Crack Prevention in Tile Work
March 11, 2012
The title of this month’s subject was supposed to be Crack Suppression for Tile Installations, something I have written about numerous times over the years. One would think when you write about the same subject matter with limited topics (thinset, grout, membranes, ect.) for 15 years you would run out of ways to write about it.
If I were solely installing or even still teaching the installation process that may be true. But, doing inspections and consulting, new examples of American ingenuity present themselves every day. Each of the subjects addressed in this article are projects I have been on or questions received within the past 30 days.
The engineering prowess of the floor covering installer never ceases to amaze me. No matter whether it works or not, some are quite ingenious in their design. Unfortunately it is rare that someone calls me to profess of their successful endeavors, salesmen and aspiring manufacturers excepted. More likely the installer or customer is lamenting on yet another fine idea gone bad and doesn’t understand why. So, for this article, instead of looking at reasons we should buy membranes, many as they may be, let’s look at ways to avoid cracks in the first place, including those caused by membranes.
Thinset bonding materials and membranes have somewhat of a love/hate relationship. All cement-based thinset mortars shrink, the question is only how much. Bonding directly to absorbent substrates such as concrete and wood does have some benefits. Chief among them is that they are solid, rigid, and absorbent surfaces to which a high degree of mechanical bond can be achieved. Because the surfaces are absorbent, initial cure times are typically shorter than they will be for a membrane. Bonding directly to a porous substrate can also achieve a much higher bond strength than that which can be accomplished with a relatively non-porous membrane. Tile bonded to non-absorbent surfaces tends to have lower bond strengths.
While many membrane products have the ability exceed the 50 PSI, shear bond required under standards it will be reduced from the published shear values for direct bond applications to regular substrates. However, lower shear bond is somewhat of a tradeoff and not necessarily a negative, because when using a crack suppression product, the bonding surface is now deformable. Given that the installation now has the ability to absorb limited amounts of movement, high shear bonds are no longer required to hold the tile in place when the installation is subjected to stress. Keep in mind that only one side of this sandwich is deformable. You still need a thinset that will remain bonded to the much more brittle tile, which is less forgiving when stresses do occur.
Another important consideration when using a membrane is thinset curing and shrinkage. The tile setter’s 24-hour period, typically the next morning, can be a very bad idea when using membranes. With the low water absorbtion of both the tile and membrane, the water in the setting material isn’t going anywhere very fast. Low temperatures of either the air or the surface may make it even slower. If you need to get back on the floor the next day, consider using a suitable rapid setting mortar. Standard thinset mortars are designed for a thickness of ¼” to 3/8” maximum after tile placement. Thicker applications require a medium bed mortar. Both need to be properly mixed and not be overwatered.
This is a conversation I have a lot. The typical call or job visit for this problem starts with, “I’ve been doing it the same way for 20 (or more) years and never had a problem,” as we look at the cracked wall or floor tile. Upon pulling the tile up you find small fractures in the thinset, indicating shrinkage cracks from either excessive water or thickness application. When installing over deformable membranes, in lieu of directly bonding to the substrate, the membrane deforms as the thinset dries. During the curing process the normal shrinkage that occurs pulls at the deformable membrane, which can cause cracking in soft tile, stone, and glass and bond loss in denser and stronger tile. When thinset is applied too thick or wet, excessive shrinkage occurs.
Fully supported tile is very important when using a membrane system. The elasticity in membranes to accommodate crack resistance also means it is compressive to a certain degree. Due to many different product variations and their ability to support tile, not all membranes are suitable for all applications. Recently on one very large high-traffic commercial shopping center, I was asked to review the tile installation. After extensive testing, the findings supported the manufacturer’s recommendation as published in their literature, that it was suitable for light commercial applications, not heavy traffic. This is NOT uncommon. On other occasions I have been called to a job where the right products were used but the tile and grout was cracking. In such instances we typically find the tile was not fully bedded in the thinset to equally distribute the load placed on the tile. While problematic in any job, not fully supporting a tile when a membrane is used carries a very high risk of an expensive failure.
One of the most popular causes of cracks is the lack of use or improper treatment of control joints. This issue alone has to be responsible for more than half of my phone traffic and job reviews. The industry position on control joints is that they must be carried through the tile work, and it has always been that way. The most recent revisions to industry documents expanded on that point, trying to add increased clarity. Thus, if you chose to relocate a control joint, you must do so using the manufacturer’s proprietary recommendations. Most popular liquid membrane systems caution against their use for control joint relocation, yet they are consistently used for that application. Most, not all, membranes require a three-tile width when they are approved for that purpose, not the 12” strip commonly found with either liquids or sheets.
Just as typical for both contractors and architects alike, because they did use a membrane in either the correct or incorrect width, they feel a soft joint (sealant) at the nearest grout joint is unnecessary. The discovery of inappropriate, insufficient product application and/or width usually leads to the, “I’ve been doing it 20 years” conversation we had previously in this article. Here is an easy way to look at this. Crack isolation/suppression membranes basically serve as shock absorbers. They transfer stress from the crack to the nearest soft joint. Unless they have the ability to move, they are basically useless. They are not designed to be energy storage devices. Granted, by their abilities to deform, they do offer a very limited amount of storage. The demonstration I use to make my point is actual vehicle shock absorbers. If you weld them solidly, it is going to be a rough ride. Give them some travel and it is a smooth ride. They are transfer devices, not storage devices.
As usual, I am just getting warmed up and out of space already for this issue. Today we have many options available for all types of service applications when it comes to crack suppression. We need to make those choices wisely based on both technical merits and pricing when we do tile work. To say the current market is competitive is a gross understatement. However, based on what I have been seeing recently, it is better to stay home poor than go out and pay for a job twice. Work smart, not desperate.