Expansive Characteristics of Hydrated Limes and the ...The hydrated limes in each series are arranged in table 1 by order of increasing percentage of calculated unhydrated oxides

U. S. Department of CommerceNational Bureau of Standards

Research Paper RP1917Volume 41, September 1948

Part of the Journal of Research of the National Bureau of Standards

Expansive Characteristics of Hydrated Limes and theDevelopment of an Autoclave Test for Soundness

By Lansing S. Wells, Walter F. Clarke, and Ernest M. Levin

A technique for preparing, curing, and autoclaving 1- by 1- by 10-inch cement-limebars was developed, and the expansive characteristics of 80 commercial hydrated limeswere determined. On the basis of chemical analysis and percentage of unhydrated oxide,the hydrated limes were classified into four series: high-calcium, regularly hydrated dolomitic,highly hydrated dolomitic, and magnesian. Data on the expansions of cement-lime barsprepared in the proportions of 2 parts cement to 1 part lime, 1 part cement to 1 part lime,and 1 part cement to 2 parts lime, by weight, and autoclaved to 295 pounds per square inchgage pressure for 3 hours, showed that bars prepared with the regularly hydrated dolomiticlimes, which had the highest percentages of unhydrated oxides, had the highest per-centages of expansion. The high-calcium limes, characterized, in general, by the lowestpercentages of unhydrated oxides, gave the lowest percentages of expansion. Most of thehighly hydrated dolomitic limes had percentages of unhydrated oxide and expansion thatwere comparable to those of the high-calcium limes. An increase in the proportion of limein the cement-lime bars was attended by an increase in expansion. The method for deter-mining the linear expansion of cement-lime bars autoclaved at a steam-gage pressure of295 lb/in.2 (equivalent to a temperature of 216° C.) was found to be reproducible, by threeindependent operators. The effect of 17 different Portland cements on the expansion ofcement-lime bars showed that the expansion values for a particular lime tended to increaseas the expansion value of the constituent cement increased. Arbitrarily subtracting the

i expansion of the neat cement from the total expansion gave the most uniform result for the"net" expansion of the lime. With an autoclave specially modified for ascertaining thebehavior of cement-lime bars during the course of autoclaving, it was found that only aslight amount of the total expansion occurred before a temperature of 150° C. was reached,but above 150° C. a rapid expansion rate was noted, which in turn tapered off before atemperature of 216° C. was reached. Retarding the rate of heating resulted in a decreasein the total expansion. Finally, from criteria set forth for a procedure for determining thesoundness of hydrated limes, a test is proposed with a suggested limit of expansion of 1.0percent.

I. Introduction

It is highly important that a mortar or plaster,after being gaged with water and setting in place,shall not undergo appreciable change in volume.Since it was known that the hydration of theconstituents of lime is attended by a markedincrease in volume, a study was made at theNational Bureau of Standards of the hydration ofmagnesia in dolomitic hydrated limes and put-

ties [I].1 The publication of the paper soon ledto consideration of the effect (if any) of unhydratedmagnesia in partially hydrated dolomitic limes,and the Bureau became interested in this problemas it relates to plaster.

Extensive surveys showed that plaster failurescharacterized by the formation of bulges (largeblisters) in the white coat, as illustrated in

1 Figures in brackets indicate the literature references at the end of thispaper.

Expansive Characteristics of Limes 179

figures 1 and 2, were widespread. Some of thesalient features of the Bureau's findings weresummarized recently [2] as follows:

"(1) Not a single instance has been found wherebulging occurred in a white-coat plaster made withhigh-calcium lime; (2) ordinary dolomitichydratedlimes contain large amounts of unhydrated MgOand little or no free CaO; (3) hydration of the freeMgO to Mg(OH)2 is accompanied by markedexpansion; and (4) after several years, the forcesset up by this hydration and attendant expansioncan and do cause the formation of bulges and theshearing of the white coat from the underlyingcoats." A more detailed paper is being preparedon this subject.

As a result of these studies, the Bureau wasrequested by other Government agencies to assistin formulating a specification that would excludepartially hydrated limes containing a high per-centage of unhydrated oxides. Accordingly, thefollowing tentative amendment to Federal Speci-fication SS-L-351 for hydrated lime [3] wasprepared: "The total free (unhydrated) calciumoxide (CaO) and magnesium oxide (MgO) in thehydrated product shall not exceed 8 percent byweight (calculated on the 'as received' basis)."A method was given for determining the quantityof unhydrated oxides.

Shortly thereafter certain of the manufacturersof regularly hydrated dolomitic lime started pro-ducing a more completely hydrated lime. Inmost instances this has been accomplished by in-stalling autoclaves to hydrate the lime at elevatedtemperatures and pressures.

The 8-percent limit of the unhydrated oxide inhydrated limes recently has been included inspecifications of the American Society for TestingMaterials [4], as well as in those of the AmericanStandards Association [5] for the hydrated limeto be used in white-coat plaster.

•Nevertheless, a limitation on the percentageof unhydrated oxide has been criticized on thescore that the quantity of unhydrated oxideitself may not be a complete indication of theinherent expansive characteristics of a hydratedlime. Such an objection would be obviated by aperformance test. Since unsoundness usually doesnot exhibit itself for several years, a feasible per-formance test must also be an accelerated one.

Consequently, the purpose of the present in-vestigation was to study more completely the ex-

CEILING

FIGURE 1. Bulges in white coat along the side of a concretebeam, caused by the expansion attending the slow hydrationof free MgO present in the hydrated lime used in preparingthe white coat.

FIGURE 2. Bulging of white coat on a side wall, involvingseparation of the finish (or white) coat from the sandedgypsum base coat within an area of about 1 sq yd.

pansive characteristics of hydrated limes and toobtain data and information that might be usedin formulating an accelerated performance testacceptable for a specification for the soundness ofhydrated lime.

II. ScopeTo procure a fair cross section of the structural

hydrated limes being produced in the United States,80 samples representing the different types wereobtained from several widely separated lime-pro-ducing centers.

Because there is no satisfactory test procedure

180 Journal of Research

for ascertaining volume-change characteristics ofhydrated limes alone, it is necessary to gage themwith some cementitious material having hydraulicproperties and then to determine the expansiveproperties of the resultant set mixture. Afternumerous experiments, portland cement waschosen as the most suitable gaging material.

It soon became evident that there were numer-ous factors that would have to be considered beforethe technique of testing cement-lime specimenscould be developed. Among these were: The pro-portion of cement to hydrated lime, the prepara-tion and curing of the specimens, and the proce-dure of autoclaving.

Since autoclaves and auxiliary apparatus fortesting the soundness of portland cement [6] arealready available in numerous testing laboratoriesthroughout the United States, attention was di-rected toward developing a procedure for testingcement-lime specimens that would utilize thisequipment.

Cement to lime mixtures proportioned 2:1, 1:1,and 1:2, by weight, and also a 1:1:5 cement, lime,sand mixture were tested. Studies were made ofthe relation of autoclave expansion values to thequantities of unhydrated oxides as determined bychemical analyses.

The reproducibility of the test results obtainedby the use of the technique developed for the testswas obtained by comparing the values obtainedby independent operators.

Autoclave tests were made of the effect of 17different portland cements on the expansioncharacteristics of cement-lime mixtures.

A special apparatus was designed to measurethe expansion of test specimens during the courseof autoclaving, which facilitated studies of theeffect of the rate of heating the autoclave to agiven final temperature, the final temperaturei tself, and the duration of autoclaving.

On the basis of the techniques developed andthe results obtained in this investigation, sugges-tions are made for a test procedure for determiningthe autoclave expansion of hydrated lime and thelimits that appear applicable.

III. Materials1. Hydrated Limes

Samples of the 80 representative hydratedlimes, in bag quantities, were received directly

from the manufacturers. Upon receipt, the ma-terials were stored in airtight containers untilused.

The chemical analyses were made in accordancewith Federal Specification SS-L-351 for HydratedLime [7, 8], and the percentage of unhydratedoxide was calculated in accordance with Pro-posed Amendment 1 to this specification [3, 8].The data are recorded in table 1, columns 2 to 10,inclusive.

On the basis of the chemical analyses, thehydrated limes were classified as high-calcium,dolomitic, and magnesian. Those containing lessthan 5 percent of total magnesia arbitrarily wereclassified as high-calcium, series A; those with morethan 25 percent as dolomitic, series B and C; andfour limes, having between 5 and 25 percent ofmagnesia, as magnesian, series D. The hydrateddolomitic limes were further subdivided into "regu-larly hydrated/' series B, where the major portionof the magnesia had been left unhydrated, and"highly hydrated/7 series C, where the major por-tion of the magnesia had been hydrated. Thedesignation highly hydrated follows the terminolo-gy already used in a recent publication [9] andseems preferable to such terms as "autoclaved" or"pressure hydrated limes", since some of the newlimes are produced by methods that do not involvesteam pressures above atmospheric.

The hydrated limes in each series are arrangedin table 1 by order of increasing percentage ofcalculated unhydrated oxides. With the exceptionof the highly hydrated limes 67, 68, 71, 74, 75, and76, which were experimentally produced, all ofthe hydrated limes were commercially available.

The subdivision of the hydrated limes into seriesA, B, C, and D and the arrangement in each seriesby order of increasing percentage of unhydratedoxides will be maintained in subsequent tables andfigures.

No distinction between "mason's" and "finish-ing" hydrated limes is being made in this investi-gation, because the kind of unsoundness char-acterized by an over-all increase in volume is ofconcern whether a lime is to be used in a mortaror in a plaster.2

2 I t should be emphasized that this kinc| of unsoundness is not to be con-fused with the kind resulting in the formation of small crater-like holes knownas "pits" or "pops", which sometimes occur on plastered surfaces.


TABLE 1. Chemical analysis and calculated percentage of unhydrated oxides for 80 hydrated limes, together with the percentagesof linear expansion of cement-lime bars prepared with these limes in the proportion of 2 cement'.l lime; 1 cement'.l lime;and 1 cement:2 lime, by weight, and subsequently autcclaved at a steam-gage pressure of 295 lb/in.2 for 3 hr

[The specimens were all prepared with cement 10, which, without the addition of hydrated lime, had a linear expansion of 0. ll percent]

Limenumber

1

Chemical analysis (oxide composition)

FreeH2O

2

Com-binedH2O

3

C O 2

4

SiO2

5

R2O3

6

C a O

7

M g O

8

Total

9

Calcu-lated un-hydrated

oxides

10

Linear expansion cement-limebars in the proportion:

2 C : 1 L

11

1 C : 1 L

12

1 C : 2 L

13

SERIES A—HIGH-CALCIUM H Y D R A T E D LIMES

12345

6789

10

1112131415

161718192021

0.57.41.24.36.29

.66

.59

.83

.54

.54

.53

.29

.75

.68

.44

.34

.58

.30

.40

.33

.70

23.7423.1823.5923.6823.90

23.4223.6623.2623.2622.96

23.0623.0022.9522.5823.07

22.8822.9923.2322.2022.4822.01

0.511.581.020.56

.39

.99

.701.100.681.56

0.70.97.67

1.270.33

1.190.65

.271.370.871.01

0.54.58.44.60.50

.56

.16

.20

.86

.42

.74

.801.141.021.09

0.581.000.701.880.861.31

0.20.38.32.42.30

.14

.55

.55

.48

; 2 4

.60

.50

.78

.64

.50

.38

.46

.46

.92

.16

.76

73.9272.8273.9773.9073.62

73.1773. 6573.4773.3173.22

72.6073.2971.3571.8972.78

73.0071.5573.6872.0271.2372.42

0.46.67.33.35.82

.60

.69

.72

.81.96

1.040.611.821.191.06

1.412.411.181.093.021.29

99.9499.6299.9199. 8799.82

99.54100.00100.1399.9499.90

99.2799.4699.4699.2799.27

99.7899.6499.8299.8898.9599.50

0.0.0.0.0.0

.0

.1. 5. 8. 8

1.01.11.11.21.6

1.71.82.22.33.23.9

0.13.12.12.12.10

.12

.14

.14

.14

.16

.14

.13

.13

.13

.12

.16

.17

.16

.13

.19

.17

0.18.19.16.15.13

.20

.18

.19

.21

.21

.19

.19

.20

.18

.18

.23

.35

.25

.24

.52

.24

1.11.10.71

.56

.24

1.41.00.90

.86

.69

1.51.30.78

.59

.97

1.02.01.31.24.11.8

SERIES B—REGULARLY H Y D R A T E D DOLOMITIC LIMES

2223242526

2728293031

3233343536

3738394041

42434445464748

0.36.19.25.30.35

.30

.13

.28

.28

.22

.15

.29

.20

.20

.19

.24

.04

.21

.17

.23

.16

.20

.22

.43

.04

.30

.50,

20.7018.2117.5119.0318.21

17.9617.9016.6317.5817.41

16.6316.9917.0117.2617.25

16.9416.5615.9616.6516.53

15.9716.2916.2815.8115.9115.4514.58

0.991.642.570.451.32

0.631.013.131.181.73

2.791.451.790.88

.50

.94

.982.090.901.02

1.591.310.751.451.050.791. 25

1.261.750.821.041.20

0.94.12.20.28.14

.26

.34

.40

.34

.54

.20

.20

.14

.20

.24

.37

.18

.80

.10

.141.421.34

0.661.070.36

.48

.44

.36

.16

.18

.22

.14

.16

.30

.28

.22

.48

.18

.10

.10

.12

.12

.53

.16

.54. .12

.18

.46

.68

44.8145.1752.5846.1447.51

46.7247.4146.9847.5847.00

46.8147.2147.0647.4347.29

47.4348.4347.7647.8347.79

48.4547.8547.6346.7848.3247.7048.27

31.1231.8325.7932.5430.64

33.0833.0932.5132.6233.30

32.9332.5133.2432.9833. 53

33.9732.9833.6934.1034.06

33.0534.0133.6335.2134.3033.8233.00

99.9099.8699.8899.9899.67

99.9999.8299.9199.7499.94

99.7399.0999.9899.3199.78

99.9099.2999.9599.9799.99

100.12100.0099.8599.9099.9499.9499.62

16.122.122.122.722.9

25.926.226.226.426.5

26.827.127.427.728.5

29.329.930.430.430.5

30.730.830.832.132.532.834.3

2.57.83.32.01.3

4.17.24.64.04.5

6.35.74.0

10.59.3

2.612.66.8

11.010.4

5.16.68.04.0

11.76.58.1

8.516.59.97V64.6

10.012.99.28.3

10.5

11.817.310.016.416.9

7.117.711.713.813.5

12.712.215.89.9

17.814.117.8

12.714.413.112.0

13.6

15.616.5

10.6

14.514.2

13.617.910.7


TABLE 1. Chemical analysis and calculated percentage of unhydrated oxides for 80 hydrated limes, together with the percentage8

of linear expansion of cement-lime bars prepared with these limes in the proportion of 2 cement: 1 lime; 1 cement:1 limey

and 1 cement:2 lime, by weight, and subsequently autoclaved at a steam-gage pressure of 295 Ib/in.2 for 3 hr—Continuecf

[The specimens were all prepared with cement 10, which, without the addition of hydrated lime, had a linear expansion of 0.11 percent]

Limenumber

1

Chemical analysis (oxide composition)

FreeH2O

2

Com-binedH2O

3

CO2

4

SiO2

5

R2O3

6

CaO

7

MgO

8

Total

9

Calcu-lated un-hydrated

oxides

10

Linear expansion cement-limebars in the proportion:

2C:1L

11

1C:1L

12

1C:2L

13

SERIES C—HIGHLY HYDRATED DOLOMITIC LIMES

4950515253

5455565758

5960616263

6465666768

6970717273

«745«776

0.04.21.16.00.51

.04

.16

.24

.09

.02

.17

.17

.23

.25

.17

.09

.19

.08

.16

.14

.09

.20

.07

.13

.09

.30

.28

.15

25.4125.2825.5225.4124.56

25.3125.1525.2225.1424.77

24.4724.8724.5824.3725.18

25.6723.6624.8223.5922.40

22.7223.3522.4123.8323.58

20.2020.6520.98

1.441.721.321.331.87

1.561.360.661.751.16

2.791.121.982.000.37

.383.120.872.303.31

3.401.183.751.060.91

2.832.643.35

1.060.24

.23

.261.56

0.18.20

1.660.261.98

0.181.530.181.451.82

0.14.16.02

1.281.64

0.222.180.14

.20

.14

4.483.440.28

0.52.18.35.36.50

.18

.20

.56

.10

.70

.14

.62

.16.62.44

.08

.20

.19

.48

.64

.16

.89

.30

.12

.26

.76

.72

.20

41.8742.5843.0142.4741.83

42.5942.7041.9942.6842.06

42.4941.9542.9142.1042.12

43.2242.7843.4142.3742.13

42.8942.8943.0243.7344.15

41.6642.3044.03

29.4529.7529.6529.8829.09

30.0829.5729.7130.0429.23

29.7529.6829.1429.4629.82

30.1729.0930.2429.5829.43

29.9329.4330.1830.5830.84

29.4029.9130.97

99.7999.96

100.2499.7199.92

99.9499.34

100.04100.0699.92

99.9999.9499.18

100.2599.92

99.7599.2099.6399.7699.69

99.41100.1299.8799.6599.97

99.6399.9499.96

1.42.22.32.32.6

2.62.72.92.93.0 ,

3.03.23.23.43.4

3.54.05.15.26.6

6.86.97.57.79.0

11.611.712.6

0.18.21.26.21.26

.22

.23

.21

.25

.25

.24

.24

.22

.22

.22

.24

.22

.34

.41

.38

.73

.88

.54

.63

.82

.48

.551.7

0.25.30.42.311.1

0.36.48.29.37.33

.36

.50

.31

.40

.36

.50

.311.22.62.4

5.17.15.23.35.6-

2.83.27.1

0.85

3.11.35.8

2.21.91.92.11.4

2.02.50.953.3

3.60.655.69.08.4

8.911.2

8.59.1

8.88.38.2

SERIES D—MAGNESIAN HYDRATED LIMES

7778

«79«80

0.38.39.99.52

21.6721.9217.6019.79

0.67.96

4.951.10

2.260.762.472.64

0.82.78

3.131.30

67.7862.9664.6157.13

6.2411.886.25

16.88

99.8299.65

100.0099.36

5.97.29.8

12.7

0.19.22.22.28

0.50.39.54.70

3.61.14.62.8

1 Failed to meet the requirement of Federal Specification SS-L-351 that the total CaO plus MgO shall be not less than 95%, calculated on nonvolatile basis.


2. Cements

Seventeen brands of portland cement were usedin the investigation. The cements conformed toFederal Specification SS-C-192, Type 1 [10].

The autoclave expansions of the cements alonewere determined in accordance with a test thathad been adopted as standard [6] and are listedin table 2.

TABLE 2. Autoclave expansion of 1- by 1- by 10-in. neatspecimens of portland cement for determination of soundnessof portland cementl

Cementnumber

123456

Linear ex-pansion

0.00.01.02.04.04.04

Cementnumber

789

101112

Linear ex-pansion

0.06.07.11.11.11.17

Cementnumber

1314151617

Linear ex-pansion

0.18.19.23.27.30

1 Standard method of test for autoclave expansion of portland cement.ASTM Designation: C151-43, ASTM Standards, part II, 14 (1946); alsoAmerican Association State Highway Officials Standard AASHO No.T107-45; as well as Federal Specification SS-C-158b-46 for Cement, Hy-draulic; General Specifications (methods for sampling, inspection, and test-ing) [6].

3. Sand

The sand used in the preliminary tests oncement-lime-sand mortars was natural silica sandfrom Ottawa, 111., graded to pass a No. 20 sieveand be retained on a No. 30 sieve. It conformedto Federal Specification SS-C-158b, paragraphF-51 [11].

IV. Preparation and Curing of TestSpecimens

All mixtures were prepared and all tests wereperformed in a laboratory in which the tempera-ture was maintained at 21° ±1° C (70° ±2° F),and the relative humidity at 55 ±5 percent. Theatmosphere of the moist closet was maintained atthe same temperature as the laboratory and at arelative humidity of 95 ± 3 percent.

The cement-lime test specimens, in the severalseries of tests, were proportioned by dry weight,namely, 300 g of cement to 300 g of lime, 200 g ofcement to 400 g of lime, and 500 g of cement to250 g of lime; hereafter designated as 1C:1L,1C:2L, and 2C:lL, respectively. A limited num-ber of mortar specimens was proportioned by dry

weight in the ratio of 150 g of cement to 150 g oflime to 750 g of sand; hereafter designated as1C:1L:5S. These quantities in each case weresufficient for the preparation of two test specimens.

To prepare cement-lime bars suitable for auto-claving, it was found necessary to modify themethod used for making portland cement bars asgiven in Federal Specification SS-C-158b.

In the, preparation of the cement-lime bars, theweighed amounts of cement and lime were shakenvigorously in a 2-qt jar until the mixture appeareduniform. The dry mix was spread on a smooth,nonabsorbent, noncorrosive plate and trowelleduntil visible agglomerates of lime were dispersed.Then, after collection of the material into a trun-cated cone, an inner and outer crater were formedin the dry mix with the end of the handle of atrowel.

For the remainder of the operation, the handswere protected with rubber gloves. During aninterval of 2 min, (1) a measured quantity of cleanwater (temperature 21° ±1.7° C) was added; (2)the dry material was folded over the exposed sur-face of the water; and (3) the dry material was cutinto the water with the edge of a trowel. The mixwas worked by kneading, squeezing, and poundingfor an additional period of 3 min. The mass wasthen allowed to stand undisturbed for 1 min andfinally vigorously reworked for 1 min.

The above procedure was found necessary be-cause of the difficulty in wetting cement-limemixes and the tendency on the part of the operatorto add excessive quantities of mixing water tomost cement-lime mixes. By following these mod-ifications, many mixes that at first appeared toodry eventually acquired the desired consistency bythe vigorous working. The advantages of thisprocedure were twofold: (1) The quantity of mix-ing water could be reduced, and consequently thebreakage of the bars resulting from shrinkage dur-ing curing was almost wholly eliminated, and (2)the quantity of water required by different limesof the same type was, with few exceptions, reducedto a narrow range.

Following completion of mixing, the Vicat pene-tration was determined according to the proceduregiven in Federal Specification SS-C-158b [6], par-agraph F-47. Although the procedure for testingportland cement requires a Vicat penetration of10 ± 1 mm, it was found that this range was toonarrow for cement-lime specimens for easy repro-


ducibility. Since it was found that small differ-ences in consistency did not materially affect theexpansion of the cement-lime specimens, the rangewas extended to 10 ± 5 mm. In a few instancesspecimens were used even though the penetrationswere not within this range, generally because therewas insufficient material to prepare additionalspecimens.

Immediately after the determination of theVicat penetration, the entire supply of putty wasremixed for approximately 15 sec. The remixedcement-lime putty was then molded into autoclavebar molds according to the procedure given inSS-C-158b, paragraph F-61. The bar molds pro-vided for test specimens having a cross section 1in. by 1 in. and of 10-in. effective gage length,SS-C-158b, paragraph F-43 [6].

Immediately after molding each specimen, itsexposed surface was covered with a strip of waxedpaper pressed down on the mold. This waxedpaper reduced surface evaporation and minimizedthe breakage, during curing, caused by shrinkagecracks. The test specimens were stored in themolds in a moist closet for at least 20 hrs, followingthe requirements of SS-C-158b, paragraph F-62.

The cement-lime-sand mortars were preparedin accordance with Federal Specification SS-C-181b [12], considering the mixture of cementand lime as masonry cement. The consistency ofall mortars was adjusted to a flow of 110 ±2 .The casting of the mortar into autoclave bar moldsand subsequent treatment were similar to that ofthe cement-lime putties.

At 23 ± 1 hr after molding the specimens, theywere removed from the moist closet, measured forlength, and placed in the autoclaves at roomtemperature. The autoclaves were immediatelyclosed, and heat was applied.

After autoclaving, the specimens for the mostpart were submerged in water and cooled to 21° Cbefore measuring the length, as required in theprocedure of autoclaving as given in Federal Speci-fication SS-C-158b, paragraph F-63f [6]. Itwas found necessary, however, to cool those speci-mens that exhibited very high expansions (10%or more) in air to 21° C, because when cooled inwater the greatly expanded specimens were aptto disintegrate.

For expansions under 2.5 percent, the com-parator described in paragraph F-45 of SS-C-158b [6] was used. For higher expansions,

methods giving values to within a hundredth ofan inch were used.

V. Autoclaves and Methods ofAutoclaving

Four autoclaves were used in this investigation—one, horizontal, gas-fired; and three, vertical, elec-trically heated. One of the latter was equippedwith a special device for the continuous observa-tion of the expanding bar specimens.

1. Horizontal Gas-Fired Autoclave

The detailed description of the horizontal, gas-fired autoclave will be omitted because it was notused in this investigation after it was found thatthe expansions obtained with it were somewhaterratic and did not agree well with the concordantresults obtained with the three vertical autoclavesthat were electrically heated.

It was found, for example, that the cement-limespecimens placed in the upper part of the chamberof the gas-fired autoclave expanded more thancompanion specimens placed in the lower part.It is believed that the reason for this is the tend-ency for the steam to become superheated in theupper portion of a horizontal autoclave that is gas-fired.

2. Vertical, Electrically Heated Autoclaves

The three electrically heated autoclaves were ofsuch a design that they met the requirements ofFederal Specification SS-C-158b [6] for an auto-clave to be used in the testing of portland cement.Since preliminary experiments showed that therate of heating cement-lime specimens markedlyaffected the resultant expansion of the bars, therate of heating each autoclave to 295 lb/in.2 wasdetermined. It can be seen from figure 3 thatwith each autoclave the pressure of the saturatedsteam in the autoclave was raised to a gage pressureof 295 lb/in.2 in 1 to 1}̂ hr, as required by thestandard method for testing cement [6].

Autoclaves designated as 1 and 2 were used in thepreliminary experiments wherein the cement-limeand cement-lime-sand specimens were autoclavedat gage pressures of 25, 120, and 295 lb/in.2 andalso in the bulk of the work where the auto-claving was done in accordance with the standardmethod for portland cement specimens (295lb/in.2).


295250 ^

ISO %

100 $

50 IZ5 X

Hea/-//7g ^ctiedu/e of ••

A Autoclave I, standardX Autoclave 2,© Autoclave 3,O Autoclave 3, spec/ally equippedf (nor/770/ heating )• Autoclave 3, * " > (refarJec/ » )

I i I • I i i0 10 20 30 40 SO 60 70 80 90 100 110 120 130 /4O ISO 160 170 180

77/775, minutesFIGURE 3. Time-temperature curves of heating autoclaves to 216° C (equivalent to steam-gage pressure of 295 lb/in.2).

The method of autoclaving at 295 lb/in.2 wasused in accordance with ASTM and FederalSpecifications for testing portland cement. Whenthe autoclaving was done at 25 and 120 lb/in.2,the heating schedule of the standard autoclaves 1and 2 (fig. 3) was followed to the required pressure,after which the pressure, or temperature, wasmaintained by manual control.

3. Special Autoclave for Continuous Observationof Expansion3

Figure 4 shows the apparatus that permittedcontinuous observation of the expansion of a bar(cement or cement-lime specimens) during thecourse of autoclaving. Figure 5 is a schematicdrawing of the specialty equipped, electricallyheated autoclave illustrated in figure 4. Thedevice consisted of a high-pressure gage-glasschamber, A, mounted on top of autoclave 3 bymeans of a short length of threaded high-pressurepipe screwed into a threaded hole tapped throughthe cover plate of the autoclave. A free-movingsquare steel rod, B, extended from the inside ofthe autoclave through the connecting pipe into thegage-glass chamber. The top of the rod waspointed in order to serve as a reference. Thelower end, which extended through the connecting

3 Special acknowledgment is due Emil Trattner for his graciousness inconsenting to the description of his ingeneously designed autoclave, whichpermitted, for the first time, the observation of a test specimen during thecourse of autoclaving.

pipe into the autoclave chamber, was providedwith a steel disk, K, that rested on the top of thetest specimen, H, which was held vertically in arack, R, placed in the chamber of the autoclave.With this arrangement, an expanding test barcaused the steel rod to rise in the gage-glasschamber, and the movement could be observedby sighting through the parallel glass windows ofthe chamber. By having the rod square, passageof steam and condensed water in and out of thegage-glass chamber was assured. A small steelspring, I, was attached to one side of the steel rodto prevent the rod from falling out of the gage-glass chamber while the cover plate was beingplaced and removed.

It was found that bars that were being contin-uously wetted during autoclaving gave erraticand unusually large expansions. Therefore, it wasnecessary to divert condensed dripping wateraway from the test specimen located directlybelow the gage-glass chamber, and hence a copperflashing, J, (fig. 5) was attached to the movablesquare steel rod directly above the steel disk thatrested on the top of the specimen. As a check onthe performance of the test bar carrying themovable steel rod, a duplicate bar was alwaysplaced in the same autoclave away from the drip-ping condensed water; after autoclaving, both barswere measured with a comparator.

In operation, the test bar was placed in a rack


FIGURE 4. Assembly of specially equipped autoclave forcontinuous observation of expanding bar-specimens duringautoclaving.

A, Gage-glass chamber mounted on top of autoclave; B, movable square steelrod extending from top of specimen into the gage-glass chamber; C, ventvalve; D, cathetometer for measuring movement of movable steel rod; E,stone base for cathetometer; F, variac; G, voltmeter.

in the autoclave, and the cover plate with theattached gage-glass chamber, together with theinserted steel rod held in place by the restrictingspring, was lowered into place in such a mannerthat the steel disk at the lower end of the rodcame to rest on the flat top of the test bar.4 Thecover plate was then bolted down and the ventvalve, C, was opened.

The position of the point of the free-movingsteel rod was located by means of the cathetom-eter, D, (fig. 4), which was firmly mounted on astone base, E.

The heaters were then turned on, and the waterin the bottom of the autoclave was brought toboiling. After the air had been displaced bysteam, the vent valve, C, was closed. The auto-clave was heated to a final pressure of 295 lb/in.2,at which pressure the thermometer reading was216° C. The rate of heating was controlled

• The steel reference point that normally would be present at the top of thetest bar, for measurement with a comparator, had been removed duringfabrication.

FIGURE 5. Details of gage-glass chamber and movable steelrod.

A, Gage-glass chamber; B, movable square steel rod; C, vent valve; H, bar-specimen; I, spring for holding movable square steel rod in place whileassembling apparatus; J, copper flashing attached to movable steel rod todivert condensing water running down rod away from specimen H; K, steeldisk attached to lower end of movable steel rod and resting on top of speci-men; R, rack for holding specimen upright.

manually by means of the several heating unitsof the autoclave. A variac, F, connected througha voltmeter, G, added flexibility to the control.

Readings of the position of the rising steel rodwere obtained easily until the vent was closed.Immediately thereafter, readings could not betaken with the cathetometer because of the con-densation of drops of water on the inside of theglass windows of the gage-glass chamber. This


condition lasted for about 1 or 2 min. However,since the gage-glass chamber acted as a condenser,the interior soon became completely filled withwater.5 From then on, the point of the steel rodwas clearly visible.

The shortest time within which the gage pressurecould be raised to 295 lb/in.2 was about 97 min.,compared to the 70 min. required to heat the sameautoclave without the gage-glass chamber. This97 min. is beyond the upper limit of l}{ hr specifiedin SS-C-158b [6]. The increase in time wascaused by the greater heat input required by thecondensing and refluxing of the water in the gage-glass chamber and by the greater heat capacity ofthe specially equipped autoclave. The normalheating schedule of specially equipped autoclave3 is shown in figure 3.

Throughout the entire course of autoclaving,readings were taken of the position of the point ofthe moving steel rod with the cathetometer, of thepressure with the steam gage, and of the temper-ature with a thermometer inserted into a thermom-eter well in the cover plate.

The design of the apparatus was such thatthermal expansions were included in the measure-ments. The thermal expansions of concern werethose of the free-moving steel rod, the test bar, andthe part of the autoclave below the lower end of thetest bar. This combined thermal expansion wasdetermined by inserting a test specimen that hadpreviously been autoclaved at a gage pressure of295 lb/in.2, setting up the equipment as in a normaltest procedure, and heating to 295 lb/in.2. Themovement of the point of the steel rod wasobserved and taken as the thermal expansion ofthe combined parts. The previously autoclavedtest bar used in this experiment was measuredwith a comparator before and after autoclavingand was found to have shown no subsequentexpansion. These results were checked by replac-ing the cement-lime test bar with a steel bar of .equal length and of known coefficient of expansion.The magnitude of the thermal expansion comparedwith the observed total expansion of the testspecimens will be discussed later.

6 To compensate for water condensed in the gage-glass chamber, 590 ml ofwater, rather than the usual 500 ml, was placed in this special autoclave at thestart.

VI. Results and Discussion

1. Preliminary Experiments

Table 3 gives data on some preliminary experi-ments on the expansion of cement-lime andcement-lime-standard sand specimens that wereautoclaved at steam-gage pressures of 25, 120,and 295 lb/in.2, respectively. It can be seen that,without exception, increasing the pressure ofautoclaving from 25 to 120 to 295 lb/in.2 wasattended by an increase in the percentage of thelinear expansion.

Also, increasing the proportion of lime tocement from 2C:lL to 1C:1L invariably ^re-sulted in a marked increase in expansion at &Mipressure of autoclaving. Increasing the propor-tion of lime in the cement-lime specimens resultsin a decrease in the strength of the specimen andtherefore a decrease in the ability of the speci-men to resist expansive forces. It is believedthat an increase in expansion attending a decreasein the strength accounts for the relatively highexpansion values of the 1C:1L:5S group. Infact, the weakness of the sanded bars and thedifficulty of their preparation and handling dis-couraged further work with sanded specimens.In addition, the possibility of a variable beingintroduced because of a probable pozzolanicreaction between the lime and sand during auto-claving further operated against their use.

Figure 6 shows the increase in expansion at-

Regu/aHy hydroylecf c/o/orrrif/c lime 26

5 7Time, hr

FIGURE 6. Increase in expansion of cement-lime barsattending increase in time of autoclaving at gage-pressureof 25 W/in2.


TABLE 3. Expansion of cement-lime and cement-lime-standard sand bars autoclaved at steam-gage pressures of 25, 120; and295 Ib/in.2 for 3 hr, respectively

[Cement 10 was used throughout]

Lime Num-ber

1

Type oflime

2

Unhydra-ted oxides

3

Mixingwater

4

Vicatpenetration

5

Flow .

6

Linear expansion after autoclaving atpressures of—

25 lb/in.2

7

120 lb/in.2

8

295 lb/in.2

9

G R O U P 1. P R O P O R T I O N OF CEMENT:LIME—2:1, BY W E I G H T

2633383955586069

(a)O(a)(•)(b)(b)(b)(b)

%22.927.129.930.42.73.03.26.8

%343435'3434293131

mm8

118

10109

108

% %0.401.53.92.10.10.06.07.23

%0.934.09.65.60.13.09.10.51

%1.35.7

12.66.80.23

.25

.24

.73

GROUP 2. P R O P O R T I O N OF CEMENT:LIME—1:1, BY W E I G H T

2633383955586069

<•)(a)(a)(•)(6)(b)(b)(»)

22.927.129.930.42.73.03.26.8

4040424143343436

1.37.76.54.80.15

.08• 1 0

1.3

3.911.614.79.30.26

.16

.223.4

4.617.317.711.70.48

.33

.505.1

G R O U P 3. P R O P O R T I O N OF CEMENT:LIME:SAND—1:1:5, BY W E I G H T

2633383955586069

(•)(•)(•)(a)(b)(6)(b)(")

22.927.129.930.42.73.03.26.8

15.815.814.815.617. 516.218.115.2

109108108112108108108110

0.724.77.04.50.06

.05

.05

.56

2.47.5

11.67.50.10

.08

.102.2

4.09.0

13.08.80.18.16.20

4.0

1 Regularly hydrated dolomitic lime.> Highly hydrated dolomitic lime.

tending' an increase in the time of autoclavingspecimens containing hydrated limes (26 and 38)at a pressure of 25 lb/in2. The expansion curvesindicate that the ultimate expansion had not beenreached after 7 hr of autoclaving. The investi-gation of the effect of time of autoclaving at apressure of 25 lb/in2. was not extended beyond7 hr, because even that is too long for a contin-uous laboratory test procedure.

It became evident from these preliminary ex-periments that the pressure of autoclaving is animportant factor with respect to the magnitudeof the expansion. It is also evident that at lowerpressures the total expansion is not reached within

a practicable time. It was found, however, thatcement-lime bars as well as portland-cement barsunderwent no further expansion at 295 lb/in2.when autoclaved for periods in excess of 3 hr. Itwas decided, therefore, to abandon further workwith lower pressures and to limit all further ex-periments to autoclaving for 3 hr at a pressure of295 lb/in 2.

2. Effect of Increasing Proportion of HydratedLime on Expansion of Cement-Lime Bars

Table 1 (cols. 11, 12, and 13) gives data on theexpansion of cement-lime bars prepared in theproportions of 2C:lL, 1C:1L, and 1C:2L,


by weighty respectively, and autoclaved for 3 hrat a gage pressure of 295 lb/in 2.

The values for the percentage of expansion re-corded in columns 11 and 13 are averages obtainedby single operators from duplicate specimens.Forty-three of the values in column 12 representthe averages for duplicate determinations of threeindependent operators; the remainder are theaverage of duplicate determinations by a singleoperator.

Figures 7, 8, and 9 show the respective ex-pansions of the 2C:lL, 1C:1L, and 1C:2L bars(cols. 11, 12, and 13 of table 1), plotted againstthe percentages of unhydrated oxides (col. 10 oftable 1) in the hydrated lime used in preparingthe bars. Several salient features are evident fromtable 1, and figures 7, 8, and 9.

From the figures it can be seen that the limesare distributed into two major groups. One iscomposed entirely of the regularly hydrated dolo-mitic limes (series B) and is characterized by con-

taining those limes having the highest percentageof unhydrated oxides and the highest percentageof expansion. The other group is composed ofthe remaining limes (series A, C, and D).

The high-calcium limes (series A) are character-ized in general by having the lowest percentagesof unhydrated oxides (0 to 3.9 percent) and thelowest percentages of expansion. This can bestbe observed from table 1.

Most of the highly hydrated dolomitic limes(series C) have percentages of unhydrated oxides(1.4 to 4.0 percent) and, with one exception, ex-pansions that are comparable to those of the high-calcium limes. These are limes 49 to 65, inclu-sive, and are all commercially available limes.The others (limes 66 to 76, incl.) have percentagesof unhydrated oxides (5.1 to 12.6) and expansionsthat are greater than those of the high-calciumlimes. Among these limes are 67, 68, 71, 75, and76, which were experimentally produced.

The four magnesian limes (series D) exhibit

>

I 9

11 s

r̂ 3

l'

i i i i i i i i i r

LEGEND

x High-ca/cium• Regukirfy hydra Me/ dofom/ficO Highly hydra/ed daio/n/YicA Mogrres/crn

ProportionCemerrt >'L/me

2:1

•• •

J L J L/O /5 2O

Unhydrafed oxides in hydrcr/ed' //'-me,25 30

I I I I35

FIGUHE 7. Linear expansion of cement-lime bars prepared with hydrated limes containing various amounts of unhydratedoxides, in the proportion of 2 cement to 1 lime (by weight) and subsequently autoclaved for 3 hr at a gage pressure of295 lb/in2.

The specimens were all prepared with cement 10, which, without the addition of hydrated lime, had a linear expansion of 0.11 percent.


id

17

16

15

14

IE —

II —

\ 10

VI 7

1 1 I 1 I 1 1 1 1 l 1 1 1 1 1 1 1

_ LE6END

x High- ca/c/'um— • f?egu/or/y hydra fed do/orrr/f/c

O Highly hydro fed do/om/h'c— A Magnes/an

ProportionCement • L/me

o o

o_ o°

° o

o

o °A A A1 1 1 1 1 1 1 1 1 t i l l

t i l l

•

1 1 1 1

1 1 I 1 J I 1 • 1 !•

• -

•

1

.I

II

-

--

1 1 1 1 1 1 1 1

9 —

6 —

5 —

4

3

2

I

JO 350 5 10 .15 20 25Unhydrafed oxides /n hydrcfed //me, percent

FIGURE 8. Linear expansion of cement-lime bars prepared with hydrated limes containing various amounts of unhydratedoxides, in the proportion of 1 cement to 1 lime {by weight) and subsequently autoclaved for 3 hr at a gage pressure of295 lb/in2.

The specimens were all prepared with cement 10, which, without the addition of hydrated lime, had a linear expansion of 0.11 percent.

unique behavior in that each magnesian limeshows a lower percentage of expansion than otherlimes with a comparable percentage of unhydratedoxides.

From table 1 it can be seen that, for the limes ofeach series, an increase in the proportion of limein the cement-lime bars is, without exception,attended by an increase in expansion. This rela-tion is also shown in figure 10, where the per-centages of linear expansion of the 2C:lL, 1C:1L,and 1C:2L bars are plotted consecutively.

Figure 10 shows, again, that as a group thehigh-calcium limes give the lowest expansions inthe three cement-lime proportions, and that amajority of the highly hydrated dolomitic limesgive expansions comparable to the high-calcium

limes. It also reveals that the expansions of theremaining 11 highly hydrated dolomitic limes,which include the six experimental limes, areintermediate between those of the high-calciumand the regularly hydrated dolomitic limes. How-ever, in the 2C:lL proportion, the expansionsof the high-calcium and all of the highly hydrateddolomitic limes fall within a narrow range—withonly one exception, under 1.0 percent. As theproportion of the lime in the bar increases from2C:lL to 1C:1L, the spread in expansion isgreatly accentuated for those highly hydrateddolomitic limes having expansions beyond therange of the high-calcium limes. The spread inexpansion is even more accentuated in the lC:2Lbars.


17

16

15

14

13

12

U

10

9

8

7

6

5

4

3

2

I I I I

X

— •

o

— A

—

-

—

—

—

—

—

—

o

—

_

X

oo

o

o

- X & x

> " x 38°/ * 5 E ^ O

x>< o

' I I I I

1 1 1 1 l 1 I l i i I i

— LEGEND —

High- cct/c/um

Regu/ar/y hydra fed do/omificHighly hydra fed do/omih'c

Magnesian

ProportionCement: L/me

1:2

o

° ° o° ° ° o

0

A

A

A •

A

L 1 I 1 I 1 I i 1 i i i i

I I I I

I I I I

I I I I • I I I I

—

•

• —

• -

* —

-

—

—

—

—

—

_

_

—

—

—

—

I I I I I I I I

30 350 5 10 15 2O 25Unhydrafed oxides in hydrafed //me, percent

FIGURE 9. Linear expansion of cement-lime bars prepared with hydrated limes containing various amounts of unhydratedoxides in the proportion of 1 cement to 2 lime (by weight) and subsequently autoclaved for 3 hr at a gage pressure of295 lb/in2.

The specimens were a 11 prepared with cement 10, which, without the addition of hydrated lime, had a linear expansion of 0.11 percent.

In figures 7, 8, and 9 the expansions of the2C:lL, 1C:1L, and lC:2L bars were plotted ineach case against the calculated percentage ofunhydrated oxide in the constituent lime. How-ever, the percentages of unhydrated oxide in the2C:1L, 1C:1L, and 1C:2L mixes are respectivelyone-third, one-half, and two-thirds of that of thepercentage of unhydrated oxide in the lime.Accordingly, figure 11 shows the relation betweenthe percentages of expansion of 2C:lL, 1C:1L,and 1C:2L bars and the corresponding percentageof unhydrated oxide in the dry mix.

Once more, it can be seen from figure 11, wherethe percentages of expansion of the 2C:1L, 1C:1L,and 1C:2L bars are plotted side by side, thatan increase in the proportion of lime in the bars is

192

accompanied by an increase in expansion. Butmore important, the three graphs making up figure11 show that the percentage of unhydrated oxidein the mix must be low if there is to be assurancethat the expansion will also be low. It is alsoevident that the limiting percentage is not thesame for each proportion of cement to lime. Thispercentage apparently narrows as the quantity ofcement decreases and the bars correspondinglybecome weaker and are less able to resist theexpansive forces attending the hydration of theunhydrated oxide. For example, with the 2C:lLbars, the total unhydrated oxide may be as. highas 4 percent and yet the expansion of the I barsremain below 1 percent. With the weaker l!C:lLbars, the percentage of unhydrated oxide} that

Journal of Research

LE6END

X High- ca/cium

• Regu/ar/y hyc/rafed c/o/omif/c

O Highly hydrafed do/orrrifie

-X

2:1 1:1 1:2Proportion of cemenf fo hydra fed //me (dy weigh f) used in preparing fhe fesf tars

10. Increase in linear expansion attending an increase in the proportion of lime in cement-lime bars auto-claved for 3 hr at a gage pressure of 295 lb/in.2

can be tolerated appears to be less, because at 4percent the expansion may exceed 6 percent.With the still weaker 1C:2L bars, the expansionof the bars with 4 percent of unhydrated oxidemay reach values in the neighborhood of 10percent. In fact, even with little or no unhy-drated oxide, the bars of the 1C:2L series showexpansions considerably greater than those of theother two mixes. It is not known whether thiscan be attributed in part to the cement itselfbecause of the inability of these weak bars toresist the expansion of some of the constituents ofthe cement. It is interesting that the uniquemagnepian limes, which show less expansion thanother limes with comparable percentage of unhy-drated' oxide, begin to show appreciable expan-sion in the 1C:2L bars.

3. Reproducibility of Determining Percentage ofExpansion of 1 Cement: 1 Lime Test Specimens

Table 4 gives data on the percentage of linearexpansion obtained by three operators with dupli-cate cement-lime bars prepared in the proportionof 1 part cement to 1 part lime, by weight, andautoclaved at a steam-gage pressure of 295 lb/in.2

for 3 hr. In addition, it gives for each operatorthe percentage of water used in fabricating theduplicate bars and the Vicat penetration of thecement-lime pastes. Forty-three samples of hy-drated limes were selected for this study. In cer-tain iristances only one of an operator's duplicatebars was autoclaved.

Although the dates at which the tests were per-formed are not given in table 4, the experiments

Expansive Characteristics of Limes798448—48- 3

193

Id

17

16

15-

14

13

12

ProportionCement': Lime

2:1

ProportionCement: Lime

1:1

\ r i i T i i i - i i r i i I i i I i I (

ProportionCement: Lime

J:2

— LEGENDX High-ca/cium• Regubr/y hydro fed' do/omi/JcO High/y hyd/rr/ed' do/omit/'cA Magnesion

I I I I I I I I I I I I I I I I I I I I

10 12 0 10 12 14 16 18 20 22

FIGURE 11.

2 4 6 6 10 12 14 16 18 0 2 4 6 QUnhydrated oxides in dry cemenf-/ime mix/c/ne, percent by tveighf

Relation between the percentage of expansion of 2C:lLy 1C:1L> and 1C.2L bars and the correspondingpercentage of unhydrated oxides in the dry mix.

purposely were not made at the same time. Thetime period between the testing by operators Band C was approximately 1 month. Operator Adid most of his testing about 1 year later.

It can be seen from table 4 that not only didthe percentages of expansion of the two duplicatebars of each operator agree closely, but that thepercentages of expansion obtained by the threedifferent operators were in good agreement.

For the purpose of a statistical treatment, thevarious hydrated limes were divided into fivenearly homogeneous groups. These include seriesA, high-calcium; series B, regularly hydrateddolomitic; series D, magnesian; and two subdi-vided groups of series C, highly hydrated dolo-mitic. One of these groups of series C includedthe highly hydrated dolomitic limes having per-centages of both unhydrated oxide and expan-sion of the same order of magnitude as those ofthe high-calcium limes, namely, limes 49, 54, 56,57, 59, 61, and 65, all of which had expansions of

less than 0.40 percent. The other group containedthe limes in series C having the high expansions:those ranging from 1.1 to 5.6 percent. A singleclassification analysis of variance [13] was made foreach group to test whether there was a signifi-cant difference among the averages obtained bythe three operators. No significant differencewas found.6

4. Effect of Different Portland Cements on Expan-sion of Cement-Lime Bars

Table 5 gives data on the expansion of cement-lime bars prepared with 17 cements (ranging inexpansion from 0.00 to 0.30 percent). The 30selected limes are grouped, as heretofore, in fourseries.

The values for the percentage of expansion ofthe 17 cements (tested in accordance with the

• The authors thank Stephen W. Benedict for his statistical analysis ofthe reproducibility of determining the percentage of expansion of cement-lime test specimens.


TABLE 4. Reproducibility obtained, by three independent operators, for the expansion of cement-lime bars prepared in theproportion of 1 cement: 1 lime {by weight), and autoclaved at a steam-gage pressure of 295 Ib/sq in.2 for 3 hr. The specimenswere all prepared with cement 10, which, without the addition of hydrated lime, had a linear expansion of 0.11 percent

Limenum-

1

12345

611

131415

2227282930

3234353637

404143444547

Calcu-latedunhy-dratedoxides

2

%0.0

.0

.0

. 0

.0

.01.01.1

1.21.6

16.125.926.226.226.4

26.827.427.728.529.3

30.430.530.830.832.132.8

Consistency (water, percent; and Vicat penetra-tion, mm) of cement-lime pastes as obtained byoperator—

A

3

%4144414238

4141444139

4140383737

3738374139

383838404239

4

mm6

1165

14

59474

]

5

B

6

c

7 8

Linear expansion of duplicate cement-lime bars asobtained by operator—

A

9 10

SERIES A—HIGH-CALCIUM HYDRATED

%4244414138

41

41444139

mm7986

10

78786

%41 .43414238

4140434139

SERIES B—REGULARLY

5

611

7

107984

575

1065

4139383737

3839384238

393838404139

121779

11

208

1313

5

116

1010

512

4138393738

3838384138

383838424239

mm5987

12

66575

%0.18

.18

.16

.16

.14

.19

.19

.20

.18

.18

%0.18

.18

.16

.16

.14

.19

.19

.20

.18

.18

]

11

LIMES

%

. . . .0.15

.13

.19

.19

- —

B

12

%0.18

.20

.15

.15

.13

.19

.19

.20

.18

.17

HYDRATED DOLOMITIC LIMES

739

118

118

1045

867

1977

9.812.59.28.3

11.610.116.317.46.9

12.015.39.3

14.2

7.610.212.59.38.4

11.610.116.217.46.9

14.013.812.014.99.1

14.2

8.89.5

8.9

— -

16.416.47.3

13.2

— -

8.99.7

13.58.98.1

12.09.8

16.416.47.3

12.612.712.416.410.114.5

<

13

%0.18

.15

.13

.21

.20'

.20

.19

.18

10.4

9.28.3

10.116.4

14.113.8

16.010.614.0

14

%0.20

.20

.16• .15

.13

.21

.20

.21

.19

.19

8.610.413.19.48.4

12.09.8

16.417.07.3

14.314.012.416.410.513.8

Aver-age oflinearexpan-

sionvalues

15

%0.18

.19

.16

.15

.13

.20

.19

.20

.18

.18

8.510.012.99.28.3

11.810.016.416.97.1

13.813.512.215.89.9

14.1

SERIES C—HIGHLY HYDRATED DOLOMITIC LIMES

4953545657

5961656768

72737475

1.42.62.62.92.9

3.03.24.05.26.6

7.79.0

11.611.7

3539403739

3937393941

38363237

85

116

613796

4

' 79

3938403739

3937404041

38373538

546

1212

989

137

88

3013

4039413739

3938394340

39383237

109487

7106

276

9118

11

0.25.76.35.24.37

.34

.30

.291.81.9

2.76.02.32.9

0.26.69.35.25.38

.35

.29

.301.91.9

2.66.32.32.8

0.241.30.37

.37

.31

.322.92.5

3.85.03.23.9

0.251.40.37

.32

.36

.37

.31

.312.92.7

3.55.02.83.5

0.261.3

.35

.36

.33

.32

— -

3.75.53.13.1

0.251.40.38

.34

.38

.37

.33

.333.72.8

3.75.72.83.1

0.251.10.36

.29

.37

.36

.31

.312.62.4

3.35.62.83.2

SERIES D—MAGNESIAN HYDRATED LIMES

777880

5.97.2

12.7

434141

696

424138

5

84

434140

6

137

0.43.35.64

0.42.36.65

0.40.69

0.54.40.68

0.57.41.75

0.53.39.76

0.50.39.70


TABLE 5. Effect of varying the cement on the expansion of cement-lime bars prepared in the proportion of 1 cement:1 lime(by weight), and autoclaved at a steam-gage pressure of 295 lb/in.2 for 3 hr

(The expansions of the 17 cements without the addition of limes ranged from 0.0 to 0.30 percent and are shown in brackets. The percentage of expansion obtainedafter subtracting the percentage of expansion of the cement from the percentage of expansion of the cement-lime bar is given in parentheses]

Limenum-ber

Unhy-dratedoxides

Linear expansion of cement-lime bars prepared with cements 1 to 17, inclusive

[0.00] [0.01]3

[0.01] [0.02] [0.04]6

[0.04] [0.06]8

[0.07]9

[0.11]10

[0.11]11

[O.H]12

[0.17]13

[0.18]14

[0.19]15

[0.23]16

[0.27]17

[0.30]

SERIES A—HIGH-CALCIUM HYDRATED LIMES

1

2

3

4

5

6

12

13

14

15

19

0.0

* .0

.0

.0

.0

.0

1.1

1.1

1.2

1.6

2.3

%

| 0.031 (-03)

f 0.051 (.05)

. . .

%

J 0.091 (.08)

% %

0.05(.03)

%

....

%

f 0.071 (-03)

f 0.071 (-03)

J 0.071 (-03)

0.10« (.06)J 0.091 (.05)

J 0.111 (.07)

Of

0.08(.02)

0.11(.05)

....

or

J 0.111 (-04)

. . .

. . .

. . .

J 0.101 (.03)

. . .

Of

0.13(.02)

. . .

0.13(.02)

0.14

(.03)

. . .

. . .

Of

0.19(.08)0.20(.09)

f 0.161 (-05)

0.15( 04)0.13(.02)0.19(.08)0.19(.08)0.20(.09)0.19

(.08)0 18(.07)0.24(.13)

Of

—

. . .

—

—

Of

. . .

_ „

. . .

. . .

Of

0.43(.25)

0.39

(.21)

0.31(.13)

0.4S(-30)

%

—

. . .

0.15(-.04)

0 17(-• 02)

0.24(.05)

%

0.26(.03)

0.29(.06)

0 32(.09)0.32(.09)

0.44(.17)

0.39

(.12)

0.30(.03)

0.62(.35)0.47(.20)0.36

(.09)0 37(.10)0.72(.45)

1

....0.29

(-.01)

0.39(.09)

SERIES B—REGULARLY HYDRATED DOLOMITIC LIMES

22

32

35

37

16.1

26.8

27.7

29.3

J 6.81(6.8)

J 13.01(13.0)J 15.11(15.1)j 5.71 (5.7)

. . . 5.9(5.9)

7.6(7. 6)

8.0(7 9)10.6

(10. 5)16.3

(16. 2)

. . .

. . _ •

. . .

. . .

8.8(8.7)12.0

(11.9)16.5

(16.4)7.3

(7 2)7.1

(7 0)

9.4(9.1)11.9

(11. 6)16.4

(16.1)


TABLE 5. Effect of varying the cement on the expansion of cement-lime bars prepared in the proportion of 1 cementil lime(by weight), and autoclaved at a steam-gage pressure of 295 Ib/in* for 3 hr—Continued

IThe expansions of the 17 cements without the addition of limes ranged from 0.0 to 0.30 percent and are shown in brackets. The percentage of expansionobtained after subtracting the percentage of expansion of the cement from the percentage of expansion of the cement-lime bar is given in parenthesis]

Limenum-ber

Unhy-dratedoxides

Linear expansion of cement-lime bars prepared with cements 1 to 17, inclusive

[0.0012

[0.01]3

[0.01] [0.02]5

[0.04]6

[0.04]7

[0.06]8

[0.07]9

[o.n]10

[0.11]11

[0.11]12

[0.17]13

[0.18]14

[0.19]15

[0.23]16

[0.27]17

[0.30]

SERIES C—HIGHLY H Y D R A T E D DOLOMITIC L I M E S

49

53

54

56

57

59

60

61

65

67

73

74

%1.4

2.6

2.6

2.9

2.9

3.0

3.2

3.2

4,0

5.2

9.0

11.6

%

. . .

JO. 2510 25)

j 1.81(1.8)

%

JO. 21l ( - 2 1 )

JO. 19{(• 18)

0.45(.44)

1(1 9)J 0.691 C68)

%

. . .

J 0.221(0.21)

2.4(2.4)4.7

(4 6)

%

0 31(.29)

JO. 161C 14)

. . .

%

....

0.21( 17)

1 8(1.8)

%

JO. 211(.17)

JO. 22!(. 18)

. . .

%

0.42(.36)

0.23(.17)

0.19( 13)

1.3(1.2)

%

JO. 201(. 13)

. . .

. . .

%

.. .

. . .

0.26(.15)

—

%J 0.251 (-14)

1.4(1.3)0.37(.26)0.33(.22)0.37(.26)0.37(.26)0.36025)0.32( 2J)0.32(.21)3.2

(3.1)5 3

(5.2)3.0

(2.9)

%

—

. . .

2.3(2.2)

. . .

%

2.9(2.7)

. . .

2.6(2. 5)5.5

(5.3)3.5

(3.3)

%

1.6(1.4)0.57(.39)0.85(.67)

%

0.40(.21)

. . .

0.32(.13)

%

0.46(.23)0.40(.17)

0.36(.13)

%0.51(.24)

0.69(.42)

0.62

V.. oo)

0.73(.46)

0.51(.24)

%

2.7(2.4)

0.57(.27)

0.47(.17)

2.9(2.6)6.1

(5.8)4.1

(3.8)

SERIES D—MAGNESIAN H Y D R A T E D L I M E S

770

78

80

5.9

7.2

12.7

JO. 26l(- 26)JO. 541(.54)

0.26(.25)

. . .JO. 211(. 19)0.24(.22)0.42(.40)

. . .0.31(.25)0.53(.47)

0.33(.22)

—

0.55(.44)0.40(.29)0.69(.58)

. . .0.44(.27)

0.68(.50)

. . .

0.95(.72)

. . . . . .0.58(.28)0.90(.60)


standard method of test for autoclave expansionof portland cement [6]) are given in bracketsdirectly below the numerical listings of thecements. The values for the percentage of ex-pansion of the cement-lime bars are listed in eachcolumn. In addition, the percentage of expansionobtained after subtracting the percentage of ex-pansion of the cements from the percentage ofexpansion of the cement-lime bars is given inparentheses.

The data are insufficient for the application ofstatistical analysis. Nevertheless, several perti-nent observations may be made. Inspection ofthe expansion values of the cement-lime bars of aparticular lime reveals that as the expansion valueof the constituent cements increases there is adefinite trend for the expansion value of thecement-lime bars to increase also. This indicatesthat the expansion of the cement used is a factorof moment, and that it is necessary to apply acorrection for the expansion contributed by thecement. By the method of trial and error, it canbe shown that arbitrarily subtracting the expan-sion of the neat cement from the total expansionof the cement-lime bar gives the most uniformresults for the "net" expansion of a lime. Netexpansion values obtained in this manner areshown in parentheses in table 5. Even the netexpansion values for a particular lime, however,tend to increase somewhat as the expansion valuesof the cements increase. For purposes of a testprocedure, the variation in net expansion valuescould be further reduced by eliminating the use ofcements of very low or high expansion. For ex-ample, good agreement of net expansion values isshown in most instances when the expansion valuesof the neat cements range between 0.05 and 0.15percent.

It is realized that the above conclusions arebased on limited data and that it would have beendesirable to have determined the expansion valuesof all the limes with all the cements. Unfortu-nately, this phase of the investigation was the lastundertaken, and a more complete study was nolonger possible because of the limited quantity ofmaterials remaining.

5. Behavior of Cement-Lime Specimens DuringCourse of Autoclaving 7

(a) Expansion of Specimens During Normal Heating Scheduleof Special Autoclave

Figure 12 shows the normal heating schedule ofthe specially equipped autoclave; the thermal ex-pansion of the parts of the autoclave, includingthe thermal expansion of the bar; and the expan-sion of bars prepared with three regularly hydrateddolomitic limes, three highly hydrated dolomiticlimes, and one neat portland cement. The thermalexpansion reached 0.4 percent at 216° C (steam-gage pressure of 295 lb/in.2). Subtracting thisvalue from that recorded for portland cementat 216° C, or 0.6 percent, gives the value of 0.2percent for the expansion of the neat portland-cement bar as compared with 0.11 as determined inaccordance with the standard method of test forautoclave expansion of portland cement. It isbelieved that this slight difference in determiningpercentage of expansion of the cement can in partbe ascribed to the method used in determining thepercentage of thermal expansion. However, inas-much as the thermal expansion is insignificantwhen compared to the expansions of bars preparedwith hydrated limes containing appreciable quan-tities of unhydrated oxide, no further studieswere made to determine whether the thermal ex-pansion could be ascertained more accurately.

The curves representing the expansion of theneat portland-cement bar and the cement-limebars shown in figure 12 involve not only the ther-mal expansion but also the expansion accompany-ing the hydration of the constituents of the testspecimens. To obtain the expansion other thanthermal, one should therefore subtract the per-centage of thermal expansion at any specifiedtime from the total percentage of expansion asshown in figure 12.

The seven curves depicting the expansion char-acteristics of the cement-lime bars shown in figure12 were chosen primarily because they were more

7 The authors are indebted to G. J. Fink and Emil Trattner for the use oftheir data on the behavior of cement-lime specimens during the course ofautoclaving, previously presented at the Twenty-Eighth Annual Conventionof the National Lime Association, March 1946.


I ' I ' I >! -L ! -

Regularly hydra fed do/ami fie lime 31 • Unhydrafed oxides 26.5 %

Regularly hydra fed dolomitic lime 30 • Unhydrafed oxides 26.4

Highly hydrafed dolomitic lime 76- Unhydrafed oxides 12.6 % -j

Highly hydrafeddolomitic lime 73 • Unhydrafed'oxides 9.0 % 7

/•Highly hydrated dolomitt lime 66- Unhydrafed oxides 5.1 %

r Highly hydrafed' dofomil/c I/me 62* Unhydrated oxides 3.4 %

1?—Neat port fond ce/rrenf 10

I , 1 . I I 1 • 1 I 1 , 1 I I I I , 1 I I I

zzo

ZIO

200

190

180

no

160

15P

140

110

100

90

30

10

60

50

40

30

2010 20 30 40 50 60 70 SO 90 100 110 120 130 140 ISO 160 I7O /8O 190 2OO 2/O 22O 23O Z4O ?5O

Time, mirrufes

FIGURE 12. Normal heating schedule of the special autoclave, thermal expansion of the bar and parts of the autoclave,* expansion of a neat portland cement bar, and expansions of cement-lime bars prepared with three regularly hydrated

and three highly hydrated dolomitic limes in the proportion of 1C:1L.

or less typical of those obtained with other hy-drated limes over this range of expansion. Nocurve showing the expansion of a cement-lime barprepared with a high-calcium lime is plotted infigure 12, since such a curve would be obscuredbecause it would very closely coincide with theone plotted for the portland cement. It can beseen from figure 12 that the curves expressing theexpansion of the cement-lime test specimens areS-shaped, with a flattening accompanying a de-crease in the total expansion. At the normalheating schedule, only a slight amount of thetotal expansion occurred before the elapse of 45min, or before a temperature of about 150° C wasreached. This was followed by a rapid rate ofexpansion which, in turn, tapered off before atemperature of 216° C was reached. Conse-

quently, there is little to be gained by continuingthe heating at 216° C for more than 1 hr.

After subtracting the percentage of expansionof the parts of the autoclave at 216° C from eachof the values of the cement-lime bars at the sametemperature and comparing the percentages ofexpansion thus obtained with those recorded intable 1 for similar 1:1 cement-lime bars, it wasfound that the expansions obtained with thespecial autoclave were for the most part appre-ciably lower than those recorded in table 1. Themost logical explanation for the lower percentageof expansion with the use of the special autoclaveappeared to be associated with the lower rate oftemperature rise during autoclaving. Conse-quently, a series of experiments was made of the


rate of expansion of the cement-lime specimenswhen heated in the special autoclave at a stillfurther retarded rate.

(b) Effect of Retarding Rate of Heating on Expansion

The results obtained with three of the limes(40, 73, and 66) are plotted in figure 13. Theretardation was started at the end of 30 min, orwhen the temperature reached 120° C. At thispoint, but little expansion had occurred in thespecimens. The rate of heating was then retardedso that the temperature did not reach 216° C untilthe end of 175 min, rather than 97 min as in thecase of the normal rate of heating. From figure 13it is evident that not only was the rate of expansionretarded at the slower rate of heating, but, moreimportant, the total expansion was decidedly less.

These experiments show the great importance ofdefining the heating schedule of an autoclave, by aspecification for the autoclave testing of expansionof a cement-lime specimen.

VII. Application of Results to Selection ofa Proposed Specification for Soundnessof Hydrated Lime

It should be recalled that the primary purposeof the investigation was to study the expansivecharacteristics of hydrated limes and to obtaindata and information that might be used in formu-lating an accelerated performance test acceptablefor a specification for the soundness of hydratedlimes.

Accelerated performance tests have been incorpo-

J 1 ' I ' 1 ' 1 ' I ' 1 ' 1 ' 1 ' 1 JJJJJJ-<Li-U_l-<LLLll?kLl.iJ>_L_LJJJ_UL^ Regu/ar/y hydrated dobmit/c //me 40 —/ Norma/ heating schedo/e

Regu/ar/y hydra/ed do/om///c //me 4O - Retarded heat/rrg schedu/c

H,gh/y hydrated do/om/t/c time 73- Norma/ heat/ng schec/a/eo ——o— •

/^Highly hydra fed do/om/t/c //me 66 - Norma/ heating schedu/e

^H/gfr/y fyyd/rr/ed do/om/Y/c //me 66 - Retarded heat/hg schedo/e.

1 I I I 1 I 1 I 1 I I ! 1 I 1 1 I I 1 I 1 I I I I , 1 .1,1, 200 tO ZO X 40 50 60 70 SO 90 100 110 120 130 /40 150 160 170 180 I9O 200 2/0 220 230 240 2S0

Time, rn/nufes

FIGURE 13. Effect of decreasing the rate of heating the special autoclave to 216° C (equivalent to gage-pressure of296 Ib/in.2) on the rate of expansion and total expansion of cement-lime bars prepared with three hydrated limes.


rated into many specifications for the selection ofvarious materials. The need of a test that bothaccelerates and amplifies the potential expansionof a hydrated lime in service is apparent fromconsiderations of a typical plaster failure involvingexpansion. Figure 14 shows a marked failurethat occurred in white coat along the side of a con-crete beam when the expansion of this coat becameexcessive. Chemical and thermal analyses indi-cated that the white coat had been prepared froma mixture of gypsum gaging plaster and a normallyhydrated dolomitic finishing lime that had con-tained considerable unhydrated MgO. The slowhydration of the MgO in the set and hardenedwhite coat to Mg(OH)2 had resulted in a cumula-tive expansion of such a magnitude that the whitecoat had sheared away from the underlying basecoat of gypsum bond plaster that had been appliedto the concrete beam. Calculations based onmeasurements of the bulge in the white coat revealedthat a linear expansion as low as 0.5 percentappears to account for a separation of the magni-tude shown in figure 14. The fact that such

FIGURE 14. Plaster failure involving the expansion of about0.5 percent in the white coat along the side of a concretebeam.

failures usually do not take place until some 5 to10 yr (or longer) after a building has been erectedshows the necessity of an accelerated test forjudging the soundness of hydrated lime, provideda performance test is to be used. Also, the factthat only a very small expansion can be toleratedwithin a set and hardened plaster or mortar with-out causing trouble shows the desirability ofamplification of the potential expansion of the

constituents composing the plaster or mortar.There is no valid reason for not accepting anaccelerated test procedure that amplifies expansionprovided it clearly differentiates hydrated limeshaving low potential expansions from all others.In fact, amplification has a definite advantagein that it facilitates this differentiation.

The factors to be considered in the selection ofa method best suited for an accelerated perform-ance test, and the selection of a plausible limit tothe percentage of expansion will now be discussed.However, it is not the purpose at this time towrite a detailed specification.

1. Selection of Method

The selection of a test procedure for a specifica-tion for soundness of hydrated lime should bebased on several criteria. Among the mostimportant of these are the following:

1. The test should clearly differentiate thehydrated limes having low potential expansionsfrom all others.

2. The procedure should be reproducible.3. It should not be necessary to repeat the test

too often because of such factors as frequentbreakage of weak test specimens in handling.

4. The time required to complete the testshould be as short as feasible.

5. The procedure, if possible, should be adapt-able to apparatus already available in numeroustesting laboratories.

On the basis of these criteria, autoclave testingat a steam-gage pressure of 25 lb/in.2 would beeliminated by criterion 4, because at this pressureultimate expansion of cement-lime bars is notattained even after 7 hr of autoclaving (fig. 5).On the other hand, the time period required forautoclaving at 295 lb/in.2 is much shorter, actuallyslightly over 3 hr, for the entire procedure ofautoclaving, divided as follows: Time of raisingpressure to 295 lb/in.2, from 1 to 1}{ hr; time ofcontinued heating at this pressure, 1 hr (see fig.12); and time of cooling autoclave, 1 hr. Further-more, autoclaving at 295 lb/in.2 seems particularlydesirable, because numerous testing and researchlaboratories throughout the country are equippedwith autoclaves so designed that the gage pressurecan be raised to 295 lb/in.2 within 1 to 1% hr andautomatically maintained thereafter at this pres-sure. Autoclaving at 295 lb/in.2, therefore, meetsboth criteria 4 and 5.


Although a test procedure very likely could bedevised for autoclaving cement-lime specimensat some gage pressure between 25 and 295 lb/in.2,it is doubtful if autoclaving at an intermediatepressure would have any advantages over auto-claving at 295 lb/in.2.

The next factor to be discussed is the selectionof that ratio of cement to lime in the test specimenthat gives the best indication of the potentialexpansion of hydrated limes. All three ratios(2C: 1L, 1C: 1L, and 1C: 2L) could be used to differ-entiate the regularly hydrated dolomitic limes,characterized by having the highest percentagesof unhydrated oxides and the highest percentagesof expansion, from the remaining hydrated limes(figs. 7, 8, and 9). There are, however, certain ofthe highly hydrated dolomitic limes that appar-ently have appreciable potential expansions whentested in the proportions of 1C:1L or 1C:2L thatdo not clearly show this expansive potentialitywhen tested in the proportion of 2C:lL (fig. 10).With a 2C:1L proportion, the percentages of ex-pansion of all the highly hydrated dolomitic limesare distributed rather uniformly over a narrowrange—with but one exception, from 0.18 to 0.88percent. Therefore, a 2C:lL proportion doesnot meet criterion 1.

In the 1C:1L and 1C:2L proportions, thehighly hydrated dolomitic limes of low potentialexpansions are easily differentiated from those ofappreciable potential expansion. But, as men-

* tioned previously, 1C:2L bars are fragile anddifficult to handle without excessive breakage.Therefore, by criterion 3, testing hydrated limesin the 1C:2L proportion is eliminated. Conse-quently, having eliminated the 2C:lL and 1C:2Lproportions, the 1C:1L proportion remains themost desirable. Since it has been shown that theprocedure of testing the cement-lime bars in the1C:1L proportion is reproducible, criterion 2 isalso met. In conclusion, therefore, testing cement-lime bars prepared in the proportion of 1C:1Land autoclaving at 295 lb/in.2 meets all of thecriteria set forth for a procedure for determiningthe soundness of hydrated limes.

2. Proposed Limit of Percentage of Expansion

Bearing in mind that the test for soundnessshould differentiate hydrated limes having lowpotential expansions from all others, and that "Nota single instance has been found where bulging

occurred in a white-coat plaster made with high-calcium lime;" [2], it would appear logical to limitthe percentage of expansion to that shown by thehigh-calcium limes. If the upper limit is placed at0.6 percent without subtracting the percentageof expansion of the cement, then all of the high-calcium limes would be included and 16 of thehighly hydrated dolomitic limes; but at the sametime those highly hydrated dolomitic limes thatshowed moderately high potential expansion wouldbe eliminated. The limit of 0.6 percent may betoo restrictive in view of the fact that certaincements might be used that give slightly higherexpansion than cement 10, which was used in thebulk of the experiments in the present investiga-tion. It is proposed, therefore, that the upperlimit of the linear expansion for a test for soundnessbe placed at 1.0 percent after subtracting the per-centage of expansion of the cement.

It is further proposed that portland cement meet-ing the requirements of Federal SpecificationSS-C-192 (type I) for portland cement (ASTMDesignation: C 150-46) be used in preparing the1C:1L test bars, and in addition that only thosecements having expansions ranging from 0.05 to0.15 be selected.

A limit of 1.0 percent should afford adequateprotection for the consumer and not work anundue hardship on the producer. All high-calciumhydrated limes and quicklimes, if properly slakedso that the unhydrated oxide is reduced to a lowpercentage, will easily meet the requirement; andapparently if the highly hydrated dolomitic limeshave been hydrated to the extent that the unhy-drated oxide is reduced to less than 5 percent, suchproducts should also meet the requirement. The16 highly hydrated dolomitic limes that meet thisrequirement were all commercially produced.The remaining 12 do not represent a true cross-section of this type of lime as now being produced.Six, as stated previously, were experimental limes,and the process of manufacture of several of theothers has been modified, so that there is goodreason to believe that r the percentages of unhy-drated oxide and the expansions are now of theorder of magnitude of the high-calcium limes.

All of the four magnesian limes with the per-centage of unhydrated oxide ranging from 5.9 to12.7 would meet the 1-percent limit on expansion.Two of these, namely, limes 79 and 80, with 9.8and 12.7 percent of unhydrated oxide, respectively,


would not, however, meet a composition require-ment that the quantity of unhydrated oxide shallnot exceed 8 percent by weight on the as-receivedbasis [3, 4]. '

On the other hand, it is obvious that certain ofthe highly hydrated dolomitic limes havingquantities of unhydrated oxide as low as 8 percentcan, nevertheless, exhibit appreciable autoclaveexpansion. In either case, since an acceleratedperformance test should be preferable to a limi-tation of the chemical composition, it is believedthat a limit of 1 percent on the expansion oflC. lL bars is better than the 8-percent limit onthe unhydrated oxides.

VIII. SummaryAn investigation of the expansive characteristics

of hydrated limes was completed, and from thedata obtained an accelerated performance testwas proposed for testing the soundness of hydratedlime.

Samples of 80 structural hydrated limes, repre-sentative of the production in the United States,were classified on the basis of chemical analysesand calculated unhydrated oxide content ashigh-calcium, regularly hydrated dolomitic, highlyhydrated dolomitic, and magnesian limes. Thepercentage of unhydrated oxide for the high-calcium class ranged from 0 to 3.9; for the regularlyhydrated dolomitic from 16.1 to 34.3; for thehighly hydrated dolomitic from 1.4 to 12.6; andfor the magnesian from 5.9 to 12.7.

On the basis of preliminary experiments, port-land cement was selected as the most suitablehydraulic material for gaging hydrated lime in thepreparation of test specimens suitable for autb-claving. A technique for preparing, curing, andautoclaving 1- by 1- by 10-in. cement-lime barswas developed. The method for determining thelinear expansion of cement-lime bars prepared inthe proportion of 1 cement to 1 lime, by weight,and subsequently autoclaved at 295 lb/in.2 for3 hr was found to be reproducible by three inde-pendent operators.

Data on the expansions of cement-lime barsprepared in proportions of 2 parts cement to 1part lime, 1 part cement to 1 part lime, and 1 partcement to 2 parts lime, by weight, and autoclavedto 295 lb/in.2 gage pressure for 3 hr, showed thatbars prepared with the regularly hydrated dolo-mitic limes, which had the highest percentages of

unhydrated oxide, had the highest percentageof expansion. The high-calcium limes, char-acterized, ia general, by the lowest percentages ofunhydrated oxide, gave the lowest percentageof expansion. Most of the highly hydrateddolomitic limes had percentages of unhydratedoxide and expansions that were comparable tothose of the high-calcium limes. The remainderhad percentages of unhydrated oxide and ex-pansions that were greater than those of the high-calcium limes. The four magnesian limes ex-hibited unique behavior in that each showed lowerpercentage of expansion than other limes withcomparable percentages of unhydrated oxide.

An increase in the proportion of lime in thecement-lime bars was attended by an increase inexpansion and by better differentiation betweenthe classes within a series. Bars of the 1 partcement to 2 parts lime series, however, were veryweak and were eliminated from consideration fora routine test method.

The effect of 17 different portland cements onthe expansion of cement-lime (1:1 by weight)bars was investigated. The neat cements rangedin expansion from 0.00 to 0.30 percent, and theywere tested in part with 30 selected limes. Theexpansion values of cement-lime bars* for a par-ticular lime tended to increase as the expansionvalue of the constituent cement increased.Arbitrarily subtracting the expansion of the neatcement from the total expansion of the cement-lime bar gave the most uniform results for the netexpansion of the lime.

An autoclave was specially modified for ascer-taining the behavior of cement-lime specimensduring the course of autoclaving. The resultsshowed that only a slight amount of the totalexpansion of cement-lime bars occurred before atemperature of 150° C was reached. Above150° C, a rapid expansion rate was noted, which,in turn, tapered off before a temperature of 216°C was reached. Since little expansion occurredafter the specimen had reached the limiting tem-perature of 216° C (equivalent to 295 lb/in.2,steam-gage pressure), it was concluded that con-tinued heating at 216° C beyond 1 hr is un-necessary for a test procedure. The results alsoindicated the importance of a controlled rate ofheating the autoclave to 216° C, as a decreasedrate of heating gave decreased expansions.

It was proposed that hydrated limes be tested


n the proportion of 1 cement: 1 lime (by weight),that the cement used have an expansion between0.05 and 0.15 percent, that the autoclaving bedone at 295 lb/in.2 gage pressure for 1 hr accord-ing to a controlled schedule and, finally, that theexpansion obtained after subtracting the percent-age of expansion of the neat cement bar from thetotal expansion of the cement-lime bar be limitedto not more than 1 percent.

The authors express thanks and appreciation toEmil Trattner and G. J. Fink, former KesearchAssociates at the National Bureau of Standards,for obtaining a great portion of the data on auto-clave tests included in this publication.

IX. References

[1] Lansing S. Wells and Kenneth Taylor, J. ResearchNBS 19, 215 (1937) RP1022.

[2] E. S. Newman, Bui. Am. Ceram. Soc. 26, No. 4,117 (1947).

[3] Proposed amendment to Federal Specification SS-L-351 for lime; hydrated (for) structural purposes.Date of amendment, Feb. 2, 1940.

[4] Tentative specifications for special finishing hy-drated lime. ASTM Designation: C206-46T.ASTM Standards, part II, 1306 (1946); and Ten-tative specifications for hydrated lime for masonrypurposes. ASTM Designation: C207-46T (TypeS—Special hydrated lime for masonry purpose)ASTM Standards, part II, 1308 (1946).

[5] Standard specifications for gypsum plastering in-cluding requirements for lathing and furring.American Standards Assn. A. 42.1—1946, J. W.McBurney, Ind. Stand. 18, No. 4, 79 (1947).

[6] Standard method of test for autoclave expansion ofPortland cement ASTM Designation: C151-43,ASTM Standards, part II, 14 (1946); also AmericanAssociation State Highway Officials StandardAASHO No. T 107-45; as well as Federal Speci-fication SS-C-158b for cements, hydraulic, generalspecifications (methods for sampling, inspection,and testing).

[7] Federal Specification SS-L-351 for lime; hydrated(for) structural purposes.

[8] Tentative revision of standard methods of chemicalanalysis of limestone, quicklime, and hydratedlime. ASTM Designation: C25-44 submittedJune, 1945, see ASTM Standards, part II, 1701(1946).

[9] G. J. Fink and Emil Trattner, Proc. ASTM 45, 723(1945).

[10] Federal Specification SS-C-192 for cements: port-land. See also Type I Cement of Standard Speci-fications for portland cement ASTM Designation:C15O-46, ASTM Standards, part II, 1 (1946).

[11] Federal Specification SS-C-158b for cements, hy-draulic; general specifications (methods for sam-pling, inspection, and testing) paragraphs F-51,F-51a, and F-51b.

[12] Federal Specification SS-C-181b, for cement, ma-sonry. Also Standard specifications for masonrycement. ASTM Designation: C91-40, ASTMStandards, part II, 7 (1946).

[13] G. W. Snedecor, Statistical methods, 4th ed. (TheIowa State College Press, 1946).

WASHINGTON, January 23, 1948.


Documents

Expansive Characteristics of Hydrated Limes and the ...The hydrated limes in each series are arranged in table 1 by order of increasing percentage of calculated unhydrated oxides