Crack and Joint Fillers: Rigid, Semi-Rigid and Flexible

Three families, three jobs

The products are usually sold as if they sat on one scale from hard to soft, with semi-rigid as a sensible middle. They do not. Each family does one job well and the other two badly, and the job is set by the gap you are filling, not by preference.

Which joint in a slab takes which product is set out on the concrete floor joints page, and whether a crack is still moving is the test on the crack telegraphing page. This page is about the materials themselves.

What “semi-rigid” actually means

The word is used everywhere and defined almost nowhere, partly because there is nothing official to point to. Metzger/McGuire’s MM-80 data sheet says plainly that there are no government or ASTM standards for semi-rigid joint fillers, and describes its own product as hard but slightly resilient. That is the working definition: stiff enough to support loads crossing the joint and protect the edges from spalling, resilient enough not to be brittle.

Stiffness is given as Shore A hardness. Euclid’s comparison guide explains the scales: Shore A is used for more rubbery materials, Shore D for harder ones. Both Euclid and Metzger/McGuire state that ACI and PCA call for a minimum of A80 for joints under vehicle traffic. We have not read that figure in ACI’s or PCA’s own text, so treat it as the industry’s stated floor rather than a quotation. The products we read sit above it:

• MM-80 (epoxy): A90 to 95 on its current data sheet. Its older guide specification asks for 85 or greater
• Euco QWIKjoint UVR (polyurea): A84 to 88
• Spal-Pro RS 88 (polyurea): A86 to 90
• Sika Loadflex-524 EZ (polyurea): A80 to 85

Metzger/McGuire says it raised its fillers to A90 to 95 in response to the trend in material handling equipment toward smaller diameter, harder wheels. A harder wheel puts its load onto a narrower strip of joint edge, which is exactly what the filler is there to carry.

What semi-rigid does not mean is “flexible enough to move with the joint.” Euclid’s guide puts numbers on it: a polyurea joint filler with a tensile elongation of 400% can expand only about 5 to 15% sideways before it splits or loses its bond to the concrete, and an epoxy filler about 5 to 8%. Its conclusion is that the high elongation polyurea offers very little expansion advantage over an epoxy. ACI 224.1R makes the same point about cracks: a low modulus, flexible adhesive in a crack will not allow significant movement, because a thin layer held between two faces of concrete behaves about as stiffly as a rigid one. So a semi-rigid filler in a joint that moves will separate, and a softer semi-rigid filler will not change that. It will only support the edges less well.

Epoxy or polyurea joint filler

Both are sold as semi-rigid, both fill the same joints, and the hardness ranges overlap. The difference is how they go in. Metzger/McGuire, which makes both, and Euclid, which also makes both, describe it the same way on most points and disagree on one.

The disagreement is moisture. Metzger/McGuire says polyureas may foam or bubble in damp or wet joints and recommends epoxy where joints are likely to be damp. Euclid says its UVR does not bubble or foam in damp joints and is less moisture sensitive than standard polyureas. Both can be true, because the second is a claim about one product. If the joints will not be dry, choose a product whose own sheet says it tolerates that.

Euclid adds one more practical difference: as a joint opens, epoxies tend to separate in a leap-frog manner, from one side and then the other, while most polyureas separate continuously along one side. Neither is a failure. Both are the filler doing what the joints page says it will.

Rigid fillers: dormant cracks only

A rigid epoxy does something neither of the other families can: it bonds the two sides of a crack back together. ACI 224.1R says cracks as narrow as 0.002 inches can be bonded by epoxy injection, and its repair bulletin, ACI RAP-1, says injection can restore structural integrity. On a floor, RAP-1 notes that horizontal cracks of sufficient width can be filled by gravity fed epoxy instead, poured and left to soak in. The sheet for one such product, Sikadur-55 SLV, limits it to non-dynamic cracks.

The catch is the same in every source. ACI 224.1R says that unless the cause of the cracking has been corrected, it will probably recur near the original crack, and RAP-1 says that if the crack is subject to subsequent movement, an epoxy repair may not be applicable. RAP-1 also rules out cracks caused by corroding reinforcing steel, because the corrosion goes on and opens new ones. Where the cause cannot be removed, ACI’s answer is to stop treating it as a crack: rout and seal it as a joint, or build a joint that takes the movement.

So rigid filler comes after the diagnosis, not instead of it. What a crack’s pattern and timing say about its future is on the crack causes page, and the routing and fill steps for a dormant crack under a coating are on the crack telegraphing page.

Flexible sealants: the joint that moves

Elastomeric joint sealants in North America are specified to ASTM C920, which covers single and multi component, cold applied sealants for buildings, plazas and decks for vehicular or pedestrian use. A sealant’s grade under C920 includes a movement class. Manufacturers’ summaries of the standard give Class 25 as able to take 25% movement in expansion and in compression, with Classes 35, 50 and 100/50 above it, and a use code of T for traffic or NT for non traffic. Read T against the standard’s own scope, which is plazas and decks that people and vehicles cross. It is not a rating for a hard forklift wheel on a joint edge, and neither sealant sheet we read claims one.

A sealant only delivers its movement class if the joint is built to let it. Three details decide that:

• No bond to the bottom of the joint. A sealant stuck to both faces and the floor of the joint cannot stretch; it tears. ACI 224.1R describes a bond breaker at the bottom of the groove so the sealant can change shape without a concentration of stress there, and Sika’s backer rod sheet calls the problem three-sided adhesion. A closed cell backer rod, sized slightly larger than the joint (Sika says 1/8 inch larger), does that job and sets the sealant depth at the same time
• Shape. Sikaflex-2c SL’s sheet gives a 2:1 width to depth ratio as the proper design, with at least 1/2 inch of sealant in horizontal joints subject to traffic. Vulkem 45SSL’s sheet sizes the joint itself at four times the movement expected, and not less than 1/4 inch wide
• Nothing rigid on top. The same sheet warns that rigid paints, coatings or primers will crack when placed over a sealant that is moving. A sealed joint is finished after the coating, not under it

Filling a joint so the filler can do its job

A semi-rigid filler carries wheels through its depth, so the filler manufacturers are consistent about filling the joint, not just its top.

• Full depth. MM-80’s sheet says full joint depth in saw cut joints, or at least 2 inches where the joint is deeper, and no compressible backer rod in saw cut joints under 2 inches deep. Euco 700 says to avoid backer rod, sand or any other fill used to save material. Metzger/McGuire’s bulletin says a filler sitting over a compressible backer rod loses its ability to carry load and protect the edges
• Sand only where the sheet allows it. Euclid’s UVR sheet permits dried silica sand in the bottom of a crack, and MM-80’s lets the installer choke off the shrinkage crack at the base of a joint with up to 1/4 inch of clean, dry silica sand. Euco 700 does not. Follow the sheet for the product in the joint
• Overfill, then shave. MM-80’s sheet says not to fill flush and leave it, because the filler settles low. Fill proud, let it cure, and razor or grind it flush. Metzger/McGuire’s bulletin says the finished profile has a direct effect on how well the edges are protected

When to fill a new slab, and why a filler that separates later has not failed, are on the joints page.

Coating over a filled joint

The filler manufacturers do not agree, so the coating manufacturer decides. What the sources say:

• Metzger/McGuire, 1996: a technical note on filling joints in coated floors recommends applying film forming coatings after the filler has cured for seven days, either stopping at the joint or running over it, with the filler surface abraded first. It warns that the joint will keep widening for a year or more, so touch-up over the filler may be needed
• Metzger/McGuire, current: the MM-80 data sheet lists joints under seamless floor coverings among the uses it is not designed for, in most settings
• Sika: Loadflex-524 EZ is not recommended as a joint filler under polymer flooring
• Euclid: coatings may not adhere to its UVR polyurea, and floors that will receive a seamless epoxy or urethane coating are referred to Euclid’s own coatings guidance

In practice that means two things. Use the joint filler your coating manufacturer names in its joint details, because it is the combination they have tested. And expect a filled joint under a coating to show eventually: the joint underneath keeps opening, and a film laid over a separating filler goes with it. Leaving the joint visible or saw cutting it back through the coating, both covered on the joints page, avoids that argument.

One that is not a floor filler: polyurethane foam

Polyurethane injection foams turn up in the same aisle and are a different tool. Sika’s SikaFix HH+, a typical hydrophobic grout, is described as designed to stop water infiltration and fill voids, and expands up to 30 times its liquid volume. That is the right product for a leaking crack in a wall or a below grade structure, and ACI RAP-1 notes that polyurethane grouts can suit cracks that do not affect structural integrity. A foam has nothing like the hardness to carry a wheel, so it has no place in the top of a joint in a floor that is going to be coated.

Which one goes where

• Dormant crack, cause dealt with: rigid epoxy, injected or gravity fed, then the coating
• Sawn contraction or construction joint with hard wheels over it: semi-rigid epoxy or polyurea, full depth, shaved flush
• Isolation joint, joint under racks, or any joint that moves and carries no hard wheels: flexible sealant over a backer rod
• Crack that is still moving: none of the above holds it shut. Honor it as a joint with a sealant, or re-cut it through the coating
• Leaking crack below grade: polyurethane injection foam, not a floor filler at all

Quantities come from the gap, not the product: the crack repair calculator sizes the fill from length, width and depth, the linear foot calculator totals the joint runs, and estimating crack filler volume covers where to measure and how much to allow. For the warehouse case, where the joints usually fail before the coating does, see the warehouse flooring page.