Look familiar? Anytime tile meets a dissimilar surface, in this case a plastic trench drain, a sealant (caulk) should be used to accommodate the differential movement that will occur. Cement grout will not provide lasting results and restricts the movement needs of the tile installation.


Writers for TILE Magazine receive the editorial calendar a year in advance to allow judicious planning for things such as research, interviews, evaluations, writing and pictures - each article must have pictures! Being summertime, distractions from our editorial obligations seemingly increase with perfect weather conditions. This coupled with a little natural procrastination by yours truly found me receiving emails from my editor followed by phone calls that this month’s issue was ready for layout and they needed my article. I had been traveling a lot recently so when I found myself at home with some time off, I was hoping a rainy day would keep me indoors so that I could start this month’s issue. The rainy day never materialized.

There I was, three days before publication and I hadn’t even looked at what topic I was going to be writing about (which means I had no pictures, had done no research, and barely had enough time to type 1,500 words: I am in trouble now!). I opened my file to see this month’s editorial subject: Expansion Joints! I was saved!

It’s a rare day when I don’t receive an email that has something to do with the lack of them. In the past 10 years, I had written 11 articles and made six presentations on the topic. I even had a 652MB folder of pictures of failed jobs on the subject alone. To top it off, I am currently monitoring the installation of expansion joints in a swimming pool after the first tiled pool failed, for among other reasons, a lack of expansion joints. As part of the tile replacement, sealant is to be installed every 12’6” in a 200-ft.-by-50-ft. pool. Once that is completed, the contractor will start replacing the tile on the adjoining 3,500 sq. ft. pool deck that has become unbonded after three years of use with nary a movement joint in sight. In case you’re curious, the total replacement cost under warranty: $342,500.

Failures due to lack of movement accommodation joints abound. With each failure, fingers point in all directions as retailers, contractors, and installers hope to find a pocket to pick other than their own. In the case above, a reputable union contractor, with whom I actually used to compete against years ago, installed the project. There was not a single expansion joint in the pool, deck, or locker room. In conversation, he like most had his share of cracked grout, loose tile, and a few tented tile jobs but never a complete failure. He had his installers put in joints when specified but if nobody asked, he would avoid the time, cost, and typical owner/end user resistance to use sealant, caulk joints or pre-made profiles whenever possible. This is the same story we hear on many occasions when problems occur. The late tile industry icon Bob Young once said most bond failures are held in place by the “three G’s: God, grout, and gravity.”

Nonetheless, these installations have lost their bond and are on their way to replacement at some point (perhaps with a nice laminate or carpet), but in many instances, never again with ceramic tile. I have said this before and I will say it again: If you do not install the appropriate expansion/movement accommodation joints in a tile installation, then the question is when it will fail and not if it will fail.

So just what necessitates this overwhelming need for movement (expansion) joints? It starts with the tile itself. Many people are quite surprised to hear that ceramic tile, a product manufactured by firing clay and other minerals at several thousand degrees, expands when exposed to direct sunlight. Porcelain tile, a dense-bodied product with a low thermal expansion rate, expands approximately .000004 inches per degree Fahrenheit. It may sounds minimal, but that same porcelain tile covering 40 linear feet exposed to a 50-degree change in surface temperature has the potential to expand approximately 1/16 to 3/32”. The adhesion of the bonding material to the tile and substrate reduces the actual amount of expansion but does not eliminate it. Types of tile other than porcelain, especially glass, are much more expansive under heat.

Thermal expansion is a two-way street: tile both expands and contracts due to thermal variation. This is one reason for my personal preference of using the term movement joints. If I were only concerned about moisture, then I could use an expansion joint. Why? Moisture is an expansive force on ceramic tile. Once it expands, it does not contract. Yes, it’s true, as my swimming pool friend above found out; tile expands when exposed to moisture. The growth rate is slow and minimal. Except for this minor amount of growth, ceramic tile is unaffected by moisture or water - it is an inert material.

When someone says they have a tile failure due to excessive moisture, what they really should be saying is that the failure was due to defective workmanship. Either they used the wrong thinset, did not allow it to fully cure prior to moisture exposure, or, there were no expansion joints and likely a few other ills. Failures due to moisture expansion can happen in as little as a few years in areas of high water exposure applications, such as an exterior deck or restaurant kitchen. However, most often they take a number of years to occur. Coastal areas such as Florida and New Orleans, which have high ground water tables, show a disproportionate amount of slab on grade moisture-related failures, most commonly due to a lack of movement accommodation joints with a good solid smattering of people still using roofing felt and paper fully adhered by multipurpose adhesive for crack suppression that didn’t hold. Papers have no tolerance for moisture, nor does multi-purpose adhesive.

For movement joints to perform their necessary task in tile installations they must have the proper width and depth. This conscientious contractor elected to use a saw to provide a clean unobstructed joint for treatment. All movement joints must extend through the setting bed without obstruction.

Less we digress, what are some other good reasons for movement/expansion joints? Wood structures. If it’s made of wood, it’s going to move. Wooden-supported floor systems all have a certain amount of deflection. To accommodate normal deflection of the floor structure, perimeter joints are critical in wood construction. If there is a beam running down the center of the room or home, then common sense tells us that is a pivot point from which the floor will deform on either side, like a board over a fulcrum. It would be prudent to put a soft joint or movement profile over that area to allow the floor to move without putting the tile assembly in stress.

Other areas of consideration are doorways. Take the example of a kitchen with an adjoining dining room. If we have a 300-sq.-ft. kitchen with a doorway to a 250-sq.-ft. dining room, both installed over backer board, then we have roughly 1,800 lbs. on one side of the wall and 1,500 lbs. on the other side, joined together at a three-linear-foot doorway, like a bar bell. Again, common sense would say this is not a good idea and real world experience will confirm that you need a movement joint.

We also have the moisture of the wood floor system to consider. Most areas of the country go through seasonal changes in humidity levels. These changes affect the dimensional stability of the supporting floor structure. This moisture-induced movement is magnified several times if the method of construction utilizes a crawl space, which subjects the supporting structure to a greater level of dimensional instability.

Concrete does not earn a free pass on stability either. Did you know that concrete spends its whole life moving? Concrete has a natural tendency to warp. Control joints are placed primarily for shrinkage during the initial curing process. Because a concrete control joint for tile is not cracked at the time of tile installation does not mean it will not crack at some point. You can be fairly well-assured it will. Movement/expansion joints must be provided at all control joints. As the concrete continues to fully cure, which can take as long as a year, some additional shrinkage will occur. Even after fully cured, there will be some minor warpage. If tile is installed without proper movement accommodation joints, then the tile may either crack or debond. Using a membrane does not eliminate the need for movement joints. For a membrane to function properly, it must be placed within the installation. Some products and methods will allow you to select an atheistically pleasing location. The physics of concrete are complicated: I have been studying cement and concrete for a number of years and still struggle with the complexities of its movement. Recently, I redoubled my efforts to better understand the dynamics due to some emerging LEEDS practices relative to changing formulations.

Bottom line: all building materials move, none at the same rate. All structures move by necessity to avoid buckling which could cause structural failure. Architects, general contractors, owners and end users are adverse to seeing caulk joints or movement profiles they feel destroys the ambiance of their building. I am from the real world of installation and understand the challenge. Awareness of the need for movement joints in tile installations needs to begin at the sale or specification stage, well before ordering products and certainly before the installer arrives on site. Asking an end user how they prefer these joints to be treated should be no different than, and is of much greater importance, then selecting grout color.

As an investment, measures should be taken to protect tile flooring. Why not take a look at your current policies regarding expansion/movement accommodation joints and see what you can do to better protect your company and those who invest in your products?