INFLUENCE OF MOLECULAR WEIGHT OF PNS SUPERPLASTICIZERS ON THE PROPERTIES OF CEMENT PASTES CONTAINING DIFFERENT ALKALI CONTENTS

Transcription

1 INFLUENCE OF MOLECULAR WEIGHT OF PNS SUPERPLASTICIZERS ON THE PROPERTIES OF CEMENT PASTES CONTAINING DIFFERENT ALKALI CONTENTS Byung-Gi Kim, Shiping Jiang and Pierre-Claude Aïtcin Dept. of Civil Engineering, University of Sherbrooke, Canada Abstract The interaction between superplasticizer and cement has not been well understood up to now due to its complexity. This study examines one of the physicochemical parameters that affect superplasticizer/cement interaction: the effect of the molecular weight of PNS (polynaphthalene sulfonate) superplasticizer on the properties of cement pastes containing different alkali content. In addition, the superplasticizer/cement interaction is discussed with respect to the adsorption of PNS on cement and its hydration. The influence of molecular weight of PNS superplasticizers on the rheological behavior of cement pastes depends greatly on cement alkali content: high-molecular-weight PNS is more effective in fluidizing pastes made with high-alkali cement than low-molecularweight PNS. However, PNS molecular weight has little effect in fluidizing pastes made with low-alkali cement. The dispersion effect of PNS superplasticizer is determined by its adsorption on cement particles, which is related to cement alkali content: the amount of PNS adsorbed is much higher with low-alkali-cements than with high-alkali-cements. In high-alkali cement, the addition of PNS superplasticizer initially retards cement hydration in the first few hours then accelerates it. Differences in PNS superplasticizer molecular weight do not appear to significantly affect cement hydration in high-alkali cement. In low-alkali cement, the induction period is increased by the addition of lowmolecular-weight PNS, but not by the addition of relatively high-molecular-weight PNS. The ionic concentration in solution is determined by the characteristics of cements, not by the molecular weight of PNS. 97

2 1. Introduction Concrete technology uses superplasticizers to improve the rheological properties of concrete. However, the interaction between superplasticizer and cement has not been well understood due to its complexity (1). The dispersing effect of a PNS superplasticizer has generally been ascribed to the development of electrostatic charges on the cement particles as determined by the electrokinetic or zeta potential. In other words, attractive forces existing between cement particles suspended in water cause flocculation. These attractive forces can be neutralized by the adsorption of PNS superplasticizer on the surface of cement particles (2). The consequences of the superplasticizer/cement interaction depend on the physicochemical parameters of the cement and the superplasticizer (2,3): chemical nature of the superplasticizer; average molecular weight and its distribution; actual dosage and admixture introduction method; cement type, including fineness, phase composition, and alkali/sulfate content; form of calcium sulfate (gypsum, hemihydrate, anhydrite, synthetic calcium sulfate). The effect of molecular weight of PNS superplasticizers on the properties of cement pastes is one physicochemical parameter of superplasticizer/cement interaction that has been thoroughly studied (4-7). Ferrari et al. (4) determined that commercial PNS superplasticizer fluidizes best when its average molecular weight is in the 6-8 range. At higher molecular weights, the fluidizing effect decreases in C 3 A-rich cement, but not in C 4 AF-rich cements. The complex relationship between adsorption, zeta potential and molecular weights, on the one hand and the fluidizing effect, on the other hand, has been extensively examined by Basile et al. (5). Their results showed that the higher the PNS molecular weight the better the fluidizing effect. Ferrari et al. (4) and Basile et al. (5), however, used only cements that were relatively high in alkalis (Na 2 O eq..68%-.82%) and did not take into account the cement alkali content. Another paper (8) pointed out that soluble alkalis play a major role in controlling the fluidity and fluidity loss of cement pastes containing PNS superplasticizer. This study investigates the effect of PNS superplasticizer molecular weight on the properties of cement pastes containing different alkali content. In addition, the superplasticizer/cement interaction is discussed with respect to the adsorption of a PNS superplasticizer on cement and its hydration. 2. Experimental Materials The chemical/phase compositions and fineness of the cements used in this study are presented in Table 1. Cements E and F are CSA Type 1 (ASTM Type I with a 5% max. limestone filler content), while Cements L and N are ASTM Type I. These 4 cements present a wide range of C 3 A composition. Their C 3 A contents vary from 6% for Cement F up to 11% for Cement N, according to Bogue composition, and their Na 2 O equivalent contents vary from.31% for Cement N to.92% for Cement E. Three sodium salts of a polynaphthalene sulfonate superplasticizer with different molecular weights were obtained from a Canadian manufacturer. The three 98

3 superplasticizers were characterized by their average molecular weight and their distribution measured by gel permeation chromatography, as illustrated in Figure 1. Other parameters were shown in Table 2. In the following notation will be used for low-molecular-weight and similarly and for medium-molecular-weight and high-molecular-weight. Table 1. Chemical and phase composition of the cements Chemical Cement Composition (%) E F L N SiO Al 2 O Fe 2 O CaO MgO SO K 2 O Na 2 O Na 2 O eq Na2O soluble eq.* C 3 S C 2 S C 3 A C 4 AF Blaine (m 2 /kg) * Measured by ICP at W/C=.35 and 3 minutes after mixing Test Methods Measurement of average-molecular-weight of PNS and its distribution Average-molecular-weight of PNS and its distribution were measured by gel permeation chromatography, using two columns system, namely Ultrahydrogel 25 and Ultrahydrogel linear columns, and refractive index flow detector (RID-1A). The columns were packed with hydroxylated polymethacrylated-based gel and its dimension was 7.8 mm 3 mm. The eluent was a.1 M NaNO3/.1 M Acetonitrile (8:2) and its flow rate was.3.8 ml/min. A Shondex Standard P-82 kit produced by Showa Denko K.K. was used as standards for the calibration of the columns. It is made up of agglomerated particles, which consist of polymaltotriose and a linear macromolecular polysaccharide having convenient values of molecular weight. A PNS sample was diluted to 1:1 with deionized water, heated for 5 minutes at 7 C, and filtrated with.45 m membrane. Chromatographic data detected were processed by a computerized system, equipped with specific programs for the calibration of the molecular weight value and the molecular weight distribution of PNS superplasticizer. 99

4 Table 2. Characteristics of the three PNS superplasticizers used Characteristic of PNS PNS Superplasticizer Average molecular weight(g/mol) Weight (Mw) Number (Mn) Dispersity (Mw/Mn) % of Mw < 1kD % of Mw > 1kD Viscosity (cps) at 25 C % solids ph (1% solution) Sulfate (%) Specific gravity Ca (mmol/l) S (mmol/l) Na (mmol/l) K (mmol/l) Paste Mixing Procedure The cement pastes were prepared at a constant water/cement ratio (W/C) of.35. Superplasticizer was added to the mixing water, then the solution was added to the cement. The cement pastes were mixed manually for 1.5 minutes, then with a high-speed hand-held mixer for 2.5 minutes in order to obtain a well-dispersed grout. The mixing procedure was carried out under controlled temperature of 25 1 C. The superplasticizer dosage is expressed as dry weight (weight PNS superplasticizer solids relative to the weight of cement):.6% for Cements E, F, and L; 1.% for Cement N. Mini Slump The mini slump test (9) was used to assess the effect of the different PNS superplasticizers on cement-paste fluidity. The procedure involves transferring the cement paste to the mini-slump cone, then lifting the cone smoothly and quickly. The area of the paste spread on a polyethylene plate is measured and expressed in cm 2. Adsorption Measurement PNS adsorption on the surface of cement particles was evaluated by measuring the amount of PNS in the solution extracted from the fresh cement pastes, expressed as a percentage (mass of PNS present in the extracted solution relative to the mass of added PNS superplasticizer). The solution was extracted centrifugally at 4 rpm at 24 C. The original extracted solution was diluted with deionized water to obtain adequate PNS concentration. The PNS concentration in solution was measured at the peak absorbance wavelength between 276 and 294 with a Hewlett-Packard diode array UV-visible spectrometer. Heat of Hydration Measurement The development of heat of hydration was monitored with an adiabatic calorimetry described by Simard et al. (1). After the cement pastes were mixed, 125 g of cement 1

5 paste was accurately weighed and placed into a polyethylene bottle, which was then set into the calorimetric vessel containing 1 g of water. The copper tube-thermistor assembly was introduced into the sample and the calorimetric vessel was sealed before placing it in the thermoregulated water bath. The temperature of the cement paste sample was monitored at one-minute intervals by reading the resistance value of the thermistor probe interfaced to a computer. Measurement of Ionic concentration in solution The ionic concentration, Ca, S, Na and K, present in solution in fresh cement pastes were measured using the ICP (Inductivity Coupled Plasma) method. The measurements were carried out 5 and 6 minutes after mixing. The solution was extracted centrifugally at 4 rpm at 24 C. The original extracted solution was diluted to 1:15 with 5% HCl. The ionic concentrations from cement were calculated by subtraction of the ionic concentrations originated from PNS superplasticizer. 3. Results and Discussion Measurement of average molecular weight and molecular weight distribution of PNS GPC analysis has been used in several studies (4,5) on superplasticizers for the purpose of the separation of polymers according to their molecular sizes and shapes. Using this technique, it is possible to determine the average molecular weight and the molecular weight distribution of polymers. The determination of the molecular weight is obtained as a result of a steric exclusion of the polymer molecules by the pores of the gel of the column. The principle of this separation is based on the fact that large molecules have less probability to access to the porosity of the gel and are eluted rapidly through gel particles. Whereas, small molecules can permeate and be retained in the pores and therefore, are eluted later. 11

6 1 8 6 Mw = 6 Mn = Differential and Integral Curves (%) Mw = 14 Mn = Mw = 16 Mn = Molecular Weight (Mw), g/mol Figure 1. Average molecular weight of three PNS and its distribution The molecular distribution of three PNS used in this study is shown in Figure 1, which is processed by computerized system from the data, detected. The superplasticizer includes 43% of low-molecular fraction (Mw < 1kD), while 21% and 19% of lowmolecular fraction (Mw < 1kD) are found in,, respectively. This lowmolecular fraction might consist of unreacted raw material, salt of sulfate (calcium and/or sodium), naphthelene sulfonate monomer, naphthalene disulfonate monomer, and low condensates of naphthalene sulfonate. The relatively high-molecular-weight PNS ( and ) present difference in both weight average-molecular-weight (Mw) and number average-molecular-weight (Mn), and they have quite different viscosity in Table 2. 12

7 Mini Slump The fluidizing effect on the different cements, shown in the mini-slump test (Figure 2), was plotted as a function of hydration time. Results show significant variance in the behavior of the tested cements. Medium () or high-molecular-weight () PNS superplasticizer is more effective than low-molecular-weight () PNS for fluidity of cement pastes made with high-alkali-cements (Cement E and F): the initial slump at 5 minutes is high and the slump loss is low up to 9 minutes. The foregoing observations can be readily understood by the fact that the larger molecules convey a higher surface charge to the particles, increasing electrostatic repulsive force between particles, on the other hand, their adsorption on the particles inhibits particle-particle contact by shorter range repulsive force due to the steric hindrance (3). Slump area (cm 2 ) Slump area (cm 2 ) Cement E,.6% PNS Time (min) Cement L,.6% PNS Time (min) Cement F,.6% PNS Time (min) Cement N, 1.% PNS Time (min) Figure 2. The effect of PNS molecular weight on the fluidity with different alkali content cements Low-alkali cements (Cements L and N) interestingly show different fluidity trends. In Cement L the initial slump was quite high, but it decreased rapidly after 15 minutes; its fluidity is independent of PNS molecular weight. Cement N recorded a very low initial slump, although the PNS dosage was 1.%. superplasticizer yields slightly better fluidity than the and varieties in Cement N. The above results will be discussed further in relation with the adsorption of different molecular weight PNS. Therefore, the findings herein correspond to those of Basile et al. (5) only in the case of high-alkali cements: the higher the molecular weight of the PNS, the better the fluidizing effect. 13

8 The results of the mini slump test indicate clearly that the influence of molecular weight of PNS superplasticizers on the rheological behavior of cement pastes depends greatly on cement alkali content; relatively high-molecular-weight (, ) PNS is more effective in fluidizing pastes made with high-alkali cement than low-molecular-weight PNS (), however, PNS molecular weight has little effect in fluidizing pastes made with low-alkali cement. Adsorption Measurement Figure 3 shows that the amount of PNS adsorbed on high-alkali cements (Cements E and F) is approximately 5-55 % with the superplasticizer and 6-65% with the and superplasticizers at 5 minutes and is increased by 5-1% at 6 minutes. As shown also in Figure 3, the amount of PNS adsorbed on cement L is approximately 6% with the superplasticizer and 65% with the and superplasticizers at 5 minutes. However this amount is significantly increased to 8 % for, 9% for and respectively at 6 minutes. In the case of cement N, the amount of PNS adsorbed is already high at 5 minutes due to high C 3 A content: 8% for, 9% for and respectively and is slightly higher at 6 minutes. It seems that the difference of amount adsorbed between 5 and 6 minutes in low-alkali-cements could be related to the rate of adsorption of C 3 A and C 4 AF (11) and not to the PNS molecular weight. C 3 A content greatly influences the adsorption of PNS on the cement particles in lowalkali-cement but not significantly in high-alkali cement. The results obtained on highalkali-cements imply that the higher amount of soluble alkali sulfate inhibits efficiently the adsorption of PNS on C 3 A and C 4 AF. On the contrary, the amount of soluble alkali sulfate in low-alkali-cements is not high enough to inhibit the adsorption of PNS on C 3 A and C 4 AF (11). This partially agrees with Bonen et al s results (12), who showed that cement with high specific surface area and high C 3 A content has a large adsorption capacity. 14

9 PNS adsorbed (%) PNS adsorbed (%) min 6 min Cement E.6% PNS Cement L.6% PNS Cement F.6% PNS Cement N 1.% PNS Figure 3. The effect of PNS molecular weight on cement adsorption Approximately 1% less are adsorbed on the cement particles than or in all tests. This may be due to the fact that the molecular weight distribution in has a higher fraction of low-molecular-weight polymers than or. This result confirms that monomers, dimers, and probably other low-molecular-weight polymers are more likely to remain in solution, while higher molecular weight polymers are adsorbed more rapidly on cement particles (6,7). The relationship between the fluidity of cement pastes and the amount of PNS adsorbed could be drawn in Figure 2 and 3 that in high-alkali-cements, the adsorption of PNS on C 3 A and C 4 AF is inhibited by alkali sulfate in the cement, this results in the higher dispersion of C 3 S by the high amount of and superplasticizer remaining in solution due to their higher electrostatic and short range repulsive force than. However, the small amount of alkali sulfate existing in low-alkali-cement does not inhibit efficiently the PNS adsorption on C 3 A and C 4 AF, therefore, C 3 S is not well dispersed at 6 minutes by the small amount of PNS remaining in solution (13). In lowalkali-cement, and do not increase the fluidity of C 3 S compared with because less amount of it exits in solution after initial adsorption on C 3 A and C 4 AF, moreover, and remaining in solution might be a lower fraction of its molecular weight distribution which has less electrostatic and short range repulsive force. 15

10 It can be concluded that the dispersion effect of a PNS superplasticizer is determined by the amount adsorbed on cement particles which is related to cement alkali content: the amount of PNS adsorbed is much higher with low-alkali-cement than with high-alkalicement. Study of Cement Hydration It is well known that the presence of a superplasticizer in a fresh concrete delays the cement hydration (14-15). Figure 4 gives the curves of heat evolution rate as a function of hydration time up to 24 hours for PNS superplasticizers with different molecular weights. These graphs indicate that the addition of a PNS superplasticizer to high-alkali cement (Cement E and F) retards hydration for a few hours then accelerates it slightly. This phenomena could be explained by the fact that the alkali sulfate present in highalkali-cements hinders the adsorption of a PNS on the aluminate phase, permitting larger adsorption on the silicate phase such as C 3 S and C2S (13), which delays the early hydration. However, alkali sulfate accelerates the hydration in the acceleration period and increases the rate of heat evolution towards its peak (16). The effect of PNS with different molecular weights on cement hydration is not significant in this case. When or superplasticizer is added to low-alkali cements (Cement L and N), the induction period is the same as with pastes not containing any PNS. But when a superplasticizer is added, the induction period increased by 2 hours with Cement L and by 4 hours with Cement N. This phenomena could be explained as follows: after early adsorption at 6 minutes, 2% of the superplasticizer remained in solution, whereas only 1% of the and remained in solution, as measured with the adsorption test. Nawa et al (11) found that superplasticizer molecules are first adsorbed onto C 3 A and C 4 AF, and then on C3S. Therefore it seems that molecules that remain initially in the solution are preferentially absorbed on C 3 S. Moreover, some authors (6,7) indicate that monomers, dimers, and probably other low molecular polymers are more likely to remain in the solution, while polymers having a higher molecular weight are adsorbed rapidly on the cement particles. The PNS molecules that remain in solution in the fresh cement pastes might be, therefore, the monomers, dimers, and oligomers found in commercial PNS. Therefore, it could be concluded that the induction period is increased by the addition of in low-alkali cements because there are more low-molecular-weight PNS in solution in the fresh cement pastes that affect C 3 S hydration. 16

11 6 5 Cement E 6 5 Cement F Temperature ( C) 4 4 Reference Reference 3 3.6% PNS.6% PNS Cement L Cement N Reference Reference % PNS 1 % PNS Time (hour) Figure 4. The effect of PNS molecular weight on the cement hydration Piotte (17) has shown with different fraction of PNS that the low-molecular-weight (3kD < NaPNS < 1kD) induces more pronounced retardation effects than the highmolecular-weight (NaPNS > 1kD).On the contrary, Kondo et al. (14) found with commercial PNS that the higher the molecular weight polymers, the greater the retardation effect. However, in this study, the effect of PNS with different molecular weights on cement hydration is not significant in high-alkali-cements, while in lowalkali cements, the induction period is increased by the addition of low-molecularweight PNS (), but not by the addition of relatively high-molecular-weight PNS ( and ). Therefore, it can be concluded that the retardation of cement hydration is determined by the amount of PNS adsorbed on C 3 S after initial PNS adsorption on C 3 A and C 4 AF which is related mainly to soluble alkali content, PNS molecular weight, and marginally C 3 A content. Ionic concentration in solution in fresh cement pastes As expected, Figure 5 shows that in high-alkali-cements (Cement E and F), the ionic concentration of Na, K and S is much higher than that in low-alkali-cement. It seems that most sulfur ions are originated from alkali sulfate because of its higher solubility than calcium sulfates. Moreover, calcium ion concentration is depressed by alkali and sulfate ions in high-alkali-cements (18): 24 mmol/l for Cement E, 22.5 mmol/l for Cement F, 43 mmol/l for Cement L and 27 mmol/l for Cement N at 5 minutes, respectively. Sulfur concentration is decreased from 5 minutes to 6 minutes probably due to the formation of ettringite, whereas Na and K concentration are slightly increased with time. The relationship between ionic strength and the flow and flow loss was studied by D. Bonen et al (12). They showed that the higher the ionic strength, the greater the flow 17

12 loss and the lower the flow at 12 minutes. Ionic strength is primarily determined by the presence of alkali sulfate or soluble calcium sulfate and alkalis (12), therefore, the ionic strength in high-alkali-cement should be higher than that in low-alkali-cement. However, as shown in Figure 2, the fluidity of pastes made with high-alkali-cement containing high-molecular-weight PNS was higher and fluidity loss was lower than pastes made with low-alkali-cement. These results, which are opposite to those of D. Bonen, show that the higher the ionic strength, the greater the initial fluidity with highmolecular-weight PNS and the lower the flow loss up to 9 minutes. As found in another paper (8), there is an optimum soluble alkali content for fluidity and fluidity loss in cements tested. It seems, therefore, that under the optimum soluble alkali content, ionic strength originating from alkali sulfate can increase the initial fluidity and decrease the flow loss, while above this optimum, it can cause a flow loss by synergetic effect resulting in the precipitation of syngenite (18, 19) or compression the electric double layer (13). As shown in Figure 5, the ionic concentrations in solution are determined rather by the characteristic of the tested cements than by the molecular weight of the PNS studied. 4. Conclusion PNS with relatively high molecular weight ( and ) yielded better fluidity in the high-alkali cement pastes than did PNS with low-molecular-weight (). PNS molecular weight had no impact on the fluidity of pastes made with low-alkali cement. The adsorption of on cement particles is about 1% less than that of and superplasticizers. The adsorption behavior of PNS is closely related to cement alkali content. The amount of PNS adsorbed on cement particles is much higher with low-alkali cements than with high-alkali cements. C 3 A content greatly influences the adsorption of PNS on the cement particles in low-alkali cements, but not significantly in high-alkali cement. The dispersion effect of PNS superplasticizer is determined by the amount absorbed on cement particles, which is related to cement alkali content. In high-alkali cements, the addition of PNS superplasticizer initially retards cement hydration in the first few hours then accelerates it slightly. Differences in PNS superplasticizer molecular weight do not appear to significantly affect cement hydration 18

13 mmol/l Cement E Ca S 5 min 6 min Na K mmol /L Cement F Ca S Na K 45 Cement L Ca S Na K mmol/l Cement N Ca S Na K mmol/l 3 15 Figure 5. The effect of different molecular weights of PNS on the ions concentration in solution in high-alkali cements. In low-alkali cements, the induction period is increased by the addition of low-molecular-weight PNS (), but not by the addition of relatively high-molecular-weight PNS ( and ). Finally the ionic concentrations in solution are determined by the characteristics of the tested cements rather than by the molecular weight of the studied PNS. 5. Acknowledgments The authors would like to thank Kyunggi Chemicals Ind., Co. Ltd. of Korea for its financial support, Mr. Doo-Ho Lee of the quality control laboratory of Kyunggi 19

14 Chemicals Ind., Co. Ltd for the measurement of molecular weight of PNS, Handy Chemicals of Canada for supplying the PNS samples of different molecular weight, and Ms. A. Lemieux, Depart. of Chemistry, Université de Sherbrooke for carrying out the ICP analysis. 6. References 1. Aitcin, P.-C., Jolicoeur, C., and MacGregor, J.G., Superplasticizers: How They Work and Why They Occassionally Don't, Concr. Int., 16 (5) (1994) Collepardi, M., Ramachandrane, V.S., Effect of Admixtures, 9 th Int. Congress on the Chemistry of Cement, 1 (1992) Jolicoeur, C., Nkinamubanzi, P.-C., Simard, M.-A. and Piotte, M., Progress in understanding the functional properties of superplasticizers in fresh concrete, 4 th CANMET/ ACI International Conference on Superplasticizers and Other Chemical Admixtures in Concrete, ACI SP-148 (1994) Ferrari, G., Basile, F., Dal Bo, A. and Mantoni, A., The Influence of the molecular weight of beta-naphthalene Sulfonate based polymers on the Rheological properties of cement mixes, Il Cemento, 83 (1986) Basile, F., Biagini, S., Ferrari, G., and Collepardi, M., Influence of Different Sulfonated Naphthalene Polymers On the Fluidity of Cement pastes, 3 rd CANMET/ACI International Conference on Superplasticizers and Other Chemical Admixtures in Concrete, ACI SP-119 (1989) Nkinamubanzi, P.-C., Effet des dispersants polymériques (superplastifiants) sur les propriétes des suspensions concentrées et des pâtes de ciment, Ph.D thesis, Université de Sherbrooke (1993), in French. 7. Ferrari, G., Cerulli, T., Clemente, P., and Dragoni, M., Adsorption of Naphthalene Sulfonate Superplasticizers by cement particles through Gel Permeation Chromatography, 5th CANMET/ACI International Conference on Superplasticizers and Other Chemical Admixtures in concrete, ACI SP-173 (1997) Jiang, S.P., Kim, B.G. and Aïtcin, P.-C., Importance of Adequate Soluble Alkali Content to Ensure Cement/Superplasticizer Compatibility, Cem. Concr. Res. (1998), in press. 9. Kantro,D.L., Influence of water reducing admixtures on the properties of cement pastes a miniature slump test, Cem. Concr. Aggr., 2 (2) (198) Simard, M.-A., Nkinamubanzi, P.-C., Jolicoeur, C., Perraton, D., and Aïtcin, P.-C., Calorimetry, Rheology and Compressive strength of Superplasticized Cement Pastes, Cem. Concr. Res., 23 (4) (1993) Nawa, T. and Eguchi, H., Effect of cement characteristics on the fluidity of cement paste containing an organic admixture, 9 th International Congress on the Chemistry of Cement, 14 (1992) Bonen, D. and Sarkar, S.L., The superplasticizer adsorption capacity of cement pastes, pore solution composition, and parameters affecting flow loss,cem. Concr. Res., 25 (7) (1995) Nawa, T., Eguchi, H., and Fukaya, Y., Effect of alkali sulfate on the rheological behavior of cement paste containing a superplasticizer, 3 rd CANMET/ACI Intern. Conf. on Superplasticizers and Other Chemical Admixtures in Concrete, ACI SP- 119 (1989) Kondo, R., Daimon, M., and Sakai, E., Interaction between cement and organic polyelectrolytes, Il Cemento, 75 (3) (1978)

15 15. Older, I. and Becker, T., Effect of some liquefying agents on properties and hydration of Portland cement and tricalcium silicate pastes, Cem. Concr. Res. 1 (3) (1987) Jawed, I. and Skalny, J., Alkalies in Cement: A Review II. Effects of Alkalies on Hydration and Performance of Portland Cement, Cem. Concr. Res. 8 (1) (1978) Piotte, M., Caractérisation du poly(naphthalènesulfonate) Influence de son contre- ion et de sa masse molaire sur son interaction avec le ciment, Ph.D. thesis, Université de Sherbrooke (1993), in French. 18. Rechenberg, W. and Sprung, S., Composition of the solution in the hydration of cement, Cem. Concr. Res., 13 (1) (1983) Odler, I. and Wonnemann, R., Effect of alkalies on portland cement hydration: II. Alkalies present in form of sulfates, Cem. Concr. Res., 13 (6) (1983)