46 Vth International Brick Masonry Conference

11-3. Brick Durability Tests and the Method of Freezing T. Ritchie

Division ofBuilding Research, National Research Council ofCanada, Ottawa, Ontario KJA OR6

ABSTRACT

The testing of bTicks fOT durability has come under critical review in recent years because of differences in the performance of bricks in the standard freeze-thaw test and in service in the walls of buildings. The difference in perfonnance is believed by the author to be due in part to differences in the process of freezing; bricks in service are frozen by the passage of a freezing. plane through the exposed face of the brick, whereas in the standard test the freezing of the brick takes place thTOUgh several of its surfaces.

The two pTocesses of fTeezing, "unidirectional" and "omnidirectional", were aPPlied to bars of elay bricks and other masonry materials, which had been fitted with reference points for measurements of length changes. The bars were subjected to freezing and thawing, with length measurements made periodically. Certain bars showed little difference in behaviour regardless of the method of freezing, but for several materiais the bars frozen omnidirectionally expanded quickly and excessively, while companion bars frozen unidirectionally were either unajJected or underwent much less expansion. The dependence of the durability of a brick on the particular freezing process emPloyed, and its implications in lhe testing of bTicks, is discussed.

Requirements for the durability of day bricks set out in Canadian and V.S. Standards 1

•2 are based on a procedure

for freezing ;:ind thawing bricks used by McBurney and Lovewell Y This procedure was to immerse half-bricks in water, then set them on edge in a shallow tray containing a half-inch of water; the tray was placed in a cold chamber to freeze the bricks, which were then thawed by immersion in water. The bricks were subjected to 50 freezings and thawings . From such tests of several thousand bricks, they conduded that certain combinations of values of the com­pressive strength, water absorption and saturation coeffi­cient could be used to predict the performance of bricks in their freeze-thaw test.

Recent failures in service of bricks that apparently met specification requirements have brought the durability testing of bricks under scrutiny. Blachere and Young" drew attention to the influence of such factors as sample saturation, size of sample, the rate of cooling and the scheduling of the freeze-thaw cyde on the results obtained. Inadequacies of the present durability test for bricks have been described by Robinson, Holman and Edwards,3 who tested a large number of bricks by the standard method with the aim of improving the accuracy of predicting brick durability from the properties of water absorption, com­pressive strength and saturation coefficient. Even with the improvements suggested by their study, however, they believed that the test method and specification would remain illlpel-fect.

Discrepancies between the performance of bricks in the standard test and in service in a wall are believed by the author to be due, at least in part, to differences in the way bricks are frozen, in particu lar, the manner in which the freezing plane passes into the brick. In the standard method of test , bricks are frozen by heat loss from ali surfaces exposed to cold air, i.e ., omnidirectionally, while in service in a wall bricks are frozen through the face only, i.e., unidirectionally . In a study6 comparing the behaviour of bricks frozen by the standard method with the behav­iour of similar bricks frozen unidirectionally, significantly different results were obtained by the two freezing meth-

ods, as assessed by visual examination of the bricks for cracking and spalling.

The present study is similar to the former investigation except that bars of masonry materiais instead of half­bricks were used, and the effects of freezing were assessed by measurements of the length changes of the samples, some of which were frozen unidirectionall y and others omnidirectionally.

MATERIALS

The materiais studied induded three day bricks, two masonry mortars, one concrete block , and one natural stone, which were made into bars one square inch (25 mm) in cross-section and five inches (127 mm) in effective length.

Bricks, concrete block and sandstone were cut into bars with a diamond saw. Holes were drilled in their ends into which were cemented, with epoxy, round-ended stainless­steel studs. The ends of the studs provided the surfaces against which the length measurements were made, using the standard comparator for cement testing. Three dif­ferent bricks were used, one formed by dry-pressing, the other two by stiff-mud extrusion. Bars of the former brick, however, had to be withdrawn early from test because cracks developed at their ends, apparently associated with the epoxy cement and studs, but this prohlem did not arise with any other of the materiais studied.

One of the mortars was prepared from portland cement, hydrated lime and sand in volume proportions of 1: I :6; the second mortar consisted of masonry cement and sand, proportioned 1: 3 by volume. The materiais were mixed with sufficient water to produce morta r of stiff con­sistency; the flows were 116 and 114 per cent, respectively, for the two mortars. The fresh morta r was cast in standard cement-testing moulds, which provide bars one square inch (25 mm) in cross-section and five inches (127 mm) effective length. Round-e nded stainless-steel studs , embedded in the mortar when cast, projected from the ends of the bar for length measurement.

Session li, Papel' 3, BTick Dumbility Tests and the Method of heezing

The concrete block used in the study had been made from lightweight aggregate; the stone was Wall ace sand­stone from j ova Scotia .

METHOD OF TEST

Six bars of each materia l were stored for two weeks in a laboratory maintained at about 50 per cent relative humid­ity and 73°F (23°C). Prior to the freezing tests, the bars were immersed in water at 73°F for four hours, removed , and the length was measured. Three bars of each material were then placed in the freezillg chambel· with ali surfaces exposed to the cold air; this provided the test condition for what is termed here , omnidirectional freezing.

The other three bars or each materia l were placed to fill an opening in one side of a box whose other five sides were closed anel heavily insulated. These bars, arranged side-by-side a nel close together, were frozen unidirection­ally when the box was placed in the freezing chamber; the freez ing plane passed from that surface of the bars exposed to the air of the freezer , to the opposite surface, across the one-inch thickness of the bars.

The bars remained in lhe freezer, where the air tem­perature was usually in the range of 0° to - IOoF (-18° to -23°C) for 20 hours , then IVere removed for thawing in water. Those bars forming part of the insulated box \Vere removed for thawing. After four hours in the thawing tank , the bars were measureel for length anel again placed in the freezer. This arbitrarily selected process of freezing and thaIVing, with length measurements made after each thawing, was continued until the tenth thaw ing, after wh ich the bars were stored in the laboratory for one week before another series of \ O freezings began. This period of time was intended to a ll ow the bars to dry, so that the gradual increase in their saturation with repeateel thawing in water did not I-esult in their reaching a high degree of saturation. The bars were subjected to a tota l of 100 freez­ings if they had not deteriorated before this number was reached.

RESULTS

The length changes of the bars, expressed as a percentage of the original length (after the first 4-hour immersion), are plotted against the number of freezings in Figures \ , 2 and 3.

Bars of three materiais, the mortar prepared from masonry cement and sand, one of the bricks , and the sandstone, showeel no significant difference in behaviour when frozen 100 times unidirectionally and omnielirec­tionally , but the other materiais tested behaved quite dif­ferently, as shown in the graphs.

The mortar bars of cement, lime and sand (Figure I) were frozen unidirectionall y about 80 times before appre­ciable expansion occurreel. The three companion bars fro­zen omn idirectiona ll y, however, expaneled a lmost contin ­uously from the Slan or test and at such a rale lhat the expansion exceeded 0.05 per cent before 50 freezings and 0. \ O per cent before 70 freezings. After 90 freezings the bars had eXpandeel O.:W per cent and were broken before \ 00 freezings.

47

The bars of brick A (Figure 2) that were frozen unidi­rectionally showed little length change in the 80 freezings to which they \Vere subjected. Their companion bars, how­ever, frozen omnidirectionall y, expanded very rapidly after \ O freezings, were cracked between \9 and 25 freez­ings , and were in such a deteriorated condition after 30 freezings that they had to be removed from test.

The ba rs prepared from concrete block (Figure 3) and frozen unidirectionally expanded slightly over the course of their 100 freezings, whereas the companion bars , frozen omnidirectionaIl y, expanded at a very rapid rate such that their length change exceeded 0.\ O per cent before 20 freezings had been carried out. The expansion exceeded 0.50 per cent after 35 freezings and the bars , markeel with a very fine pattern of cracks, were so fragile that they had to be removeel from test.

DISCUSSION

The results confirm those of a previous study" that bricks frozen by heat loss from severa l surfaces simul taneously behave quite differently rrom bricks frozen unidirection­ally.

Another study' of unidirectional fl-eezing indicated that the water in a brick is forced away from the advancing freezing plane and may be forced from the brick, a proc­ess that may provide relief from stress. But no such pos­sibility of stress relief can operale when freezing planes pass simultaneously into severa I of the brick surfaces, wh ich probably accounts for the differences in the effects of freezing by the two processes, although this remains to be proven.

The apparent relief of freezing stresses has been noted by the author in the freezing of cylindrica l specimens (1-1/8 in. , 29 mm, diameter and of length equal to brick height) cored from various bricks. Three such samples from a brick were compared with three other samples from the same brick in which a 1/4 in. (6.3 mm) hole was drilleel down the centre. The samples were saturated, fro­zen by exposure in a cold chamber and thawed in water. Man y of the solid samples crackeel in the first freezing and others after the second or third freezing. The samples with centra l holes, however, IVere unaffecteel by many freezings, presumably beca use under essentia ll y similar rates of freez ing and moisture content there was no con­finement of the freezing water.

CONCLUSIONS

The behaviour of certain bricks and othe r masonry mate­I-ials frozen unidirectionall y, such as occurs in service in a wall , differed significantly from their behavioul- when frozen through severa l surfaces . This difference is believed to account in part for the reported difficu lties in relating the results of freeze-thaw tests or bricks by the standard (A.S.T.M.) method to their performance in service. The stanelard method or freeze-thaw lesting or bricks accord­ingly should be changed to correspond more closel)' to the actual cond itions af lhe freezing of bricks in service in a wall, which is unielirectional.

48

This paper is a contribution from the Division of Bui ld­ing Research, National Research Counci l of Canada and is published with the approval of the Director of the Division.

REFERENCES

I. American Society for Testing Materiais , Specification C62-75a, "Building Brick (Solid mason ry units made from day or shale)." 2. Canadian Standards Association, Standard A82.1-1977 , "Burned day brick ." 3. McBurney, j.W. and Lovewell, C.E. "Strength, water absorp­tion and weather resistance of bui ld ing bricks produced in the U nited States." Proceedings, A.S.T.M., Vol. 33, Part 11 , 1933. 4. Blachere, R., and Young, j.E. "Freezing and thawing tests and theories of frost damage." A.S.T.M. journal or Testing and Eval­uation , Vol. 3, No. 4, july, 1975 . 5. Robinson, C.C ., Holman, j.R., and Edwards, j.F. "Relation between physical properties and durabi lity of commercia lly mar­ke ted brick." Bulletin or lhe American Ceramic Society, Vol. 56, No. 12, 1977. Ritch ie, T. "The lest method of freez ing bricks: its influence on their durability." journal or the Canad ian Ceramic SocielY, Vol. 44, 1975. 7. Ritchie, T. "Freeze-thaw action on brick." journal or the Canadian Ceramic Society, Vol. 4 1, 1972.

I o 11 BR IC K A

"

I NOTE , BA R S CP.ACKEO BETWEEN 19

Z ANO 15 FREEZ I NGS, ANO ~ 0.10

wERE REMOVE D FROM TE S T AFTER 30 FREEZ INGS ..

~ o . OB

t:> EACH POINT REPRESENTS O O. 06 TH E AVERAG E VALUE FOR

Z .-. TH REE BARS O

2 .04 .......... OM N I D IREC TIO NA L FREEZ I NG

" Z

:: .. UN IOIRECT IONAL FREEZ I NG

Z ~ 0.02

<>--o- ___ l-A---__ ---0----0 U l ~

~

" 0.02 z .. I u

.02 ;:

" z ~

0.02 o 10 20 30 40 50 60 70 BO 90

NUMBER OF FREEZINGS

FiguTe 2. Length changes of brick samples frozen unidirectionally and omnidirectiona ll y

100

I

;; Z

~

.. ~ t:> ~

O ~

O

" .. Z

u ~

~

w

" z .. I U

Vth International Brick Masonry Con ference

o . 14 7 I MORT A R 1 , 1 ,6 CEMENT . LIME , SAND 0.30%

EXPANS ION o .12 • AFTER 90

/ FREEZ I NGS

0.10 EACH PO INT R E PR E S E N T S THE AVERAGE VALUE FO R TH R E E BA R S

0.08 """" I

O MN I D I REC T IO NA L ______ o I

0.06 FREE Z I NG~ • I I

.y I Z I

I 20 .04 .~N I DI RECT lONA L FREEZ ING

, ~ I Z o .. .-0--- \ I

I ~ o .02

,(0 ___ -0----0-___ 0 ____ 0_---0----0.._ // X

l . '-,!'

0.02.---.----r---r---;----r---~--;_--_.--_.--_,

... __ MORTA R 1,3 MASONRY CEME N T '7 -="'::"'::i"="'::"::":Q=-== -=--=..::0 __

OM N I DI RECT 10 NA L FR EE Z I Ny-·--------. U NID IRECT IONA L FREEZI N G -=~~

0.02 L-__ ~ __ ~ ____ ~ __ ~ __ ~ __ ~~ __ ~ __ ~ __ ~ __ ~

o 10 20 30 40 50 60 70 80 90 100

NUM8ER OF FREEZINGS

FiguTe 1. Length changes of mortar bars frozen unidirectionally and omnidirectionally

;: " Z ~

.. Z

t:> ~

O

O

" .. Z

U ~

~

" z .. I u I

" Z

~

o . 14

o 11

O. 10

O.OB

0.06

Z

2 0.04 Z .. ~

x o .02

l 0.02

CONeRETE!

BLOCK;. ~~;~~;~O~R//Z~~Eg~ 0.50%

EACH PO I NT REPRESENTS THE AVERAGE VALUE FOR THREE BARS

. I-----OMN I D I REC T IONAL FREEZ I NG

UN I DI RECT IONAL FREEZ IN G

/0 \ _-0----<>----0--_ /"o..._O""_~-().. - --o----~-

0.02~ __ ~ __ ~ __ -L __ ~ __ ~~ __ L_ __ ~ __ ~ __ ~ __ ~

o 10 20 30 40 50 60 70

NUMBER O F FREEZ I NGS

Fig'uTe 3. Length changes of bars of concrete and sandstone frozen unidirectio nally a nd omnid irectionally