Effect of recycled powder on the yield stress of cement paste with varied superplasticizers

The influence of superplasticizer on the yield stress of cement pastes with recycled powder (RP) was examined in the study. Four superplasticizers were used to obtain the similar fluidity by adjusting the dosage. The results show that the 10% RP decreases the yield stress of paste compared to the reference paste at the same fluidity, but 20% and 30% RP increases the yield stress, ranging from 11 to 599%. The superplasticizer with adsorptive group of phosphate-type minimizes the yield stress of paste than that of polycarboxylate -type, but it made a significant increment in yield stress as the incorporating of RP increased. Besides, the polycarboxylate superplasticizer with the higher molecular weight of side chain and charge density led to lower yield stress. Based on the Yodel model, the yield stress of paste with RP was analyzed by the polymer adsorption and particle packing density of particles to reveal the influence of RP with different superplasticizers on the colloidal interaction and contact network among the particles. The packing density of particles with recycled powder was a little higher than the reference paste, but the higher fraction of fine particles made a stronger PSD effect, which improved the particle contact interaction. On the other hand, due to the higher polymer adsorption of recycled powder than cement, especially for superplasticizer with phosphate group, the average surface coverage was increased, which extended the separation distance, so that colloidal interaction among particles was weaken.


Introduction
Nowadays, sustainability of building industry has been focused on the recycling of construction and demolition (C&D) wastes because of the depletion of natural resources and the increasing generation of construction waste.In the last decades, the reuse of recycled aggregate is the most effective way to recycle C&D waste.And there are large researches on the improvement of the properties of recycled aggregates [1][2][3].Whereas, a large quantity of recycled powder, producing from the crushing of C&D waste, have not been utilized effectively because of its impurity content and higher water absorption [4].The working performance of cement-based materials incorporating with recycled powder (RP) will be weakened and affect the quality of the project directly.Poor workability will affect the strength and durability of cement-based materials after hardening, reducing the service life of structures.On the other hand, compared to cement particles, recycled powder has more specific surface area and finer particle, leading to a fine micro filling effect and volcanic ash effect in the cement-based materials [5].Therefore, it is gradually used as a mineral admixture to replace cement.
The workability of paste with RP should be adjusted by superplasticizer.Numerous studies have shown that [6][7][8][9][10] varied superplasticizers have significant differences in adsorption behavior on the surface of different particles.These differences are mainly affected by the charge of the powder, specific surface area and the adsorption group of the superplasticizer [6].When there is much ineffective adsorption, the amount of superplasticizer will be increased, and even the excellent dispersion effect will be weakened [11].The workability of paste can be described by yield stress more accurately, which is closely related with the flocculation state of the solid particles in paste.Some studies concluded that the RP containing different mineral compositions exhibits dissimilar surface charge, which affects the adsorption of superplasticizer on suspending particle surface [12,13].Therefore, the adsorption has a significant influence on the yield stress of paste.However, there are few investigations focus on the superplasticizer adsorption of RP and its influence mechanism on the yield stress of paste.
In order to understand the mechanism of various superplasticizer on the yield stress of paste with recycled powder, four superplasticizers with different molecule structures were investigated in the study.For a given fluidity, superplasticizers have varied influence on the yield stress and polymer adsorption of recycled powder.The affecting mechanism of the recycled powder and superplasticizers on the yield stress was analyzed by investigating the change of attractive colloidal interaction and contact interaction.

Materials
Recycled powder from recycled concrete with original strength of 40 Pa was chosen in this study, which was collected from the particles finer than 75 μm during the production of recycled aggregate in the factory.Besides, P•II 52.5cement was also used in this study.The physical performance of cement (C) and recycled powder (RP) was listed in Table 1.RP10, RP20 and RP30 indicate that the replacement rate of recycled powder for cement is 10%, 20% and 30% respectively.The content of crystalline phases in the cement and recycled powder was tested by X-ray diffraction with CuKα source.And the sample was scanned from 5º to 70º.The particle size distribution (PSD) of particles was measured by laser particle size analyzer(PSA1190LD) with a refractivity of 1.68.The PSD and XRD of particles were shown in Figs. 1 and 2.
Four superplasticizers were used in this study, whose adsorptive group and molecular weight of side chain was list in the Table 2.The same adsorptive group of SP-1, SP-2 and SP-3 is poly-carboxylate, but SP-4 has phosphonic acid group.In addition, the molecular weight of side chain for SP-3 is larger than SP-1, and the charge density of SP-2 was higher than SP-1.

Mix proportions
Mixtures with W/C of 0.29 were designed in the study to investigate the effect of superplasticizer, which has the same flow spread with (190 ± 5) mm, (210 ± 5) mm, (230 ± 5) mm and (250 ± 5) mm.The RPs were added into paste with the mass replacement with 10%, 20% and 30%, respectively.
The mixture with cement and recycled powder was premixed in a rotated mixer for better uniformity, and then added into the water mixed with the superplasticizer to prepare paste.The paste was firstly mixed at slow speed of (140 ± 5) rpm for 120 s, and then at high speed of (285 ± 5) rpm for another 120 s after resting for 15 s.The  2000 materials were placed in the preparation room at least before 24 h, the temperature was kept at (20 ± 2) °C.

Fluidity and yield stress
The flow spread and rheological test were conducted immediately after the mixing of fresh paste.The flow spread was carried out conforming to the Chinese standard GB/T 8077.A mini-slump cone was used, its upper diameter, lower diameter and height was 36 mm, 70 mm and 60 mm respectively, which was shown in Fig. 3.While the rheology of paste was tested by using an Anton paar rheometer (MCR 302), the measurement procedure was shown in Fig. 4. In order to test at a uniform state, the paste was firstly pre-sheared at a speed of 100 s −1 for 60 s.After a rest period of 30 s, the shear rate increased by step from 0 to 100 s −1 and then decreased to 0 s −1 , the two steps were completed respectively within 60 s.The shear stress change with the shear rate of 5-25 s −1 in descending stage was used to calculate the yield stress by using H-B model.

Polymer adsorption measurement
The adsorption of superplasticizer on the particle surface was determined by using a total organic carbon analyzer (Multi N/C3100, Analytikjene AG, Germany).Firstly, the paste was centrifuged at 10,000 rpm for 5 min to obtain the aqueous phase.Then the supernatant is immediately collected by a filter with pore size of 50 μm.In order to eliminate the effect of carbonation and inorganic carbon, the equal mass hydrochloric acid solution with a concentration of 1 mol/L was added into the sample.Before the measurement, the supernatant diluted dozens of times with deionized water to satisfy the measurement range, and the reference superplasticizer solution was diluted 100 times.As shown in Eq. ( 1), the concentration difference before and after superplasticizer contacting with the powders ( C 0 and C t ) was assumed to be the adsorbed polymer.
P: adsorption ratio of superplasticizer, %, C 0 : the organic carbon of superplasticizer solution with the same dosage in the paste, mg/L, (1) C c : the organic carbon of supernatant of the paste with- out superplasticizer, mg/L, C t : the organic carbon of supernatant of the paste with superplasticizer, mg/L, V : the volume of water in the cement paste, L, M : the total mass of cement and RP in the paste, g,a: the dosage of superplasticizer, %.

Packing density measurement
The packing density of particles in cement paste was tested by the minimum water demand for cement paste [14,15], which can just make cement particles transition from a solid powder state to a paste state.The mixture was firstly mixed at slow speed of (140 ± 5) rpm for 60 s and then at high speed of (285 ± 5) rpm for 60 s.And then the paste was mixed at high speed for another 300 s after a rest of 15 s.where, ρ w is the density of water(g/cm 3 ), ρ c is the density of mixture of cement and RP(g/cm 3 ), m w is the minimum water demand (g), and m c is the mass of water and particles (g).

Flow spread
The dosage of different superplasticizer for keeping the same flow spread of cement pastes incorporating different displacement of RP are shown in Fig. 5.The dosage of superplasticizer increased steadily for keeping higher flow spread.Compared with the cement paste, RP obviously increases the dosage of superplasticizer.In addition, for achieving the same flow spread, the dosage of SP-4 is much higher than other three kinds of superplasticizer. (2)

Yield stress
Figure 6 shows the yield stress of paste with different kind of superplasticizer at the same initial flow spread.When the paste with different kind of superplasticizer gives the same flow spread, the yield stress presents different values.It can be seen clearly that the difference between the paste increases with the decreasing of the flow spread.The yield stress of paste reduces with the increasing of flow spread, but the change exhibits lower for the pastes with SP-4.
Figure 7 shows the effect of the replacement of RP on the yield stress of paste at the same flow spread with different superplasticizer.A lower yield stress of paste with 10% RP can be achieved by the controlling of SP-1, SP2 and SP3.But when the displacement of recycled powder increases from 10 to 30%, the yield stress of paste increases gradually for all superplasticizer.Additionally, the increment of yield stress with the increase of RP for Fig. 8 Adsorption of polymer in cement pastes with different displacement of RP and varied kind of SP SP-4 is higher than the other superplasticizer.When the flow spread is 190 mm, the yield stress of paste with 10%, 20% and 30% recycled powder is higher than referent paste 200.59%, 304.76% and 599.40% respectively with SP-4.However, the value for paste with SP-1, SP-2 and SP3 is 23.08%, 27.15% and 17.08% respectively when the displacement of recycled powder is 30%.

Polymer adsorption
Figure 8 shows the adsorption of superplasticizers on the particles in the paste incorporating different replacement of RP with varies superplasticizer, respectively.It can be seen clearly that the polymer adsorption by particles increases with the rise of flow spread for pastes.Besides, RP can increase the polymer adsorption with the increasing of replacement.Furthermore, SP-1 and SP-4 play the greatest and lowest effect on the polymer adsorption, respectively.

Particle packing density
Figures 9 and 10 show the packing density of the mixtures with different superplasticizer and incorporating varied displacement of RP.It can be seen that the packing density of the paste with RP is larger than that of the reference paste obviously, and it increases with the rise of RP replacement.In addition, the packing density increases steadily with the increment of superplasticizer dosage, which is due to the gradual decrease of flocculation of the mixtures with the increase of superplasticizer.Besides, SP-1 makes the highest and SP-4 makes the lowest packing density respectively.

YODEL model for yield stress 4.1.1 Colloidal interactions
Some studies [16][17][18] have shown that the yield stress of paste is related to flocculation state or the internal particle network, which are mainly affected by the colloidal interaction and physical contact effect among solid particles.In the following, the influence of different superplasticizer and replacement of RP on the two interactions was investigated to analyzed the mechanism of yield stress of paste with RP.
The particles in the colloidal particle suspensions will be agglomerate by attractive van der Waals force [19,20].However, it will be weakened by the electrostatic repulsion and spatial hindrance force, which is attributed to the polymer adsorption on particle surface [21].Therefore, the total interparticle force (F) can be described as follows: where A 0 is the Hamaker constant (J), a * and H is the curvature radius and separation distance between particles at "contact" point (m), respectively.For cement and hydrated particles, A 0 varies between 0.5 × 10 -20 and 3.85 × 10 -20 J [16,22].
The separation distance between particles is related to the average surface coverage of particles with superplasticizer ( θ , −): (3)

Fig. 9 density of cement without SP
where H p is twice the adsorbed layer thickness of superplasticizer when the surface coverage is up to 1 (m), H 0 is the separation distance between particles without superplasticizer when the surface coverage is 0(m), which is range from 1 to 2 nm in the previous study [17].In this study, H 0 is taken as 1.5 nm.By assuming the adsorption by superplasticizer on cement particle surfaces is uniform single-layer adsorption, it conforms to the Langmuir adsorption model [23], so the average surface coverage of particles can be obtained as the following equation: (4) where Q ads is the adsorption amount of superplasticizer on cement particles (mg/g), Q sat is the saturated adsorption amount of superplasticizer C sol is the concentration of superplasticizer in the supernatant solution (mg/L), and k is the temperature dependent equilibrium constant (L/mg).

Contact interactions
The network structure of particle suspensions is decided by the strength of colloidal interaction and particle contact interaction [16,17,24].According to the first principle analysis, a YODEL model is proposed by Flatt and Bowen for describing the yield stress of paste [16], which is described in Eq. ( 6).(6)  where τ 0 is the yield stress of paste (Pa), F σ ,� is the size distribution function of particles(unitless), R v,50 is the median volume radius (m), ϕ is the volume fraction of solid particles (unitless), ϕ max is the maximum packing density of particles (unitless), ϕ perc is percolating volume fraction.
The particle contact interaction depends on volume fraction of particles and number of contacting particles.Once the value of ϕ reaches ϕ perc , a stable particle net- work will be formed due to the indirect contact occurs between particles.The number of contacting particles in the suspension is related to the PSD of particles [25,26].Therefore, the F σ,∆ /R v,50 in the YODEL model can be used to describe the influence of PSD of particles on the solid network structure of paste [27].However, A 0 and ϕ perc are hard to be acquired by experiment, so the YODEL model cannot be simplified as follow in this study:

Attractive colloidal interactions
The average surface coverage of particles is shown in Fig. 11. the average surface coverage increases steadily with the increasing of flow spread.Compared with the reference paste at the same flow spread, the average surface coverage of paste incorporating RP increases.Therefore, the increasing of yield stress can be ascribed to the improvement of attractive colloidal interactions of paste incorporating with RP. for the paste using different superplasticizer, the average surface coverage varies significantly.The average surface coverage with SP-1 is higher than that of other pastes, but the value with SP-4 is lowest.(7)  The value of 1/H 2 of particles is shown in Fig. 12.At the fixed flow spread, the average surface coverage of the suspending particles in the cement paste increases with the displacement of RP, so that the colloid effect between particles decreases.With the incorporation of RPs, the value of 1/H 2 is increases with the range of 4.52% to 150% with the increase of the incorporation.Besides, for the paste with different superplasticizer to achieve the similar flow spread, the value of 1/H 2 varies for each other.For paste with SP-4, the increment of 1/H 2 compared to reference paste is lower than paste with other superplasticizer, and the value with SP-2 is the highest.Such decrease in 1/H 2 for paste with SP-4 is generally amplified by the significant increasing the superplasticizer dosage for the same flow spread.For paste with 10% RP, the yield stress of paste with SP-4 is higher than that with other superplasticizer, which cannot be explained by the average separation distance.The reason for this phenomenon is that the yield stress of paste is decided both by the colloidal interaction and particle contact effect.

Contact interactions
The contact interactions should be related to the volume fraction and the packing density of solid particles, which can be ascribed as the PSD effect of the particles in the YODEL model.Figure 13 reveals the volume fraction function (φ 3 /φ max (φ max -φ)) in Eq. (7).Because the density of RP is lower than cement, the volume fraction of particle in paste with RP is larger than that of reference paste.The value of φ 3 /φ max (φ max -φ) for paste with 10% RP is a litter lower than that of reference paste.When the RP dosage increases from 10 to 20% and 30%, φ 3 /φ max (φ max -φ) becomes higher than that of reference paste.It can be drawn that the variations Fig. 12 Inverse square of average separation distance of solid particles in pastes of φ 3 /φ max (φ max -φ) are well conformed to the change of yield stress with the increasing of RP replacement.Besides, for the paste with different superplasticizer, the value of φ 3 /φ max (φ max -φ) varies significantly.The value with SP-1 is the largest, SP-4 is the smallest, and the results of SP-2 and SP-3 are close.

Yield stress
From Fig. 14, the difference between average surface coverage ratio of suspending particles with different displacement of RP is related to the specific dosage of superplasticizer for keeping the same flow spread.For different superplasticizer, the average surface coverage ratio varies significantly, SP-4 has the lowest average surface coverage ratio, and SP-1 has the highest value.
The particle contact interactions of the paste with different displacement of RP at a fixed fluidity of (210 ± 5) mm is shown in Fig. 15, which was described as F σ,∆ /R v,50 φ 3 /φ max (φ max -φ).The contact interactions among the particles in the paste increases with the increasing of RP.For SP-1, SP-2 and SP-3, the contact interaction of reference paste is higher than the paste with 10% RP.With the increasing of the displacement of RP, the dosage of superplasticizer increases for the same flow spread of paste.The enhanced PSD effect will be helpful for increasing the packing density of paste, leading to the improvement of the particle contact interaction.Therefore, the particle contact interaction of SP-1 is stronger than that of paste with the other superplasticizer.For paste with SP-4, the particle contact interaction is weaker than that with other superplasticizers, but is higher slightly than the reference paste.The general influence of superplasticizer structure on particle contact interactions is consistent with that on yield stress increase:SP-1>SP-2>SP-3>SP-4.Generally, the yield stress of paste will be strengthened due to the improvement of the colloidal interaction and contact probability among the particles in the network structure.From the above discussions, a relationship between the yield stress of paste and the average separation distance, PSD function and volume fraction function of particles is shown in Fig. 16.Generally, yield stress of paste with different superplasticizer appeared to be well correlated with the and the calculated value of the For different kind of superplasticizer, more fine particles in RP can enhance the PSD function F σ,� /R v,50 moderately.

Conclusions
The effect of the kind of superplasticizer on the yield stress of cement paste with RP was investigated in this study.And the mechanism was analyzed by studying the polymer adsorption and packing density of the mixtures with cement and RP.Some conclusions can be drawn as follows: (1) For a fixed flow spread, the incorporation of the RP increased the dosage of superplasticizer.The yield stress of paste increased with the increasing of RP for all superplasticizer when the incorporating rate was higher than 10%.And the superplasticizer with phosphonic acid group could achieve lower yield stress but more significant increment with the increase of RP than the other superplasticizer with poly-carboxylate.(2) The higher adsorption of superplasticizer by RP resulted in an increase of the average surface coverage of particles in paste, so that the colloidal interaction among particles was mitigated.(3) The packing density of the mixtures increased with the increment of with RP and the dosage of superplasticizer, so that the particle contact interaction was improved.(4) Under the synergism of colloidal interaction and particle contact effect on the particle network structure in paste, the displacement of RP had different influences on the yield stress.The superplasticizer plays a major role when the displacement of RP is 10%, while the particle contact will be more significant on the yield stress when the displacement is higher than 20%.

Fig. 1 Fig. 2
Fig. 1 PSD of the cement, recycled powder and the mixtures

Fig. 5 Fig. 6 Fig. 7
Fig. 5 Dosage of different superplasticizer of pastes for keeping the same flow spread with different displacement of recycled powder

Fig. 10
Fig. 10 Packing density of cement pastes with varied kind of SP

Fig. 11
Fig. 11 Average surface coverage of particles in pastes

Fig. 13
Fig. 13 Value of the volume fraction function

Fig. 14 Fig. 15
Fig.14 Yield stress and average coverage ratio of particles for paste with different SP and fixed fluidity of (210 ± 5) mm

Table 1
Physical performance of cement and recycled powder

Table 2
Physical performance of four superplasticizers