Rheology of low‐clinker suspensions: Multiscale comparison of the effect of PCE on thixotropy

Addressing climate change requires reducing the carbon footprint of cement‐based building materials on a global scale. One promosing approach involves decreasing the clinker content by incorporating fine supplementary cementitious materials (SCM), which, however, can increase the stickiness of fresh concrete. Along with viscosity, thixotropic structural build‐up is a critical rheological parameter that impacts the stickiness of the SCM‐rich suspensions. Thus, it is essential to gain a precise understanding of how the molecular structure and mode of addition of polycarboxylate ethers (PCE) affect the thixotropic structural build‐up in low‐clinker suspensions. To address this issue, we investigate key parameters affecting thixotropy, specifically PCE adsorption and early hydration kinetics at the cement paste and mortar level, using a multiscale approach. In an application‐oriented concept, we maintain a constant yield stress by varying the PCE dosage to achieve constant workability. Our results demonstrate that higher thixotropy correlates with increased PCE charge density and direct addition in cement pastes and mortars. Moreover, it is indicated that increasing specific surface area and ettringite contents promote additional thixotropy‐enhancing mechanisms.


Introduction
Polycarboxylate-based superplasticisers (PCE) are commonly used in cement-based suspensions due to their high dosing efficiency and the ability to tailor their molecular structure to the desired rheological requirements, such as reducing yield stress [1; 2].Although the effect of the molecular structure on yield stress has been investigated well [3][4][5][6], its impact on thixotropy is not fully understood.Thixotropy is particularly important in low-clinker binders with a high content of fine additives such as limestone powder, blast furnace slag or calcined clay, as it increases the yield stress and viscosity at low shear rates, making the concrete more difficult to process [7][8][9].
Thixotropy is a rheological phenomenon that describes the reversible decrease of yield stress and viscosity under shear, with a corresponding increase when the shear is removed [10; 11].One way to control thixotropy in cement-based suspensions is to use PCE superplasticisers [12].As the effect of PCE molecular structure on the thixotropy of suspensions with low-clinker binders is not fully understood, we investigate the effect at cement paste and mortar scale.The investigation at the mortar level is crucial as the presence of larger inert particles increases the local shear rates, thereby affecting the rheological properties of the cement paste [13; 14].In addition to the molecular structure, the timing of adding the PCE superplasticisers is also important.Direct addition can affect the early aluminate reaction by favouring ettringite precipitation over crystal growth [15; 16], which can have an impact on thixotropy [17].In contrast, delayed addition does not significantly modify the aluminate reaction [18].
To better control the thixotropy of low-clinker cementbased suspensions using PCE, it is important to understand the relationship between molecular structure and addition mode on the underlying mechanisms of thixotropy, including key parameters like PCE adsorption and very early hydration kinetics.This article, therefore, examines the effects of two polymer structures on the increase in static yield stress, correlating it with PCE adsorption (colloidal interactions) and specific surface area (contact interactions).
behaviour leads to a time-dependent increase in static yield stress [10; 11].The structural parameter λ or Athix can represent this increase, as they both indicate the thixotropic structural build-up when the material is at rest (1).
In cement-based suspensions, thixotropy is caused by microstructural flocculation and deflocculation, which is determined by colloidal interactions and early hydration kinetics [7; 10; 19].This article investigates the thixotropic structural build-up at rest, which depends on the initial colloidal flocculation within a few seconds, the interparticle bridging due to the precipitation of early hydration products and the subsequent strengthening of the rigid microstructure.

Early hydration kinetics
Early hydration is a crucial factor in the thixotropic structural build-up as it controls two mechanisms.On the one hand, there is interparticle bridging by the formation of early hydration products at contact points and on the other hand, the precipitation of early hydration products leads to an increase in cohesive contact points [19][20][21].The thixotropic structural build-up is enhanced by higher fineness, higher solid volume fraction and lower maximum packing density of the cement-based suspension [12; 22].
The reactivity of cement significantly influences the thixotropic structural build-up.During the early hydration of cement, C3A and the sulphate carrier dissolve rapidly, leading to continuous ettringite precipitation, followed by a pronounced silicate reaction at the start of the acceleration phase.Both ettringite and CSH are assumed to play a role in bridging and precipitation.However, the separation of their contributions to thixotropy remains an active area of research beyond the scope of this article, hence, we refer to other studies [19][20][21].
This article discusses ternary binder mixes with calcined clay and limestone powder, which exhibit hydration kinetics comparable to Portland cement in the very early stage (< 2h) [23; 24].Hou et al. observed higher thixotropy in these types of binders compared to Portland cement pastes, attributed to more pronounced flocculation resulting from higher fineness [8].However, further research is necessary due to varying boundary conditions when these low-clinker binders interact with admixtures.

Polymer structure and dispersing properties
Polycarboxylate ethers (PCE) are a type of comb-shaped co-polymers that consist of a backbone of carboxylate groups with polyethylene glycol side chains [1; 2].PCE are known to adsorb onto solid particles and disperse extremely efficiently due to the steric hindrance created by the side chains protruding into the liquid phase.Another significant advantage of PCE is that their molecular structure can be tailored to meet the respective rheological requirements.The molecular structure of PCE is often simplified using three parameters: N-1, P and n [25; 26].N-1 represents the number of charged groups per side chain, which indicates the adsorption affinity of the adsorbing backbone.The parameter P represents the length of the side chain, while n represents the number of repeating units in the molecule that include a side chain.In addition, the anionic charge density is an essential parameter that should be taken into account when examining the properties of PCE.It is defined as the number of elementary charges per repeating unit divided by the molar mass of 1 mol of polymer [27], as shown in equation (2).
where σ denotes the theoretical anionic charge density of a polymer, N-1 is the anionic charge ratio, z is the number of elementary charges per anchoring group (e.g.carboxylic groups with one elementary charge) and Mu is the molar mass of the repeating unit according to equation 3.
where MBB represents the molar mass of the monomer in the polymer backbone, for example, acrylic acid, MSC refers to the molar mass of the monomer in the side chain, such as ethylene glycol (EO), while ML represents the molar mass of the monomer that forms the linkage between the backbone and side chain.
In cement-based suspensions, steric interactions resulting from adsorbed PCE dominate the dispersion mechanism, with other repulsive particle interactions playing a secondary role [25; 28].Accordingly, particle dispersion scales with adsorption and the resulting surface coverage.Therefore, when the surface coverage is complete, the average distance between particles should correspond to twice the thickness of the polymer layer [25].However, in practical applications, PCE are used only at low surface coverage, as higher surface coverage can result in a significant delay in hydration or sedimentation [1; 18].Steric interactions scale with the amount of adsorbed PCE when there is incomplete surface coverage [29].The most appropriate representation for this is the adsorbed mass with respect to the specific surface area of the particles (4), as the interaction between PCE and the hydration reaction can affect the kinetics and thus the specific surface area [30].
where mPCE,adsorbed represents the amount of adsorbed PCE on the solid particles, while SSACEM,hydr denotes the specific surface area determined from hydrated particles.

Direct and delayed addition mode
PCE superplasticisers are commonly used in two ways: direct addition to the mixing water or delayed addition after water and cement have been mixed.Direct addition of PCE leads to an earlier onset of the aluminate reaction and a delay of the silicate reaction [18].This phenomenon is often accompanied by an increase in specific surface area due to ettringite precipitation during very early hydration [31].
Delayed addition, on the other hand, means that the majority of initial ettringite has already precipitated and passivation is initiated before the PCE is added [18].It is worth noting that delayed addition could also affect the specific surface area, although to a lesser extent than observed with direct addition [31].

Effect of PCE superplasticiser on the thixotropic structural build-up
The presence of PCE has a significant effect on the thixotropic structural build-up of cement-based suspensions, along with various other parameters [12; 32].The steric repulsion introduced by PCE affects interparticle interactions, leading to a decrease in the thixotropic structural build-up as particle separation increases [12].Consequently, flocculation and interparticle bridging of hydration products slow down, causing a slower macroscopic structural build-up.Previous studies have revealed that the thixotropic structural build-up depends not only on the adsorbed amount but also on the molecular structure [33].
In particular, an increase in the structural build-up at constant yield stress has been observed with increasing grafting density and decreasing side chain length.However, further research is needed to gain a better understanding of the underlying mechanisms, particularly for cements that contain calcined clay and limestone.This is especially important concerning the competition of steric interactions, which lead to an increase in interparticle distance (colloidal interactions), i.e. a decrease in thixotropy, and a possible increase in thixotropy through an increase in specific surface area (contact interactions). 3

Concept of investigations
In an application-oriented approach, we investigate the effect of PCE molecular structure on thixotropic structural build-up at a constant spread flow of 270 ± 5 mm, i.e. dynamic yield stress of 9.3 Pa acc. to [34], since concrete is ordered and produced with a given consistency.Furthermore, PCE is either added directly to the mixing water or with a slight delay to reduce the effect on the initial hydration kinetics.Both modes of addition are practically relevant and are therefore investigated.Rheological and analytical investigations are carried out both at the level of paste and mortar to identify potential cross-scale similarities and differences, with a focus on the dominant colloidal (mainly steric) interactions and contact interactions (mainly due to the formation of fine hydration products).
The objectives of the investigation include 1.) measuring the thixotropy of PCE-superplasticized cement pastes and mortars, 2.) investigating PCE adsorption and early hydration kinetics (SSABET, Ettringite formation) and relating it to thixotropy and 3.) identifying the dominant mechanisms with respect to PCE molecular structure and addition mode.

4
Materials and methods

PCE superplasticiser
We investigate two non-commercial PCE superplasticisers, one with high and one with low charge density (Table 1).
The copolymers are based on acrylic acid and vinyloxybutyl-polyethylene glycol and their charge density is calculated according to Eq. 2. To simplify our designation of the polymers, we reduce them to side chain length P and anionic charge ratio N-1, resulting in the notation "PCE [P]. [N-1]".

Cement
We used a low-clinker binder mix which has been characterized in the context of the DFG Priority Program 2005 [35].The main constituents are summarized in Table 2 and additional analytical information can be found in the reference paper cited before.The average volumetric particle size is xv,50 = 9.99 μm micrometres.We used quartz sand ranging from dmin = 0.2 mm to dmax = 2.0 mm to match the mortar fraction described in [7].The particle size distribution was characterized by sieving according to [36], as shown in Figure 1.The constituents of the pastes and mortars are listed in Table 3 and Table 4.The mortar composition, which consists of 60 Vol.-% paste and 40 Vol.-% sand, was selected to be similar to that in previous studies [7; 12].

Paste and mortar production
To ensure a well-defined reference state before rheological characterization, it is essential to define the production and shear history of the mixed suspension in detail before the rheological experiments.For this purpose, the procedure is summarized in Table 5.The cement pastes and mortars were produced in a standard EN 196-1 mixer [37], with 1.3 L of material for each experiment.Before each paste production, the binder mix was homogenized by tipping and rolling the bucket and the PCE dosages were adjusted to match the designated workability range, compare chapter 3.After the mixing procedure, the suspension temperature was recorded and the mixing container was covered with an acrylic glass plate to prevent evaporation until the preshearing step.The thixotropic structural build-up of the pastes was measured with a rheometer (Anton Paar MCR 502, plate diameter = 50 mm) equipped with a sandpaper surface having an average roughness of 25.8 ± 1 μm.The rheometer has a Peltier element and a temperature hood to maintain a stable temperature of 20 ± 0.01 °C during the measurement.The shear rate was calculated at a decisive radius of 3/4R [38; 39].The mortars were measured using a 6-bladed wide-gap vane-in-cup geometry with a stationary outer cylinder and a rotating vane probe.To minimize wall slip, 24 vertical lamellae were fixed at the outer wall.
Further details and dimensions of the rheometer setup can be found in [40].
After allowing the early hydration reaction to subside, the suspensions were presheared to obtain a common reference state with sufficient structural breakdown.This was accomplished by using a standard EN mixer set to level 2 for 60 seconds, but only in the case of paste, 12 minutes after water-cement contact.The mortar, on the other hand, was presheared at the same time, but with a helix paddle attached to a screwdriver in the rheometer cup, rotating at 1700 rpm (Bosch GSR 18 V EC FC2), to minimize shear-induced particle migration that typically occurs during preshearing in wide-gap rheometer geometries [41].
The plate-plate rheometer measurements started 14 minutes after initial water-cement contact.To induce structural breakdown, a constant shear rate of 90.3 s -1 was applied during the preshear period for 30 s (only for paste), as illustrated in Figure 2. The thixotropic structural build-up was then measured as the growth of the static yield stress at rest after various rest intervals.To minimize the impact on the built-up particle structure, a shear rate of 0.130 s -1 and a measuring interval of 6 s were selected to obtain a measurable static yield stress.The thixotropic structural build-up Athix was calculated as the slope of the linear regression of the static yield stresses against resting time using Equation 1[12; 42], also compare red circles in Figure 2. The spread flow of the cement paste was also determined simultaneously with the first static yield stress measurement at 15 min, as shown in Figure 2 and discussed in Section 4.2.3.The mortar was characterized at the same time intervals, with a constant rotation speed of n = 0.1155 min -1 and the shear stress was calculated using the measured torque, as given by equation (5).
where T is the measured torque, RV is the radius and hV is the height of the vane paddle.
Figure 2 Shear profile of the preshearing (only plate-plate, mortar preshearing at 12:00 min using helix paddle outside of rheometer, compare section 4.2.2) and rheometric determination process used to measure the static yield stress as a function of resting time and evaluation method for the thixotropic structural build-up

Spread flow and yield stress determination
The suspension was presheared in a standard EN mixer [37] at level 2 for 30 seconds, 13:45 min after water-cement contact.Subsequently, the spread flow test was conducted on a dry and smooth glass plate with a truncated cone shape, following the guidelines of EN 1015-3 (cone dimensions: height: 60 mm, bottom diameter: 100 mm, top diameter: 70 mm) [43].The cone was lifted 15 minutes after water-cement contact and the diameter was measured two times in perpendicular directions after the flow stopped, approximately 90 seconds after lifting the cone.Finally, the dynamic yield stress was calculated using an equation proposed by Roussel [34]: ρ refers to the bulk density of the sample, g denotes gravity, V represents the volume of the tested sample and R is the radius of spread flow.

PCE-Adsorption: TOC depletion method
To determine PCE adsorption, we employed the TOC depletion method.We measured the difference between the dosage and the remaining amount of PCE in the solution, which represents the quantity absorbed.Although strictly speaking this is the amount consumed, however, if the C3A/SO4 2-ratio exceeds 0.75 it is reasonable to assume complete adsorption [44; 45].
We determined the PCE adsorption in the liquid phase of the identical suspension that was measured in the rheometer.After 16 minutes from the water-cement contact, we extracted the aqueous phase using pressure filtration in a stainless steel cartridge at 6 bar.The extracted phase was passed through a 0.2 µm syringe filter and subsequently diluted 1:10, 1:20 or 1:40, depending on the PCE dosage, to match the calibration range of the TOC analyser.The remaining dissolved organic carbon was determined by high-temperature oxidation at 680°C using a TOC analyser (Shimadzu TOC L-CPN) and recorded as non-purgeable organic carbon (NPOC).To correct the results for the organic content in the mixing water and cement, we also measured a cement paste and mortar with identical composition but without PCE and subtracted its results from the results obtained with PCE.The adsorbed amount of superplasticiser is considered the difference between the added PCE amount and the corrected TOC measurement.

Specific surface area determination using the BET method
The specific surface area of cement pastes and mortars was determined using the solvent exchange method after the rheometer measurement, 15 minutes after water-tocement contact.We used an isopropanol-to-sample ratio of 10:1, following Mantellato et al. [46; 47] and Lowke [7].
After allowing coarser particles to settle for 24 hours, the turbid supernatant was passed through a 0.20 µm pore size filter.The resulting sediment and filter residue were dried at 20 °C, combined and homogenized.Before measuring the specific surface area using the BET method, the cement pastes and mortars were degassed under vacuum at 40°C for 16 hours to preserve ettringite and sulfate microstructure.The nitrogen sorption measurements were conducted with a 5-point isotherm in a relative pressure range of 0.05 to 0.30 at 77.3 K.

Quantitative X-ray diffraction (XRD)
We used the same samples from the BET method to quantify the phase composition with XRD, compare chapter 4.2.5.For mortars, we obtained an additional subsample by sieving under vibration to ensure the limited maximum particle size of less than 0.5 mm.The hydration was stopped in the same procedure.
We used a Bruker D8 Advance device with Bragg-Bentano geometry for our XRD measurements.The quantification of the phase composition was done using the software TOPAS and the G-Factor quantification method [48].Cu Kα radiation was used, voltage and current were set to 40 kV and 40 mA.The step size of each measurement step was set to 0.011 °2θ, the counting time for each step was 0.6 seconds/step and the detection range was 6-70 °2θ.
In the stopped powder samples, the mass attenuation coefficient (MAC) is unknown, because an unknown amount of water is already bound in ettringite after 15 minutes, and some gypsum is already dissolved.Thus, the H2O content was determined by thermogravimetric analysis (TGA).We heated the samples to 1000°C with a rate of 10°C per minute in an STA 449 F5 Jupiter device from Netzsch.The obtained results were then combined with RFA analysis of the raw powder to determine the chemical composition and to calculate the correct MAC of each sample.In the next step, the phase quantities from our analysis had to be corrected to resemble the composition of the cement and mortar paste, where liquid water is present.To do so, we performed TGA on reference samples that were hydrated for at least 28 days in sealed sample carriers.TGA measurement properties were the same as for the stopped samples.We then subtracted the H2O content from the stopped samples from the contents of the hydrated samples and used the difference as a correction factor.This gave us the final results of the quantitative XRD analysis to determine the phase composition of the pastes after 15 minutes of hydration.Effect of the PCE molecular structure and addition mode on the thixotropic structural build-up Figure 3 illustrates the thixotropic structural build-up Athix of low-clinker binder pastes with constant dynamic yield stress and varying PCE molecular structure in direct or delayed addition mode.For the highly charged PCE S.Hi, there is a distinct difference between direct and delayed addition.While thixotropy with direct PCE addition is at 1.85 Pa/s, it is considerably lower at 0.13 Pa/s for delayed addition.Binder pastes with PCE L.Lw also exhibit a significant difference between direct and delayed addition, with 0.51 Pa/s for direct addition and 0.40 Pa/s for delayed addition.Comparing the thixotropy of PCE S.Hi and L.Lw with direct addition, it is evident that the highly charged PCE S.Hi exhibits significantly higher thixotropy than PCE L.Lw.However, when considering delayed addition, the trend reverses and PCE S.Hi demonstrates a significantly lower thixotropy compared to PCE L.Lw.It becomes evident that both the molecular structure and the mode of addition significantly affect thixotropy.The reasons for these findings will be explored in detail in the subsequent chapters.In line with the cement paste results, Figure 4 presents the thixotropic structural build-up Athix of the mortars in response to variations in PCE molecular structure and addition mode.Qualitatively, the trend observed for PCE S.Hi is consistent, with high thixotropy in the case of direct addition and low thixotropy when the addition is delayed.PCE L.Lw displays subtle differences as a result of the addition mode, exhibiting a higher thixotropy of 0.56 Pa/s compared to 0.46 Pa/s with delayed addition.Notably, PCE S.Hi exhibits the highest mortar thixotropy value of 1.69 Pa/s in direct addition mode and the lowest value of 0.37 Pa/s in delayed addition mode.
This, the results on the mortar level confirm the qualitative trends observed at the binder paste level.Consequently, it is essential to investigate whether the same underlying mechanisms are responsible for the rheological findings.5.2 PCE Adsorption and its impact on the thixotropic structural build-up In Figure 5, the adsorbed amount of PCE related to the hydrated surface after 15 minutes is plotted against the thixotropy of the binder pastes.Consideration of the specific surface area is crucial, as it can vary depending on the molecular structure and dosage, potentially providing more sorption sites for PCE [18; 30; 31].
Taking PCE S.Hi (Delayed), PCE L.Lw (Direct) and PCE L.Lw (Delayed) into consideration, it becomes apparent that, with an increase in adsorption, thixotropy decreases.Resulting of higher steric interactions, the observed correlation can be attributed to the increase in average particle distance that counteracts colloidal flocculation and particle bridging by hydration products [12].An exception in this context is the cement paste with directly added PCE S.Hi, which exhibits both higher thixotropy and higher adsorption compared to other PCE.This relationship of increasing adsorption correlating with increasing thixotropy, however, is not causal, suggesting that contact interactions could potentially be modified as well, which will be discussed in subsequent chapters.Paste be attributed to the higher particle separation in the mortar due to steric repulsion.Interestingly, the mortar with directly added PCE S.Hi exhibits the same trend as the paste produced with PCE S.Hi in direct addition.This mortar shows again a higher thixotropy despite increased PCE adsorption.This finding underlines the need of considering an additional thixotropy-modifying mechanism, besides PCE adsorption.
It can furthermore be stated that the effects of the PCE polymer structure on the thixotropic build-up observed at the paste level are fully present at the mortar level as well.

5.3
Effect of polymer structure and addition mode on the ettringite content and specific surface area Besides colloidal interactions, contact interactions can also contribute to increased thixotropy [12].For PCE S.Hi with direct addition, this could explain the higher thixotropy observed in Figures 5 and 6.An increase in contact interactions can be attributed to a reduction in particle size [49; 50], which is experimentally detectable by higher specific surface area.To link the differences in specific surface area with early hydration products when varying polymer structure and addition mode, Figure 7 displays the ettringite content of the investigated pastes and mortars against the specific surface area, determined by quantitative XRD.
Both superplasticisers exhibit a higher specific surface area due to direct PCE addition, with PCE L.Lw ranging from approximately 2.85 to 3.15 m 2 /gCEM, while PCE S.Hi ranges from about 3.7 to 4.6 m 2 /gCEM.The increased surface area can be attributed to reduced passivation of C3A and/or preferential precipitation of ettringite crystals as opposed to growth.
Ettringite is the primary hydration product at the investigation time for all samples as no other hydration products were found.The data reveal that both in pastes and mortars, a rising specific surface area correlates with an increasing ettringite content.It is worth noting that specific surface areas in pastes are higher than in mortars at slightly lower ettringite levels, which is due to the coarser sand particles in the mortars, reducing the mass-based surface area.
Regardless, the observed trends in both suspension types confirm that polymer structure and PCE addition mode affect early hydration kinetics.PCE with a higher charge density and direct addition, as opposed to delayed addition, produce more ettringite, leading to an increase in specific surface area.

5.4
Effect of the specific surface area on the thixotropic structural build-up Figure 8 shows the observed thixotropy plotted against the measured specific surface area of the cement pastes stopped at the time of investigation.
It becomes evident that both for pastes with PCE S.Hi and L.Lw, a higher specific surface area correlates with increased thixotropy.Notably, the pastes with PCE S.Hi display a larger surface area and especially when PCE S.Hi is added directly, the pastes exhibit the highest specific surface area and thixotropy.From these findings, we can deduce that the increasing trend of specific surface area and thixotropy scales with higher charge and direct addition of the PCE.The more pronounced increase with a higher charge density of the PCE and direct addition is supported by the correlation of the specific surface area and the ettringite content, compare Figure 7. Hence, we can conclude that pronounced ettringite precipitation is likely to affect thixotropy as well in this context.As in pastes, direct PCE S.Hi addition results in the largest specific surface area among the studied mortars.Also, a consistent trend appears for both PCE types, with direct PCE addition leading to a higher specific surface area compared to delayed addition.These findings support the idea that the rise in thixotropy for PCE S.Hi in direct addition mode may be related to less effective adsorption or stronger contact interactions, also which is significantly correlating with increasing thixotropy both in pastes and mortars.In this paper, we investigated the effect of PCE molecular structure and addition mode on the thixotropic structural build-up in an application-oriented approach, maintaining constant spread flow by adjusting the PCE dosage for binder pastes and mortars.The suspensions contained a low-clinker binder based on Portland cement combined with calcined clay and limestone powder.We analysed the thixotropic structural build-up at rest as the increase in static yield stress, determined PCE adsorption, measured the specific surface area and quantified phase contents of early hydration products.Based on our findings, we draw the following conclusions: -for both, binder pastes and mortars, higher thixotropy correlates with higher PCE charge density and direct addition.
-for the highly charged PCE S.Hi (Delayed addition) as well as the lower charged PCE L.Lw (Direct and delayed addition) an increase in surface-related adsorption correlates with a decrease in thixotropy.
-it was therefore possible to show that the effects of the PCE polymer structure on the thixotropic structural buildup observed at the paste level are fully present at the mortar level as well.
-in the case of the highly charged PCE S.Hi, there is a comparatively higher increase in specific surface area and ettringite content, suggesting that other thixotropy-enhancing mechanisms like ineffective adsorption or increased contact interactions due to the formation of fine ettringite should be considered.

Figure 1 4
Figure1Particle size distribution of the quartz sand characterized acc.to[36]

Figure 3
Figure 3 Thixotropic structural build-up Athix of low-clinker binder pastes for different PCE molecular structures and addition modes at a constant spread flow (P = side chain length, N-1 = number of charges per repeating unit n, nSamples = 2)

Figure 4
Figure 4 Thixotropic structural build-up Athix of low-clinker mortars, showing the effect of the PCE molecular structure and addition mode at a constant spread flow (P = side chain length, N-1 = number of charges per repeating unit n, nSamples = 2)

Figure 5
Figure 5 Adsorption as consumed PCE amount in binder pastes related to the specific surface area 15 minutes after water addition plotted against thixotropy at a constant dynamic yield stress with variation in polymer structure and addition mode; trend line is not fitted through PCE S.Hi direct (P = side chain length, N-1 = number of charges per repeating unit n, nSamples = 2)As in the case of the binder paste, Figure6presents thixotropy as a function of surface-related adsorption for the investigated mortars.A correlation of decreasing thixotropy with increasing adsorption is observed for most PCE, except for PCE S.Hi once again.This relationship can again

Figure 6
Figure 6 Adsorption represented by the consumed PCE amount in mortars related to the specific surface area 15 min after water addition plotted against thixotropy at constant dynamic yield stress while varying polymer structure and addition mode; Note that the trend line does not fit through PCE S.Hi direct (P = side chain length, N-1 = number of charges per repeating unit n, nSamples = 2)

Figure 7
Figure 7 Correlation between the specific surface area of the investigated pastes and mortars and the ettringite content 15 minutes after water addition (P = side chain length, N-1 = number of charges per repeating unit n, nSamples = 2)

Figure 8 Figure 9
Figure 8 Thixotropic structural build-up of the investigated binder pastes as a function of specific surface area 15 minutes after water addition (P = side chain length, N-1 = number of charges per repeating unit n, nSamples = 2)

Figure 9
Figure 9 Representation of the thixotropic structural build-up of the mortars, correlated with the specific surface area at 15 minutes after water addition (P = side chain length, N-1 = number of charges per repeating unit n) (nSamples = 2) 6 Conclusion

Table 2
Components of the low-clinker binder mix

Table 3
Mixture proportion of the cement pastes (Temperature after mixing was adjusted by tempering the water to reach 20 ± 1°C)

Table 4
Mixture proportion of the mortars per m 3 (Temperature after mixing was adjusted by tempering the water to reach 20 ± 1°C)

Table 5
Paste and mortar production procedure (the comments in brackets indicate variation in case of mortar and delayed as opposed to direct PCE addition)