Is this crack structural? How to read a crack in concrete

The three properties that settle it

Width tells you how much has moved. Pattern tells you what is moving. Movement over time tells you whether it has stopped. A crack cannot be classified from any one of them alone, which is why a photograph on its own is rarely enough to price a repair from.

The order matters as well. A hairline crack in a diagonal pattern that is still opening is a more serious finding than a wide crack in a vertical pattern that stopped moving a decade ago. Width is the property people fixate on and it is the least informative of the three.

Width, and what the standard actually allows

Measure it with a crack comparator card — a printed plastic card with graduated lines, available for a few dollars — or a feeler gauge, not by eye. Estimates by eye are routinely out by a factor of two, and the bands below are narrow enough that a factor of two changes the answer.

AS 3600, the Australian concrete structures standard, allows for crack widths broadly in the 0.1 to 0.4 mm range at serviceability, depending on the element and its exposure classification. That is worth knowing, because it means fine cracking in concrete is a designed-for condition rather than a defect. Concrete cracks as it cures and as it carries load; the standard accepts it and specifies reinforcement to control it.

Pattern: what the shape of the crack is telling you

Pattern is the most diagnostic of the three properties and the easiest to read from a photograph, provided the photograph is wide enough to show the whole element rather than a close-up of the crack.

• Vertical, roughly uniform width, often near the middle of a panel or wall. Usually shrinkage. The most benign pattern and the most commonly injected.
• Horizontal, at the line where a wall meets a slab. Usually a cold joint rather than a crack at all — the line where one pour met the next. It is the most common water path in a basement or undercroft and it is treated as a construction joint rather than as cracking.
• Diagonal, particularly running from the corner of an opening. Movement. Diagonal cracking indicates shear or differential settlement, and it is an engineer’s question rather than a waterproofer’s.
• Stepped through the mortar joints of blockwork. Movement in the wall or its footing. On a retaining wall this is a stop sign, not a repair scope.
• Widening toward the top. Rotation — the wall is leaning. On a retaining wall this warrants immediate structural attention, because retaining wall failures are sudden.
• Map or crazing patterns across a surface. Usually surface phenomena — plastic shrinkage, drying shrinkage, or alkali-silica reaction in older concrete. Rarely a defined water path, which means rarely injectable.
• Following the line of the reinforcement, with rust staining, spalled edges, or concrete that sounds hollow when tapped. Corroding steel. The crack is a symptom of expansion and sealing it hides an accelerating problem.

Movement: the property everybody skips

A crack is either active — still opening, closing or shearing — or dormant. It is the single most important property for deciding a repair, and it cannot be established from one visit, which is precisely why it gets skipped.

The reason it matters is that repair materials have almost no movement tolerance. Cured epoxy elongation is effectively zero: it welds the two crack faces into one element, so a structure that is still moving cracks again a few millimetres away, usually within a season. Polyurethane foam is semi-rigid rather than elastomeric, so continued cyclic movement tears it open. A rigid repair in an active crack is a repair with a known expiry.

Establishing it takes tell-tales or crack monitors — a two-part plate fixed either side of the crack with a graduated scale, or a simple glass or plaster tell-tale — read over 4 to 12 weeks. The reason for that duration is not caution for its own sake: it has to span a full moisture and temperature cycle. In South East Queensland, where reactive clay drives an annual shrink-swell cycle in the ground, a crack read only through a dry month can look perfectly dormant and open again in February.

Two months of monitoring costs a fraction of injecting the same crack twice, and it is the step that most reliably separates a contractor who diagnosed the problem from one who priced the symptom.

When it stops being a waterproofing question

There is a defined set of findings where the correct professional answer is that no contractor should price a repair until a structural engineer has assessed the element. They are worth knowing because a waterproofer being pushed to quote around them is being asked to take on somebody else’s risk.

• Any lean, bulge or bow in a wall, measured rather than eyeballed.
• Stepped cracking through mortar joints, or diagonal cracking from the corner of an opening.
• A crack that widens toward the top of the element.
• A change in floor level, doors or windows binding, or a gap opening between a wall and a ceiling.
• Cracking that is visibly progressing between visits, or that opens and closes measurably with the weather.
• Any crack in a post-tensioned element, where the additional question is not whether it can be sealed but whether it can be drilled at all — which requires a ground-penetrating radar scan and written engineer approval before a drill touches it.

How the answer decides the repair

Once width, pattern and movement are known, the repair follows fairly mechanically. This is the logic a competent scope is built on.

What to photograph before anybody attends

A crack can be assessed a long way from photographs, provided the right ones exist. Four shots do most of the work, and taking them dated and from fixed positions means they can be repeated later to answer the movement question directly.

1. The whole elementFar enough back to show the floor, the ceiling and both ends. Pattern is only readable at this distance, and pattern is the most diagnostic property.
2. The crack close up, with scaleA comparator card, a coin or a tape in frame. Width claimed without something for scale is an estimate, and estimates by eye are routinely out by a factor of two.
3. Both ends of the crackWhere a crack starts and stops says more about the cause than its middle does. A crack that runs out of the corner of an opening and a crack that runs the full height of a panel are different findings.
4. The other side, and the ground outsideWhether the crack passes through the element, and what the ground, drainage and falls are doing behind it. This is the photograph nobody takes and it frequently answers the question.

Common questions

How wide does a crack have to be before it is a problem?

There is no single width that makes a crack a problem, which is why width alone is the least useful of the three properties. AS 3600 allows for crack widths broadly in the 0.1 to 0.4 mm range at serviceability depending on the element and its exposure classification, so fine cracking in reinforced concrete is a designed-for condition. A 0.2 mm crack in a diagonal pattern that is still opening matters more than a 1 mm vertical shrinkage crack that stopped moving twenty years ago. Measure the width, then read it alongside the pattern and the movement.

How can anyone tell whether a crack is still moving?

By measuring it over time, and there is no shortcut. Tell-tales or crack monitors are fixed either side of the crack and read over 4 to 12 weeks, which is long enough to span a full moisture and temperature cycle. The duration is the point: in reactive clay country a crack read only through a dry month can look dormant and open again after the first sustained rain. A cheaper version of the same test is a dated photograph from a fixed position with a scale in frame, repeated monthly.

Can a crack that is still moving be sealed at all?

Not with a rigid material, and not before the cause has been established. Cured epoxy has effectively zero movement tolerance and welds the crack faces together, so the element cracks again alongside the repair. Polyurethane foam is semi-rigid and tears under continued cyclic movement. Where the movement is small and seasonal, a flexible resin grade, an acrylate gel or a movement-capable sealant can accommodate it. Where the movement is real structural movement, the sequence is engineer first, cause corrected second, seal third — and skipping to the third step is money spent twice.

What does a stepped crack through the mortar joints mean?

Movement in the wall or in whatever it is bearing on, following the path of least resistance through the mortar rather than through the blocks. On a retaining wall it is a stop sign: retaining wall failures are sudden and they take whatever is in front of them. It is an engineer’s assessment before any contractor prices anything, and a contractor willing to quote a resin injection over stepped cracking is quoting the wrong scope.

The builder says it is only shrinkage cracking. Is that plausible?

Often, yes. Concrete shrinks as it cures and continues to shrink for a long time afterwards, and vertical cracks of roughly uniform width near the middle of a panel are the classic signature. It is also the explanation of first resort, so it is worth testing against the other two properties: shrinkage cracking is dormant, so it should show no measurable movement across a monitoring period, and it makes a vertical rather than a diagonal or stepped pattern. If either of those does not hold, the shrinkage explanation does not either.

Does a crack that leaks have to be structural to matter?

No, and conflating the two causes a lot of unnecessary alarm. Most leaking cracks are entirely benign structurally — a shrinkage crack that happens to pass water is a waterproofing problem and nothing more. The reason the structural question is asked first is not that leaks are usually structural; it is that the two repairs are incompatible. Sealing a moving crack fails, so the movement question has to be settled before the water question can be answered.

Is a crack in a post-tensioned slab different?

The assessment is the same but the constraint is not. A post-tensioned slab contains steel tendons stressed after the concrete cured and still under enormous load, and they are invisible from the surface. Whether the crack can be sealed is secondary to whether the element can be drilled at all, which requires a ground-penetrating radar scan, the tendons marked, and written engineer approval nominating both where drilling is permitted and the maximum injection pressure. Where nobody can confirm whether a slab is post-tensioned, it has to be treated as though it is.

Where to next

• fifteen situations where injection is the wrong repair
• polyurethane or epoxy, and the question that decides it
• leaking retaining walls, where movement matters most
• run the leak diagnostic