The effect of joint thickness and other factors on the compressive strength of brickwork
5.-The Effect of Joint Thickness and Other Factors on the Compressive Strength of 8rickwork by A. J.
and L. E.
University o[ Melbourne
ABSTRACT A mechanism for the compressive failure of brickwork is developed quantitatively, and is shown to be capable of explaining the infiuence
that certain variables have on the compressive strength. lt is shown experimental/y and theoretically that the strength of four-brick prisms declines as the joint thickness increases and as the lateral tensile strength of the bricks diminishes in relation to their compressive strength. The e./fect of other well-known parameters is explained in quantitative terms.
b b d
e m x,y, z E P
L'Effet de l'Epaisseur des Joints et d'Autres Facteurs sur la Résistance à la Compression de la Maçonnerie en Briques Un mécanisme pour la rupture à la compression de la maçonnerie en brique est développé de façon quantitative, et on montre qu'il est capable d'expliquer l'inf/uence qu'exercent certaines variables sur la résistance à la compression. II est montré de façon expérimentale et théorique que la résistance de prismes de quatre briques décroit avec I'augmentation de I' épaisseur du joint et à mesure que la résistance à la traction latérale des briques diminue par rapport à leur résistance à la compression.
L'effet d'autres parametres bien connus est expliqué de façon quantitative.
has been made, with partial success, to describe in quantitative terms the mechanism of the process of compressive failure. The present paper contains an account of a simple theoretical model, and some experimental work on the effect of bed joint thickness in four-high stackbonded prisms which appears to support the theory put forward. The model also explains a number of the features of the compressive failure of brickwork.
NOTATION width of brick (as suffix) brick length of brick strain (as suffix) mortar axes of reference modulus of elasticity load tb/tm Eb/Em
1.2 Model of Compressive Failure of a Short Stackbonded Prism If a short prism of bricks bonded with mortar (Figure 1 (a» is loaded in axial compression in a testing machine the mortar joints above and below a brick sufficiently remo te from the restraining influence of the platens of the testing machine tend to expand laterally more than the brick itself, since the modulus of elasticity of the mortar is normally much lower than that of the bricks. Because of the mortar-brick bond and the frictional resistance to slip between the bricks and mortar at the interfaces, slip will not occur at the interfaces. Lateral tension is, therefore, induced in the brick, and lateral compression in the mortar. Vertical splitting, due evidently to lateral tension, is usually present in a compressive failure of brick walling. The criterion of failure of a brittle material like brick under a condition of vertical compression plus biaxial lateral tension is not known, but failure will certainly occur at a lower compressive stress than would be required in the absence of lateral tension, or if the lateral stresses were compressive. The prism shown in Figure l(a) is subjected to an axial compressive stress a y • The lateral stresses induced
stress compressive stress to cause failure of brick in absence of lateral tensile stress compressive stress to cause failure of brick in presence of lateral tensile stress lateral tensile strength
EinflujJ der Fugendicke und anderer Factoren auf die Druckfestigkeit von Ziegelmauerwerk Ein Zerstorungsmechanismus von Ziegelmauerwerk unter Drucklast ist quantitativ entwickelt worden. Es wird gezeigt, wie er sich zur Erkliirung des Einf/usses verschiedener Grossen auf die Druckfestigkeit eignet. ExperimenteI! und theoretisch ist bewiesen, da.fJ die Festigkeit von Prismen aus je vier Ziegeln mit grosser werdender Fugendicke abnimmt und da.fJ die seitliche Zugfestigkeit der Ziegel im selben Verhii/tnis wie ihre Druckfestigkeit geringer wird. Die Wirkung anderer gut bekannter P 64
308 fl 277
Modulus of elasticity Eb (lbf;in 2)
3·8 x IG6
2·95 X 106
by volume). 1·25 parts of water, also by volume, were added, after trial mixes for workability. The compressive strength, obtained on mortar cubes accordi ng to AS A123 , was as follows : Mean sfrengfh Coeff. of (lbf/in 2 ) variation ( %) 927 8·6 Other mean values used were: Em 0·20 X 106 1bf/in 2 Vm 0·25*
• Polished faces . Masonry saw-cut faces.
2.2.4 Four-Brick Prisms In each type of brick, six prisms were made with mortar joints approximately either 0 ·4-in. or 0 ·6-in . thick, and four prisms of each brick type with l-in.-thick joints, including in each case layers of mortal' of the same thickness at top and bottom of each prism . Four prisms of each brick type were also made with as fhin joints as possible. The prisms were cured in air in the laboratory and tested in a Denison 200-ton compression testing machine between cardboard sheets at the age of 14 days. On a number of the prisms, longitudinal and transverse strains were measured with a 2-in . Demec gauge. The values of Eb, Vb and Em quoted above and in Table 2 were determined from these measurements, the latter from measurements across the joints. Three solid brick prisms and two perforated brick prisms were also prepared and tested dry. In these the contacting surfaces of the bricks were polished in a geological polishing machine until they were plane to a high degree of accuracy. Four piers with solid bricks and three with perforated bricks, whose upper and lower faces had been trimmed fiat with a masonry saw, were also tested dry. The frogs in the solid bricks were filled * Value supplied by Division of Building Resean;h. CSIRO, Melbourne.
Average joint thickness (in .)
(lbfl in 2 )
Average joint Ihickness (in.)
5430* 8760* zero (dry)
* Polished faces . t Masonry saw-cut faces.
A. J. Francis, C. B. Horman and L. E. Jerrems spa lling of brick faces. The solid bricks were obviously less brittle. One of the dry piers with polished contacting faces (perforated brick) developed the remarkably high strength of 9760Ibf/in 2, failing with explosive force. The dry piers with saw-cut contacting surfaces, however, were much weaker, and it was noticed that slight cracking and spalling began under very low stresses (of the order of 300 Ibf/in 2 in some cases). This was probably caused by uneven bearing between the dry surfaces, which indicates the important part played by the mortar joints in reducing or eliminating these stress concentrations, as had been demonstrated by WEST and others. 3 The compressive strengths of the prisms are plotted against the mean joint thickness in Figures 6 and 7. So l id Four
Br ic" Pr i sms Bricks H igh
35 Br ick Prisms
Four Br i c" s HI9h
760 I bf/ i n'
":'--O~II' 5 600 Jbf/i n'
o Dr y Pr i sms ; Po li shQd FOCQS • Dry Pr i 5ms , Mosonry SowCul Foces
Vori ous Jo i nl Th ick ness.s A verogQ CurvQ
Dr y Pr isms , Pol i shQd FOCQS
Dry Prismsí Mosonry So",Cul FO CQS V orious Jo i nl Thi cknesses
--- Averog. Curve
O Av.roge Jo inl Th ickn.ss , Im [in]
FIGURE 7- Variation of prism compressive strength with morta r joint thickness- solid bricks. TABLE 4-PARAMETER VALUES
AvorogQ Joinl Thi cknQss , trJl
3-80 = 19.0 2'95 = 14 '75 0·20 0 ·20
FIGURE 6-Variation of prism compressive strength with mortar joint thickness- solid bricks.
lt was hoped to obtain a 'ull from the tests on the prisms with dry joints, but it was obvious that the strengths of these prisms depended very much on the degree of planeness of the contacting surfaces; furthermore, the values bore no consistent relation to the trend of the results for various joint thicknesses. A mean line was drawn, therefore, through the results for the prisms for each brick type with mortar joints and the values at zero joint thickness taken as a'u/t: these were 4150 and 5600 Ibf/in 2 for solid and perforated bricks respectively. With the values of E, 'I, a't, etc. given earlier, eqn. (12) can be plotted for the two types of brick. The values of the relevant parameters are given in Table 4. The theoretical and experimental values of p( = ao/t/ a 'u/t) are shown in Figure 8. The most striking aspect of these results is the effect of joint thickness on the strength of the prisms. This phenomenon has been observed before in the Iiterature. 4 Obviously, the thinner the joints the stronger the brickwork, and this is the reason for the requirement in the SAA Brickwork Code that in structural brickwork (i.e. brickwork requiring engineering design to the
4150 = 11040 5600= 18'80 364 298
requirements of the Code) the joints must not be more than 1--in. thick, and shall preferably not exceed i in. in thickness. (The ancient Egyptian, Greek and Roman engineers, in whose stone temples, aqueducts, and other major structures the joints are usually extremely thin, often with no morta r at all, evidently had an intuitive sense of good practice in masonry construction.) Another interesting aspect of the results is the greater loss of strength with increase in joint thickness for the perforated bricks compared with the solid bricks. It seems certain that a major cause of this is the lower ratio of >, i.e. the greater weakness in lateral tensile strength in the perforated bricks caused by the presence of the holes. The theory indicates a less pronounced difference between the performance of the two types of brick than was found experimentally. The reason for this discrepancy probably lies in the method of estimating the lateral tensile strength, which is admittedly crude and probably only gives a rough approximation to the true value of this parameter.
36 The Effect of Joint Thickness and Other Factors on the Compressive Strength of Brickwork or thicker waIling to have a lower compressive strength than 4t -in. walling, because of the presence of vertical joints in both directions. This is again confirmed by experiment: Swiss results quoted by MONK5 are given below.
501 i d
So li d Bricks
AV Qrog Q Joint Th i ck n Qss , tm [i n] FIGURE
Single brick width Single brick width Single brick width Multiple brick width
8- Variation of p with joint thickness: theoretical and experimental resu lts compared.
Further, the fail ure criterion adopted in the theory is only conjecturaI. The value used in eqn. (12) for Poisson 's ratio of the mortar also has an important bearing on the result. The figure ofO ·25 was not measured but was suggested by the Division of Building Research CSIRO, Melbourne. 3. FURTHER DISCUSSION OF THE MECHANISM OF COMPRESSIVE FAILURE On the basis of the model proposed above, other observed phenomena associated with the compressive strength of brickwork can be explained. 3.1 Strength of Short 4t -in. WalIs The presence of vertical joints, which have a much lower lateral tensile strength than even the bricks, may be expected to reduce further the compressive strength under axial load, and the greater their frequency in the brickwork, the lower should be the compressive strength. Thus we would expect the four-high prisms used in the present investigation, and standard brickwork compressive strength specimens in the SAA Code, to be stronger than 4t-in. walling of small slenderness ratio. This is in agreement with experimental results, and Rule 22.214.171.124 in the Code requires that the basic strength of brickwork (F'm) for the purposes of establishing design stresses is to be taken as 0 ·75 of the minimum prism strength. If an average bond strength of 60 Ibf/in 2 is assumed at the vertical joints, then since a tensile crack will pass through joint and brick in alternate courses the mean value of a' l for the solid bricks in this study would be (364 + 60)/2 = 212Ibf/in 2 . Eqn. (J 2) then yields a value for the strength of 4t-in. walling which is in reasonable agreement with the test results given by FRANCIS.4 3.2 Strength of Short 9-in. WalIs Continuing the above reasoning, we should expect 9 in.
5 6 7- 10 10- 15
1·00 0 ·89 0 "80 0·68
Account is taken ofthis in the Swiss regulations governing structural masonry.6 3.3 Brick and Mortar Properties The theory indicates that the strength of brickwork should increase with the compressive strength of the bricks, and with increase in the compressive strength (and therefore in the modulus of elasticity) of the mortar. These two effects are well known and allowed for in all modem codes dealing with the form of construction. Poor lateral tensile strength in the bricks compared with their compressive strength is known to have a deleterious effect on the strength of brickwork, not only from the work reported in this paper but also from the extensive test programme recently conducted on storey-height walls by the British Ceramic Research Association,7 and from tests on storey-height walls and on wallettes for the Brick Development Research Institute. 8 Provision for this effect has not been made in codes in the past, but it may be desirable to place some limit on the reduction in cross-sectional area of perforated bricks to be used in highly stressed construction. 3.4 Bond Strength Since the lateral strength or vertical joints depends mainly on the bond strength between bricks and mortar, it is obvious that the axial compressive strength of brickwork must be improved if good bond strength is achieved. Bond strength is of COLme, of paramount importance wherever bending or eccentricity of load causes tensile stresses. 3.5 Joint Reinforcement Steel reinforcement placed in the bed joints, even if only light-gauge, will substantially increase the compressive strength of the mortar and particularly the effective value of Em and the Poisson's ratio of the jointing. Tests by HENDRy9 showed that the compressive strength of 4t -in . storey-height walls was increased by over 60 % when every course was reinforced with a patent woven mesh, but that if only every fifth COLme was reinforced there was no strengthening effect. 4. CONCLUSIONS The mechanism of compressive failure developed in this paper certainly does not take account of all the relevant factors of importance, but it is a start in the right direction. It appears to be capable of explaining a number of well-known but apparently unconnected phenomena associated with the behaviour of brickwork in compression, in particular the effect of the thickness of joints and of the lateral tensile strength of the bricks.
A. J. Francis, C. B. Horman and L. E. Jerrems ACKNOWLEDGEMENTS
Thanks are due to Mr 1. C. McDowall, formerly Director, Brick Development Research Institute, Melbourne for help with the project, and to Mr C. Tonta for assistance in the testing. REFERENCES 1. BRADSHAW, R. E ., Structural Ceramics. Conference on Industrial Building and the Structural Engineer. Inst. Str. Eng., 1966. 2. HILSDORF, H. K., Investigation into the Failure Mechanism of Brick Masonry Loaded in Axial Compression. ' Designing, Engineering and Construction with Masonry Products.' Edited by F . B. Johnson. Houston, Texas, Gulf Publishing, 1969. pp. 34-41. 3. WEST, H. W . H. , EVERILL, J. B. and BEECH, D. G ., The Testing of Bricks and B10cks for Load-bearing Brickwork. Proc. X Int. Ceram. Congr., Stockholm, 1966. pp. 559- 565 .
4. FRANCIS, A. J., The S.A .A. Brickwork Code: The Research Background . Inst. Engrs. Aust. , Civil Eng. Trans., October 1969. pp. 165-176. 5. MONK, C. B. Jr., Old and New Research on Clay Masonry Bearing Walls. 1st Nat. Brick and Tile Bearing Walls Conf., Pittsburgh (May, 1965). 6. SCHWEIZERISCHER INGENIEUR UND ARCHITEKTEN, Technical Note No. 113. Verein, SIA, Oct. 1963. 7. WEST, H . W. H ., HODGKINSON, H. R. and DAVENPORT, S.T.E., The Performance of Walls Built of Wire-cut Bricks with and Without Perforations. Trans. Brit. Ceram. SOCo 67, (10) , 434, 1968. 8. McDoWALL, L c., McNEILLY, T . H. and RYAN, W. G., The Strength of Brick Walls and Wal1ettes. Special Rpt. No. I, Brick Development Research Ins!., Melbourne, Nov. 1966. 9. PRASAN, S., HENDRY, A. W. and BRADSHAW, R. E. Crushing Tests on Storey-height Wal1s 4·Hn. Thick. Proc. Brit. Ceram . Soe. (4), 67, 1965.