Effects Of Aggregate, Water/Cement Ratio, And Curing On .

44TRANSPORTATION RESEARCH RECORD 1335Effects of Aggregate, Water/CementRatio, and Curing on the Coefficient ofLinear Thermal Expansion of ConcreteGABRIEL D. ALUNGBE, MANG TIA, AND DAVID G. BLOOMQUISTA study to determine the coefficient of linear thermal expansionof concrete was conducted using concrete mixtures prepared withthree types of coarse aggregates: porous limestone, dense limestone, and river gravel. A Type II portland cement was used atcontents of 508, 564, and 752 lb/yd 3 and water/cement ratios of0.53, 0.45, and 0.33, respectively. The concrete specimens weremoist-cured and tested at 28 and 90 days. The concrete madewith the porous limestone had the lowest coefficient of linearthermal expansion (5.42-5.80 x 10-6 in. 2/ F), whereas the concrete made with river gravel had the highest (6.49 - 7.63 x 10·6in. 2/ F). The concrete made with dense limestone had an intermediate coefficient (5.82 -6.14 x 10-6 in. 2/ F). The watersaturated concrete specimens had a lower coefficient of linearthermal expansion than the oven-dried specimens. The coefficientof oven-dried concrete decreased with moist-curing time. No significant difference between the 28-day and 90-day moist-curingwas observed in the water-saturated concrete specimens.The coefficient of linear thermal expansion of concrete islisted in the literature as varying from 4 to 8 millionths/ F (7 .2to 14.4 millionths/0 C). Data on Florida concretes, which aremade of predominately porous limestone, are lacking. In absence of actual data, a value of 6 millionths/ F is usually assumed. This could result in an error of 30 percent in thecoefficient of linear thermal expansion and errors of morethan 100 percent in the computed maximum thermal-loadinduced stresses in concrete pavements. In view of the aforementioned reasons, the research reported here was started toobtain the needed data to be used in modeling and analysisof concrete pavement response and performance.The coefficient of linear thermal expansion of concrete hasbeen reported to be affected primarily by factors such as typeand amount of aggregate and moisture content, as well as bythe type and amount of cement and concrete age (J-3). However, the effect of the moisture content on the coefficient oflinear thermal expansion applies only to the paste componentas reported by one researcher (4,5). The variation of thecoefficient of linear thermal expansion of cement paste ismuch greater than that of concrete (6). Orchard (7) pointedout that the coefficient of linear thermal expansion is notaffected by drying wet-cured concrete specimens. Meyers (8)observed that the coefficient of linear thermal expansion increases with an increase in the cement content of the concretemix and that the coefficient is dependent on the quantity oftricalcium silicate in the cement. He reported that cementwith low tricalcium silicate had a low volume change withtemperature variation. The type of cement does not greatlyaffect the coefficient of linear thermal expansion of concrete(7,8). Orchard (7) reported that the time of wet-curing (i.e.,in water) has little effect on the coefficient of linear thermalexpansion, whereas the time of dry-curing (i.e., in air) haslittle effect up to 3 months and tends to reduce the coefficientslightly between 3 months and 1 year.MATERIALSThe following materials and variables were used to preparethe concrete specimens for the study:1. Three types of aggregates were used: a porous limestone(Florida Brooksville aggregate), a river gravel, and a denselimestone (Alabama aggregate). The aggregates had a maximum size of % in. because of the small size of the concretetest specimens. Although a maximum size of% in. might alsohave been used, the% in. size was used to minimize the effectsof specimen size and facilitate consolidation of the fresh concrete in the specimen molds. The physical properties of theaggregates are presented in Table 1. The fine aggregate usedfor all mixtures was a fine sand from Goldhead, Florida. Thephysical properties for th

thermal expansion than the oven-dried specimens. The coefficient of oven-dried concrete decreased with moist-curing time. No sig nificant difference between the 28-day and 90-day moist-curing was observed in the water-saturated concrete specimens. The coefficient of linear