Polyurethane or epoxy? It depends on one question

The question that decides it

Stand at the crack and ask what you are actually trying to achieve. There are only two honest answers, and they lead to different materials.

If the answer is stop the water, you need a sealant. Polyurethane injection resin is a water-reactive foam or gel. It cures on contact with water, expands into the crack, and forms a plug that blocks the path. It has some flexibility, it works in a wet crack, and it contributes essentially nothing to the strength of the element. That is not a shortcoming, it is the design.

If the answer is restore the strength, you need an adhesive. Structural epoxy is a two-part resin and amine hardener at 100 per cent solids with no solvent, no expansion and negligible shrinkage. It welds the two faces of the crack back into one monolithic element and restores the load path. A correct epoxy injection produces a repair where the bond exceeds the tensile strength of the parent concrete, so if the element is loaded to failure afterwards, the concrete breaks beside the old crack rather than in it.

Those are two different outcomes. Neither material delivers the other one. The most common failure in this trade is not bad workmanship, it is a correct installation of the wrong material.

If the crack is leaking, epoxy is the wrong answer. Full stop.

This is the flat statement the rest of the internet hedges. A crack with water moving through it cannot be repaired with structural epoxy, and it does not matter how good the epoxy is.

The mechanism is simple and it is physical. Epoxy cures by chemical reaction between resin and hardener at a fixed ratio. Uncured epoxy sitting in a crack with water running past it gets emulsified and washed out, and what remains behind is off-ratio and contaminated. It does not cure into a hard, bonded, load-transferring solid. It cures into a soft, weak, useless plug that looks like a repair from the outside and has none of the properties anyone paid for. Days later, when the plug has failed or the water has simply found its way around it, the leak is back and the crack now has a soft mass in it that has to be dealt with before anything else can be injected.

There is a second, quieter problem even where the water is not running. Epoxy is an adhesive, and adhesives need a clean, dry substrate to bond to. A film of water on the crack face acts as a bond breaker. Moisture-tolerant epoxy grades exist and they genuinely help with a damp crack, but tolerating dampness is not the same as tolerating flow, and no grade of epoxy is designed to cure inside moving water.

Polyurethane is the opposite. It needs water. Water is the reactant that cures a hydrophobic polyurethane, and if a crack is bone dry the correct procedure is to flush water into it first so the resin has something to react with. A wet, actively leaking crack is not a problem for polyurethane. It is the condition it was designed for.

Side by side, with the numbers

The differences below are not preferences. They are material properties, and each one has a practical consequence on site.

What epoxy actually does, and the four conditions it needs

Epoxy injection is the only injection material that restores structural continuity. Foam cannot do it and gel cannot do it. At a typical cured compressive strength in the 60 to 90 MPa range, with a bond that exceeds the tensile strength of the concrete around it, a correct epoxy injection returns a cracked element to something close to monolithic behaviour. That is a genuinely valuable outcome and there are jobs where nothing else will do: cracked beams, columns and corbels, delaminated sections, tilt panels after movement has ceased, and slabs on ground where load has to transfer across the crack.

It also has a second, less obvious role on wet jobs. A stiff epoxy paste is frequently used as the surface seal capping a crack so that something else can be injected through ports without blowing straight back out of the crack face. Seeing epoxy on a materials list does not automatically mean epoxy injection; it may be doing that job instead.

What epoxy needs is non-negotiable, and it is four things.

• A dry, non-flowing crack. Moisture-tolerant grades exist and they widen the window, but running water will emulsify or wash out uncured resin and leave a soft plug with none of the properties above.
• A dormant crack. Movement tolerance is effectively zero. If the element is still moving, the epoxy holds and the concrete cracks again alongside it, which looks like a failed repair to the client and is very hard to explain after the money has been spent.
• Time and temperature. Cure takes 24 to 72 hours and slows badly below about 10 degrees Celsius. Unheated basements, plant rooms and car park soffits on a cold Queensland winter night sit closer to that threshold than people expect.
• Sound steel behind it. Epoxy is the wrong material for corrosion-driven cracking. Sealing a crack over rusting reinforcement locks in a problem that keeps expanding, and the correct scope is concrete repair with break-out, steel treatment and reinstatement.

What polyurethane actually does, and where it is genuinely weak

A hydrophobic polyurethane is a single-component MDI prepolymer that cures on contact with water. The water is a reactant only and is not built into the finished polymer, which is why the cured foam is dimensionally stable through wet and dry cycling. The reaction releases carbon dioxide, which blows a closed-cell foam. Gel time is dialled in on site by dosing an accelerator at roughly 1 to 10 per cent by volume, giving anything from about 20 seconds for a gusher to several minutes for a fine crack that needs time to travel.

The honest weaknesses matter as much as the strengths, and they are rarely published. Bond to a wet substrate is poor, because the water film acts as a bond breaker. The seal is mechanical, a plug keyed into the geometry of the crack, not an adhesive bond to the concrete. Elongation is low, so it is a semi-rigid to flexible foam rather than an elastomer, and it is not a movement joint solution. Injected too fast or too hot, the cells collapse and you get a weak, friable mass. And foam can vanish into a void or a drainage layer behind the element without ever sealing the face, which is why disciplined contractors set a hard volume budget per packer, commonly 1 to 2 litres, and stop rather than pumping a drum into the ground.

The expansion figure deserves its own warning, because it is the most commonly misrepresented number in the category. Manufacturer free-rise figures of 8 to 30 times are unconfined laboratory numbers measured in an open container. Inside a confined crack under back-pressure, real expansion is far lower, and field practice assumes roughly 3 to 8 times. Estimating resin volume from a free-rise figure on a data sheet is how a job runs out of material at four in the afternoon.

The second question: is the crack still moving?

Once the water question is answered, movement is the next gate, and it is the one that most often disqualifies epoxy on a job where the crack is otherwise dry.

A dormant crack has stopped opening and closing. An active crack is still moving, with temperature, with moisture, or because the building is still moving. Thermal movement alone is enough to disqualify epoxy in some elements: precast and tilt panels move every day with the sun, which is why epoxy in a panel crack that is still cycling will simply re-crack alongside the repair.

Where movement is uncertain, the correct answer is not to guess. Fit tell-tales, which are simple gauges fixed across the crack with a scale on them, and read them over 4 to 12 weeks so the reading spans a temperature and moisture cycle. It is cheap, and it prevents the most expensive mistake available in this trade, which is a rigid repair in a moving crack. It also distinguishes the two cases that look identical on the first visit: a shrinkage or thermal crack, which is dormant and injectable, and a flexural crack caused by load, which is a structural warning sign that should not be sealed over at all.

There are cracks nobody should be pricing on the first visit. Diagonal cracking, cracking that widens toward the top, any measurable lean, bulge or step, and cracking in beams, columns or other primary members are all engineer territory. So is any drilling into a post-tensioned slab, which requires scanning and written engineer approval before a hole goes in, because cutting a tendon is not a recoverable mistake.

The jobs that legitimately use both

A structural crack that is also leaking is not a contradiction, it is a sequencing problem. It happens on basement walls, on cracked suspended slabs and on tank structures, and the correct approach uses both materials in a specific order.

Polyurethane first, to kill the water. A hydrophobic resin with a fast accelerator dose, injected through mechanical packers bottom to top, stops the flow within minutes and holds the crack dry. Where flow is heavy, that is preceded by a hydraulic cement plug or oakum packing to knock the volume down so the resin gels in the crack instead of being washed out of it.

Epoxy second, once the element has genuinely dried and movement has been established as dormant. Only then does the structural repair have the conditions it needs.

Two practical points decide whether this sequence is realistic on your job. The first is that the order cannot be reversed: once epoxy has cured in the crack there is no path left for foam to travel along, so anything that has to be done for water has to be done first. The second is that below-ground elements may never dry through. A basement wall under a water table can hold moisture indefinitely, and if the crack cannot be dried and kept dry, epoxy stays off the table and the structural requirement has to be met another way, by strengthening design rather than by injection.

When the honest answer is neither

There are common situations where the polyurethane against epoxy question is the wrong argument entirely, because the correct material is a third thing or the correct scope is not injection at all.

Acrylate gel is the answer where the path is too fine or too diffuse for polyurethane. It is a two to four component water-based system with a viscosity of roughly 10 to 50 mPa·s, essentially that of water, so it permeates soil, capillaries and hairline cracks that no polyurethane will enter. Gel time is controllable from a few seconds to around an hour. It has zero structural strength and it is highly elastic, with elongation often in the hundreds of per cent. It is the correct material for curtain injection into the ground behind a wall, for honeycombed and porous concrete, and for structures with real elastic movement.

Microfine cement grout is the answer for large void volumes. Filling honeycombing or a void network with resin is expensive by volume and adds nothing structurally, whereas cement in cement matches modulus and will not creep, embrittle or degrade. Its hard limit is particle size: as a field rule the crack or pore needs to be at least three to five times the d95 particle size, which in practice means it will not reliably enter cracks below roughly 0.3 to 0.5 mm, and flowing water washes the fines out before set, so the flow has to be killed first.

And then there are the cases where no injection material is correct. A designed movement joint filled with resin defeats its purpose and transfers the movement into the adjacent concrete, which then cracks; that is a backing rod and movement-capable sealant scope. Cracking that follows the reinforcement with rust staining is corrosion and needs a concrete repair scope. Broad-area seepage through the body of a wall or slab with no defined line has nothing to grip and needs tanking, a drained cavity or curtain injection. Damp with no rain correlation in a still, humid space is often condensation. And a blocked ag line or a rendered-over weep hole behind a leaking wall is a drainage problem that sealing the concrete will make worse, because it raises the water head against the structure.

The two methods do not look alike on site

If you are watching the work or reading a method statement, the two materials produce visibly different procedures. Knowing which one you should be seeing is a useful cross-check on whether the right decision was made.

A polyurethane leak-sealing scope drills 13 mm holes at 45 degrees to the face, alternating either side of the crack, offset from the crack line by roughly half the element thickness so a 45 degree hole meets the crack at mid-depth. Holes go two-thirds to three-quarters of the way through and never break out the far face, or all pressure is lost. Mechanical packers go in at 150 to 400 mm centres, with a common rule that spacing should not exceed the element thickness. Each packer is water-flushed before injection, which proves the crack is hydraulically connected from one packer to the next, clears laitance and salt deposits that would block resin, and pre-wets a dry crack. Injection runs bottom to top on a vertical crack at roughly 35 to 140 bar, so that resin displaces water upward and out rather than trapping it. Packers stay in for a minimum of 24 hours before being cut back below the surface and plugged.

An epoxy scope frequently looks nothing like that. Ports are often glued to the face over the crack with a stiff epoxy paste capping the crack between them, and injection runs at low pressure, roughly 0.7 to 3.5 bar and commonly around 10 to 12 psi. Crucially, a crack is never water-flushed before epoxy unless it can then be dried with oil-free compressed air or a genuinely moisture-tolerant grade has been specified, because the flush water is exactly what the epoxy cannot cure through. Cure then takes 24 to 72 hours before anything is cut back.

One method uses water as a tool. The other treats water as the enemy. That is the whole difference, expressed as a procedure.

How to tell, from a quote, that the wrong resin has been chosen

You do not need to be technical to catch most of these. Each one is a mismatch between what the quote says the problem is and what the quote says the material will be.

1. Epoxy specified on a scope that also describes waterIf the same document says active leak, water ingress, weeping crack, damp patch or efflorescence, and then specifies epoxy injection, ask how the resin is expected to cure with water in the crack. Running water emulsifies uncured epoxy into a soft plug, and moisture-tolerant grades tolerate damp, not flow.
2. Surface ports on a wall that is currently wetThe epoxy paste surface seal that surface ports depend on will not bond to wet concrete. If the wall is leaking and the method statement says ports glued to the face rather than mechanical packers drilled at 45 degrees, the method does not match the condition.
3. No mention of flushing on a polyurethane scopeWater-flushing proves the crack is connected between packers before any resin is bought. A leak-sealing scope that does not mention it is a scope that will probably need doing again.
4. Either resin specified across a crack with rust staining along its lineCracking that follows the reinforcement with rust staining, spalling or drummy sounding areas is corrosion driven. Sealing it with anything traps an expanding problem behind the repair. The correct scope is break out to sound concrete behind the bar, treat or replace the steel, and reinstate with a compatible repair mortar.
5. Injection specified across a designed movement jointExpansion joints, control joints and panel-to-panel joints are built to move. Filling one with resin defeats its function and pushes the movement into the adjacent concrete, which then cracks. That work is joint preparation, correct backing rod and a movement-capable sealant, and it should not appear on an injection quote at all.
6. An epoxy scope with no assessment of whether the crack is still movingEpoxy has effectively zero movement tolerance. On anything where movement is uncertain, the honest sequence is tell-tales read over 4 to 12 weeks before injection, not a decision made in the first ten minutes on site. A quote that commits to epoxy on the first visit, on an element that might still be moving, has skipped the step that protects you.

Questions we actually get asked

Moisture-tolerant epoxies exist. Do they solve the wet crack problem?

They widen the window, they do not remove the rule. A moisture-tolerant grade will bond to a damp substrate that a standard grade would fail on, which is genuinely useful in a below-ground element that never quite dries. What no grade of epoxy can do is cure inside water that is moving. Flow emulsifies and washes out uncured resin, leaving an off-ratio, contaminated, soft plug. If water is running, weeping or beading, the sequence is polyurethane first to kill the flow, and epoxy only afterwards if the element dries and the crack is proven dormant.

Does temperature really matter for epoxy in South East Queensland?

Less often than in a cold climate, but it is not irrelevant. Epoxy cure takes 24 to 72 hours and slows badly below about 10 degrees Celsius. A shaded basement, an unventilated plant room or a car park soffit on a winter night can sit close to that, and the element temperature matters more than the air temperature because concrete holds its overnight temperature well into the morning. The practical effect is a longer cure and a longer wait before anything is cut back, not a failed repair, provided the programme allows for it.

If epoxy is so much stronger, why not just use it everywhere?

Because strength is not what most of this work is for. At 60 to 90 MPa compressive with a bond that beats the tensile strength of the parent concrete, epoxy is the strongest option available and it is completely useless in a leaking crack, because it cannot cure in water and it has effectively zero movement tolerance. It is also worth being precise about what it achieves: epoxy restores the capacity the element was designed to have. It does not add capacity. An element that is failing because it is under-designed or overloaded needs a strengthening design, not a stronger glue.

Does polyurethane make the concrete stronger?

No, and anyone who says otherwise is overselling it. Polyurethane injection resin is a foam or gel. It fills the crack, blocks the water path and tolerates a little movement, and it contributes essentially nothing to load transfer across the crack. That is exactly what a waterproofing repair is supposed to do. If the element genuinely needs its strength restored, that is a separate decision requiring a dry, dormant crack and, in most cases, an engineer.

How do I find out whether my crack is still moving?

Fit tell-tales, which are simple gauges fixed across the crack with a printed scale, and read them over 4 to 12 weeks so the readings span a temperature and moisture cycle. It is cheap, it is non-destructive, and it is the difference between a repair that lasts and a repair that cracks again alongside itself. Movement assessment is the step most often skipped by contractors who want to inject on the first visit, and skipping it is what turns a one-off cost into a recurring one.

Our engineer has specified epoxy but the crack is wet. What now?

Raise it before anyone drills, and get the answer in writing. There are three legitimate paths. Seal the water first with polyurethane, allow the element to dry, and then inject epoxy as a second, separately programmed operation. Or specify a moisture-tolerant grade and dry the crack with oil-free compressed air if the moisture is residual dampness rather than flow. Or, where the element will never dry because it sits under a water table, accept that epoxy is not available and meet the structural requirement with a strengthening design instead. What should not happen is epoxy going into a flowing crack because that is what the specification said.

How would anyone actually prove the epoxy filled the crack?

Honestly, on most jobs nobody does, and it is worth knowing that. During injection, resin appearing at the next port along is confirmation that it has travelled. The only genuine verification of penetration through the element is coring across the repaired crack and examining the core, which is destructive and expensive and should only be done where an engineer specifies it on a structurally critical repair. For water sealing the verification is different and much more practical: observation over at least 24 to 48 hours and ideally through a rain event, moisture readings at the same marked points before and after, and comparable photographs taken from the same positions before and after the work.

Where to next

• find out which resin your crack needs
• polyurethane crack injection
• how concrete crack injection works
• hydrophilic or hydrophobic resin
• what crack injection costs in Australia
• when injection is the wrong fix entirely