Masonry facades

The previous pages covered some of the mechanisms by which movement can occur in a masonry facade. It is appropriate to consider the effects of those movements and the bowing or crack patterns that they create. The diagnosis of building movements is not necessarily straightforward, and sometimes cannot be fully understood until a pattern of movement has been established after a period of monitoring.

In theory, brickwork that is totally unrestrained ought to be free to expand and contract according to moisture and temperature changes without harm. In historic construction, lime-based mortars created a degree of flexibility that would allow brickwork to move without necessarily cracking; this is not the case in contemporary construction where brittle Portland cement mortars are used. However, brickwork is not necessarily homogenous anyway and since it is generally very weak in tension, cracking can occur as a result of shrinkages, or, according to location, expansion.

Putting aside cracks that occur as a result of structural or chemical changes, cracking in brickwork of 0.5mm to 5mm or so is often symptomatic of expansion or shrinkages due to temperature or moisture effects. Vertical cracks usually occur at right angles to the forces that cause them, while horizontal cracks (usually finer) indicate the slip plane. It is usually possible to distinguish between shrinkage cracking and thermal expansion:

• with expansion there is usually some form of displacement, for example, brickwork sliding off a damp-proof course, cracking at corners or evident disruption;
• shrinkage cracking does not generally produce these manifestations.

Assuming a brick panel expands, say, by 2mm as a result of temperature or moisture, the compressive forces generated may be sufficient to overcome friction and create a slip plane, for example, along a damp-proof course. When the temperature drops, or moisture content reduces, tensile forces are set up as the wall tries to return to its original size. This time the frictional forces along the slip plane serve to place the wall in tension. It is probable that the strength in tension of the panel is less than the frictional restraint and so the only way in which the stress can be relieved is by the development of a crack.

This type of problem can often be seen in contrasting brick string courses, particularly if these are of a darker colour than the adjoining brickwork. The darker band absorbs more heat than the adjoining brick and so expands more. Friction prevents or reduces the ability of the band to return to its original length and cracking ensues.

If the brickwork were free to move along the slip plane and take up the width of the crack as it expands again, all well and good. In practice, the crack will probably fill with small pieces of debris, mortar, etc. that will prevent this from occurring. The result will be progressive movement as the cracks gradually 'ratchet' up.

Tension cracks in feature courses

Spalling of concrete slab due to expansion of brickwork and missing slip plane where DPC stops short of the corner

Example of cracking arising from rotation of short return (see figure below)

Cracking on line of short return due to expansion of long panel of brickwork

Vertical cracking can often be seen at the vertical end of a cavity wall where it returns. Commonly, brick spandrel panels beneath ribbon glazing exhibit this type of problem - usually caused by the restraint provided by wall ties placed close to the return and by the effects of expansion in the longer section of the spandrel panel.

Moisture and temperature effects can also become manifest in brickwork that is arranged in panels of different aspect ratios, such as the section between windows and the longer spandrel section (as shown in the figures below). The relative expansion and contraction of the 2 sections can result in horizontal cracking.

Cracking due to expansion of long brick panels and shorter panels between windows - expansion occurs vertically as well as horizontally

Alternative crack pattern due to expansion of brickwork in long run

Occasionally, cracking and displacement of brickwork can occur at ground level, particularly above and below dpc level. The problem is often associated with the brickwork oversailing the dpc. Such movement could indicate thermal and moisture movement, but could also indicate that the floor slab is expanding, possibly as a result of expansive reactions such as sulphate attack.

By comparison with thermal and moisture movements, damage resulting from structural problems such as lack of support, foundation failure, impact damage, etc. can be up to 25mm or so in width. Cracks created by structural movements may be vertical (tension) or diagonal, being reflective of shear (slicing) forces. Such movements often affect both leaves of a cavity wall rather than simply the outer leaf.

Movement control in walling is not the only issue. Thought needs to be given to the use of brickwork in conjunction with elements of structure that could provide a restraining influence. Any form of construction that will prevent free movement will allow tensile forces to build up and cracking to form. For example, concrete columns or walls cast up against brickwork will create a restraining influence unless separated from the brickwork by a slip membrane. The propensity of concrete frames to shrink over time is well known, and provision for this needs to be allowed in construction - a requirement that is probably more important with clay brickwork than with calcium silicate bricks.

In all of the above cases, adequate restraint needs to be maintained and so the use of appropriate sleeved anchors and ties must be considered.

In 1986, CIRIA published the results of an investigation into cracking problems experienced on a sample survey of 85 buildings constructed of both clay brickwork and concrete blockwork (Movement and cracking in long masonry walls, Special Publication 44). The study examined aspects of the structure and fabric that were most vulnerable to cracking and concluded that for unbroken wall lengths exceeding the limits given in BS 5628 moderate level cracking was no more likely to occur than no or slight cracking. The study went on to consider the influence of structural form and details, with the following features being vulnerable to cracking:

• Short returns to a long straight wall: Returns that are less than 900mm are vulnerable owing to expansion of longer length of wall. The shorter the length of the return, the greater the risk of cracking.
• Spandrel walls: Spandrel walls are, or can be, relatively unrestrained permitting circumstances where the amount of free movement is greater.
• Long parapets: Parapet walls are also relatively unrestrained, although expansion of all 4 sides of a square or rectangular roof can lead to bowing or displacement of the parapet.
• Stronger mortars: Strong mortars are brittle and weak in tension. Lime mortars can accommodate movement by forming many tiny cracks rather than several larger ones. Mortars with a lime or low cement content possess some plasticity.
• Discontinuous movement joints: A movement joint reduces the tendency of the masonry to crack in other locations, it effectively permits what would otherwise be a crack in a specified location. If the movement joint is discontinuous, the notional crack will continue as an actual crack through the masonry.
• Discontinuous DPC: The DPC acts as a slip plane, permitting a small degree of movement. If the DPC is discontinuous, frictional resistance is increased in a localised position.
• Brick slips: Brick slips and the propensity for damage resulting from creep of concrete frames.
• Incompressible joint fillers: Fillers such as bitumen impregnated fibreboard do not compress particularly well and can act as a restraint to expansion.
• Changes of vertical load: The abrupt change of vertical load, for example at the junction of a low rise and high rise building, will give rise to cracking unless the structures are isolated.
• Changes in shape, thickness and height of wall: Changes in the aspect ratio of brick panels mean that the amounts of thermal and moisture movement will be dissimilar.
• Eccentrically confined walls: For example, projecting brick panels confined within a concrete frame.
• Bonding to dissimilar materials, for example clay bricks with calcium silicate backing: Different propensity for moisture or thermal movement, for example, expansion on one face coupled with contraction on the reverse.

Case study

The building was a late 1970s office located within the conservation area of a small provincial town. Accommodation was arranged on 2 floors with a pitched roof terminating in a series of rendered gable walls. The roof structure was constructed of standard fink trusses.

Horizontal cracking (0.5-1mm) had occurred along the base of the gable and a dispute arose between landlord and tenant as to the cause. Concern had been expressed by the tenant that the cracking was symptomatic of a lack of wall ties to the gable, but investigations revealed that while the positioning of ties was not strictly in accordance with BS 5628 they were sufficient in number not to be of concern.

Horizontal cracking to base of gable wall

Measurements with a plumb bob revealed a slight inward lean of the gable wall on the south-west elevation and a corresponding slight outward lean on the north east gable. Internally, an inspection of the roof void revealed that while galvanised steel restraint straps had been installed between the roof trusses and the end gables, the work was defective because:

• the straps should have been taken over at least 3 sets of rafters (they had been fixed to 1 or in some cases 2);
• solid bridging had been placed between the 2 end rafters;
• no bridging had been provided between the inside of the gable wall and the nearest rafter; and
• diagonal bracing to the fink trusses was insufficient.

Lack of solid bridging to restraint strap between gable wall and nearest adjacent trussed rafter. The metal strap was also laid on top of the rafter rather than on the underside, and hooks over a cut block rather than a full block, thus reducing its effectiveness.

The lack of solid packing between the trusses and the gable walls meant that there was no immediate restraint against inward movement of the wall. The gable walls needed to be firmly anchored to a braced roof structure; they were not. The effects of positive wind pressure on the south west gable, which was acting as an unrestrained cantilever, were sufficient to cause a small amount of movement in the wall and overcome the tensile strength of the mortar bedding and render. Similarly, negative pressure was sufficient to influence the slight outward movement of the north east gable.

While fink trusses are virtually standard in housing, they are somewhat less common in commercial buildings since spans tend to be greater. Up to a span of around 12m detailing can be compared with domestic construction but beyond this special measures will be needed. An additional problem with the subject building was that conventional lay-in grid suspended ceilings had been installed at the first floor, meaning that the stiffness that would normally be provided by a plasterboard ceiling was missing. (Further details of bracing may be found in Timber Engineering Guidance Document 8, Bracing for non-domestic timber trussed rafter roofs, TRADA, 1999.)

The restraint problem could be resolved by providing new straps and solid bridging, ensuring that the rigidity of the 3 trusses nearest the gable wall were sufficient to provide proper lateral restraint.