Epoxy and polyurethane concrete crack injection field guide

Crack injection, and the two jobs it does

Crack injection is filling and bonding a concrete crack from inside the section instead of skimming over the top of it. You get the material down into the full depth of the crack under pressure, where a surface patch never reaches. There are two versions of the job, and they are not interchangeable. One welds the crack back together so the concrete acts as a single piece again. The other plugs the crack so water stops coming through. The same hole on the wall can want either one, and the difference is the whole decision.

Structural epoxy injection is the welding version. A low-viscosity epoxy resin penetrates the crack, cures rigid, and bonds the two faces so the section transfers load across the crack again. Done right, the cured epoxy is stronger than the concrete around it, and the next crack, if there is one, forms somewhere else, not back along the old line. That is the repair for a dormant structural crack you want to make monolithic again.

Polyurethane injection is the water-stop version. A reactive polyurethane resin meets the water in a leaking crack, foams, expands, and seals the path the water was taking. It stays flexible, so it tolerates a little movement, but it does not restore structural strength and it is not pretending to. That is the repair for an active leak in a wall, a slab, or a tank.

Which one you reach for is not a preference. It comes out of the diagnosis: is the crack structural or just leaking, is it moving or dormant, is it wet or dry. Get the diagnosis wrong and the best injection technique in the world fails, because you put the wrong material in the crack.

Find the cause before you inject anything

The crack is a symptom. Before you fill it, you find out what opened it, because a crack that is still being driven by an active cause will crack again no matter what you inject. This is the rule that separates a repair that holds from one that comes back in a season.

Cracks fall into three buckets, and the bucket decides everything. A dormant crack happened once and is done moving, from early shrinkage, a one-time overload, settlement that has stopped. That one is a candidate for rigid epoxy. An active or moving crack is still working, opening and closing with temperature, traffic, or seasonal movement, or still widening because a footing is settling or a slab is curling. Fill that with rigid epoxy and the epoxy holds while the concrete cracks again right beside it. A leaking crack has water moving through it, which rules epoxy out on its own terms because epoxy will not bond a wet face.

Some causes have to be addressed before injection, not after. A crack from ongoing settlement gets the foundation stabilized first, or you are injecting a crack that is going to keep growing. A crack over corroding reinforcing is a corrosion problem wearing a crack, and that is the territory of the spall and corrosion repair guide, not this one: inject the crack and leave the rusting bar and you have sealed the cause inside the wall. A crack from a hydrostatic head outside a basement wall needs the drainage looked at, because you can seal the crack and watch the water find the next one.

The blunt version: name the cause and decide whether it is still active before you order material. If a crack might affect structural capacity, or you cannot tell whether it is still moving, that is a call for the engineer of record before anyone mixes resin.

Epoxy or polyurethane injection?

Epoxy injection is for a dormant structural crack in dry conditions, where you want to restore load transfer and make the section monolithic again. Polyurethane injection is for an active leak, a wet crack, or a crack that moves, where you want a flexible seal against water and do not need structural strength. That single sentence covers most of the decision, and the rest is reading which case you actually have.

Epoxy is rigid and bonds hard. It restores the concrete's ability to carry load across the crack, and the cured resin is typically stronger than the parent concrete. But it needs a dry or only slightly damp crack to bond, and because it is rigid it cannot tolerate continued movement: a moving crack shears the epoxy bond or simply cracks the concrete next to it. Epoxy is the structural answer and the wrong answer for water and movement.

Polyurethane is flexible and water-driven. The resin reacts with water and expands into a foam that fills and seals the crack, so it works on a wet, actively leaking crack that epoxy could never bond, and the flexibility lets it ride small movements without losing the seal. The trade is that it does not weld the section back together. It stops water; it does not restore strength.

So the decision tree is short. Structural and dormant and dry, use epoxy. Leaking or wet or moving, use polyurethane. A crack that is both structural and leaking is a sequencing problem: you often stop the water with polyurethane first, or control it, then make the structural repair, and that is a detail for the engineer to set.

The structural epoxy: welding the crack back to monolithic

Structural epoxy injection works because the resin is thin enough to go where you cannot. A low-viscosity injection epoxy has a viscosity low enough to penetrate fine cracks, commonly down to about 0.002 in, which is roughly 0.05 mm, and follow the crack through the full depth of the section under modest pressure. It wets both faces, fills the void, and then cures into a rigid solid bonded to the concrete on both sides.

Once it cures, the section behaves as one piece again. Load that used to stop at the crack now transfers across it, because the epoxy is bonded face to face and is typically stronger than the surrounding concrete. The practical proof of that strength is where the next crack goes: a properly injected dormant crack does not reopen along the old line, because the old line is now the strongest part of the wall. If the concrete cracks again, it cracks somewhere new.

That strength is exactly why epoxy is the wrong tool on a moving crack. The bond does not yield, so if the crack is still working, the movement has to go somewhere, and it goes into fresh concrete beside the repair. Epoxy assumes the crack has stopped moving. Confirm that it has before you trust the resin to hold it.

Injection epoxies are commonly specified to ASTM C881, the specification for epoxy-resin bonding systems for concrete, with the type and grade chosen for injection and for the crack width and orientation. The low-viscosity grade goes into tight cracks; a thicker grade or a gel holds in a wide or overhead crack. ACI 224.1R covers epoxy injection as a structural repair method, and the ACI field guidance, RAP-1, Structural Crack Repair by Epoxy Injection, walks the procedure. The material data sheet governs the actual product, mix, and cure.

The polyurethane: a flexible water-stop for a leak

Polyurethane injection seals a leaking crack by using the water against itself. The resin is reactive with water, so when it meets the moisture in the crack it foams and expands, filling the void and the side channels the water was running through, then cures into a flexible foam that stays put and stays watertight. Expansion is large, often many times the liquid volume, which is what lets it chase a leak into voids you cannot reach with a rigid resin.

There are two families, and the names get used loosely on the jobsite. Hydrophobic polyurethanes react with water but cure into a dense, more durable foam and are less sensitive to drying out, which suits a crack that wets and dries. Hydrophilic polyurethanes absorb and bond with water and stay more gel-like and flexible, which suits a joint or crack that has to keep tolerating movement and stays damp. Which one fits depends on the leak, the movement, and whether the crack ever dries, and the manufacturer's selection guide is the place to settle it.

The point that gets missed is that polyurethane does not restore strength. It is a seal, not a weld. On a structural crack that also leaks, polyurethane stops the water but leaves the structural question open, so you do not get to call a structural crack repaired because the leak stopped. It also wants the water there to react properly, which is the opposite of epoxy, so a crack that is bone dry is a poor candidate for polyurethane and a good one for epoxy.

Polyurethane injection resins for leak sealing are specified to their own product specifications and water-reaction characteristics, not to the epoxy bonding spec, so do not carry the C881 number over to the polyurethane. Read the polyurethane product's data for the reaction, the expansion, and the conditions it is rated for.

Is the crack active or dormant?

Before you pick a material you measure the crack and decide whether it is still moving, because that one answer drives the choice between rigid epoxy and a flexible seal. A dormant crack is finished moving and can take rigid epoxy. An active crack is still opening and closing and needs flexibility or a movement joint, not a rigid weld.

You tell active from dormant with time and a monitor, not with a guess. A crack monitor, the simple two-plate gauge or a tell-tale glued across the crack, read over weeks through a temperature swing, shows whether the crack is breathing. A smear of brittle material across the crack that cracks again tells you the same thing cheaply. If the crack opens and closes with the day or the season, it is active. If it sits still through the cycle, it is dormant. When the schedule will not give you weeks, the honest move is to treat an uncertain crack as active and use a flexible repair, or get the engineer to make the call.

Width matters next. Epoxy injection has a workable range, commonly from about 0.002 in, near 0.05 mm, up to roughly 0.5 in, near 13 mm, with the resin grade matched to the width: thin resin for tight cracks, thicker grade or gel for wide ones. A crack too tight to take resin gets a different method; a crack too wide or too broken up is a different repair entirely.

Then read the wet and the deep. Is water in the crack now, after a dry spell, during rain? Wet rules out plain epoxy and points to polyurethane. And is it a through-crack, all the way through the section, or surface-deep? A full-depth crack in a structural member is a structural question; a shallow surface crack on a slab is a sealing or a gravity-feed job. Pull all four readings, the movement, the width, the moisture, and the depth, before you decide, because each one can change the material on its own.

Surface ports and the paste-over seal

For most wall and vertical work the method is surface-mounted ports and a paste cap. You glue injection ports along the crack at the surface, then trowel an epoxy paste over the crack face between and around the ports. The paste is the dam. It seals the crack opening so that when you inject, the resin cannot escape out the face and instead builds pressure and travels down into the crack and across to the next port.

Port spacing follows the section thickness. A common rule of thumb is to space the ports about an inch apart for each inch of wall thickness, so an 8 in wall gets ports roughly every 8 in, which on typical work lands ports somewhere around 6 to 12 in apart. The logic is geometry: the resin entering one port has to reach the full depth of the crack before it travels to the next port, and the spacing that lets it do that scales with how deep the crack runs. Wider spacing on a thick wall, tighter on a thin one.

The paste-over has to actually hold pressure, so it is not a skim. It goes on as a band over the crack, commonly on the order of 1/8 to 3/16 in thick and at least an inch wide, mounded a bit around each port, and given time to cure before you inject so it does not blow off under pressure. The one detail crews ruin: do not let the paste plug the bottom of the port itself, because then the resin has nowhere to go into the crack. The port stays open to the crack; the paste seals everything around it.

On a crack you can reach from both sides, port both faces where the detail calls for it. On a thick or one-side-only wall, the ports and the pressure have to carry the resin all the way through, which is the reason the spacing and the seal matter more, not less.

How do you inject a concrete crack?

You inject low and slow, port to port, watching the resin travel. After the paste-over has cured, you start at one end of the crack and pump resin into the first port until it weeps out of the next port along, then cap the one you are on, move to the port that is now flowing, and repeat down the line. The resin showing at the next port is the signal that the crack is full between the two, which is the whole reason you watch for it.

On a vertical crack you work bottom-up. Start at the lowest port and move upward, because the resin fills the crack and rises, pushing air ahead of it, and chasing it from the bottom keeps you from trapping a void you would have to inject around later. When clean resin runs from the port above, that section is full, so you cap and climb.

Keep the pressure low. Low-pressure injection, commonly somewhere in the range of about 20 to 40 psi on these systems, gives the resin time to penetrate the full depth and lets you actually see it reach the next port. Push the pressure up and the resin runs to the path of least resistance, the wide part of the crack, and blows past the tight sections without filling them, or it pops the paste-over and you are chasing leaks instead of filling a crack. Slow and steady beats fast and forceful on every crack that matters. There are high-pressure systems and packers for special cases, but the low-pressure port method is the bread and butter for structural epoxy on walls.

The hardware ranges from a hand-held dual-cartridge gun with a static mixer for short cracks to a metered two-component pump for long runs, and the resin has a pot life that the gun or pump has to respect. Mix ratio is not negotiable: an off-ratio epoxy does not cure to strength, so the metering, whether it is the cartridge or the pump, has to deliver the ratio the data sheet calls for.

Curing, knocking the ports off, and grinding the cap

When the crack is full and capped, you let it cure before anyone touches it. Epoxy cure time runs with temperature and the product, faster when it is warm, slower when it is cold, so the data sheet sets the wait, not the clock on the wall. Inject in the cold and the resin can be slow to gel and slow to gain strength, which matters most on a structural repair where the bond is the whole point.

Once it has cured, the ports and the paste-over come off. You knock or cut the surface ports off and grind the cured paste band back flush, leaving the crack filled inside the section and the face cleaned up. On a finished or exposed surface that grind-back and any patching is part of the job; on a hidden foundation face it is cosmetic and sometimes skipped.

Resist the urge to grind too early. A paste cap or a port broken off before the resin inside has cured can pull the seal and leave a void at the face. Let it cure, then clean it up.

How do you know the crack actually filled?

The honest answer to whether the crack filled is that you watched the resin reach each port, and on a critical repair you take a core to prove it. Watching pure resin run from the next port as you inject is the real-time check that the crack is full between ports, which is why low pressure and that visual cue matter so much: if resin never showed at the next port, that segment is not full, and capping and moving on hides a void.

For acceptance on a structural repair, the standard verification is to core across the repaired crack and look at the depth of penetration. A core through the injected crack shows whether the epoxy reached the full depth and bonded both faces, or whether it filled the front of the crack and left the back empty. The project specification sets whether cores are required, how many, and the acceptance, because that depends on the structure and the engineer's tolerance for risk. Where a core comes out and the epoxy is full depth and bonded, the repair did what it was supposed to.

The quality of the work shows up in a short list: the ports spaced right for the thickness, the resin seen at every port, the full-depth fill confirmed, the cure given its time, and then the crack not reopening along the old line. That last one is the real test of a structural injection. A crack that comes back where you injected it means the diagnosis was wrong, the crack was still moving, or the fill was incomplete. The crack that stays closed while the wall takes its next thermal cycle is the one that was done right.

Can you inject a leaking crack?

Yes, and a leaking crack is a polyurethane job, not an epoxy one. Epoxy will not bond a wet face, so injecting epoxy into water-charged concrete gives you resin sitting on a film of water with no bond, which is a non-repair. Polyurethane is built for the wet crack: it reacts with the water and foams to seal it, so the water that defeats epoxy is what makes polyurethane work.

The catch on a heavy leak is controlling the water before and during injection. A crack actively gushing carries the resin out before it can set, so the field move is to inject from the bottom of the leak upward, let the resin react and build a plug against the flow, and work with the water rather than fighting a full head. On a high hydrostatic head you sometimes inject deep packers into drilled ports rather than surface ports, so you are sealing inside the section where the pressure is, and the choice of packer versus surface port goes with the wall and the water.

Stopping the leak does not close the structural question. If the leaking crack is also structural, sealing the water with polyurethane buys you a dry surface but leaves the load-transfer issue unaddressed, and that may need an epoxy repair or another structural method once the water is handled. Sequence it: stop or control the water, then make the structural repair the engineer specifies. Do not let a stopped leak get logged as a structural crack repaired.

The moving or expansion crack: not a job for rigid epoxy

A crack that keeps moving is not a candidate for rigid epoxy, full stop. Epoxy bonds the faces hard and does not yield, so a crack that is still opening and closing either shears the epoxy bond or, more often, just cracks the concrete in a fresh line right beside the repair. You did not stop the movement; you moved the crack over an inch.

When a crack is going to keep moving, you have two honest options. Seal it with a flexible material that can stretch with the movement, which on an injection job means a flexible polyurethane, or on a surface job means routing it out and filling it with a flexible sealant. Or, if the movement is real and ongoing, stop pretending the crack is a defect and treat it as a joint: build a proper movement joint that is detailed to open and close, so the movement has a designed place to happen instead of tearing the concrete.

The mistake is treating an expansion or working crack like a one-time structural crack. A crack at a re-entrant corner, at a slab that curls and moves with temperature, at a long wall with no working joints, those move with the seasons, and rigid epoxy in them is a callback you scheduled yourself. Read whether the crack is dormant or active first, the way the assessment section lays out, and let that decide between the rigid weld and the flexible seal.

Route and seal: the non-structural alternative

Route and seal is the simple answer when the crack is not structural and you only need to keep water and debris out of it. You widen the crack into a clean groove with a saw or a crack chaser, commonly a V or a U section, blow it clean, prime it if the sealant calls for it, and fill the groove with a sealant. It does not restore structural strength and it does not pretend to. It keeps water out and gives a crack that moves a flexible joint to move in.

The reason to route is the same reason you flex-seal a moving crack: the groove and a flexible sealant can stretch as the crack opens and closes, where a rigid fill would tear. It also gives the sealant a real cross-section to bond to and hold in, instead of a knife-thin crack. For a non-structural crack on a wall or a slab where the goal is sealing, not welding, route and seal is faster and cheaper than injection and often the right call.

It is the wrong call when the crack is structural. Routing the surface and sealing it does nothing for load transfer through the section, so a structural crack treated with route and seal is still a structural crack with a tidy surface. Match the method to the job: route and seal for sealing a non-structural or moving crack, injection for restoring a structural one or sealing a deep leak.

Gravity feed for a horizontal slab crack

A crack in a horizontal slab or deck can often be filled without ports or a pump at all, by gravity. You V-notch the crack along its length with a grinder to open a channel, clean it out, then flood a very low-viscosity resin into the notch and let it sink in under its own weight. No pressure, no surface ports, no paste-over: the resin runs down into the crack because it is thin and the crack is below it.

The materials for this are the thinnest ones made, an ultra-low-viscosity epoxy or a high-molecular-weight methacrylate, with viscosities low enough, typically well under 100 centipoise, to penetrate cracks down to around 0.002 in, near 0.05 mm, without being pushed. You pour neat, mixed resin into the V-notch and keep topping it off as it sinks until the crack will not take any more. On a deck you sometimes pre-treat by ponding the resin over the crack and letting it draw in.

Gravity feed only works where gravity helps, which means horizontal and near-horizontal surfaces. On a slope the thin resin runs off before it soaks in, so you dam it or switch methods. It is the go-to for floor slabs, bridge decks, and pier caps where the crack is accessible from the top. ACI publishes field guidance on the methacrylate flood-coat approach, RAP-13, and the resin manufacturer's data sets the viscosity, the cure, and the conditions.

Stitching and other structural alternatives

Injection is one method on a longer menu of crack repairs, and on some cracks it is not the right one. ACI 224.1R lays out the range, and knowing the alternatives keeps you from forcing injection where another method fits better.

Stitching is the one to know. You drill holes on each side of the crack and set U-shaped metal staples, the stitching dogs, across it, anchored in grout or epoxy, so the steel carries the tension across the crack the way a stitch holds a tear. It is used where a crack needs tensile capacity restored across it and injection alone will not do it, often paired with injecting the crack to seal it. Stitching adds strength across the crack; it does not make the section monolithic the way epoxy does, so the two solve different parts of the problem.

Other structural methods show up by case. Drilling and plugging suits a crack that runs in a reasonably straight line you can follow with a drilled key. External prestressing closes and compresses a crack across a member where the geometry allows it. Adding reinforcement and overlaying, or full removal and replacement, takes over when the cracking is too far gone for any fill. Which one fits is the engineer's call on a structural member, and it turns on what the crack is doing to the structure, not on what is easiest to install.

Where this lives: basement walls, slabs, beams, and tanks

The method changes with the structure, even when the crack looks the same. A leaking poured-foundation wall is the classic polyurethane job: water under hydrostatic head, a crack that has to seal and tolerate the seasonal movement of the wall, and a homeowner who wants a dry basement, not a structural analysis. Polyurethane injection from surface ports or deep packers stops the water and flexes with the wall. If that same wall crack is structural, from settlement or bowing, the structural cause gets addressed first and the structural repair follows; sealing the leak is not the structural fix.

A structural beam or column crack is the opposite end. There the question is load transfer, the crack is dormant once the cause is dealt with, conditions are usually dry, and epoxy injection to restore monolithic behavior is the textbook repair, set and verified by the engineer with cores where the spec calls for them. A floor slab crack splits the difference: if it is a non-moving crack in a slab on grade you often gravity-feed it with thin epoxy, and if it is a moving joint masquerading as a crack you flex-seal or convert it to a joint.

Industrial and data center slabs raise the bar on what the fill has to take. A cracked or spalled joint or crack under hard wheel traffic and tight equipment loads gets a rigid, high-strength repair so the arris carries the load instead of raveling under the next pallet jack or server-row cart, and the flatness the operation runs on is part of the acceptance, not an afterthought. Liquid-retaining structures, water tanks, and pits, are their own case: the crack has to seal against pressure from the inside, the material has to suit potable or process exposure, and the spec, not habit, picks the resin. The diagnosis and the technique carry across all of them. What changes is what you are protecting against and what counts as a pass.

Monitoring the repair after it cures

An injection repair is not done the day it cures; it is done when the crack proves it stays closed. The cheap insurance is to leave or set a crack monitor across a repaired structural crack and read it through a temperature cycle, so if the crack is still moving you find out from the gauge instead of from a fresh crack beside the repair.

On a leak repair, the test is the next wet season. A polyurethane water-stop that holds through the rain or the high water table it was sealed against is a repair; one that weeps somewhere new means the water found another path, which is common on foundation walls and is why the drainage and the head outside the wall belong in the conversation. Watch the repaired wall after the first heavy rain, not just on the day you injected it dry.

On a structural repair, the durable signal is simple and binary: did the old crack line reopen. It should not, because the injected line is now the strong part of the section. A reopened crack along the repair is the structure telling you the diagnosis missed something, usually that the crack was still moving or the cause was still active, and that is a reason to go back to the cause, not to inject the same line again.

What to document

Document each crack the way you would defend it later, because if the wall leaks again or the crack reopens, your injection record is what shows the repair was sound. The record is also what tells the next person whether this line was a structural weld or a water seal, which changes what a new crack beside it means.

Capture the crack location and length, the measured width, whether it was active or dormant and how you decided, wet or dry, the material and product used, the port spacing and the pressure, and how the fill was verified. If it was structural, record the engineer's direction and any cores. The table below is the spine of a crack injection record.

Common mistakes

• Injecting rigid epoxy into a crack that is still moving, so the concrete cracks again right beside the repair.
• Filling the crack without finding and addressing the cause, so the active cause reopens it.
• Injecting epoxy into a wet or leaking crack, where it cannot bond the water-filmed face.
• Pushing the pressure too high, so the resin runs past the tight sections or blows the paste-over off.
• Spacing the ports wrong for the wall thickness, leaving the crack unfilled between ports.
• Capping and moving on before resin shows at the next port, sealing a void inside the section.
• Letting the paste-over plug the bottom of the port so the resin cannot enter the crack.
• Running an off-ratio mix from a tired cartridge or unmetered pump, so the epoxy never reaches strength.
• Grinding the ports and cap off before the resin has cured, pulling the seal at the face.
• Treating a polyurethane leak seal as a structural repair on a crack that was both.

Field checklist

Standards and references

The engineer of record governs any crack that affects structural capacity, and the material manufacturer's data sheet governs the product, the mix, and the cure. Everything below is the framework those two work within, and where the project specification is stricter, it wins. The exact document numbers and editions shift over time, so confirm them against the version in force on your project before you cite one.

ACI is the body for the concrete side. ACI 224.1R, Causes, Evaluation, and Repair of Cracks in Concrete Structures, is the document that sorts cracks by cause and lays out the repair methods, including epoxy injection, routing and sealing, stitching, and the structural alternatives. The ACI field guidance, the RAP series, covers the procedures: RAP-1 for structural crack repair by epoxy injection and RAP-13 for the methacrylate flood-coat gravity approach. ACI 546R, the guide to concrete repair, covers the broader method selection. ICRI publishes complementary repair-industry guidance on the techniques.

ASTM gives the material specifications. Epoxy injection resins are commonly specified to ASTM C881, the specification for epoxy-resin bonding systems for concrete, with the type and grade selected for injection and for the crack. Polyurethane injection resins for leak sealing follow their own product specifications and water-reaction characteristics, so do not carry the epoxy spec over to the polyurethane. Name the standard that controls the point, let the engineer set the structural call, and let the data sheet set the material.

Units, terms, and conversions

Crack injection borrows from materials science, structural work, and waterproofing, so the same idea reads differently across a spec, a product data sheet, and a condition report. Keep the units straight between them.

Crack widths run in thousandths of an inch and in millimeters, where 0.002 in is about 0.05 mm and a 1/2 in crack is about 13 mm. Resin viscosity is in centipoise, cps. Injection pressure is in psi or bar. Bond and compressive strengths read in psi or MPa. The terms below are the ones that travel across the whole job.

Crack injection

Filling a crack from inside the section under pressure or by gravity, to bond it or seal it

Epoxy injection

Rigid, structural injection that welds a dormant crack back to monolithic strength

Polyurethane injection

Flexible, water-reactive injection that foams and seals an active or leaking crack

Dormant vs active crack

A crack that has stopped moving versus one still opening and closing, which decides rigid versus flexible

Surface port

A fitting glued over the crack to inject resin, spaced roughly to the wall thickness

Paste-over / cap seal

The epoxy band troweled over the crack face so injection builds pressure instead of leaking out

Gravity feed

Filling a horizontal crack with a very low-viscosity resin that sinks in under its own weight

Route and seal

Widening a non-structural crack into a groove and filling it with a flexible sealant

EOR

Engineer of record, who controls any crack repair that affects structural capacity

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• Why does spalled concrete happen?
• What is the minimum concrete cover for rebar?
• Why does concrete crack?
• How do you install an adhesive anchor?
• What is a concrete admixture?