The Working Mechanism of Starch and Diutan Gum in Cementitious ...

52903-1 Applied RheologyVolume 23 · Issue 5

The Working Mechanism of Starch and Diutan Gum inCementitious and Limestone Dispersions in Presence of

Polycarboxylate Ether Superplasticizers

Wolfram Schmidt*1, H.J.H. Brouwers2,Hans-Carsten Kühne1, Birgit Meng1

1 Division 7.4 Technology of Construction Materials, BAM Federal Institute for MaterialsResearch and Testing, Unter den Eichen 87, 12205 Berlin, Germany

2 Eindhoven University of Technology, Eindhoven, Netherlands

* Corresponding author: [email protected]: x49.30.81041717

Received: 29.3.2013, Final version: 25.6.2013

© Appl. Rheol. 23 (2013) 52903 DOI: 10.3933/ApplRheol-23-52903

Abstract:Polysaccharides provide high potential to be used as rheology modifying admixtures in mineral binder systemsfor the construction industry such as concrete or mortar. Since superplasticizers have become state of technol-ogy, today, concrete is more and more adjusted to flowable consistencies. This often goes along with the risk ofsegregation, which can be effectively avoided by adding stabilising agents supplementary to superplasticizers.Stabilising agents are typically based on polysaccharides such as cellulose, sphingan gum, or starch. Starch clear-ly distinguishes in its effect on rheology from other polysaccharides, mainly due to the strong influence of amy-lopectin on the dispersion and stabilisation of particles. Based on rheometric investigations on cementitious andlimestone based dispersions with different volumetric water to solid ratios, the mode of operation of modifiedpotato starch is explained in comparison to a sphingan gum. It is shown that the stabilising effect of starch ina coarsely dispersed system is mainly depending upon the water to solid ratio and that above a certain particlevolume threshold starch mainly affects the dynamic yield stress of dispersions, while plastic viscosity is affect-ed only to a minor degree. Sphingans operate more independent of the particle volume in a coarsely dispersedsystem and show significantly higher effect on the plastic viscosity than on the yield stress. In systems incor-porating superplasticizers, influences of both stabilising agents on yield stress retreat into the background, whileboth observed polysaccharides maintain their effect on the plastic viscosity.

Zusammenfassung:Polysaccharide weisen als Rheologiemodifizierer für in der Baustoffindustrie verwendete mineralische Binde-mittelsysteme wie Beton oder Mörtel ein hohes Anwendungspotential auf. Seit Fließmittel gängige Zusatzmit-tel in der Baupraxis geworden sind, kann Beton heutzutage verstärkt hinsichtlich seiner Fließfähigkeit spezifi-ziert werden. Erhöhte Fließfähigkeit geht allerdings häufig mit einer erhöhten Entmischungsneigung einher,der durch zusätzliche Zugabe von stabilisierenden Zusatzmitteln effektiv entgegengewirkt werden kann. Sta-bilisierer haben üblicherweise Polysaccharide als Grundstoff, z. B. Zellulose, Sphingan oder Stärke. Hierbei unter-scheidet sich Stärke in ihrem Einfluss auf die Rheologie deutlich von anderen Polysacchariden, was durch denstarken Einfluss ihres Amylopektinmoleküls auf die Dispersion und Stabilisierung von Partikeln begründet wer-den kann. Anhand rheometrischer Untersuchungen an Zement- und Kalksteinmehlsuspensionen mit unter-schiedlichen volumetrischen Wasser-Feststoffverhältnissen werden die unterschiedlichen Wirkungsweisen vonmodifizierter Stärke im Vergleich zu Sphingan erklärt. Es kann gezeigt werden, dass der stabilisierende Effektder Stärke in grobdispersen Systemen im Wesentlichen vom Wasser-Feststoffverhältnis abhängt und dass ober-halb eines Partikelvolumen grenzwertes vor allem die dynamische Fließgrenze und weniger stark die plastischeViskosität beeinflusst. Der Wirkungs mechanismus von Sphinganen in grobdispersen Systemen ist deutlich weni-ger abhängig vom Partikelvolumen. Hier kann ein deutlich größerer Einfluss auf die plastische Viskosität aus-gemacht werden. In Systemen, die Fließmittel enthalten, treten Einflüsse der Stabilisierer auf die Fließgrenze inden Hintergrund, während ein deutlicher Effekt hinsichtlich einer höheren plastischen Viskosität ausgemachtwerden kann.

Résumé:Les polysaccharides présentent un grand potentiel dans des applications de mixtures modifiant les propriétésrhéologiques de liants minéraux pour la construction civile comme le béton ou le mortier. Puisque les super plas-tifiants sont devenus une technologie à la mode, de nos jours la consistance du béton est de plus en plus ajus-tée afin de présenter des caractéristiques de fluidité. Celles-ci s’accompagnent souvent d’un risque de ségréga-tion qui peut être évitée de manière effective en additionnant des agents stabilisant en plus des super plastifiants.

1 INTRODUCTION

1.1 IMPORTANCE OF POLYSACCHARIDES INCONCRETE TECHNOLOGYSuperplasticizers can be considered as the key tomost significant innovations in cement and con-crete technology during the last decades sincethey allow optimising the rheology withoutchanging the water content, which is relevant forthe strength and durability of concrete. Innova-tions such as Self-Compacting Concrete (SCC),Ultra-High Performance Concrete, or ReactivePowder Concrete were only achievable due to theintensive use of superplasticizers (SPs). Today, forhigh performance applications, polycarboxylateethers (PCE) are typically used but often theyneed to be supplemented with stabilising admix-tures (STA) in order to avoid uncontrolled segre-gation. A major problem of flowable concrete issegregation, fostered by the wide range of diam-eters of mortar and concrete constituents rang-ing in size from nm to cm. In order to avoiddynamic segregation upon flow, flowable con-crete types typically contain higher contents offines than normal concrete. However, this can-not effectively avoid segregation at rest, which isparticular a problem induced by the retardingeffect of superplasticizers. Furthermore the ma -terial costs as well as the shrinkage tendency areincreased. Therefore, concepts with lower pow-der contents, incorporating PCE superplasticizersand supplementary rheology modifying admix-tures are becoming more and more popular andimportant.

According to a milestone paper by Khayat [1]for mortar and concrete applications, commonlypolysaccharides are in use as polymeric STAs. Thevariety of these products is huge. Often cellulose,gums from plants or microbes as well as starchesare used as basic components. Typically, the modeof operation of STAs in mortar systems is explainedby their capability to be adsorbed on water.Increasing molecular weights yield higher waterretention properties [2, 3]. The cementitious poresolution at fresh state exhibits pH-values higherthan 13. It is therefore typically as sumed that mostpolysaccharides may be incompatible with thehydration of cement due to degrading in the highalkaline surrounding of the cement paste or loss ofeffectiveness due to shrinkage in presence of met-al ions [2, 4, 5]. Izumi [4] investigated a number ofpolysaccharides and showed that from one specif-ic polysaccharide, the majority of STAs exhibitdecreasing viscous behavior in solution withincreasing ion concentration. Pourchez foundadverse results with regards to the stability of dif-ferent cellulose derivatives, which showed highstability in alkaline surrounding [6]. In general, thethreat of a high ionic solution is well known to theproducers of commercial STAs, so that the avail-able products on the market can be considered assufficiently stable

1.2 STARCH ETHER AND SPHINGAN GUMSTABILISING AGENTSStarch is a polysaccharide, which typically con-sists of two types of macromolecules composedof differently linked glucose monomers. Though,

52903-2Applied RheologyVolume 23 · Issue 5

Les agents stabilisant sont typiquement basés sur des polysaccharides tels que la cellulose, la gomme de sphin-gan, ou l’amidon. L’amidon se distingue clairement des autres polysaccharides par son effet sur la rhéologie quiest principalement du à une forte influence de l’amylopectine sur la dispersion et la stabilisation des particules.Sur la base d’études rhéologiques de suspensions calcaires et de ciment possédant différents ratios volumé-triques d’eau et de solide, le mode d’opération de l’amidon de pomme de terre modifié est expliqué et compa-ré à celui de la gomme de sphingan. On montre que l’effet stabilisant de l’amidon sur une dispersion de grossesparticules dépend principalement du ratio eau/solide et qu’au-delà d’un certain seuil de volume de particule,l’amidon affecte principalement la contrainte seuil dynamique des dispersions, tandis que la viscosité plastiquereste moins affectée. L’effet de la gomme de sphingam dépend moins du volume en particules et présente unplus grand effet sur la viscosité plastique que sur la contrainte seuil. Dans les systèmes incorporant des superplastifiants, les influences des deux agents stabilisants ont peu d’impact sur la contrainte seuil, tandis que lesdeux agents maintiennent leurs effets sur la viscosité plastique.

Key words: cement, limestone, rheology, stabilising agent, coarsely dispersed systems, diutan gum, starch ether

Potato starch Diutan gum Amylose Amylopectin

Content ~ 20%/wt ~ 80%/wt -Molecular mass [u] 50000 - 10000000 - 2900000 - 500000 100000000 5200000Radius of gyration [nm] ~30 50 - 500 N/A

sodium mono chloracetate reactions. Such sta-bilisation is typically made for the reason toreduce the tendency for retrogradation and tominimise intermolecular interactions [9].

Diutan gum is a microbial polysaccharide.The backbone consists of the repeated configu-ration of rhamnose, glucose, glucoronic acid, glu-cose units [10]. A carboxylate group attached tothe glucuronate gives anionic charges to thebackbone of diutan gum [5]. The side chains ofdiutan gum, which consist of two linked rham-nose units, are attached to each second glucoseunit [5, 10]. These are considered to stericallyshield the carboxylic acids, thus avoiding cross-linking by calcium ions. The structure of the poly-mer is given in Figure 1 in comparison to starch.Diutan belongs to the group of sphingans, whichall exhibit the same backbone structure. Howev-er, only welan and diutan are considered to becompatible with the cement hydration since theyare stable in high pH systems. They distinguish intheir main chain length and side chain geometry[5, 11]. Since only few reports about diutan gum incementitious systems are available, the reviewpart discusses experiences with welan gum aswell. Due to their structural similarity experi-ences with welan can be qualitatively assigned todiutan as well.Based on the data provided by Swinkels [12], themolecular weight of starch is in the range of 2 to2.5 million g/mol mainly caused by the amy-lopectin molecules with a degree of polymeri -sation of 2,000,000. The average molecularweight of commercial diutan lies between 2.88and 5.18 million Da [5, 10]. Starch ethers and diu-tan gum distinguish significantly in their ap -pear ance (Figure 1). Diutan gum has a linearmain chain with regular side chains. Starch con-sists of the two aforementioned differently sizedmacromolecules. Diutan incorporates anioniccharges, while typical starches for cementitioussystems can be considered as non-ionic (butthere are also carboxymethyl-starches on themarket). As a result of these differences, it canbe assumed that the modes of operation of theSTAs differ greatly.

Khayat attributed effects of sphingans tothe binding of water, which in return increasesthe viscosity of the cementitious system [1].Despite the chemical similarity between starchand cellulose, in cementitious systems, the per-formance of cellulose compares much more to


chemically identical to cellulose, the starchstrands as well as the macromolecular structurediffer in a pronounced way. The glucose units ofcellulose are arranged in alternating order, whilethey are arranged identically for starch. Further-more, cellulose strands are mostly linear, whilestarch consists of two types of macromolecules,amylose and amylopectin. Amylose is a largelylinear polymer with regularly repeated hydrox-ide bonds as shown in Figure 1. For the more com-plex macromolecule amylopectin, the majorityof bonds are constituted by hydroxide bonds aswell. However, hydroxymethyl bonds are report-ed every 12 to 17 glucose units [7] or every 15–30glucose units [8]. As a result, a tree like structureas shown in Figure 1 forms in solution. Amylosemolecules occur 200 to 1000 times more oftenthan amylopectin but amylopectin has anapproximately 1000 times higher degree of poly-merisation than amylose, thus the radius of gyra-tion of amylopectin is about 10 times higher thanthat of amylose (see Table 1). For typical starches(apart from waxy maize starch, which has 100%amylopectin), about 80 % by weight are consti-tuted by amylopectin. For construction materialsstarches have to be made cold-water soluble. Fur-thermore, for application in the high pH-envi-ronments of cementitious systems, starchesneed to be stabilised by ether or ester bond in thehydroxyl groups. Typical modifications are con-ducted by ethylene oxide, propylene oxide, or

Figure 1:Chemical and macromolec-ular structures of starch anddiutan gum.

Table 1:Typical technical specifica-tions of starch and diutangum based on [5, 8-12, 32].

the performance of sphingans. As will be dis-cussed later, this is likely to be linked to the lin-ear structure of the cellulose as well as to the factthat often cellulose derivatives used in construc-tion materials are amended by carboxylic groups.For sphingan and cellulose, Khayat distinguishedbetween three modes of operation, which occurin coarsely dispersed cementitious systems de -pen ding upon the concentration [13]:

� Adsorption: Polymers adhere to and immo-bilise water.

� Association: Molecules form a network, caus-ing gellation, which blocks the water motion.

� Intertwining: At low shear rates, polymerchains intertwine and entangle. This effect isbasically limited to high addition amounts.

Sonebi observed marked shear thinning behav-ior in mixes incorporating SP and sphingan gums[14, 15]. This effect was more pronounced withdiutan gum than with welan gum, which wasattributed to the higher molecular mass. Thestudy shows that diutan gum significantlyincreases the apparent viscosity particularly atlow shear rates. It is assumed that the long chainsof diutan entangle and intertwine at low shearforces. At higher shear rates, the polymers directinto the flow direction, which lowers the appar-ent viscosity again. Rajayogan [16] and Terpstra[17] published studies that show the suitability ofstarch ether as a STA for cementitious materials.According to Simonides and Terpstra [18], the sta-bilising mechanism of starch ethers differs great-ly from diutan gum. As shown in Figure 2, themode of operation is basically attributed to theeffect of the amylopectin, which spreads out insolution into its tree-like structure, thus keepingparticles in distance, to avoid segregation. Theeffect therefore mainly affects the yield stress atrest and does not show strong influence on theplastic viscosity. Based on this, it is assumable,that differing from cellulose or diutan gum, themolecular weight of starch may play a minor rolefor the stabilising effect and that the ratiobetween amylose and amylopectin may be of sig-nificantly higher importance.

The former observations, however, clearlycontradict to observations of Rajayogan et al.,who observed a large increase of the plastic vis-cosity due to starch [16]. A major differencebetween the mixtures that were observed in

these studies were the amounts of starch in themixture compositions, the water to powderratios as well as the binder to aggregate ratios.Simonides and Terpstra were using self-com-pacting concrete (SCC) with a powder content of400 kg/m³ and a water to powder ratio of 0.47.The starch dosage was 0.06 % of the powder.Rajayogan et al. were using much higher pow-der contents between 550 and 650 kg/m³ withwater to powder ratios ranging between 0.34and 0.40. The starch dosages ranged between0.11 % and 0.23 % of the powder. Hence, the sys-tems ob served by Rajayogan et al. had signifi-cantly higher particle packing and higher starchcontents. It can be assumed that the effectdemonstrated in Figure 2 is the major effect ofstabilisation in relatively fluent systems asobserved by Simonides and Terpstra. At higherpowder densities and lower water volumes andhigher starch contents, the space filling effect ofstarch might be outweighed by intertwining andassociation of polymers. It is, however, ques-tionable, if stabilising agents are required at allin high powder systems.In a former comparison between water-STAsystems and water-cement-STA systems underthe assumption of Bingham behavior, it wasshown by Schmidt et al. that diutan generated asignificant yield stress in water, while starch onlyshowed relatively negligible effects. Upon addi-tion of particles, however, the starch stronglyincreased the yield stress [19]. It can therefore beconcluded that both admixtures work on differ-ent modes of operations. Diutan gum mainlyaffects both, viscosity and yield stress throughbinding high amounts of water. The thus in -creased apparent viscosity of the fluent phasefinally affects the rheology of the overall system.In contrast to that, starch mainly affects the yieldstress in presence of particles and much lesser theviscosity. It does not bind high amounts of waterbut can affect the rheology upon addition of par-ticles significantly. Hence, starch and diutan inmortar like systems can be distinguished be -tween (a) stabilising of particles (starch) and (b)stabilising of the fluent phase (diutan gum).


Figure 2:Stabilising mechanism ofstarch ether and effect onyield stress after Simonidesand Terpstra [18].

1.3 POLYSACCHARIDES IN PRESENCE OF PCEModern flowable concrete and mortar types con-tain high amounts of SPs, typically PCE. PCEs arecomb polymers consisting of a polycarboxylicbackbone and polyethylene oxide graft chains.Their dispersing mechanism is based on theiradsorption on positively charged surfaces and asteric hindrance of the particles upon adsorption.Since a cement grain consists of multiple phasesexposed to the pore solution exhibiting differentzeta potentials in its environment, PCEs adsorbpreferably on the aluminous and ferrous phases ofthe cement clinker. Furthermore, they adsorbpreferably on newly formed hydration phasesupon addition of water such as ettringite andmonosulfate, which occur by reaction of water,gypsum and calcium aluminate phases. Theireffect on yield stress is very pronounced, while theyonly show little influence on the plastic viscosity.

Some polysaccharides such as cellulose, we -lan and diutan gum are known to provide anion-ic charges, which makes them adsorb on particles[1, 5, 20, 21]. It is reported that the adsorption ofpresent SPs avoids the adsorption of diutan gum[5], however, there are also results that suggestthat adsorption still takes place to certain amountin presence of superplasticizers [14]. The anioniccharges generally suggest that interactionsbetween the polymers can take place or thatadsorption occurs competitively between bothtypes of polymers. Khayat showed that smallchanges of naphthalene based SP dosages hadlarge effects on the apparent viscosity in welanstabilised systems [1]. With increasing content ofWelan gum, the robustness against variations inthe SP dosage could be significantly increased. Inreturn, higher amounts of SP were required toachieve similar reduction of the apparent viscos-ity. Yammamuro et al. investigated the interac-tions of SPs with adsorptive and non adsorptiveSTAs [3]. The non adsorptive STA did not at allaffect the adsorption of SPs, while increasingamounts of adsorptive STA reduced the adsorp-tion of SPs significantly. The authors concludedthat competitive adsorption takes place between

SPs and adsorptive STAs. Results by Phyfferoen etal. shown that indeed diutan gum STA shows atendency to adsorb on cement particles. Thiseffect, however, shall be eliminated in presenceof SPs [5, 22]. Interactions are assumed to be hin-dered by steric shielding of the anionic charge bythe double rhamnose side chain. Also starch isreported to be found ad sorbed on cementitiousparticles, however, to significantly lesser extent.By adding PCE, the adsorption could be avoided[23], and a strong influence of starch on the yieldstress in presence of PCE could be observed. As apossible explanation for this effect, the authorssuggest depletion forces due to unadsorbed poly-mers.

PCEs and STAs distinguish strongly from eachother. PCEs may have molecular weights between10,000 and 200,000 g/mol and typically gyratoryradii between 5 and 150 nm [24 – 28], while for dif-ferent polysaccharide STAs for application incementitious systems molecular weights be tween300,000 and 5,000,000 g/mol and gyratory radiibetween several tenth and 500 nm are reported [4,5, 12, 20]. Due to the enormous variety of productsof SPs and STAs, it is difficult to detect generallyvalid laws regarding possible interactions be -tween these polymers. However, a good under-standing is important, since an increasing trend touse both types of admixtures in parallel can beobserved on the side of the concrete engineers aswell as of blenders of admixtures for the market.

1.4 MOTIVATION OF THE INVESTIGATIONSThe use of polysaccharides can effectively avoidsegregation and increase the robustness againstunavoidable scatter in the constituents’ quali-ties. The performance of polysaccharides in mor-tar can depend upon the water to solid ratio. Fur-thermore, interactions between su per plas ti -cizers and STAs are possible and understandingthe combined effects becomes increasinglyimportant for the concrete industry. However,interactions between STAs and PCEs have notbeen matter of intensive research yet. Based onrheometric experiments, the characteristic dif-ferences between non-adsorptive and adsorp-tive stabilising agents based on starch and diu-tan gum shall be discussed without and in pres-ence of superplasticizers and in systems with var-ied water to solid ratios.


Name Commercial name Description Form Bulk density pH

PCE Glenium Sky 591 Polycarboxylic backbone 30% solids 1.07 ± 0.02 g/cm² 6.5 ± 1 with polyethylene grafts in water ST1 Foxcrete S 100-F Starch ether Powder 400 kg/m³ 11 - 12 (100 g/l H2O)ST2 Kelkocrete DG-F Diutan gum Powder n/a n/a

Table 2:Technical specificationsaccording to the producerdata sheets of the PCE andthe STAs used for the inves-tigations.

2 MATERIALS AND METHODS

2.1 ADMIXTURESThe STAs that were used are commercially avail-able polysaccharides. The STA abbreviated ST1 isa starch derivative and has been subject ofresearch in the field of SCC in the past years [17,18, 29]. ST2 is a diutan gum and has also beensubject of research in the field of cement andconcrete technology [5, 14, 15, 30]. Commercialpure polycarboxylate ether with a solid contentof 30 % was used as superplasticizer. Technicalspecifications of the admixtures are givenTable 2.

2.2 MATERIALS AND MIXESInvestigations were conducted on water-lime-stone systems and water-cement systems withdifferent water to solid ratios. The powders’ par-ticle size distributions and oxide compositionscan be found in Figure 3 and Table 3, respectively.Figure 3 also shows results of slump flow testsof the used cement and limestone when sup-plemented by water at their respective waterdemand according to the Puntke method [31] atincreasing PCE solid additions. The flow induc-tion of LSF happens at significantly lower dos -age and the maximum spread flow diameteroccurs at lower dosage and is wider than withcement, which indicates that significantly high-er amounts of PCE can be adsorbed on cementand that the dispersing forces of PCE are lowerin cementitious systems than in LSF-systems.Nevertheless, the basic mechanism of disper-sion can be observed for both powders, so thatdepending on the observation type, limestonefiller systems can possibly replace cementitioussystems without the negative effect of the rapidchange of rheology due to the ongoing cementhydration.

The STAs were observed in water only, andin water to powder systems at medium and highpowder content. For all investigations, STA wasfirst dissolved in water. Due to the differentwater demands of LSF and cement, their volumeswere varied between 0, 33.3 % and 50.0 %, and 0,25.0 % and 40.0 %, respectively. Furthermore thecombined influence of PCE and STA was studiedfor each powder at the highest observed powderconcentration. In these investigations, the PCEsolid contents were varied for LSF between


Water [vol %] + ST* 100 66.7 50.0 50.0 50.0LSF [vol %] - 33.3 50.0 50.0 50.0PCE solids [% bwo LSF] - - - 0.06 0.30

Water [vol %] + ST* 100 66.7 50.0 50.0 50.00 50.00Limestone filler [vol %] - 25.0 40.0 40.0 40.0 40.0PCE solids [% bwo cement] - - - 0.06 0.30 1.20

0.06 % and 0.30 %, which approximately are thedosages required to induce flow and to achievemaximum spread flow according to Figure 3(right). For cement, the same dosages were cho-sen and supplemented by a dosage of 1.20 %.Here the first dosage represents a very low PCEdosage, and the latter dosages represent theonset and the maximum. In the LSF investiga-tions, the STA content was fixed at 0.5 % of thewater. However, due to the significantly higherefficiency of ST2, for the cement tests more di -verse dosages were chosen. ST1 was varied be -tween 0.5 % and 5 %, ST2 was varied between0.05 % and 0.5 %, each percentage related to thewater content. In flowable cementitious systems0.5 % for ST1 and 0.05 % for ST2 represent typicaldosages for use as stabilising agent, while sig-nificantly higher dosages would only be used forspecial applications. The mixture proportioningfor the LSF tests and the cement tests is given inTable 4 and Table 5, respectively.

Oxides Cement Limestone (CEM I 42.5 R)1 filler2

CaO 62.80% 90.68%SiO2 20.56% 1.47%Al2O3 4.36% 0.46%Fe2O3 2.27% 0.40%MgO 2.14% 0.61%Mn2O3 0.03% -Na2O 0.28% 3.27%K2O 0.95% 0.54%Ti2O 0.20% 0.05%P2O5 - 2.19%SO3 3.45% 0.34%

Figure 3:Grading of the CEM I andlimestone filler used for theinvestigations and slumpflow diameters of cementand limestone filler ofpastes with Punkte-waterdemand [31] at varying PCEadditions.

Table 3:Oxide compositions of theused cement and limestonefiller( 1based on wet chemistryaccording to DIN EN 196-2,2determined with XRF).

Table 4:Conducted tests with STAsand LSF at varied powdercontents(* ST1: 0.5 % by weight ofwater, ST2: 0.05 % byweight of water).

Table 5:Conducted tests with STAsand cement at varied pow-der contents(* ST1: 0.5 % by weight ofwater, ST1: 5.0 % by weightof water, ST2: 0.05 % byweight of water, ST2: 0.5 %by weight of water).

2.3 RHEOMETRIC EQUIPMENT AND MEASURE-MENTSA Couette type viscosimeter was used with a dou-ble gap cell as shown in Figure 4. A network struc-tured grid as shear body induces a flow that islargely based on the cohesion between the fluidlayer on the wall and the fluid. This minimiseswall slip effects, which typically limit the use ofstandard geometries for coarsely dispersed sys-tems, and allows measurements up to a maxi-mum grain size of 1 – 2 mm. A ramp profile wasestablished as shown in Figure 5 and the valuesat decreasing shear rates were used to calculateyield stress τ0 and plastic viscosity hpl based onthe assumption that the fluids largely show Bing-ham behavior. A Bingham interpretation isassumed to not truly render the performance ofSTAs, particularly at low shear rates and in thewater-STA systems only. However, the strengthof the Bingham model is the clear differentiationof the yield stresses, and the effect on yield stress

is assumed to be one of the major distinctionitem between ST1 and ST2. Furthermore, theBingham approach is the most commonly usedmodel in cement and concrete technology andthe observed systems showed Bingham behav-ior over a wide range of shear rates.

Since shear forces strongly affect the cementhydration and the adsorption of polymers, forcementitious systems it is of utmost importanceto keep the measurement time as short as possi-ble. The applied profile is considered to be a rea-sonable compromise between precision andcompactness. Due to the initial formation ofettringite and monosulfate, the performance ofPCE in cementitious systems can change rapidlyduring the first 3 – 5 minutes. In order to makesure that the cementitious systems were stablethe measurements were conducted not earlierthan 10° minutes after water addition.

3 RESULTS AND DISCUSSIONFigure 6 shows yield stresses and plastic viscosi-ties for ST1 and ST2 for different water to solidratios compared to identical water-limestonesuspensions without stabilising admixture. Itcan be clearly observed that ST1 does not strong-ly affect τ0 compared to a mixture without STA,when added to water only and to a mixture withonly 33.3 % solids. At the same time, with ST2 inwater only a clear yield stress can be observed,which increases only slightly at low solid content.For mixtures with high solid content, ST1 increas-es τ0 strongly, while the effect of higher powdercontents is only small for ST2. Despite the strongeffect of ST1 on τ0 upon addition of high solid con-tents, hpl is only slightly affected. A pronouncedincrease of hpl can be observed at high solid con-tents with ST2, although at the same time theincrease of τ0 is small.

Regarding the yield stress, at 0 % solids, thestarch system does not distinguish from the con-trol system with water, and starch only increaseshpl slightly. At 50 % of solids, τ0 increases promi-nently, while the increase of hpl is relatively low.This clearly indicates that for ST1 a threshold forthe solid particle amount exists, which triggersthe rapid increase of τ0, which cannot solely beexplained by the higher particle volume, sincethen hpl would exhibit at least a similarly strongincrease. This is clearly not the case in Figure 6.Since the polymer concentration in water is main-


Figure 4 (above):Wide gap basket cell, asused for the rheometricinvestigations (sketch bySchleibinger).

Figure 5:Measurement for the obser-vation of the influence ofsolid particles with samplemeasurement results (left)and sample flow curves andBingham approximations(right).

tained for all tests, intertwining of polymers doesnot seem to be the driving force for the signifi-cant yield stress increase. Without doubt theeffect of depletion forces as described in [23]induced by non-adsorbed starch molecules can-not be neglected in the discussion. However, dueto the large molecule size of the amylopectin,depletion forces may be mainly induced by the amultiple smaller amylose. Another mechanismthat can explain the observed effect would berather a particle-amylopectin-particle latticeeffect, which is activated by the closer particle dis-tance of LSF, when added above a thresholddosage. The cause is similar to that of depletion,large particles cannot access a zone in betweentwo particles, however, the forces are lesserinduced by osmotic pressure but rather by thepure size of the molecule. This means, the hugeamylopectin structures act like springs – or ratherlike deformable particles – between the finestparticles, as shown in Figure 2. In this context itmay be negligible whether the amylopectin canbe found adsorbed or non-adsorbed on particles.Adsorption of starch on cement is reported in lit-erature [23]. As starch can be attracted on sur-faces with positive zeta potentials [33], it isassumable that it can be adsorbed on limestonefillers as well. The adsorption of small moleculessuch as amylopectin would add a steric repulsivecomponent, which contradicts to the observa-tion, though such effect cannot be excluded. Incase amylopectin adsorbs, this would add a bridg-ing component to the mechanism. However, con-sidering the tree-like structure of amylopectin aswell as its enormous size in the order of magni-tude of finest particles, the number of adsorbedbranches would be small compared to the non-adsorbed branches, which would make bridgesrelatively flexible. Furthermore, similar behaviorwas observed by the authors for a high numberof differently modified starches regardless of thepresence of different charges, which also indi-cates that the effect of starch on rheology is notpredominantly induced by adsorption mecha-nisms.

Due to the flexibility, upon shear, the inter-particular mobility remains good so that low vis-cosity can be maintained despite high yieldstresses. Nevertheless, the rapid increase of τ0 athigh solid content also indicates that starch poly-mers might entangle at dense particle packing.The latter would explain why Rajayogan et al. [16]

found – differing from the results presentedhere – high viscosity induced by starch at low w/p(volumetric water to powder ratio). The diutangum based ST2 strongly immobilises water andforms a network. Thus, it strongly stiffens the flu-id even at no or low solid contents. At higher par-ticle dosages, the network maintains stable sothat additional solids do not significantly in -crease the yield stress. Nevertheless, the strongincrease of viscosity at high solid contents pointsout that also entanglement of the polymers orbetween polymers and particles takes place,which reduces the mobility of particles.

In the cementitious systems (Figure 7), ST1shows similar behavior as in the limestone sys-tem. At low cement volume, no significant effecton τ0 can be observed. For the low dosage of ST1,τ0 is even lower than in the reference systemwithout STA. Also hpl is slightly lower than in a sys-tem without STA, however, at a 5 % dosage of ST1hpl is strongly increased due to higher particle vol-umes. It can be assumed that at such a highdosage entanglement of the starch moleculesacts against the motion of particles. While inabsence of PCE the effect of ST1 is similar in LSFand cement systems, the behavior of ST2 with dif-ferent solid types needs further discussion. Theinfluence of the particle solid content on hpl is sim-ilar to the LSF system. In case of 0 and 25 %cement, ST2 also shows similar behavior in termsof τ0 as in the LSF system at 0.5 % ST2. At lower


Figure 6 (above):The influence of the solidcontent of LSF on yield stressand plastic viscosity whenmixed without stabilisingagent, with ST1 and withST2.

Figure 7:The influence of the solidcontent of cement on yieldstress and plastic viscositywhen mixed without stabil-ising agent and differentamounts of ST1 and ST2.

dosage of ST2, no significant effect on τ0 can beobserved. However, at both dosages at 50 % par-ticle content, τ0 is increased significantly towardsvalues above the yield stress induced by ST1. Thisis a clear distinction between the LSF system andthe cement system. The reason for this might beexplained by the stronger attraction forces of thecement (and monosulfate and ettringite in par-ticular) causing more adsorbed ST2 polymers andstronger bonds, which bridge the particles. It canthus be concluded that the stabilising mecha-nism of ST1 may be more complex than that ofST2, since the enormous size and the special tree-like structure of the amylopectin supplement thestabilising mechanism by an additional particle-polymer interaction, which is triggered by athreshold particle density. ST2 mainly interactswith the fluent phase of a dispersed system, how-ever, with major effect on hpl at increasing solidconcentrations. In cementitious systems at high-er solid concentration, the anionic character ofthe polymer may cause a strong tendency toadsorb and increase τ0 by bridging.

In presence of PCE the behavior of both STAsdiffers (Figure 8 and Figure 9). Already smallamounts of PCE significantly reduce τ0 of thepastes regardless of the STA type. Already at thelow dosage of 0.06 % of PCE in LSF and cementsystems, the yield stresses are approximatelybisected. It is noteworthy that differing from theLSF system, in the cementitious system 0.5 % of

ST1 exhibits the same yield stresses as cementi-tious systems without any STA. With furtheraddition of PCE the yield stress values approxi-mate the values of systems without PCE. Abovea PCE dosage of 0.3 % in the cementitious system,only the system with 5 % ST2 can maintain a sig-nificantly higher yield stress than the water sys-tem without STA. In the LSF system, the additionof PCE causes an increase of hpl at 0.06 % PCE,while a significant drop of hpl can be observed forST2. The latter effect may be induced by the com-petitive adsorption of PCEs. Further addition ofPCE does not significantly modify hpl values forboth STAs in LSF.

In comparison to the LSF systems, the addi-tion of PCE causes a significant drop of hpl of thestarch systems. It is assumable that in the LSFsystem only little or no adsorption of ST1 takesplace, while in the cementitious system more ST1can be found adsorbed without PCE. Therefore,effects of competitive adsorption occur morepronounced in the cement system. For the sys-tems with ST2 the influence of the PCE is similarin the LSF and cement system with the differencethat in the cement system the slope for the lossof hpl is smaller at low dosages than in the LSFsystem, which underlines the observation thatcement gives a stronger tendency for the ST2 toadsorb. For all mixes with STA hpl remains high-er than in the reference system without water.At the same time with the exception of ST2 at0.5 % dosage at high PCE dosages the effect ofthe STAs on τ0 is small. Different dosages of ST1only show little effect on τ0, while the increaseof hpl at higher dosages is very pronounced. High-er dosages of ST2 affect both τ0 and hpl towardssignificantly higher values. At dosages of 0.5 %for ST1 and 0.05 % for ST2, which can be consid-ered as typical values for the stabilisation ofcementitious systems, the behavior of both STAsin presence of PCE is very similar. However, dif-fering behavior could be observed for systemswithout PCE.

Without doubt the rheometric investiga-tions can only indicate effects, and due to the sys-tems’ complexity, it may be impossible, to clear-ly separately observe single parameters thatcontrol the effect of stabilising agents. The latteris particularly valid for the starch stabilisingagents due to the extremely differing character-istics of amylose and amylopectin. Based on thepresent results the following main mechanisms


Figure 8 (above):The influence of the PCEcontent on yield stress andplastic viscosity of LSF-watersystems in presence of ST1and ST2, and without stabil-ising agent.

Figure 9:The influence of the PCEcontent on yield stress andplastic viscosity of cement-water systems in presenceof different amounts of ST1and ST2, and without stabil-ising agent.

are suggested to account for the observed rheo-logical effects. For ST1 it is concluded that theamylopectin is the main driving parameter for thestabilisation (although supplementary mecha-nisms may overlap). The stabilising mechanism isconsidered to be similar to a filler effect with thedifference that the filling medium amylopectinremains deformable under shear forces (Fig-ure 10). This mechanism needs a sufficient vol-ume of particles in order to become active. ThePCE causes better dispersion of the solids, thusreducing τ0 significantly. Since the amylopectinmolecules remain in the solution, they continuehindering motion between particles, which caus-es that hpl is largely unaffected for LSF and lesserthan the yield stress for cement by PCE (Figure 10).

It can be assumed that the major effect ofST2 on yield stress is caused predominantly bywater immobilisation, which is the reason, whyST2 increases τ0 and hpl significantly withoutpresence or at low solid particle volumes. At high-er particle volume, a rapid increase of τ0 can beobserved for cementitious systems, which isassumed to be caused by bridging effects causedby adsorption (Figure 11). Upon addition of SP,indeed, as suggested by Phyfferoen [5], the ad -sorption of ST2 seems to be reduced. Therefore,the stabilising mechanism in presence of PCE isvery similar to that of ST1 (Figure 11).

4 SUMMARYInvestigations were conducted to demonstratethe difference in the stabilising mechanism ofSTAs based on potato starch (ST1) and diutangum (ST2) on coarsely dispersed systems. It wasshown that it is very important to distinguishbetween flowable systems with and withoutadsorptive superplasticizers. While diutan gumeffectively stabilised water at low solid particlesystems, starch required a certain threshold par-ticle volume in the fluid to significantly affect theyield stress. At high solid content starch effec-tively increased the yield stress. The same effectcould be observed for diutan gum in presence ofcement but much lesser in presence of LSF. Theaddition of PCE significantly reduced the yieldstress and the plastic viscosity regardless of theSTA. Beyond dosages of 0.06 and 0.3 % for LSFand cement, respectively, an increase of the PCEdosage did not cause any further reduction ofyield stress and plastic viscosity.

In limestone fillers systems ST1 and ST2 wereshown to only partly render the behavior incementitious systems. At low powder volumesand without presence of PCE, limestone filler sys-tems showed qualitatively similar behavior.However at higher powder content and in pres-ence of ST2 the cementitious systems exhibitedsignificantly higher yield stresses, which is mostlikely caused by stronger bridging effects. In pres-ence of PCE similar observation could be madefor the effect of the PCE dosage on performanceof ST1 and ST2 for the yield stress. For the plasticviscosity, however, qualitative differences wereobserved for ST1. The performance differences ofthe two stabilising agents appeared more pro-nounced in absence of PCE. In presence of PCE, atdosages of 0.5 and 0.05 % for ST1 and ST2, respec-tively, both STAs showed similar behavior. Athigher dosages of starch the effect on yield stresswas small, while a significant increase of theplastic viscosity could be observed. For highdosages of diutan gum yield stress and plasticviscosity could be significantly increased.

The results underline that in coarsely dis-persed systems particle interactions significant-ly contribute to the stabilising mechanism ofstarch, while diutan functions by immobilisingthe fluent phase between the particles and bybridging. Upon addition of high amounts of PCE,however, the effects induced by both STAs are


Figure 10 (above):Suggested stabilising mech-anism of amylopectin with-out and with PCE.

Figure 11:Suggested stabilising mech-anism of diutan gum with-out and with PCE.

similar. Compared to diutan gum, starch is morecomplex for the use in cementitious systems,since its performance depends on factors such asthe particle size distribution and the water to sol-id ratio. However, since its influence seems to bemore independent of the adsorption of particles,it might less interfere with PCE adsorption.

The obtained results are based on the as -sumption that a Bingham approach can suffi-ciently describe the stabilising agents’ rheologi-cal properties. However, particularly at low shearrates, is can be assumed that the observed sta-bilising agents show clear non-linear behavior.Furthermore, it was observed that particularlythe stabilising effect of starch can be reducedover the course of time (e.g. when, due to coldtemperatures, the stability is not supported byhydration). Therefore, for future research, it isimportant to put focus on the behavior particu-larly at low shear forces.

ACKNOWLEDGEMENTSThe authors would like to thank the reviewers fortheir thorough and constructive input, includingsuggestions for supplementary observations,which surely increased the value of this study.

REFERENCES[1] Khayat KH: Viscosity-enhancing admixtures for

cement-based materials – An overview, CementConcrete Composites 2 (1998) 171 – 188.

[2] Pourchez J, Peschard A, Grosseau P, Guyonnet R,Guilhot B, Vallee F: HPMC and HEMC influence oncement hydration, Cement Concrete Res. 2 (2006)288 – 294.

[3] Yammamuro H, Izumi T, Mizunuma T: Study ofnon-adsorptive viscosity agents applied to self-compacting concrete, 5th CANMET/ACI Interna-tional Conference on Superplasticizers and otherchemical Admixtures in Concrete (1997) 427 – 444.

[4] Izumi T, Dikty S, Yamamuro H: Properties of NewPolysaccharide as a Thickener for Concrete, 8th

CANMET/ACI International Conference on Super-plasticizers and other chemical Admixtures inConcrete (2006) 105 – 116.

[5] Phyfferoen, A, Monty, H, Skaggs B, Sakata N,Yanai S, Yoshizaki M: Evaluation of the biopoly-mer, Diutan gum, for use in self-compacting con-crete, First North American Conference on theDesign and Use of Self-Consolidating Concrete(2002) 141 – 146.

[6] Pourchez J, Govin A, Grosseau P, Guyonnet R, Guil-hot B, Ruot B: Alkaline stability of cellulose ethersand impact of their degradation products on

cement hydration, Cement Concrete Res. 36(2006) 1252 – 1256.

[7] Bechthold A, Seibel, J: Stärke (RD-19-03654) (2012)www.roempp.com.

[8] Lander V: Amylopektin (RD-16-03559) (2012)http://www.roempp.com.

[9] BeMiller JN, Huber KC: Starch – Ullmann's Ency-clopedia of Industrial Chemistry, Wiley-VCH Ver-lag (2000).

[10] Coleman R, Patel Y, Harding N: Identification andorganization of genes for diutan polysaccharidesynthesis from Sphingomonas sp. ATCC 53159,Journal of Industrial Microbiology & Biotechnol-ogy, 4 (2008) 263 – 274.

[11] Sonebi M, McKendry D: Effect of Mix Proportionson Rheological and Hardened Properties of Com-posite Cement Pastes, Open Construction Build.Technol. J. 2 (2008) 15 – 23.

[12] Swinkels JJM: Composition and Properties ofCommercial Native Starches, Starch – Stärke 37(1985) 1 – 5.

[13] Khayat KH, Ghezal A: Effect of viscosity-modify-ing admixture-superplasticizer combination onflow properties of scc equivalent motar, 3rd Inter-national RILEM Symposium on Self-CompactingConcrete (2003) 369 – 385.

[14] Sonebi M: Rheological properties of grouts withviscosity modifying agents as diutan gum andwelan gum incorporating pulverised fly ash,Cement Concrete Res. 9 (2006) 1609 – 1618.

[15] Sonebi M: Rheological characterization of cement-based materials pastes containing Diutan Gumand limestone powder, 12th International Congresson the Chemistry of Cement (2007).

[16] Rajayogan V, Santhanam M, Sarma BS: Evaluationof hydroxy propyl starch as a viscosity-modifyingagent for self compacting concrete, 3rd Interna-tional RILEM Symposium on Self-Compacting Con-crete (2003) 386 – 394.

[17] Terpstra J: Stabilizing self-levelling concrete withpolysaccharide additive, SCC'2005-China: 1st Inter-national Symposium on Design, Performanceand Use of Self-Consolidating Concrete (2005)207 – 213.

[18] Simonides H, Terpstra J: Use of innovative starchethers for paving blocks and other concrete prod-ucts, Concrete Plant + Precast Technology 9(2007) 38 – 45.

[19] Schmidt W, Brouwers HJH, Kühne, H-C, Meng B:Influence of Environmental Temperatures on thePerformance of Polymeric Stabilising Agent inFresh Cementitious Materials, Key EngineeringMaterials 466 (2011) 97 – 104.

[20] Brumaud C, Baumann R Schmitz M, Roussel N:Influence Of High Molar Mass Polymers On TheRheological Behavior Of Fresh Cement Pastes,ICCC 13th International Congress on the Chemistryof Cement (2011).


[21] Fugmann B, Lützen A: Cellulose (RD-03-00833)(2012) http://www.roempp.com.

[22] Phyfferoen A, Skaggs CB: Optimizing the use ofwelan gum in cementitious and gypsum-basedsystems, ConChem – International Exhibition &Conference (1997) 193 – 202.

[23] Palascios M, Flatt R, Puertas F, Sanchez-HerenciaA: Compatibility between Polycarboxylate andViscosity-Modifying Admixtures in CementPastes, 10th International Conference on Super-plasticizers and Other Admixtures in Concrete(2012) 29 – 40.

[24] Zingg A, Winnefeld F, Holzer L, Pakusch J, BeckerS, Figi R, Gauckler L: Interaction of polycarboxy-late-based superplasticizers with cements con-taining different C3A amounts, Cement ConcreteComp. 3 (2009) 153 – 162.

[25] Plank J, Bassioni G, Dai Z, Keller H, Sachsenhau-ser B, Zouaoui N: Neues zur Wechselwirkung zwi-schen Zementen und Polycarboxylat-Fließmit-teln, 16th International Conference on BuildingMaterials – ibausil (2006) 579 – 598.

[26] Plank J, Hirsch C: Impact of zeta potential of ear-ly cement hydration phases on superplasticizeradsorption, Cement Concrete Res. 4 (2007)537 – 542.

[27] Giraudeau C, D’Espinose De Lacaillerie J-B,Souguir Z, Nonat A, Flatt RJ: Surface and Interca-

lation Chemistry of Polycarboxylate Copolymersin Cementitious Systems, J. Am. Ceramic Soc. 11(2009) 2471 – 2488.

[28] Winter CAJ: Untersuchungen zur Verträglichkeitvon Polycarboxylaten mit den Hydratationsver-zögerern Citrat und Tartrat und zur Wirkung derCaseinfraktionen α-, β- und κ- Casein im ternä-ren Bindemittelsystem Portlandzement-Toner-deschmelzzement-Synthetischer Anhydrit, TU-München, München (2007).

[29] Terpstra J: Stabilizing low-viscous self-compact-ing concrete, 2nd North American Conference onthe Design and Use of Self-Consolidating Con-crete (SCC) and 4th International RILEM Sympo-sium on Self-Compacting Concrete (2005) 47 – 53.

[30] Zhang J, Weissinger EA, Peethamparan S, SchererGW: Early hydration and setting of oil well cement,Cement Concrete Res. 7 (2010) 1023 – 1033.

[31] Puntke W: Wasseranspruch von feinen Kornhauf -werken, Beton 5 (2002) 242 – 248.

[32] Burchard W:Polysaccharides – Structural diversi-ty and functional versatility, Marcel Dekker(2005).

[33] Izaguirre A, Lanas J, Álvarez JI: Behaviour of astarch as a viscosity modifier for aerial lime-basedmortars, Carbohydrate Polymers 80 (2010)222 – 228.


The Working Mechanism of Starch and Diutan Gum in Cementitious ...

Documents