Freezing and thawing tests of high-strength concretes

Freezing and thawing tests of high-strength concretes

CEMENTand CONCRETERESEARCH.Vol.21, pp. g44-852, 1991. Printedin the USA. 0008-8846/91. $3.00+00, Copyright(c) 1991PergamonPressplc. F R E E Z I N G A...

575KB Sizes 0 Downloads 11 Views

CEMENTand CONCRETERESEARCH.Vol.21, pp. g44-852, 1991. Printedin the USA. 0008-8846/91. $3.00+00, Copyright(c) 1991PergamonPressplc.

F R E E Z I N G AN D T H A W I N G T E S T S OF H I G H - S T R E N G T H C O N C R E T E S

by Michel Pigeon(i), Richard Gagn6(2), Pierre-Claude Ai'tcin(2) and Nemkumar Banthia(1) Centre de Recherche Interuniversitaire sur le B6ton (1) Universit6 Laval, Ste-Foy, Qu6bec, Canada G1K 7P4 (2) Universit6 de Sherbrooke, Sherbrooke, Qu6bec, Canada, J1K 2R1 (Communicatedby M. Moranville-Regourd) (ReceivedMarch25, 1991)

ABSTRACT Seventeen high-strength concretes were made using Portland cement (with and without silica fume) and tested for frost resistance (using the procedure "A" (freezing and thawing in water) of ASTM Standard C 666) to analyze the influence of various parameters on the limiting value of the water to binder ratio below which air entrainment is no longer required for good freezing and thawing cycle durability. The parameters included the type of cement, the type of aggregate and the length of the curing period. The results of these tests, as well as previously published data, indicate that this value can be higher than 0,30 in certain cases, but equal to or lower than 0,25 in others, depending particularly on the characteristics of the cement. More research is needed before these values can be used as guidelines, since field exposure conditions differ from laboratory testing conditions, and because the air-void spacing factor of non-air-entrained field concretes could be significantly higher than that of laboratory made concretes.

High-strength concretes have a very low water to binder ratio (usually lower than 0,3) and are often made with a "high performance" Portland cement. They also often contain silica fume because this increases the strength level for a given water to binder ratio (1, 2). These concretes therefore only contain a very limited amount of freezable water, and it has been shown that it is possible to produce high-strength concretes that contain such a small quantity of freezable water that they are completely immune to frost even if they are not airentrained (3). What is the limiting value of the water to binder ratio below which air entrainment is no longer required for frost resistance in high-strength concrete? The answer to this question is very important, since the use of air entrainment significantly reduces strength and also because air entrainment can be very difficult to achieve in very low water to binder ratio concrete mixtures. It can be safely hypothesized, from fundamental considerations, that this value is mainly a function of the cement characteristics and of the length of curing, both of which directly influence the capillary porosity of the paste. It could also be influenced to some extent by the aggregate characteristics because of the higher porosity of the paste at the aggregate-paste interface (4). This paper presents the results of freezing and thawing 844

Vol. 21, No. 5

FREEZING-THAWING, HIGH STRENGTH CONCRETES, SILICA FUME

845

tests carried out to investigate the influence of these three parameters on the limiting value of the water to binder ratio.

Z ?r_ogram In this investigation, seventeen high-strength concretes (forming four distinct series) were prepared and tested for freezing and thawing cycle durability in accordance with Procedure "A" of ASTM Standard C 666 (rapid freezing and thawing in water). The basic characteristics of the seventeen concretes tested are presented in Table 1. Four (forming series A) were made with a dolomitic limestone coarse aggregate and a Type HI cement at a water to binder ratio of 0,30, four (forming series B) were made with a granite coarse aggregate and the same Type HI cement with silica fume at a water to binder ratio of 0,30, six (forming series C) were made with the limestone aggregate and a Type I cement with silica fume at a water to binder ratio of 0,30, and three (forming series D) were made with the same limestone aggregate and the same Type I cement with silica fume at a water to binder ratio of 0,26. For series A, B and C, half of the concretes were prepared using an air entraining agent (at a dosage sufficient to obtain a spacing factor of approximately 200 Ixm or less) and half without. For series D, only non-air-entrained concretes were made. Table 1

Mix Description Mix SERIES A C. Type III Limestone W/C = 0,30 Curing (days)

Spacing Factor High Low

SERIES B C. Type III+SF Granite W/(C+S) -- 0,30

SERIES C C. Type I+SF Limestone W/(C+S) = 0,30

Spacing Factor High Low

Spacing Factor High Low

SERIES D C. Type I+SF Limestone W/(C+S) --- 0,26 Spacing Factor High

1

AH1

ALl

BH1

BL1

CH1

CL1

....

3

AH3

AL3

BH3

BL3

CH3

CL3

DH3

CH7

CL7

DH7

7

. . . . . . . . . . . .

28

. . . . . . . . . . . . . . . . . .

DH28

These four series of concretes were chosen on the basis of previously published test data (3) which had shown that, for concretes made with the same limestone aggregate and Type m cement, but with silica fume used as partial cement replacement, air entrainment was not required at a water to binder ratio of 0,30 even if the freezing and thawing tests were carded out after only one day of curing. The objective of series A was thus to analyze the influence of the use of the Type III cement (without silica fume), that of series B the use of a different coarse aggregate, and that of series C and D the use of a Type I cement with silica fume instead of the Type III. For the concretes made with the type III cement (series A and B), the curing periods that were chosen (again on the basis of previous test results (3)) were 1 and 3 days respectively. For those made with the Type I cement, because of the slower hydration rate, the following curing periods were selected: 1, 3 and 7 days for series C and 3, 7 and 28 days for series D.

846

M. Pigeon et al.

Vol. 21, No. 5

Materials. Mixture Characteristics and Exnerimental Procedures

The chemical analysis and some of the physical properties of the two cements that were used are given in Table 2. The two most important differences between the Type I and the Type HI cements are due to the high C3S content (63%, calculated on the basis of Bogue's equations) and the very high Blaine fineness (545 m2/kg) of the Type III cement. The silica fume that was used contains more than 90% SiO 2 and was obtained from a plant in BEcancour, Qu6bec. Table 2 Cement Characteristics

Cement Type I

Type III

2,3 2,1 62,4 21,5 2,9 4,6 0,80

1,8 2,9 64,2 20,9 2,6 4,1 0,71

51 23

63 12

C3A (~)

8

8

C4AF (%)

7

5

340

545

Fe203 (%) SO3 (%) CaO (%) SiO2 (%) MgO (%) A1203 (%) Alkalies (Na20 equivalent) (%) C3S (%) C2S (%)

Blaine Specific Surface (m2/k~)

A granitic sand containing particles of feldspar, quartz and hornblend with a fineness modulus of 2,3 was used for all concretes. In spite of the fineness of this sand (Table 3), all mixtures had normal workabilities and cement contents for that type of concrete. The coarse aggregate for the series A, C and D mixtures was a dolomitic limestone containing Table 3

Aggregate Gradings

Sieve Size 1/2" (12.5 mm) 3/8" (9.5 mm) #4 (4.75 ram) #8 (2.36 ram) #16 (1.8 ram) #30 (600 ~tm) #50 (300 grn) #100 (150 ~tm) ,#200 (75 orn)

Percentage retained, cumulative Fine Coarse Aggregate Aggregate Dolomitic Limestone Granitic Gravel 0 0 0 4 12 42 76 93 97,4

0 31 96 99 100

0 0 80 100

Vol, 21, No. 5

FREEZING-THAWING, HIGH STRENGTH CONCRETES, SILICA FUME

847

more than 98% of hard dolomite, and for the series B mixes, a crushed granitic gravel. The basic properties of these aggregates are the following: absorption, 0,58% for the fine aggregate, 0,70% for the dolomitic limestone aggregate and 1,0% for the granitic gravel aggregate, density, 2,63 for the fine aggregate, 2,77 for the dolomitic limestone aggregate and 2,69 for the granitic gravel aggregate, and loss on the sulfate soundness test, 4,3% for the fine aggregate, 3,0% for the dolomitic limestone aggregate and 1,3% for the granitic gravel aggregate. The aggregates were used dry in all mixtures. The proportion of the constituents and the fresh concrete properties are summarized in Table 4 for all mixtures made. Only seven mixtures were required to prepare the seventeen concretes since, in many instances, the curing period is the only parameter that differentiates one concrete from another (see Table 1). In the silica fume concretes, 6% of the total weight of the cement was replaced by silica fume. All concrete mixtures were prepared using a naphthalene formaldehide condensate type superplasticizer. The liquid to solid ratio in this superplasticizer is 0.59 by weight and the amount of water it contains was of course taken into account in the calculation of the water to binder ratio. Relatively high dosages (24 to 34 mL/kg of cementitious material) were required to maintain a minimum slump of 40 mm. This is common in high-strength concretes with low water to binder ratios. The higher dosages (31 to 34 mL/kg of cementitious material) were used in the mixtures made with the Type III cement (which has a very high fineness). To obtain spacing factor values of approximately 200 lain or less in the air-entrained mixtures, a synthetic detergent type air entraining agent was used. Because of the very low water content and the high fineness of the Type III cement, the effectiveness of this admixture was highly reduced and more than six times the manufacturer's recommended dosage for normal concrete was used in the series A and B mixtures.

Table 4 Mix Characteristics i

Mix Water I

i

Cement Silica kg,/m3

C. Ag. F. Ag.

I

SP AEA ml/k~of(C+S)

Slump

] (mm) I

Air (%)

AH

149

487

0

1061

805

34

0

40

2,8

AL

133

437

0

953

723

33

2,5

160

11,0

BH

144

449

26

1035

785

31

0

150

3,4

BL

128

398

23

917

696

31

1,9

230

9,2

CH

143

448

27

1033

784

28

0

210

4,8

CL

131

435

27

1003

762

24

0,7

175

7,8

DH

144

543

33

1036

700

24

0

60

ii

2,8 i

The same mixing procedure was used for all concretes. The superplasticizer was In:st added to the mixing water, and the cement, the silica fume and the water were mixed to obtain a uniform paste. If an air entraining agent was needed, it was added at this stage, premixed with approximately 200 mL of the mixing water. The coarse and fine aggregates were then placed in the pan-mixer and the concrete was mixed for three minutes. The required number of 75 x 100 x 400 mm prisms (one for the determination of the air void characteristics of the hardened concrete and two for each of the freezing and thawing

848

M. Pigeon et al.

Vol. 21, No. 5

tests (one test for each curing period selected, see Table 1)) were cast from each mixture. Concrete was consolidated in the molds with the use of a vibrating table. Six cylinders (100 x 200 mm) were also cast from each mixture for compressive strength measurements. All prisms were cured (for the selected period of 1, 3, 7 or 28 days) by simply sealing in a number of layers of plastic film. This type of curing is closer to normal field curing than water curing and was that utilized for the previous tests used as a basis of comparison (3). The freezing and thawing tests were carried out, immediately after curing, according to ASTM Standard C 666, procedure "A" (rapid freezing and thawing in water), in a Logan freezing and thawing unit. The freezing rate in this apparatus is approximately 8°C/h. In most cases, the tests were continued beyond the recommended number of cycles (300). The characteristics of the air void system of the hardened concretes were determined using the ASTM C 457 "modified point count" method.

Test Results Air Void Characteristics The basic characteristics of the air void system of the hardened concretes are presented in Table 5, together with the air content measured on the fresh concrete. In the four non-airentrained mixtures, the specific surface varies approximately between 7 mm -1 and 10 mm-1 and the spacing factor between 500 Ixm and 1 mm. These values are fairly typical of non-air-entrained concretes. The lowest value, 503 l~m, corresponds to an air content of close to 5 %. This is probably due to the high dosage of the superplasticizer that was used. In the three air-entrained mixtures, the values of the specific surface are all higher than 20 mm -1 and those of the spacing factor are all approximately equal to, or lower than, the usually recommended value of 200 I.tm. One value is particularly low (78 ~tm), due to a very high value of the air content measured on the hardened concrete (15,0%, as opposed to 9,2% obtained on the fresh concrete). Such discrepancies between the air content measured on the fresh and on the hardened concrete are quite common when the air content measured on the fresh concrete is high, but there is no general agreement on the reasons behind this phenomenon (5).

Table 5 Air Void Characteristics

Mix

AH AL BH BL CH CL DH

Air Content, A Fresh Hardened (%) (%) 2,8 11,0 3,4 9,2 4,8 7,8 2,8

2,8 7,4 4,8 15,0 4,9 7,1 2,8

Specific Surface

Spacing Factor

mm-1

grn

6,9 28,3 7,4 21,7 10,0 20,7 10,2

948 135 691 78 503 203 690

Vol. 21, No. 5

FREEZING-THAWING, HIGH STRENGTH CONCRETES, SILICA FUME

Compressive Strength Table 6 gives the results of the compressive strength tests that were carried out after 28 days of moist curing. For each of the seven values of the compressive strength (one per concrete mixture made), the corresponding value of the air content measured on the hardened concrete is also shown in the Table. The compressive strength of the non-air-entrained concretes ranges between 74,8 MPa and 89,1 MPa. The highest value was obtained with the Type HI cement and 6% silica fume (at a water to binder ratio of 0,30). At the same water to binder ratio of 0,30, a value of 79,2 MPa was obtained with the Type III cement without silica fume, and a value of 74,8 MPa with the Type I cement and 6% silica fume. Although a different coarse aggregate was used for series B (made with the Type III cement and 6% silica fume), these results do point out the influence on strength of silica fume as well as that of the cementcharacteristics. The compressive strength of the air-entrained concretes are significantly lower, due to the relatively high air contents. Table 6

Compressive Strength Mix All AL BH BL CH CL DH

Air Content

Compressive Strength, (28d)

%

MPa

2,8 7,4 4,8 15,0 4,9 7,1 2,8

79,2 60,1 89,1 69,5 74,8 65,0 82,3

Freezing and Thawing Test Results The results of the freezing and thawing tests carried out on the seventeen concretes are summarized in Table 7 where the results of the mass, ultrasonic pulse velocity and length change measurements after 300 cycles and at the end of the tests (after more than 500 cycles in most cases) are shown. It is clear from the results in this Table that only two groups of concretes were significantly damaged by the freezing and thawing cycles: the three non-air-entrained concretes of series C, and two of the three concretes of series D (also non-air-entrained). For all the other concretes tested, the value of the pulse velocity at the end of the cycles is higher than the initial value, and the length change is lower than 400 I.tm/m; both types of measurement thus clearly show that no significant microcracking has occurred in these concretes. For the three non-air-entrained concretes of series C, only the seven day cured concrete showed signs of deterioration after 300 cycles. But after more than 600 cycles, all showed definite signs of internal microcracking, i.e. an important decrease of pulse velocity and a length change of more than 2000 I.trn/m. For two of the three concretes of series D (also non-air-entrained), the results indicate that some deterioration has occurred, but this deterioration is much smaller than that of the non-air-entrained concretes of series C. For the three day cured concrete, the length change is 1029 I~m/m after 961 cycles and there is a small increase in pulse velocity. For the twenty-eight day cured concrete, the length change is 831 ~tm/m after 306 cycles, and there is a 1% decrease of pulse velocity.

849

850

M.Pigeonet al.

Vol.21, No. 5

The main objective of this investigation was to obtain information on the influence of the cement characteristics on the limiting value of the water to binder ratio below which air entrainment is not required for good resistance to freezing and thawing cycles. For the Type III cement that was used, it is evident from the test results for the series A concretes that air entrainment is not required for good frost resistance at a water to binder ratio of 0,30, even after only one day of curing. Concrete AH-1 with a spacing factor of 948 ~tm withstood close to 500 freezing and thawing cycles without any deterioration. If this result is compared to the previously published results (which have shown that non-air-entrained concretes made with similar materials but containing silica fume had an excellent freezing and thawing resistance after one day of curing (3)), it is clear that, at a water to binder ratio of 0,30, the use of silica fume does not increase significantly the frost resistance of the concretes made with this Type HI cement. Table 7

Freezing and Thawing Test Results Mix(l)

L

Nb of Cycles

~m)

Variation of Mass (%M0) after end of 300 cycles cycles

Variation of Pulse Velocity (%V0) after end of 300 cycles cycles

Variation of Length (xl0 -6) after end of 300 cycles cycles

AH1 AH3

948 948

483 464

100 100

100 100

106 104

107 105

123 186

133 219

ALl AL3

135 135

438 401

100 100

100 100

107 104

107 104

141 261

194 176

BH1 BH3

691 691

724 714

100 100

100 100

105 101

108 100

27 180

40 299

BL1 BL3

78 78

724 632

100 100

100 100

107 103

110 105

106 88

146 121

CH1 CH3 CH7

503 503 503

691 679 817

100 100 101

98 100 90

104 102 87

<75 <75 <60

269 288 2014

2572 3622 7930

CL1 CL3 CL7

203 203 203

1270 547 524

100 100 100

98 100 100

109 105 103

110 106 104

116 204 226

225 207 252

DH3 DH7 DH28

690 690 690

961 940 306

100 100 101

100 100 101

104 103 99

102 102 99

209 96 831

1029 381 831

(1) The number in the mix code indicates the length of the curing period in days. It is also evident, if the test results for the series B concretes are compared to the previously published results, that the use of the granitic gravel aggregate instead of the dolomific limestone aggregate had no significant influence on the freezing and thawing resistance.

Vol.21. No. 5

FREEZING-THAWING, HIGHSTRENGTHCONCRETES.SILICAFUME

After only one day of curing, concrete BH-1 with a spacing factor of 691 ~tm withstood more than 700 cycles without any deterioration. This result confirms that, for the Type III cement with 6% silica fume that was used, air entrainment is not required at a water to binder ratio of 0,30 for good resistance to freezing and thawing cycles. Both the dolomitic limestone and the granitic gravel are high quality aggregates (with a very low absorption) and were dry when they were incorporated into the various concrete mixtures. Lower quality aggregates could, of course, have a negative influence on the freezing and thawing durability and the use of saturated aggregates could also influence the test results. For obvious reasons, however, high-strength concretes are generally made with high quality aggregates and it is probable, although this remains to be verified, that self-dessication (which is very important in low water to binder ratio pastes) tends to decrease significantly the degree of saturation of the aggregates. The results for the concretes made with the Type I cement with 6% silica fume (series C and D) are very different from those obtained with the Type III cement (with and without silica fume). These results indicate that, for this Type I cement with 6% silica fume and even after seven days of curing, air entrainment is required for good freezing and thawing resistance at a water to binder ratio of 0,30. For the concretes of series C, the critical value of the spacing factor is obviously higher than 203 ktm (concretes CL-1, CL-3 and CL-7 all showed an excellent durability) and lower than 503 lxrn (concretes CH-1, CH-3 and CH-7 were all severely damaged by the freezing and thawing cycles). This range of spacing factors generally corresponds to air-entrained concretes. Two of the three concretes of series D (with a water to binder ratio of 0,26 and cured for 3 and 28 days respectively) were damaged during the tests. This damage, however, was not very important and only occurred after a very large number of cycles. This indicates that the spacing factor of these concretes (690 ~tm) is very close to the critical value. For the Type I cement with 6% silica fume that was used, the limiting value of the water to binder ratio below which air entrainment is not required is probably thus of the order of 0,25. It is interesting to note that Foy et al. (6), using the same Type I cement without silica fume (at a water to binder ratio of 0,25) and a lower quality aggregate found that the critical spacing factor was of the order of 700 I~m when the concretes were cured for 7 days. Hammer and Sellevold (7), using low temperature calorimetry, determined the amount of freezable water in various high-strength concretes and then measured their frost resistance. They found that non-air-entrained concretes (with a water to binder ratio of 0,26) containing almost no freezable water were severely damaged by freezing and thawing cycles. They suggested that this damage was due to the difference between the coefficients of thermal expansion of the paste and of the aggregate, and not to freezing effects. This explanation cannot be valid for the test results described in this paper which indicate very clearly that there exists, for the concretes of series C and D (the only two series for which a certain number of specimens were damaged by freezing and thawing cycles), a critical value of the air-void spacing factor. If such a critical value exists, then freezing effects are most probably the cause of the observed damage. The values in Table 7 show that, for the non-air-entrained concretes of series C and D (most of which were damaged by the freezing and thawing cycles), curing had very little influence on the test results. In certain instances, curing even seemed to have a slightly negative influence. This can not be fully explained at the present time, but could be due in part to the already low freezable water content after one day of curing and to the positive influence of permeability on frost resistance (8).

On the basis of this investigation and of previously published data by the same authors, it can be concluded that, for the Type III cement that was used (with or without silica fume),

851

852

M. Pigeonet al.

Vol.21, No. 5

the limiting value of the water to binder ratio below which air entrainment is not necessary for adequate frost protection is most probably higher than 0,30, even after only one day of curing. For the Type I cement that was used (with or without silica fume), the results indicate that the limiting value is of the order of 0,25 (in this case, because of the slower hydration rate of Type I cement, and although only one day of curing was sufficient for the air-entrained concrete to become frost resistant, a seven day curing period is nevertheless recommended). It is thus clear that cement has a large influence on the frost resistance of high-strength concretes. It is also clear that guidelines for the use of such concretes under severe winter conditions will have to be very safe, because it is not evident that all Type III cements, for instance, will have such a good performance and because laboratory tests can not perfectly reproduce actual field exposure conditions. The influence of deicer salts must also, of course, be investigated. As the test results show, the spacing factor of the air voids in non-air-entrained concretes is usually in the 600 ~tm to 1,0 mm range. In certain cases, however, it can also be as high as 1,5 ram. This could be particularly true with high-strength concretes designed for maximum strength and thus an air content as small as possible. It remains therefore to be verified if the suggested limiting values of the water to binder ratio below which air entrainment is not necessary for good frost resistance would still be applicable for concretes with extremely high values of the spacing factor (in other words, the possibility of a critical spacing factor being in the 1,0 mm to 1,5 mm range must be investigated). Acknowledgements

The authors are grateful to the Natural Sciences and Engineering Research Council of Canada for its financial support for this project. References

(1) Aitcin, P.C., 1987, "La technologic des bttons/t tr~s haute performance en Amtrique du Nord", Materials and Structures/Mattriaux et Constructions, Vol. 20, pp. 180-189. (2) de Larrard, F., 1988, "Formulation et proprittts des bttons de hautes performances", Rapport de recherche LPC No. 149, Laboratoire Central des Ponts et Chausstes, Pads, 335 p. (3) Gagnt, R., Pigeon, M. and Aitcin, P.C., 1990, "Durabilit6 au gel des bttons de hautes performances mtcaniques", Materials and Structures/Mattriaux et Constructions, Vol. 23, pp 103-109. (4) Maso, J.C., 1980, "La liaison entre les granulats et la p~te de ciment hydratt", 7th International Symposium on the Chemistry of Cement, Paris, Vol. 1, VII-l, pp. 1-14. (5) Saucier, F., Pigeon, M. and Cameron, G., 1990, "Air Void Stability, Part V" Temperature, General Analysis and Performance Index", ACI Materials Journal (in press). (6) Foy. C., Pigeon, M. and Banthia, N., 1988, "Freeze-Thaw Durability and Deicer Salt Scaling Resistance of a 0,25 Water-Cement Ratio Concrete", Cement and Concrete Research, 18(4), pp. 604-614. (7) Hammer, T.A. and Sellevold, E.J., 1990, "Frost Resistance of High-Strength Concrete", Second International Symposium on Utilization of High-Strength Concrete, Berkeley, 17p. (8) Powers, T.C., 1949, "The Air Requirement of Frost Resistant Concrete", Proceedings of the Highway Research Board, Vol. 29, pp. 184-211.