74

13 0.00 6.72

14 0.00 7.88

15 0.00 9.24

16 0.00 10.8

17 0.00 12.7

18 0.00 14.9

19 0.00 17.4

20 3.44 20.4

21 9.32 24.0

22 12.9 28.1

23 16.0 32.9

24 21.3 38.6

25 26.1 45.2

26 35.0 53.0

27 48.7 62.1

28 73.6 72.8

29 110 85.3

30 155 100

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.


4.2.2. Results and discussion for reference pastes

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.

reference

0

500

1000

1500

2000

2500

3000

3500

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

Flo

w r

esis

tan

ce (

Pa/s

)

CEM I 52.5 R LA

CEM I 42.5 RR

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.


reference

0

5

10

15

20

25

30

35

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

10

se

c.

ge

l s

tre

ng

th (

Pa

)

CEM I 52.5 R LA

CEM I 42.5 RR

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.

reference

0

50

100

150

200

250

300

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

10 m

in. g

el str

en

gth

(P

a)

CEM I 52.5 R LA

CEM I 42.5 RR

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.


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 - reference

0

0.5

1

1.5

2

2.5

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

Time (hours)

Ra

te o

f h

yd

rati

on

hea

t (m

W/g

)

1.00 % CN

0.75 % CN

0.50 % CN

0.25 % CN

0.00 % CN

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 - reference

0

0.5

1

1.5

2

2.5

3

3.5

4

1 3 5 7 9 11 13 15 17 19

Time (hours)

Ra

te o

f h

yd

rati

on

hea

t (m

W/g

)

1.00 % CN

0.75 % CN

0.00 % CN

0.25 % CN

0.50 % CN

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.


4.2.3. Results and discussion for sodium lignosulphonate

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).

Calcium nitrate [%] Flow resistance

[Pa/s] 0.00 0.25 0.50 0.75 1.00

Reference 2283 2253 2515 2418 2372

0.15% NLS 1552 1973 1815 2060

0.15% NLS DA 683 618 727 839

0.30% NLS 353 651 819 1030 1201

0.50% NLS 147 287 528 671 881

Table 4.5 Gel strength after 10 seconds of rest (Pa) for CEM I 52.5 R LA cement paste

(w/c=0.40).

Calcium nitrate [%] 10 sec. gel

strength [Pa] 0.00 0.25 0.50 0.75 1.00

Reference 18.9 22.2 30.5 30.5 30.5

0.15% NLS 22.2 35.8 30.5 35.8

0.15% NLS DA 5.3 3.9 4.5 6.2

0.30% NLS 2.4 4.5 6.2 8.6 13.1

0.50% NLS < 1 3.3 6.2 7.3 10.0

Table 4.6 Gel strength after 10 minutes of rest (Pa) for CEM I 52.5 R LA cement paste

(w/c=0.40).

Calcium nitrate [%] 10 min. gel

strength [Pa] 0.00 0.25 0.50 0.75 1.00

Reference 73.6 104 114 161 271

0.15% NLS 52.1 87.6 95.6 35.6

0.15% NLS DA 20.1 20.1 26.0 30.9

0.30% NLS 7.1 15.5 30.9 52.1 67.5

0.50% NLS 3.9 11.9 36.8 47.7 73.6


Table 4.7 Flow resistance (Pa/s) for CEM I 42.5 RR cement paste (w/c=0.50).

Calcium nitrate [%] Flow resistance

[Pa/s] 0.00 0.25 0.50 0.75 1.00

Reference 2788 3161 2644 3099 3160

0.50% NLS 2138 2884 2614 2542 2364

1.00% NLS 231 425 416 492 581

Table 4.8 Gel strength after 10 seconds of rest (Pa) for CEM I 42.5 RR cement paste

(w/c=0.50).

Calcium nitrate [%] 10 sec. gel

strength [Pa] 0.00 0.25 0.50 0.75 1.00

Reference 22.2 30.5 22.2 26.1 26.1

0.50% NLS 22.2 30.5 26.1 30.5 26.1

1.00% NLS < 1 8.6 10.0 10.0 13.1

Table 4.9 Gel strength after 10 minutes of rest (Pa) for CEM I 42.5 RR cement paste

(w/c=0.50).

Calcium nitrate [%] 10 min. gel

strength [Pa] 0.00 0.25 0.50 0.75 1.00

Reference 67.5 104 80.3 104 95.6

0.50% NLS 33.7 52.1 52.1 56.8 61.9

1.00% NLS 7.7 14.2 16.9 5.9 33.7

The flow resistances for CEM I 52.5 R LA cement pastes are also shown in Figure 4.10. In

case of the reference no significant influence of the addition of calcium nitrate on the flow

resistance could be measured. When sodium lignosulphonate (NLS) was added, however,

calcium nitrate had a clear increasing effect on the flow resistance as can be seen in Figure

4.10. The values found for the flow resistance are nevertheless still far below those of the

reference. From Figure 4.11, which shows the increase in flow resistance relative to the flow

resistance of the respective reference without calcium nitrate, it can be seen that the increasing

effect of calcium nitrate on the flow resistance became more pronounced when higher dosages

of sodium lignosulphonate were used. An interesting observation for the flow resistance was

that simply delayed addition of 0.15% sodium lignosulphonate makes it in excess of 50%

more effective as plasticizer than when it is added with the mixing water. This effect is

attributed to less intercalation of lignosulphonate in the early hydration products of cement,

leaving more lignosulphonate available to function as plasticizer through physical absorption

on the grain surface.


CEM I 52.5 R LA - w/c = 0.40

0

500

1000

1500

2000

2500

3000

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

Flo

w r

esis

tan

ce (

Pa/s

)

Reference

0.15% NLS

0.15% NLS DA

0.30% NLS

0.50% NLS

Figure 4.10 Flow resistance for CEM I 52.5 R LA cement pastes (w/c = 0.40) for different

dosages of calcium nitrate.

CEM I 52.5 R LA - w/c = 0.40

0

100

200

300

400

500

600

700

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

Flo

w r

esis

tan

ce (

%)

Reference

0.15% NLS

0.15% NLS DA

0.30% NLS

0.50% NLS

Figure 4.11 Increase in flow resistance relative to the flow resistance of a reference without

calcium nitrate for CEM I 52.5 R LA cement pastes (w/c = 0.40) for different dosages of

calcium nitrate.


Figure 4.12 shows the flow resistances for CEM I 42.5 RR cement pastes. Neither clear

increasing nor decreasing effect of calcium nitrate on the flow resistance could be denoted in

case of the reference or in case of the pastes prepared with 0.50% sodium lignosulphonate.

The pastes prepared with 1.00% sodium lignosulphonate, however, again show the increasing

trend also observed for the CEM I 52.5 R LA pastes.

When comparing CEM I 42.5 RR cement paste (w/c = 0.50) and CEM I 52.5 R LA paste

(w/c = 0.40), one can see that higher dosages of plasticizer were required to achieve

comparable reductions in flow resistance in spite of the higher water-to-cement ratio. The

tendency of increasing flow resistance with increasing calcium nitrate dosage is less

pronounced for CEM I 42.5 RR than for CEM I 52.5 R LA paste. This may be associated with

calcium nitrate being a less effective accelerator for this cement compared to the other

according to the mineralogy: CEM I 42.5 RR cement has a lower belite and a higher alkali

content than CEM I 52.5 LA cement (see section 2.5 and Table 3.2).

As only a limited number of plasticizer concentrations were studied in case of CEM I 42.5 RR

and as the effect of 0.50% sodium lignosulphonate on the flow resistance was rather small, no

noteworthy conclusions can be drawn from Figure 4.13, which shows the increase in flow

resistance relative to the flow resistance of the respective reference without calcium nitrate.

CEM I 42.5 RR - w/c = 0.50

0

500

1000

1500

2000

2500

3000

3500

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

Flo

w r

es

ista

nc

e (

Pa

/s)

Reference

0.50% NLS

1.00% NLS

Figure 4.12 Flow resistance for CEM I 42.5 RR cement pastes (w/c = 0.50) for different

dosages of calcium nitrate.


CEM I 42.5 RR - w/c = 0.50

0

50

100

150

200

250

300

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

Flo

w r

es

ista

nc

e (

%)

Reference

0.50% NLS

1.00% NLS

Figure 4.13 Increase in flow resistance relative to the flow resistance of a reference without

calcium nitrate for CEM I 42.5 RR cement pastes (w/c = 0.50) for different dosages of

calcium nitrate.

Figure 4.14 shows the gel strengths after 10 seconds of rest for CEM I 52.5 R LA cement

pastes. An increasing effect of calcium nitrate on the gel strength can be seen. This increasing

effect on the gelling tendency may be beneficial in some cases since tendencies to segregation

will be reduced.


CEM I 52.5 R LA - w/c = 0.40

0

5

10

15

20

25

30

35

40

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

10 s

ec. g

el str

en

gth

(P

a)

Reference

0.15% NLS

0.15% NLS DA

0.30% NLS

0.50% NLS

Figure 4.14 Gel strength after 10 seconds of rest for CEM I 52.5 R LA cement pastes (w/c =

0.40) for different dosages of calcium nitrate.

The gel strengths after 10 seconds of rest for CEM I 42.5 RR cement pastes are depicted in

Figure 4.15. Only in case of the pastes prepared with 1.00% sodium lignosulphonate a clear,

increasing, trend can be seen.

Figure 4.16 and Figure 4.17 show the gel strengths after 10 minutes of rest for

CEM I 52.5 R LA and CEM I 42.5 RR cement pastes, respectively. For all mixtures an

increasing effect of calcium nitrate on the 10 minutes gel strength was measured.


CEM I 42.5 RR - w/c = 0.50

0

5

10

15

20

25

30

35

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

10

se

c.

ge

l s

tre

ng

th (

Pa

)

Reference

0.50% NLS

1.00% NLS

Figure 4.15 Gel strength after 10 seconds of rest for CEM I 42.5 RR cement pastes (w/c =

0.50) for different dosages of calcium nitrate.

CEM I 52.5 R LA - w/c = 0.40

0

50

100

150

200

250

300

0.00 0.25 0.50 0.75 1.00

Calcium nitrate (%)

10 m

in. g

el str

en

gth

(P

a)

Reference

0.15% NLS

0.15% NLS DA

0.30% NLS

0.50% NLS

Figu

Combining Plasticizers/Retarders And Accelerators · high shear rate (i.e. relevant for mixing) and...

Documents

Transcript of Combining Plasticizers/Retarders And Accelerators · high shear rate (i.e. relevant for mixing) and...