Available online at www.sciencedirect.

com

Cement & Concrete Composites 30 (2008) 444454 www.elsevier.com/locate/cemconcomp

Resistance to external sodium sulfate attack for early-opening-to-trac Portland cement concrete
Nader Ghafoori
a

a,*

, Hamidou Diawara a, Shane Beasley

Department of Civil and Environmental Engineering, University of Nevada, Las Vegas, 4505 Maryland Parkway, Box 454015, Las Vegas, NV 89154-4015, USA b Department of Civil and Environmental Engineering, Tennessee Tech University, P.O. Box 5015, Cookeville, TN 38505, USA Received 31 July 2006; received in revised form 5 April 2007; accepted 4 May 2007 Available online 18 May 2007

Abstract The results of a study on sulfate resistance of early opening-to-trac concretes with accelerating admixture, also known as fast-track concretes (FTC), are presented. A total of 11 concretes made with four dierent cement factors and three dierent cement types were investigated at both opening and maturity (28 days) ages. Upon curing, the test samples were immersed in a 5% sodium sulfate solution. Length change, mass loss, and compressive strength were monitored for a period of 270 days to evaluate the performance of the test specimens exposed to severe sulfate attack. The inuence of immersion period, curing age and cement type and factor on bulk characteristics and sulfate resistance were evaluated. The study produced FTC with excellent sulfate resistance. Length change of FTC incorporating accelerating admixture increased with increasing immersion age and stabilized within 3 months from initial contact. No mass of concrete residues nor notable strength loss were found in any opening time fast track concretes used in the investigation. 2007 Elsevier Ltd. All rights reserved.
Keywords: Fast track concrete; Accelerating admixture; Sulfate resistance; Gypsum; Ettringite; Strength; Curing; Opening time; Maturity

1. Introduction Sulfate attack occurs when components of cement paste come into contact with sulfate ions introduced from external sources into the matrix. The result is a chemical reaction that can have detrimental eects on the concrete. Sulfates are found in a variety of sources. Groundwater, high clay-content soils, seawater, organic materials in marshes, mining pits, and sewer pipes all have potential to contain sulfates. Common sulfates include calcium, magnesium, sodium, potassium, and ammonium [1]. There are two ways in which sulfate attack can compromise the integrity of concrete: expansion of the cement matrix, and progressive loss of strength and mass [2].
*

Expansion due to sulfate attack is a three-step process. First, ettringite forms in the cement matrix from the result of reactions between the tricalcium aluminate in the Portland cement and sulfate ions from internal or external sources or both. More specically, calcium sulfate dihydrate (gypsum) combines with C3A to form 6-calcium aluminate trisulfate hydrate (ettringite) [1]. The reaction is shown below: 3CaO Al2 O3 3CaSO4 2H2 O 26H2 O
" C3 A " Gypsum

! 3CaO Al2 O3 3CaSO4 32H2 O

Corresponding author. Tel.: +1 702 895 3701; fax: +1 702 895 3936. E-mail address: [email protected] (N. Ghafoori).

Next, monosulfoaluminate is formed. Specically, sulfate ions from the ettringite react with the remaining C3A to form tetracalcium aluminate monosulfate-12-hydrate

0958-9465/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.cemconcomp.2007.05.003

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454

445

(monosulfoaluminate) [4]. The chemical reaction is given below:

In the third step, ettringite is formed again when the monosulfoaluminate is brought into contact with a new source of sulfate ions [1]. This reaction is shown below: 4CaO Al2 O3 SO3 12H2 O 2CaSO4 H2 O 16H2 O ! 3CaO Al2 O3 3CaSO4 32H2 O
" Ettringite

The continued formation of ettringite within conned solids causes signicant internal pressure leading to expansion and cracking [1]. Progressive loss of strength and mass occurs when concrete is exposed to external sources of highly concentrated sodium sulfate (Na2SO4) or magnesium sulfate (MgSO4). Calcium hydroxide from the cementitious material reacts with sulfate ions to form gypsum [1]. The two reactions with sodium sulfate and magnesium sulfate are shown below: Na2 SO4 CaOH2 2H2 O ! Gypsum 2NaOH Mg2 SO4 CaOH2 2H2 O ! Gypsum MgOH2 When magnesium sulfate attack occurs, the deterioration of the concrete is further enhanced by the decomposition of calcium silicate hydrate to hydrated magnesium silicate, which has no binding properties [1]. This reaction is shown below: 3MgSO4 3CaO 2SiO2 3H2 O 8H2 O ! 3MgOH2 2SiO2 H2 O The extent of sulfate attack is also dependent on the quantity of the liberated hydrolyzed calcium-containing compounds of Portland cement. The continued hydrolysis of cement paste produces an environment in which most of the calcium hydroxide is leached away, thus exposing the other cementitious materials to chemical decomposition. The process eventually results in loss of strength of Portland cement paste [3]. Cements constituents also play a major role in sulfate vulnerability. The amount of C3A is an indicator as to how much ettringite can form. To limit potential expansion due to sulfate attack, the amount of C3A should be restricted [1]. While concretes with air-entrainment perform much better against sulfate attack than concretes without air entrainment, perhaps the most important factor aecting expansion due to sulfate attack is permeability. Permeability controls the rate of diusion of sulfate ions into the concrete. Reducing the amount of sulfate in the pore structure reduces reactions with C3A and calcium hydroxide. Low

permeability may be attained by using a low water-tocementitious materials ratio and proper compaction [1]. The use of accelerating admixtures (i.e., calcium chloride, CaCl2) can also inuence the ability to resist sulfate attack [2]. The use of calcium chloride in Portland Cements has been shown to increase strength as far out as 28 days compared to concretes without CaCl4 . However, in situa2 tions where the durability of concrete may be compromised by external sources such as sulfate, the use of calcium chloride is cautioned [5]. Sulfate attack is dicult to measure. The main problem in assessing concretes resistance to practical levels of sulfate is that deterioration may not occur in a reasonable amount of time. For this reason, the rate of sulfate attack is accelerated in the laboratory. Techniques used in shortterm methods of testing include increasing the reactive surface, using highly concentrated solutions, percolating the attacking solution, and raising the temperature of the aggressive medium. Laboratory tests are not very accurate at predicting the service life of concrete in a sulfate-rich environment. Rather, the tests are more applicable in comparing the relative performances of dierent concrete mixtures [1]. There is no universally accepted criterion for measuring failure of laboratory specimens exposed to sulfate. Common methods of evaluation include strength loss, change in dynamic modulus of elasticity, expansion, loss of mass, and visual inspection. Miller and Manson proposed a 0.02% expansion limit to classify failure of specimens exposed to a 1% solution of sodium sulfate [6]. Stark devised a visual rating scale for evaluation of concrete. Ratings were based on a scale of 1.06.0, with the upper limits indicating failure [7]. Mehta based an evaluation on strength loss, a drop of more than 25% indicating failure [2]. Other failure criteria have been proposed, each based on dierent variables such as type of specimen and type of exposure. 2. Background on early-opening-to-trac concrete The development of fast-track concrete (FTC) grew out of the need for engineers and material scientists to devise a way to reduce congestion and public dissatisfaction as it pertained to roadway construction and repair. The result was a material capable of being placed and opened to the traveling public in the same day. Fast-track concrete has been demonstrated to meet opening-time strength requirements in less than 24 h and occasionally in as little as 4 h. This requires a high rate of early strength gain obtained by using cements with high tricalcium silicate and tricalcium aluminate contents, high cement content, low water-to-cement ratio, and accelerating admixtures [8]. An important feature in FTC is the curing methods employed to aid in the development of early strength. The most commonly used curing method involves the application of curing compounds and placement of insulating

446

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454 Table 1 Chemical composition of Portland cements Chemical compositions Silicon dioxide (SiO2) Aluminum oxide (Al2O3) Ferric oxide (Fe2O3) Calcium oxide (CaO) Magnesium oxide (MgO) Sulfur trioxide (SO3) Loss on ignition Tricalcium silicate (C3S) Dicalcium silicate (C2S) Tricalcium aluminate (C3A) Tetracalcium alumino-ferite (C4AF) Available alkalis as Na2O Type I Portland cement (%) 21.88 4.37 2.84 62.62 4.50 2.70 1.10 49.90 27.00 6.90 8.64 Type III Portland cement (%) 20.61 5.67 2.33 64.62 0.94 3.46 1.60 55.00 17.00 11.00 7.00 Type V Portland cement (%) 22.02 3.90 4.58 64.17 2.00 2.15 0.50 58.00 21.70 2.90 13.90

blankets. Curing compounds, in conjunction with the use of curing blankets, enable the concrete to retain the heat of hydration and promote the rapid gain of strength [9]. The rst major application of fast-track concrete occurred in Iowa in 1986 [10]. Since then, a variety of placement methods, mixture proportions, and curing methods have been used. Other states using FTC technology include Colorado, Kansas, Michigan, Montana, Nebraska, North Carolina, Pennsylvania, Virginia, Wyoming, and Wisconsin [11]. 3. Scope of the work The scope of the investigation was to determine the behavior of laboratory-made fast-track concretes incorporating accelerating admixture in a severe sulfate-rich environment. A large number of specimens were cast, cured in insulating boxes for the designated opening ages, immersed in a 5% sodium sulfate solution, and tested to generate and analyze specic data on the sulfate resistance of opening-time and mature (28-day moist cured) fast-track concretes. The inuences of cement type and factor, immersion period, and curing age on resistance to sulfate attack were also examined. The fresh properties included slump, air content, bleeding, setting times, and adiabatic temperature. The bulk characteristics of interest included demolded unit weight and compressive strength. Sulfate durability was ascertained through linear expansion, mass loss, and reduction in strength. 4. Experimental program 4.1. Raw materials and proportions The matrix constituents used in the investigation included cementitious binder of ASTM Types I, III, and V, siliceous ne aggregate, crushed limestone coarse aggregate, chemical admixtures, and tap water. The three cement types had an average specic gravity of 3.15. The chemical properties of the Portland cements are shown in Table 1. The ne aggregate (neness modulus of 2.56, oven-dry specic gravity of 2.6, saturated surface-dry specic gravity of 2.63, and absorption of 1.10) had a wellgraded size distribution with particles that were dense, smooth in texture, and rounded in shape. The crushed limestone coarse aggregate (nominal maximum size of 19 mm) had an oven dry specic gravity, saturated surface dry specic gravity, absorption, and unit weight of 2.64%, 2.67%, and 1.20%, and 1567 kg/m3, respectively. Both ne and coarse aggregates satised ASTM C33 gradation requirements. The mixing tap water used throughout the investigation was pre-heated to a temperature of 49 1 C. High range water-reducer and air-entraining admixtures were used to achieve the desired workability and air content, respectively. To enhance early opening-time strengths, an acceler-

0.55

0.69

0.43

ator, namely calcium chloride (CaCl2), was used in all trial mixtures. The dry quantities of the concrete constituents and water-to-binder ratio are shown in Table 2. Four cement factors, namely: 386, 446, 505, and 564 kg/m3 were used. Based on the physical properties and gradations of the coarse and ne aggregates, a uniform coarse aggregate weight of 1059 kg/m3 was found suitable for the selected concretes. Water-to-cement ratios raging from 0.325 to 0.4 were used. The quantity of the water-reducer used was determined by various trials until the desired slump of 127 6 mm was attained. Similarly, various amounts of the air-entraining admixture were used to achieve 6 1% by volume air content for the freshly mixed matrices. All concretes contained an accelerator by 2% weight of Portland cement. 4.2. Opening-to-trac time A minimum compressive strength of 20.7 MPa, corresponding to a exural strength of 4.5 MPa at opening-time, was required for all fast track concretes used in the investigation. 4.3. Casting and curing Two types of laboratory-made FTC specimens were used in the investigation, namely: 102 102 356 mm prisms and 102 204 mm cylinders. A total of six beams

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454 Table 2 Mixture constituents of fast-track concretes (kg/m3) Mixture type Cement content (kg/m3) Actual water (kg/m3) Designed w/c Fine aggregate (kg/m3) Coarse aggregate (kg/m3) Chemical admixtures WRDA-19 (kg/m3) 7.62 4.21 3.99 4.46 5.18 4.57 5.58 13.42 8.44 8.80 11.63 Daravair (kg/m3) 0.11 0.62 1.76 1.96 0.81 1.55 1.76 0.14 0.62 1.76 2.74

447

I-386 AA I-446 AA I-505 AA I-564 AA III-386 AA III-446 AA III-505 AA V-386 AA V-446 AA V-505 AA V-564 AA

386 446 505 564 386 446 505 386 446 505 554

131 153 173 179 151 175 185 131 153 173 179

0.350 0.350 0.350 0.325 0.400 0.400 0.375 0.350 0.350 0.350 0.325

729 627 525 458 679 569 492 729 627 525 458

1059 1059 1059 1059 1059 1059 1059 1059 1059 1059 1059

and nine cylinders were cast for each mixture in accordance to ASTM C 31 and cured in insulating boxes for a pre-designated period of time sucient to attain the above-noted minimum opening-time compressive strength. The insulation boxes were made of styrofoam with an internal insulation blanket having an R value of 2.45 105 h m2 C/J. The amount of curing time was dependent on the openingtime classication of the fast-track concretes. Upon removal from the insulation boxes, each cylinder was weighed, and each prism was measured to determine its initial length. Half the specimens were placed directly in a 5% sulfate solution, and the remainder were placed in a lime saturated water tank at a temperature of 23 2 C for a period of 28 days prior to being immersed in a 5% sulfate solution. The volumetric ratio of sulfate solution-toconcrete was kept constant at 3.0. When the sulfate content dropped to a level of roughly 32,000 ppm, the old solution was replaced with a fresh 5% sulfate solution. This occurred every 34 months. The immersion age of all test specimens was kept uniform for a period of 9 months. 4.4. Testing procedure The slump, air content, and setting times of freshly mixed fast-track concretes were measured using ASTM C 143, ASTM C 231, and ASTM C 403, respectively. The unit weight and compressive strength of the cylindrical samples were obtained using ASTM C 138 and ASTM C 39, respectively. Length change at dierent immersion ages was calculated using ASTM C 1012. 5. Results and discussion 5.1. Fresh properties Fresh properties of the trial fast-track concretes are shown in Table 3. A uniform slump of 127 6 mm was achieved by using various dosages of the water-reducing admixture. Similarly, dierent dosage rates of the air-

entrainer were used to attain the target air content of 6 1%. The maximum dosage rate of the air-entrainer was limited to 142 ml per 45 kg of Portland cement to avoid extreme dosages that were not practical. In examination of the matrix constituents and proportions, no bleeding was observed in any of the selected fast rack concretes. The setting times of the trial matrices were investigated and the results are shown in Table 3. For mixtures of the same cement type, Type III concretes had initial set times lower than Type I and Type V by 21% and 22%, respectively. Type III concretes also had nal setting times shorter than Type I and Type V by 20% and 24%, respectively. This can be attributed to the chemical compositions of the dierent types of cement. Type III cements have a higher percentage of C3A, the chemical responsible for early strength gain and subsequently faster setting times. For concretes of the same cement factor, those with higher cement contents had initial and nal set times lower than those with lower cement contents. Increases in cement content of 60 kg/m3 from 386446 to 505564 kg/m3 reduced initial set times by 20%, 6%, and 15% respectively. For the same increases in cement content, nal setting times were reduced by 19%, 10%, and 15%, respectively. The results for the adiabatic temperature of the selected fast-track concretes are documented in Table 3. Freshly mixed temperatures had little variation, ranging from 28 to 32 C. The peak temperatures reached during the insulated curing period ranged from 48 to 69 C. On average, Type III cements reached higher peak temperatures than Type I and Type V by 20% and 27%, respectively. Moreover, Type III cements reached their peak temperatures faster than Type I and Type V by 28% and 29%, respectively. Additionally, concretes with a higher cement contents produced higher peak temperatures when compared to the mixtures of lower cement contents. On average, peak temperatures increased 3.6 C for each 60 kg/m3 increase in cement factor, whereas the time to reach the peak temperature reduced by an average of 0.9 h for each 60 kg/m3 increase in cement content.

448

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454

Table 3 Fresh characteristics of fast-track concretes Mixture type I-386 AA I-446 AA I-505 AA I-564 AA III-386 AA III-446 AA III-505 AA V-386 AA V-446 AA V-505 AA V-564 AA Opening time (h) 8 8 8 8 6 6 6 8 8 8 8 Slump (mm) 121 127 127 121 127 133 133 114 127 127 121 Measured air content (%) 6.50 5.50 4.90 4.90 6.50 4.40 4.50 5.40 5.00 4.60 5.10 Initial time of setting (h) 5.47 3.03 2.81 2.58 2.91 2.73 2.56 4.38 3.50 3.15 2.70 Final time of setting (h) 6.21 3.82 3.52 3.13 3.61 3.28 2.93 5.75 4.20 3.85 3.43 Freshly mixed temperature (C) 28 29 31 32 28 29 32 29 31 32 32 Peak temperature (C) 48 52 55 59 59 64 69 46 48 53 54 Time to peak (h) 12.00 9.38 9.25 9.00 9.25 8.38 8.25 13.00 11.78 10.50 9.75

5.2. Ettringite-based expansion Table 4 presents the sulfate expansion test results for all trial mixtures at various immersions and curing ages. The discussion of the ettringite-based expansion pertaining to the inuence of immersion age, cement factor, and cement type is devoted to the results obtained for the opening-time specimens. The impact of curing age on sulfate expansion of the opening-time and 28-day cured samples is also discussed. 5.2.1. Inuence of immersion age A total of 21 length-comparator readings were taken on each specimen during the 9-month immersion in the sodium sulfate solution. A typical average linear expansion as a function of immersion age is shown in Fig. 1 for the opening-time fast-track concretes containing 505 kg/m3 cement factor. The graph displayed is representative of the expansion behavior for the mixtures of all cement factors, i.e., the rate of expansion was highest during the rst 1215 weeks of submersion, and then slowly leveled o during the last two-thirds of the immersion period. For Type V, Type I, and Type III concretes displayed in Fig. 1, after just 12 weeks of immersion, each had reached 79%, 80%, and 77% of their ultimate 9-month expansion,
Table 4 Linear expansion of fast-track concretes (opening-time and 28-day curing) Mixture type Linear expansion (%) Immersion age (months) 0.5 OT I-386 AA I-446 AA I-505 AA I-564 AA III-386 AA III-446 AA III-505 AA V-386 AA V-446 AA V-505 AA V-564 AA 0.005758 0.005758 0.004848 0.004242 0.005758 0.005758 0.005152 0.003333 0.00303 0.002727 0.002424 28 days 0.005152 0.003939 0.004848 0.003636 0.005152 0.005152 0.004242 0.00303 0.00303 0.002424 0.002121 1 OT 0.009394 0.008485 0.007576 0.006364 0.010000 0.009697 0.008182 0.005152 0.004545 0.003939 0.003636 28 days 0.007879 0.006364 0.007576 0.006061 0.009091 0.008788 0.006364 0.003939 0.004545 0.003333 0.002727 3 OT

respectively. While Fig. 1 exhibits, in detail, the expansion rates for opening-time mixtures of three distinct cement types all containing the same cement factor, Fig. 2 displays the rate of expansion for all trial opening-time FTC for the immersion period of 0.5, 1, 3, 6, and 9 months. As can be seen, the average expansion rate decreased steadily over time. After the rst month of sulfate exposure, the rates of expansion (given as percentage of total expansion) for Type I, III, and V matrices were approximately 48%, 43%, and 47%, respectively. During the next 2 months of immersion, from 1 to 3 months, the Type I, III, and V mixtures expanded an additional length equal to 31%, 34%, and 33%, respectively, of their ultimate 9-month expansions. During the last 6 months of immersion, the additional lengths expanded were just 20%, 23%, and 21%, respectively, of the ultimate 9-month expansions for each cement type. The steady decrease in the rate of expansion may be explained by (1) continued conversion of the C3A to ettringite compound and (2) the buildup of sulfate reaction by-products, which typically occupy a greater volume than the compounds they replace, within the pores of the test specimens. The latter process can eectively lower the permeability of the pore structures, thus restricting further permeation of the sulfate ions and limiting the potential expansion.

6 28 days 0.013636 0.011212 0.011515 0.008788 0.016667 0.015455 0.011212 0.006970 0.006970 0.005455 0.004848 OT 0.019697 0.015758 0.014545 0.010909 0.022727 0.020606 0.016364 0.011212 0.009394 0.007576 0.006970 28 days 0.015758 0.014545 0.010909 0.022727 0.020606 0.016364 0.015455 0.011212 0.009394 0.007576 0.006970

9 OT 0.021212 0.017273 0.015455 0.011818 0.024545 0.022424 0.017879 0.011818 0.010000 0.008182 0.007273 28 days 0.016061 0.013939 0.012727 0.010000 0.019697 0.018485 0.014515 0.008788 0.008182 0.006667 0.005455

0.017273 0.013636 0.012424 0.009091 0.018485 0.017879 0.013333 0.009091 0.007879 0.006667 0.005758

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454

449

0.015

Linear Expansion (%)

where (SExp)OT = sulfate expansion at opening-time (%); IAw = immersion age (week). The regression variables R2, S, Prob(t) and Prob(F) are given in Table 6. The calculated values are indicative of a strong relationship between the dependent variable (beam expansion) and the independent variable (immersion age) for all three cement types. 5.2.2. Inuence of cement factor The inuence of cement factor on sulfate-induced expansion of opening-time fast-track concretes was also examined and the results are documented in Fig. 3. As can be seen, for mixtures of the same type, average linear expansion decreases with increases in cement factor. Using the 386 kg/m3 cement factor as the basis for comparison, on average, opening-time fast-track concretes with cement factors of 446, 505, and 564 kg/m3 expanded 14%, 28%, and 41% less, respectively. For Type I mixtures, as the cement factor was increased from 386 to 446, 505, and 564 kg/m3, the reduction in expansion was 19%, 27%, and 44%, respectively. For Type V concretes, the reductions in linear expansion were 15%, 31%, and 38%, respectively. For opening-time FTC containing Type III cement, an increase in cement factor from 386 to 446 and 505 kg/m3 resulted in roughly 9% and 27% less expansion, respectively. The preceding trends suggest that the increase in cement content results in a decrease of expansion, although sucient dierences in expansion exist among dierent cement types. Based on the chemistry of the reactions, concretes with higher cement content, which also contain a greater amount of C3A (by mass), are potentially more vulnerable to external sulfate attack by producing more ettringite. The fact that this did not occur suggests that the paste quality, and its eect on permeability and strength, is the main deterrent to expansion caused by

0.010

0.005

0.000 0 5 10 15 20 25 30 35 40

Immersion Age (Weeks)

0.02

Linear Expansion (%) /

0.015

0.01

0.005

I-386 I-564 III-505 V-505

I-446 III-386 V-386 V-564

I-505 III-446 V-446

0 0 1 2 3 4 5 6 7 8 9 10

Immersion Age (Months)

A statistical program was used to determine the best-t predictive equations for sodium sulfate expansion test results. Analyses were conducted at 95% condence level. The predictive equations were tested for accuracy using R2 (the coecient of multiple determination) and S (average standard deviation). Correlations between the data predicted from the regression equations and the actual results obtained from expansion test results were evaluated using F and T tests. The relationship between the sodium sulfate expansion and the immersion age is as follows: 0:0186IA Type I cement: SExpOT IAw 5:301 0:0253IA Type III cement: SExpOT IAw 6:723 0:0109IA Type V cement: SExpOT IAw 6:364 1

Type I

Type III

Type V

0.025

Linear Expansion (%) /

0.02

0.015

0.01

0.005

2 3

386

446

505

564

Cement Factor (kg/m3)

450

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454

sulfate reactions. Fast-track concretes with higher cement contents contain a smaller void structure and are thus more eective in restricting the inltration of the sulfate solution. Additionally, the improved paste quality also oers greater resistance to the internal pressure induced by the accumulation of reaction by-products, thereby limiting potential sulfate expansion. The best-t predictive equation dening the relationship between the sodium sulfate expansion and the cement content is as follows:
Type I cement: SExpu OT 0:01486 3:77 1011 CF3
11 3

1247:69 CF2

4
Type III cement: SExpu OT 0:03027 9:44 10 3:00 1010 CF3 CF

5 6

Type V cement: SExpu OT 0:02066 8:90 10 CF

2:5

expanded 31% more, while Type V mixtures expanded 43% less. On average, Type III mixtures expanded 2.3 times that of Type V mixtures. This trend can be explained by the chemical composition of each cement type. Type V cements, designed specically to resist ettringite-induced expansion, contain a smaller amount (less than 5%) of tricalcium aluminate which gives it the ability to resist expansion [3]. Type III groups, designed specically for early strength gain; contain more tricalcium aluminate which increases its vulnerability to expansion. Type I FTC is chemically similar to Type III matrices, but is not ground as ne, thus reducing the rate of expansion by ettringite attack. The beams expansion due to 5% sodium sulfate attack was also evaluated under the combined action of immersion age and cement factor. The following expressions were found: Type I cement SExpOT 0:311 2:351 104 IAw 453:830 CF 200; 837:529 3:912 105 IA2 w CF2 IA 9:831 107 IA3 1:027 w CF 30; 596; 154:498 IAw 52:078 3 CF2 CF 1:528 102 Type III cement SExpOT 0:050 3:045 103 IAw 2:296 104 IA2 w 9:643 106 IA3 2:066 107 IA4 w w 1:756 109 IA5 2:679 104 CF w IA2 w CF 7

where (SExpu)OT = 9-month sulfate expansion at openingtime (%); CF = cement factor (kg/m3), with 386 kg/m3 6 CF 6 564 kg/m3. The regression variables R2, S, Prob(t) and Prob(F) are given in Table 6. The calculated values are indicative of a strong relationship between the dependent variable (ultimate beam expansion) and the independent variable (cement factor) for all three cement types. 5.2.3. Inuence of cement type The inuence of cement type on the linear expansion of opening age samples is shown in Fig. 4. For mixtures of the same cement factor, Type V mixtures performed the best in their ability to resist expansion due to ettringite formation, followed by Type I and Type III. Using Type I FTC as the basis for comparison, on average, Type III mixtures

0.03
386 kg/m3 446 kg/m3 505 kg/m3

3:414 107 CF2 Type V cement

0.025

SExpOT 0:071 2:226 103 IAw 4:651 104 CF

Expansion (%)

4:667 105 IA2 9:849 107 CF2 w 4:035 106 IAw CF 4:271 107 IA3 w 6:861 1010 CF3 2:477 109 IAw CF2 2:810 108 IA2 CF w

0.015

0.01

0.005

0 V I III

Cement Type
Fig. 4. 9-Month linear expansion of fast-track concretes (opening-time): eect of cement type.

where (SExp)OT = sulfate expansion at opening-time (%); IAw = immersion age (week); CF = cement factor (kg/ m3), with 386 kg/m3 6 CF 6 564 kg/m3. The regression variables R2, S, Prob(t) and Prob(F) are given in Table 6. The calculated values are indicative of a good relationship between the dependent variable (beam expansion) and the independent variables (immersion age and cement factor) for all three cement types.

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454

451

5.2.4. Inuence of curing age To gauge the inuence of the curing age on sulfate expansion, additional FTC prisms were allowed to be water-cured for 28 days, immediately after the designated blanket curing time, prior to the placement into a 5% sulfate solution. The results are shown in Table 4. On average, the expansion of opening-time fast-track concretes was roughly 1.3 times that of the 28-day mixtures. The abovementioned behavior can be attributed to the improved paste quality, resulting in an enhanced impermeability and increased tensile strength for the trial fast-track concretes. The relationship between expansion at 28 days and the expansion at opening-time for various immersion ages and cement factors is as follows: Type I cement SExp28 0:796SExpOT 9:728 105 IAm 1:832 105 CF 7:937 103 Type III cement SExp28 0:911SExpOT 1:922 104 IAm 1:099 106 CF 4:271 104 11 10

Type V cement SExp28 0:978SExpOT 1:119 104 IAm 2:112 106 CF 1:276 103 12

where (SExp)28 = sulfate expansion at 28 day (%); (SExp)OT = sulfate expansion at opening-time (%); CF = cement factor (kg/m3), with 386 kg/m3 6 CF 6 564 kg/m3; IAm = immersion age (month). The regression variables R2, S, Prob(t) and Prob(F) are given in Table 5. The calculated values are indicative of a good relationship between the dependent variable (beam expansion at 28 day) and the independent variables (beam expansion at openingtime, immersion age, and cement factor) for all three cement types. 5.3. Gypsum-based strength reduction and mass loss In addition to expansion, loss of mass and reduction in strength resulting from gypsum formation were also monitored. After 9 months exposure to a sulfate-rich environment, no material residue was collected and the FTC samples were in the same condition as the beginning of the test, indicating little or no softening. Table 6 highlights the compressive strength data for the opening-time

Table 5 Statistical regressions variables Equation (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) Description Inuence of immersion age on sulfate expansion, Type I cement Inuence of immersion age on sulfate expansion, Type III cement Inuence of immersion age on sulfate expansion, Type V cement Inuence of cement factor on sulfate expansion, Type I cement Inuence of cement factor on sulfate expansion, Type III cement Inuence of cement factor on sulfate expansion, Type V cement Inuence of immersion age and cement factor on sulfate expansion, Type I cement Inuence of immersion age and cement factor on sulfate expansion, Type III cement Inuence of immersion age and cement factor on sulfate expansion, Type V cement Expansion(28) vs. expansion(OT), immersion age and cement factor, Type I cement Expansion(28) vs. expansion(OT), immersion age and cement factor, Type III cement Expansion(28) vs. expansion(OT), immersion age and cement factor, Type V cement Expansion vs. compressive strength for OT Inuence of water-cured and sulfate immersed compressive strength on sulfate expansion for OT Inuence of compressive strength and cement factor on sulfate expansion for OT Expansion vs. compressive strength for 28 days Inuence of water-cured and sulfate immersed compressive strength on sulfate expansion for 28 days Inuence of compressive strength and cement factor on sulfate expansion for 28 days Coecient of multiple determination, R2 (%) 99.8 99.9 99.7 98.6 98.2 99.8 98.2 97.6 98.7 66.0 98.3 91.8 82.8 99.9 99.1 79.5 99.7 81.1 Standard deviation, S (%) 1.8.1 104 1.9 104 1.4 105 7.9 104 6.5 104 1.6 104 7.2 104 1.04 104 3.3 104 3.1 103 8.4 104 8.1 104 4.5 103 7.6 105 1.8 103 3.9 103 8.1 104 4.6 103 Prob(t) 0.00 0.00 0.00 <0.45 <0.09 <0.13 <0.02 <0.16 <0.01 <0.394 <0.879 <0.581 <0.171 <0.041 <0.153 <0.000 <0.225 <0.466 Prob(F) 0.00 0.00 0.00 0.12 0.09 0.05 0.00 0.00 0.00 0.000 0.000 0.000 0.296 0.009 0.225 0.365 0.125 0.566

452 Table 6 Bulk properties of fast-track concretes Mixture type Opening time (h)

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454

Demolded unit weight (kg/m3)

Compressive strength (MPa) Curing age (h) 6 8 17.0 21.2 22.8 27.3 32.8 37.3 41.5 15.8 20.8 21.6 25.7 12 27.9 28.7 29.5 33.4 25.7 26.9 28.6 31.2

Compressive strength of water-cured specimens (MPa) Curing age (days) 28 55.4 56.6 57.6 63.5 56.5 59.1 62.2 52.6 54.9 56.1 62.8 270 62.8 66.5 72.4 79.8 57.9 64.9 68.3 64.0 66.2 67.6 68.0

9-Month sulfate-immersed compressive strength (MPa) Sample type Opening time 60.0 64.1 70.4 78.4 52.6 60.9 64.2 62.3 65.2 67.2 67.7 28 days 64.0 66.0 72.0 79.6 55.0 61.9 65.6 64.0 66.0 67.5 67.9

I-386 AA I-446 AA I-505 AA I-564 AA III-386 AA III-446 AA III-505 AA V-386 AA V-446 AA V-505 AA V-564 AA

8 8 8 8 6 6 6 8 8 8 8

2366 2368 2370 2379 2348 2357 2370 2362 2365 2369 2372

16.4 22.3 26.4 33.6 37.1 15.3 19.5

and 28-day specimens immersed in a 5% sulfate solution for dierent periods. As can be seen, mixtures with the same cement content, Type V fast-track concretes performed the best, followed by Types I and III. The average percent strength loss for Type I, III, and V matrices was 3.2%, 7.0%, and 1.3%, respectively. For concretes of the same cement type, those with higher cement factors performed better than the ones with lower cement contents. The average strength reductions of the FTC containing 386, 446, 505, and 564 kg/m3 were 5.5%, 3.7%, 3.1%, and 1.1%, respectively. In the absence of any softening, as evident by the lack of noticeable material loss and strength reduction, it can be concluded that the bulk of the sulfate attack occurred in the form of expansion due to the formation of ettringite compounds.

SExpu OT 2818:468 67:644CSw

254; 796:618 CSs 7; 590; 040:104 2 0:536CSw CS2 s CSw 4116:962 1:407 103 CS3 w CSs 74; 334; 349:388 CSw 62; 034:831 3 CS2 CSs s 16:459 CS2 w CSs 14

SExpu OT 1392:555 37:071CS 1070:746 logCF 0:441CS2 286:508 logCF2 21:332CS logCF 9:790 104 CS3 25:417 logCF3 2:816CS logCF2 0:102CS2 logCF For 28-day specimens, cement types I, III and V SExpu 28 1:346 107 CS7 6:309 105 CS6 1:264 102 CS5 1:405CS4 93:518CS3 3726:803CS2 82; 346:587CS 778; 236:405 SExpu 28 14; 984:019 340:017CSw 1; 468; 394:160 2:562CS2 w CSs 47; 742; 212:644 CSw 22; 302:746 CSs CS2 s 514; 565; 724:455 6:418 103 CS3 w CS3 s 364; 270:177 CSw CS2 84:331 w 2 CSs CSs 17 16 15

5.4. Relationship between sulfate expansion and compressive strength Table 6 gives the compressive strength of water-cured and sulfate-immersed specimens for dierent cement factors and cement types. It can be seen from that table that an increase in cement factor leads to an increase in compressive strength and reduction in expansion. The best-t predictive equations of linear expansion as a function of compressive strength for the specimens exposed to 5% sodium sulfate during 9-month (270 days) period were determined at 95% condence level. The relationship can be expressed as follows: For opening-time specimens, cement types I, III and V SExpu OT 8:605 109 CS7 3:773 106 CS6 7:063 104 CS5 7:318 102 CS4 4:531CS3 167:722CS2 3434:949CS 30; 027:279 13

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454

453

SExpu 28 1077:700 82:173CS 2:496CS2 3:775 10 CS 2:843 10 CS

8:532 107 CS5 1:589 103 CF 2:892 106 CF2 1:753 109 CF3 18

where (SExpu)OT = 9-month sulfate expansion for the opening-time (%); (SExpu)28 = 9-month sulfate expansion for the 28-day curing (%); CSw = 9-month water-cured compressive strength (MPa); CSs = 9-month sulfateimmersed compressive strength (MPa); CF = cement factor (kg/m3), with 386 kg/m3 6 CF 6 564 kg/m3. The regression variables R2, S, Prob(t) and Prob(F) are given in Table 6. The calculated values are indicative of a good relationship between the dependent variable (ultimate beam expansion) and the independent variable (ultimate compressive strength and cement factor) for both the OT and 28 days. Finally, in order to ascertain the performance of the selected opening-time fast-track concretes, the results were compared to the applicable expansion limits as noted in the introductory segment of the manuscript. Based on the 6-month expansion limit recommended by ASTM C1012, all opening-time fast-track concretes displayed expansion below 0.05% and, thus, can by classied as having high sulfate resistance. Moreover, Mehtas recommended 25% strength loss limit for sulfate-resistant concretes was also met by all selected opening-time fast-track matrices. 6. Conclusions The following conclusions are drawn based on the experimental results presented in the manuscript. 1. During the 9-month immersion in a 5% sulfate solution, the opening-time test samples experienced continued length change, but did not exhibit any signs of softening. Moreover, the highest rate of expansion occurred within the rst 3 months of exposure to sulfate. 2. An increase in cement content of 60 kg/m3 resulted in reduced 9-month linear expansions of roughly 19%, 27%, and 44% for the opening-time Type I concretes with cement factors ranging from 386 to 564 kg/m3. For the equivalent Type V groups, the corresponding reductions in expansion were 15%, 31%, and 38%. The opening-time fast-track concretes containing Type III Portland cement ranging from 386 to 505 kg/m3 showed reductions in sulfate-induced expansion of 9% and 27% with an incremental increase of 60 kg/m3 of cement. 3. The opening-time fast-track concrete specimens containing Type V Portland cement displayed the least amount of expansion followed by the mixtures made with Type I, and III cement. After 9 months exposure in a sulfate-rich environment, on average, the opening-time

Type V fast-track concretes produced linear expansion results which were 45% and 54% less than the companion concretes containing Type I and Type III Portland cements, respectively. 4. Fast-track concretes allowed to mature prior to immersion in a sulfate solution resisted expansion and strength loss more eectively than those exposed to a sulfate-rich environment immediately at the opening-time. On average, the opening-time concretes displayed 9-month linear expansions that were roughly 1.3 times higher than those of the equivalent 28-day water-cured specimens. 5. When applying the failure criteria for expansion, the trial opening-time fast-track concretes performed remarkably well. With 9-month linear expansions well below the limit of 0.05%, the selected openingtime matrices are considered to have high sulfate resistance. 6. The results of the statistical regression studies revealed the presence of strong relationships among the sodium sulfate expansion (dependant variable), immersion age, cement factor, and compressive strength (independent variables) for the fast-track concretes made with dierent cement types. Conversion factors 1 mm = 0.0394 in. 1 MPa = 145 psi 1 kg/m3 = 1.684 lb/yd3 C = (5/9)(F 32) 1 h m2 C/J = 20,408 h ft2 F/Btu Acknowledgement The writers are grateful to Tennessee Technological University for supporting the research project through a faculty research grant. The contributions of a number of material suppliers are greatly appreciated. References
[1] Mathis Richard P. Freezing and thawing, deicing salt, and sulfate resistance of concrete pavers. Graduate thesis, Universtiy of Southern Illinois, 1991. [2] Mehta PK. Concretestructure, properties, and materials. Englewood Clis, NJ: Prentice-Hall, Inc.; 1986. p. 450. [3] Biczok. Concrete corrosion and concrete protection. New York: Chemical Publishing Company, Inc.; 1967. 291 pp. [4] Kosmatka SH, Panarese WC. Design and control of concrete mixtures. 13th ed. Skokie, IL: Portland Cement Association; 1988. p. 779. [5] Neville AM. Properties of concrete. Bath, Great Britain: Pittman Publishing; 1975. p. 779. [6] Miller DG, Manson PW. Tests of 106 commercial cements for sulfate resistance. In: Proceedings, American Society for Testing and Materials, vol. 41. 1996. p. 9881001. [7] Stark D. Durability of concrete in sulfate-rich soils. PCA Research and Development Bulletin RD097, Portland Cement Association; 1990. p. 114.

454

N. Ghafoori et al. / Cement & Concrete Composites 30 (2008) 444454 [10] Grove J. Blanket curing to promote early strength concrete. Research Project MLR-87-7, Iowa Department of Transportation, 1989. [11] American Concrete Institute. Accelerated techniques for concrete paving (ACI 325.11R-01). ACI Committee 325, Detroit, MI, 2001.

[8] Whiting D, Nagi M, Okamoto PA. Early strength gain of rapid highway repair concrete. Concrete International, August 1994. [9] Pearson, Robert I. Fast track concrete paving. Concrete International, August 1988.