Crack Injection and Patch Repair Technique

Why this matters

A crack repair that looks finished on handover and reopens by spring is worse than no repair at all, because now the customer has paid for a fix that failed on top of the original defect. The gap between those two outcomes is almost never the resin choice. It is whether the resin actually filled the crack from top to bottom or only bridged the visible face, and whether a patch was bonded to sound material or feathered onto a dusty edge that was always going to let go. Concrete Repair Methods in this library tells you which damage category calls for injection versus a surface patch, and the crack severity trees tell you whether a given crack is dormant enough to inject at all. Neither one tells you how to actually run the port, the pump, or the trowel, which is where a repair earns or loses its warranty period.

Confirm dormancy and dryness before you pick a resin

Injection assumes the crack is not actively moving during the work and, for most structural epoxies, that the crack face is dry. Both assumptions fail silently. Pull the job's crack monitor readings if one is already running per the hairline crack condition tree, or take a baseline width reading at three points with a comparator card the day you plan to inject; a crack still cycling with temperature will pump full of epoxy and reopen along a new plane the next cold snap, because rigid epoxy does not stretch. With no monitoring history and no time to wait out a cycle, treat the crack as active by default and route to polyurethane, since a flexible fill tolerates the movement an audit later proves was there.

Dryness is a separate check. Press a paper towel into the crack mouth at the widest point and watch it for a minute. Damp or wicking moisture defeats a standard structural epoxy's bond, since its low-viscosity resin rides on top of trapped water instead of displacing it and leaves voids you cannot see from the surface. A crack that stays damp, or one that is actively conveying water, is a polyurethane job regardless of how dormant it otherwise reads.

Choosing epoxy or polyurethane, and why the choice is not interchangeable

Two-part structural epoxy is a rigid adhesive. It restores some load transfer across the crack face and is the only choice when the repair's job is structural continuity, not just keeping water out. It needs a dry substrate, it does not tolerate ongoing crack movement, and viscosity has to match the crack: low-viscosity epoxy for cracks at or under about 1/8 inch, where capillary action does most of the filling work under modest pump pressure, and a gel or paste epoxy for anything wider, where a thin resin would simply run out the back of an unbacked crack before it sets.

Polyurethane is flexible once cured and, in its water-reactive foaming formulations, expands on contact with moisture, which is exactly what makes it the right call on an active crack or one carrying water. It does not restore structural continuity the way epoxy does, so a crack needing both a strength restoration and a water stop is a case for the engineer, not a field call between two resins, since a repair sold as structural when it is really a flexible seal is a warranty claim waiting to happen. Read both data sheets before the truck leaves the shop; pot life, viscosity grade, and moisture tolerance vary enough between manufacturers that "epoxy" and "polyurethane" are categories, not specifications.

Setting injection ports

Surface-seal the crack face between where the ports will go with a fast-setting paste before any port is installed, so pumped resin cannot blow back out through the visible crack line instead of traveling down into the section. Set the ports themselves at a spacing tied to the thickness of the member you are injecting, which is the guidance both ACI 224.1R and ICRI's repair materials guideline build on: on a thick wall the resin has to travel a long lateral path to reach the far face, so ports go in wider apart; on a shallow residential slab, four inches of thickness gives the resin almost nowhere to hide, so in practice ports on flatwork run closer together than the thickness rule alone would suggest, commonly every six to eight inches, because the labor cost of an extra port is trivial next to the cost of a section that never got filled. Seat each port over the crack, not beside it, and let the surface seal cure to the product's stated time before pumping. Wrong looks like ports crowded at the wide, visible part of a crack and skipped where it narrows toward the ends, which leaves exactly the section that is hardest to fill unfilled and easiest to miss because nobody put a port there to prove it.

Pumping the injection

Start at the lowest port on a vertical crack, or at one end on flatwork, and pump at low pressure until resin appears at the next port over rather than by a calculated volume, since the crack's internal geometry, not a formula, decides how much resin it holds. Cap the port you just filled the moment resin shows at its neighbor, move the pump to that neighbor, and repeat down the line. This breakthrough method is the actual verification that the crack is filling continuously rather than pooling at the first port and leaving the rest of the run dry; a volume estimate can look correct on paper while half the crack stayed empty. Keep the gun's hose and coupler rated for the resin's working pressure, check the coupler seat before every port, and never sight down a port while it is under pressure, since a blown fitting sends resin and a fair amount of force back at whoever is looking at it. Watch the product's stated pot life against how many ports remain; a batch that starts to gel in the line stops flowing well before it stops being sticky, and forcing a gelling batch through the pump is how a coupler blows.

Patch repair for chips, corner breaks, and small spalls

Where the damage is a missing chunk rather than a crack, injection is the wrong tool; Concrete Repair Methods in this library covers which patch material fits which damage category, and this is the placement technique underneath that choice. Saw-cut a clean perimeter to sound concrete, going a little past where the eye first sees sound material, since concrete degraded enough to spall is usually degraded a bit past its visible edge too. Never feather a patch to a knife edge; the thin margin is exactly where a patch debonds first under a tire or a boot, so undercut the perimeter instead, narrower at the surface and wider at depth, which locks the patch in mechanically as well as by bond. Chip out loose material with a hand tool rather than a heavy chipping hammer, since an aggressive tool telegraphs a shock crack past the boundary you just cut. Confirm the substrate condition the patch material's data sheet specifies, dry for most epoxy and polymer mortars, saturated surface dry for straight cementitious ones, apply a bonding agent where called for, then place the patch to the manufacturer's stated thickness, consolidating it against the substrate rather than dropping it in loose.

Matching color and texture

Pigment blended into the patch mortar gets close on a fresh patch, but every patch reads slightly different from the surrounding slab for the first several months regardless of pigment, because the surrounding concrete has already weathered and the patch has not; set that expectation before you leave, not after the customer calls. For texture, work the same finishing tool the original surface used, broom or trowel, in the same direction, and match the timing of the pass to how the patch material is setting rather than to the clock, since patch mortars often set faster than the field mix that surrounds them.

A worked example: injecting a dormant slab crack at a garage door threshold

A 4 inch attached garage slab has a diagonal crack running about 6 feet from a re-entrant corner at the man-door pocket out to the vehicle threshold, where the slab edge is visible and confirms the crack runs full depth rather than sitting only at the surface. The customer's monitor readings, kept per the hairline crack condition tree over the prior 90 days, show no measurable width change and the crack faces are flush, so this is a dormant, structural-depth crack in a dry interior slab. The paper towel test at the widest point, about 3/32 inch, comes back dry.

Dormant, dry, and needing depth-fill rather than a shallow route-and-seal calls for low-viscosity structural epoxy. Surface-seal goes on the full 6 foot length and cures per its data sheet. Ports go in at 6 inch centers along the 72 inch run, 13 total counting one at each end, tighter than the member-thickness guideline strictly requires but justified on a slab this shallow, where the resin has almost no lateral distance to travel and an extra port costs a few minutes.

Pumping starts at the threshold-end port and moves down the line, capping each port the moment resin shows at its neighbor and shifting the pump there. Port 9, near the 48 inch mark by the man-door corner, takes noticeably more volume before port 10 shows anything, consistent with the crack widening slightly through that stretch, matching the field width reading taken there earlier. All 13 ports show breakthrough by the far end, and the last port is capped once resin appears at the slab edge itself, confirming the fill reached the point where the crack was first identified as full-depth.

Two days later, once cure time has passed, the port stubs are ground flush, the surface seal is removed, and the injected line is checked for hardness at three points with a thumbnail press, solid at all three. The ground port locations get a thin cementitious skim matched to the surrounding broom finish.

Verifying the repair

Breakthrough at every port during pumping is the primary evidence the fill is continuous; a repair with a port that never showed breakthrough is not confirmed and needs a second pass at that segment before anyone calls the job done. After cure, a hardness check along the injected line and a visual pass for any resin sheen indicating a skipped section are the field verification available without destructive testing; coring to confirm fill on a residential repair is rarely proportionate, and is reserved for cases where the crack was carrying real structural load and the engineer wants direct evidence. For a patch, verification is a sounding check, tapping the patch and its perimeter with a hammer or a length of chain and listening for the hollow return that means the patch has not bonded, done a few days after placement rather than immediately, since a fresh patch can sound solid before full cure and still debond within the week.

References

• ACI 224.1R, Causes, Evaluation, and Repair of Cracks in Concrete Structures, for injection resin selection by crack width and movement.
• ICRI Guideline 210.1R, Guide for the Evaluation of Materials for the Repair of Concrete, for port spacing tied to member thickness.
• 29 CFR 1926.1153, respirable crystalline silica in construction, for the water-fed or shroud-and-collection control when saw-cutting a patch perimeter or grinding a port flush.
• The resin and patch material manufacturer's data sheet and SDS for pot life, moisture tolerance, cure schedule, and the respiratory and skin protection specific to that product.
• See related: Concrete Repair Methods for damage categorization, the hairline crack and crack width and pattern condition trees for the repair-now-versus-monitor decision, and the silica dust control SOP.