Effects of Saccharide Set Retarders on the Hydration of Ordinary Portland Cement and Pure Tricalcium Silicate Linghong Zhang, z Lionel J. J. Catalan, w,z Raymond J. Balec, y Andrew C. Larsen, y Hassan Haji Esmaeili, z and Stephen D. Kinrade y z Department of Chemical Engineering, Thunder Bay, ON P7B 5E1, Canada y Department of Chemistry, Lakehead University, Thunder Bay, ON P7B 5E1, Canada The effects of aliphatic sugar alcohols (e.g., threitol, xylitol, sorbitol) on the hydration of tricalcium silicate (C 3 S) and ordi- nary portland cement (OPC) were investigated and compared with those of sucrose, a well-established cement set retarder. Only sugar alcohols which contain threo diol functionality re- tarded the setting of C 3 S and OPC, their efficacy increasing with the number of threo hydroxy pairs and, to a smaller extent, with the overall population of hydroxy groups. None, however, were as effective as sucrose. The initial and final setting times increased exponentially with the concentration of saccharide, although the hydration of OPC was less inhibited than that of C 3 S. Saccharides function as ‘‘delayed accelerators,’’ that is, cement hydration is first inhibited and then proceeds faster than in saccharide-free cement. This behavior is consistent with the theory that the induction period is controlled by slow formation and/or poisoning of the stable calcium silicate hydrate (CSH) nuclei. The early inhibiting influence of saccharides on CSH precipitation is apparently stronger than on the growth of crys- talline calcium hydroxide. Saccharides did not negatively affect the degree of hydration and compressive strength of fully set OPC paste; on the contrary, sorbitol yielded modest increases. I. Introduction S ET-RETARDING admixtures (‘‘retarders’’) are used to delay the onset of cement hydration without affecting the material’s long-term mechanical properties, that is, to compensate for high temperatures or for delays between cement mixing and place- ment. Common organic retarders include sugars, lignosulfon- ates, and hydroxycarboxylic acids. Sucrose is one of the most effective; addition of 0.075 wt% to ordinary portland cement (OPC) increases the set time from approximately 2.5–31 h. 1 Taplin 2 observed that organic retarders typically possess one or more pairs of oxygen atoms that are located on adjacent or closely neighboring carbons and thus capable of complexing metal ions. The retarding effect of sugar acids increases with the number of a- or b-hydroxycarboxylic groups and, in the case of aromatic molecules, with the number of vicinal dihydroxy pairs. 3 As retarders are effective at very low concentrations, they are generally thought to function via adsorption at active nucleation/growth sites on the cement grains rather than by complexation of solution species. Indeed, EDTA, a good che- lator of Ca (aq) 21 , exhibits little set-retarding influence. 4,5 More- over, cement dissolution continues even while the retarder inhibits the formation of hydration products. 1,4–11 Thomas and colleagues 4,5 proposed that organic set retarders adsorb onto hydrated tricalcium silicate (C 3 S) surfaces at pendant – Ca(OH) groups to form monodentate RO y Ca(OH) ‘‘half salts,’’ thus poisoning product nucleation sites (although biden- tate interaction seems more probable 2 ). Pure C 3 S is often used as a model for studying cement hydration as it is usually the dominant clinker constituent and exerts, by far, the greatest in- fluence on the strength and other characteristics of the hydrated paste. Young et al. 12 and Milestone 13 preferred the term ‘‘delayed accelerator’’ over ‘‘retarder’’ because, once the inhibition period has ended, hydration proceeds at a faster rate than when no set retarder has been added. The influence of organic retarders is generally diminished in Al-rich cements owing to their dispro- portionate adsorption on aluminate phases. 1 For example, su- crose causes a maximum set delay for C 3 S when it is added directly to the mixing water 14 and, for OPC, when it is added 2–4 min after mixing (i.e., once most of the aluminates have reacted with gypsum). 1 Sorbitol, a neutral six-carbon sugar alcohol, was recently shown to act as a set retarder for OPC; 0.40 wt% sorbitol in- duced a set delay of 2 days, equivalent to the effect of 0.15 wt% sucrose. 15 Unlike sucrose, however, sorbitol conferred a slight increase in the final degree of hydration and compressive strength. Sorbitol is currently used as a water-reducing plastic- izer 16 (an admixture that confers flowability to concrete paste so that excess water can be kept to a minimum, thus decreasing final porosity and thereby ensuring high mechanical strength and durability). Cody et al. 16 observed that sorbitol also pre- vents nucleation and growth of ettringite crystals. The chemical mechanism underlying the role of organic set retarders is not well understood. To determine explicitly the in- fluence exerted by different possible aliphatic hydroxy group configurations, we compared the effects of sugar alcohols, a hitherto unexplored class of set retarders, with those of sucrose on the hydration of C 3 S and OPC pastes. II. Experimental Procedure (1) Preparation of Pastes The cements used in this study were ASTM C 150 Type I OPC (CEMEX, Charlevoix, MI) and pure tricalcium silicate (CTL Group, Skokie, IL). The manufacturers’ specifications for each product are listed in Table I. The additives—threitol, erythritol, adonitol, arabitol, mannitol, xylitol, sorbitol, sucrose, and cat- echol—were obtained from Sigma-Aldrich (Oakville, Canada) (all  99%). Kilogram amounts of OPC were mixed with precooled (101C) deionized water (18 MO  cm) at a water-to-cement mass ratio R 5 0.40 in a plastic bowl over an ice-water bath. The mixtures were stirred with a plastic spoon for about 7 min until they were L. Struble—contributing editor Based in part on the thesis submitted by L. Zhang for the M.Sc. degree in Environ- mental Engineering, Lakehead University, Thunder Bay, ON, 2007. This work was supported by the Natural Sciences and Engineering Research Council of Canada. w Author to whom correspondence should be addressed. e-mail: lionel.catalan@ lakeheadu.ca Manuscript No. 25897. Received February 23, 2009; approved August 14, 2009. J ournal J. Am. Ceram. Soc., 93 [1] 279–287 (2010) DOI: 10.1111/j.1551-2916.2009.03378.x r 2009 The American Ceramic Society 279