Jim was right, I was wrong.

This web site has a 10,000 character per post limit, so I might have to break this into two posts.

Jim: You suggested I investigate and I found out you're correct. ALL fired clay products gradually expand, and that expansion starts before they're even taking out of the kiln. And, it's not just true of ceramic tiles, it happens in every fired clay product, even your dining room dishes. It's just that it's a very small expansion that occurs over a period of many years and normally only causes problems with building materials like bricks and ceramic tiles.

Since my disagreement with Jim the other day on whether or not ceramic tiles "swell" due to temperature and humidity or not, I've been spending my spare time snooping around web sites on the internet that deal with something called "moisture expansion" in ceramic tiles. There's a lot of info on the phenomenon, so I decided to do some investigating to learn more about it.

Moisture expansion is something that happens to ALL products made from fired clay; from terracotta to ceramic tiles to bricks to "whiteware products" (like dishes and teapots) and presumably even plumbing fixtures like toilet tanks, bowls and bathroom sinks.

Whenever any clay product is fired to harden it, the moisture is driven out of the clay. But, as soon as the fired clay product starts to cool (while it's still inside the kiln) the fired clay begins to expand. This expansion is called "moisture expansion" and is attributed to the uptake of moisture from the air back into the clay. Unlike thermal expansion due to temperature changes, moisture expansion is permanent. The rate at which the clay will absorb moisture from the air is dependant on a lot of variables, including the firing temperature and the nature of the firing cycle. How much the clay expands depends largely on the temperature and humidity it is exposed to in service, the composition of the clay and the moisture absorbtion of the clay, with clays of lower water absorbtion generally exhibiting lower moisture expansion. That moisture expansion of the fired clay is also associated with a proportional increase in it's mass. It's well accepted that the expansion is due to "rehydration" of the clay as it absorbs water from the air, but why the firing temperature or the nature of the firing cycle affect the subsequent moisture expansion of the clay is not understood.

The practical problem caused by moisture expansion is that causes compressive stresses in ceramic tiled floors and walls as the tiles expand. Consequently, the amount of room needed to accomodate that expansion is of importance.

On this "Google Answers" web page:

Google Answers: Moisture expansion of ceramic tiles

the question was asked about the amount and rate of moisture expansion in terracotta ceramic tiles. The respondent quoted from a research paper by Richard Bowman entitled "The Importance of the Kinetics of Moisture Expansion" to explained why no one single chart of moisture expansion in ceramic tiles was possible. In that paper, Richard Bowman says that ceramic tiles fired at lower temperatures generally have a very rapid expansion at first, but then subsequently drop to very low expansion rates. Tiles fired at higher temperatures have lower initial expansion rates, but the rate of moisture expansion diminishes much more slowly over a much longer period of time. Consequently, the moisture expansion with time graph will vary with the firing temperature. Clay composition also affects the way the tiles expand, but research papers that deal with the most technical aspects of moisture expansion like that are only available to purchase online. Much of what I found out about moisture expansion was obtained by reading the abstracts of papers, which are free.

That paper by Richard Bowman also states that in the 1960's it was observed that the amount of moisture expansion of ceramic tiles seemed to follow a linear relationship with the log of time, so that you'd get about the same incremental amount of moisture expansion between the first and 10th day after firing as you would between the 10th and 100th day, and again between the 100th and 1000th day, etc.

However, in this paper: (only the abstract of which is available free)

Moisture expansion and mass gain in fired clay ceramics

it was found that the moisture expansion of fired clay materials occurs in a two stage process; one very short period of rapid expansion and weight gain lasting less than two hours after the clay starts to cool inside the kiln, followed by a very much longer period of slower expansion and weight gain. During both periods, the amount of expansion and weight gain follows another simple time relationship; the amount of expansion (and weight gain) is proportional to the fourth root of time. So, according to this relationship, you'd see the same amount of moisture expansion during the first and 16th day after firing as you would between the 16th and 81st day, and that would be the same as the incremental expansion between the 81st and 256th day, etc. (1^4=1, 2^4=16, 3^4=81, 4^4=256, 5^4=625, etc.) And, the authors note: "The strain (expansion) is found to show the same linear relationship with mass throught both stages providing evidence that the underlying physical process is the same in each stage." He's saying that there's good reason to believe it's the same physical and chemical changes at work in both stages because the expansion to weight gain ratio is the same in both stages.

Reading information written by tiling professionals about moisture expansion leads one to believe that problems associated with it occur mostly during the first dozen years after installation, and mostly during the first half dozen. One Q&A forum post by Dave Gobis (who is well respected in the ceramic tiling community) was surprised that moisture expansion would still be at work to cause a ceramic tile floor failure 18 years after installation. However, that Journal of Physics paper cited immediately above says that moisture expansion continues at progressively slower and slower rates for centuries. It appears that it's only a lack of tile floor failures after a dozen years that leads the ceramic tiling community to believe the tile eventually stops expanding.

All this is fine and good, but the real problem is determining the the ultimate amount of moisture expansion to plan for. Unfortunately, about the only way of accurately predicting the amount of moisture expansion in any clay fired product is by lab testing. They can boil the tile in water for 24 hours or steam it for 24 hours or expose it to high humidity at elevated pressure and temperature in an autoclave for 4 hours. Then, they measure the expansion and use that logarithmic or fourth root of time relationship to predict the ultimate expansion potential of the clay. But, examples cited in the literature vary considerably. That research paper by Richard Bowman says that the expectation is that the best quality ceramic tiles should not exhibit more than 0.03 percent moisture expansion in the first 5 years (presumably after installation), but also covers his butt by making the statement that accurate prediction of moisture expansion "is not a simple matter". Under the heading of "Failures", that same paper by Richard Bowman cites a case where terra cotta tiles came loose from the substrate and were subsequently lab tested and found to have an expansion potential of 0.11 percent over 15 years, or about 4 times that 0.03 percent guideline. In another case, glazed ceramic tiles were tested over a three year period to show moisture expansion of 0.046 percent, but fully 2/3 of that expansion occurred in the first two months. In this paper:

http://www.irbdirekt.de/daten/iconda/CIB1110.pdf

a Czech research team investigating failure of floors in which concrete was poured over fired clay slabs set on steel I beams to make floors in multilevel buildings. The investigation revealed that the failure of the floor was due to the moisture expansion of the clay combined with shrinkage in the concrete as moisture evaporated from it as it cured. That Czech paper cites a French research paper which found that the irreversible moisture expansion of clay minerals can reach more than 5mm per meter.", but it's not clear that by "minerals" they mean clay or the particular compounds that make up clay.

I simply couldn't find any concensus of opinion on how much space to allow for moisture expansion, except that providing expansion joints in the tiling should be more than enough to accomodate it.

It should be noted that moisture expansion is typically a significantly larger dimensional change than thermal expansion. Thermal expansion of ceramic tiles is quoted as being anywhere from 2.5 to 9 parts per million per degree Celsius. In an indoor setting, you don't see much of a temperature change from summer to winter, but dark tiles exposed to direct sunlight can be warmed by radiant heat from the Sun. Assuming a temperature change of 20 degrees Celsius, you could potentially see a dimensional change in the tile of 0.005 to 0.018 percent, which is considerably smaller than the 0.03 percent moisture expansion expected of "the best quality tiles" and much less than the 0.5 percent moisture expansion suggested in the Czech research paper.

Of course, the only way to accomodate thermal and moisture expansion of ceramic tiles is to provide for expansion joints. Some sources recommend expansion joints every 20 to 25 feet, but if the tiles are exposed to direct sunlight or frequent moisture (like kitchens and baths), that spacing should be reduced to 8 to 12 feet. The Ceramic Tile Institute of America:

The Tile Doctor - Field Report #42

(which is presumably an organization started by an individual tiling contractor) suggests: "For interior joints, the 3/8" minimum joint width is required at 24' to 36' on center. However interior timework subject to moisture or direct sunlight require spacing as for exteriors." ...and suggests 3/8 joints on 12 foot centers or 1/2 inch joints on 16 foot centers. This web page describes the way of making expansion joints (if tiling over tile backer board) as leaving a grout line empty and filling it with a foam backer rod and then caulking over the foam with either a polyurethane or silicone caulk that matches the colour of the surrounding caulk. This web page suggests using polyurethane caulk for high traffic areas because it's harder and stronger, but notes that silicone doesn't deteriorate from exposure to moisture or UV light from the Sun. Polyurethanes aren't as resistant to UV light as silicone.

The problem is that the web site suggests "checking" the adhesion of the caulk to the tile edges by cutting part of the caulk out and pulling on it to see if the caulk breaks or it pulls off the tile edge. The web site suggests that the caulk should break before pulling off the tile edge. The author then goes on to say: "Sealant may be replaced in test area easily, by merely applying more sealant in the same manner it was originally installed (assuming good adhesion was obtained.)" what he's not realizing is what everyone who's ever tried to re-caulk a bathtub knows... nothing sticks to silicone caulk, not even silicone caulk. The silicone caulk you put back in that test area won't stick to any residual silicone film still on the tile edges, and you need to use a trick to ensure that there's no residual silicone on the tile edges. Consequently, I think that web site should make people aware not to test the adhesion of silicone caulk in a conspicuous place or traffic lane.

The papers I read suggest installing expansion joints every 20 to 25 feet, but to reduce that to every 8 to 12 feet in areas exposed to direct sunlight or potentially wet areas like kitchens. In my conversations with a local tile contractor, he says he sets expansion joints every 12 to 16 feet, and this means that you normally don't need expansion joints in typical residential settings like kitchens and baths. He also uses prefabricated expansion joints from Schluter because it saves time and results in a cleaner appearance than caulked joints.

Also, that Ceramic Tile Institute of America web page advises to cut any mortar bed the tiles are being laid on at the expansion joint so that there's complete separation of the mortar bed and tile layers together across the expansion joint. I phoned Custom Building Materials tech support to find out why that's not done with setting tile over Wonderboard, and found out why Custom Building Products recommends leaving a gap of 1/8 inch between Wonderboard panels but then advises that you fill those joints with thin set. Those 1/8 inch gaps between panels aren't meant to be expansion joints at all. Apparantly, the fiberglass mesh that extends around the panel actually sticks out further at the edges than the cement. The purpose of leaving the gap and then filling it in is to embed the fiberglass along the edges in the thin set used to fill the joints between panels, thereby connecting all the panels structurally together. Also, the 1/8 to 1/4 inch gap recommended around the perimeter of the room is not to allow for the exapansion of the Wonderboard panels; it's to allow for expansion of any wood walls around the room. Custom Building Products presumes no expansion of the Wonderboard panels whatsoever in their installation instructions.