For the Want of a Horseshoe Nail: Identifying causes of tile failure

by Katie Daniel | March 1, 2017 11:20 am

[1]
All photos courtesy Custom Building Products

by Howard Jancy, CSI, CDT
A folk rhyme often attributed to Benjamin Franklin describes the unintended outcome of simply omitting a horseshoe nail—namely, the loss of a kingdom:

For the want of a nail the shoe was lost,
For the want of a shoe the horse was lost,
For the want of a horse the rider was lost,
For the want of a rider the kingdom was lost,
And all for the want of a horseshoe-nail.

This proverb illustrates a basic truth—quick and poorly thought-out decisions have consequences beyond the initial myopic moment. Though the rhyme is not modern in origin, its maligned nail is often the culprit in today’s construction failures. The ‘nail’ in this narrative, for example, is expansion and movement joints, and the unintended outcome is the failure of a building’s tiled exterior.

A project located in a Midwestern city provides an excellent example to illustrate this point. The complex consists of three residential buildings considered upscale apartments. The exteriors are clad with a brick veneer, 450 x 450-mm (18 x 18-in.) glazed porcelain tiles, and 600 x 600-mm (24 x 24-in.) unglazed, through-body porcelain tiles. The structures were completed in 2014. One year later, after exposure to a winter’s and summer’s temperature extremes, the larger porcelain tiles started to loosen and fall. Technical data for these light brown tiles indicated the tile was applicable for commercial and outdoor use, and was freeze-thaw-, thermal-shock-, and frost-resistant. Thermal cycling seemed to be the catalyst for the failures, yet the question remained—why? Why would construction materials and methods that were used successfully in the past suddenly exhibit such extensive failures?

The search for the proverbial nail begins.

A closer look
Tile failure was clearly evident on each building, occurring randomly. Though there was a clear indication that the failed tiles were not properly bonded during installation, this did not illuminate the initiating factor that pulled the tiles from the structure. A closer inspection of the tile installation revealed extensive cracking and debonding of the cement grout.

On further investigation, it became clear that no movement joints had been installed to accommodate expansion and contraction of the tile during freeze-thaw cycling in winter or thermal expansion during the summer. As the tiles expanded and contracted, correctly located and installed silicone soft joints in the assembly would have absorbed their movement without adhesive or cohesive failure, but not so with rigid cement grout. Once the grout cracked, moisture readily traveled behind the tiles, further deteriorating the already-tenuous adhesive bond between tile and mortar (Figure 1).

The Tile Council of North America (TCNA) Handbook for Ceramic, Glass, and Stone Tile Installation contains general recommendations for the use and installation of movement joints. Its section EJ171, “Movement Joint Guidelines for Ceramic, Glass, and Stone,” provides 12 details illustrating the correct design and placement of various types of movement joints. It states:

Because of the limitless conditions and structural systems on which tile can be installed, the architect or designer shall show the specific locations and details of movement joints on project drawings. Preparation of openings left by the tile contractor and installation of backup strip and sealant should be specified in the Caulking and Sealant section of the job specification.

More specifically:

On some portions of the tile project, the grout was not even installed, readily allowing moisture behind the tiles. The extent of water intrusion behind the tiles can be clearly mapped by the darker, damp areas on the tiles’ surface (Figure 2).

[2]
Adhesive and cohesive failure of cementitious grout. The grout cracked from expansion and contraction of the tile. Properly installed movement joints utilizing an elastomeric sealant within the tile assembly would have prevented this damage, but the cracks allowed water to travel behind the tiles, causing failure.
[3]
Grout was not installed in some of the tile joints, notice the open joints around the tile. These gaps allowed for even greater moisture infiltration into the setting materials. The moisture eventually wicked into the tile, darkening its surface, as shown.

Canary in a coal mine
Lack of movement joints and the subsequent failure of the grout was just the first of many problems plaguing this tile installation. Once intrusive moisture had saturated the mortar, thermal cycling of the tile compromised the adhesive bond between the two. It was clearly evident from examining the failed tiles they had been negatively impacted by insufficient mortar coverage and spot-bonding. With these factors combining with excess moisture in the mortar and temperature swings from season to season, tile failure was likely inevitable (Figure 3).

Spot-bonding is not recognized as an acceptable installation method in the tile industry. Despite being quicker and cheaper than uniform directional troweling due to its lower labor and material requirements, this method generally results in inadequate mortar coverage, which can cause tiles to shear away from the mortar or crack from impact. Figure 3 shows approximately 60 to 70 percent mortar coverage and uncollapsed mortar ridges. For exterior installations, the standard is mortar coverage ≥ 95 percent, with all corners of the tile supported.

American National Standards Association (ANSI) A108.5 2.5.4, Installation of Ceramic Tile with Dry-set Portland Cement Mortar or Latex-Portland Cement Mortar, states when 95 percent coverage is specified in the project specifications, it is best to:

Back butter each tile with bond coat; or select a notch trowel sized to facilitate the proper coverage, key in the mortar into the substrate with the flat side of the trowel, and comb with the notched side of the trowel in one direction. Embed the tile in the mortar by beating in, pushing in a direction perpendicular to the combed ridges or other means to achieve the specified coverage. The method used should produce maximum coverage with corners and edges supported. Periodically remove and check a tile to assure that proper coverage is being obtained.

For this particular project, looking at the backs of the loosened tiles was also enlightening, and further emphasized the importance of utilizing the correct methods and materials when setting tiles (Figure 4). If the tile had been correctly back buttered, if the contractor used the right type of mortar, if the mortar had been uniformly combed in sufficient quantity, or if the installer periodically removed a tile during installation to assess mortar coverage and correct accordingly, then might this failure have been avoided?

The tile in Figure 4, if correctly set, would have retained quite visible mortar residues over 95 percent of the tile back with few, if any, uncollapsed mortar ridges. The white residue on the tile’s back is kiln release from the tiles’ manufacture, which can act as a bond breaker. That possibility could have been eliminated if the tile had been thoroughly back buttered (Figure 5).

Mortar coverage revealed
As already discussed, adequate mortar coverage can be ensured when the proper materials and methods are employed, along with periodic inspection of the back side of the tile during installation. A better understanding of tile mortar failure can be gained if one looks behind a tile after installation—something that can be accomplished by installing transparent tiles (i.e. small glass panels) with different mortar application techniques.

Figure 6 illustrates an example of a spot-bonded tile. An attempt to fully embed the tile with completely collapsed mortar ridges is prevented by the thicker spots of mortar, which slightly elevate the tile above the bonding material. This means it is only secured to the substrate with five points of contact, and the gaps and hollows underneath it will lead to cracking or delaminating. The unsupported corners and edges are particularly prone to damage.

The glass tile in Figure 7 was embedded in mortar that was swirl-applied rather than uniformly troweled in one direction. This method also results in uncollapsed mortar ridges and therefore voids under the tile, leading to cracking.

Finally, Figure 8 shows an example of a properly embedded and bonded tile. The mortar was combed in one direction, allowing for near-complete coverage under the tile once it was embedded, with edges and corners completely supported. This tile is less likely to crack from impact or delaminate from shear stresses across the bond line.

Mortar selection
Several types of cementitious mortars are available to facilitate tile installation and ensure long-term performance of tiled floors or walls. They include:

The latter category—LHT mortars—is another component of the apartment project’s failed 600 x 600-mm (24 x 24-in.) tile installation.

The TCNA handbook describes LHT mortars (formerly referred to as medium-bed mortars) as thin-set bonding mortars formulated to minimize slump and facilitate a thicker bond coat (from 3 to 13 mm [3/32 in. to 1/2 in.] thick) after the tile is embedded. LHT mortars are useful for setting:

There are no ANSI standards for this type of mortar. However, LHT mortars must meet the requirements of existing ANSI standards for the installation of ceramic tiles, which include:

These and other tile-specific standards can be found in the TCNA publication ANSI A108, A118, and A136–American National Standard Specifications for the Installation of Ceramic Tile.

Applying a mortar too thin will likely lead to inadequate coverage, particularly if there is warpage in the tile. Lack of coverage creates voids under the tile, ultimately leading to cracking and delamination. Applying mortar too thick is also problematic. Mortars applied in excess of their formulated and functional thickness will shrink as they dry and cure. This can also create voids under the tile, and can pull the tile from the bottom as the mortar shrinks, cracking it.

[10]
Gaps or hollows between the tile and substrate can lead to tile failure. The 13-mm (1/2-in.) gap indicates the thin-set mortar has been applied beyond the 6-mm (1/4-in.) functional thickness. This can cause the mortar to shrink as it cures, and to pull away from the tile. A medium-bed or large and heavy tile (LHT) mortar should have been used for these large-format tiles.

A contractor may apply an excessively thick mortar bond coat if:

Typically, the maximum applied thickness for a standard thin-set mortar is 6 mm (1/4 in.). In the case of this project, the installer incorrectly used and over-applied this type of mortar by spot-bonding. It would have been much more effective to uniformly trowel an LHT mortar at maximum applied thickness of 13 mm (3/4 in.) before tile embedment (Figure 9).

In hindsight
Many problems are associated with this tile installation, any of which could have caused the tiles to loosen and fall eventually. The lack of movement joints and resultant water infiltration through the cracked grout joint simply accelerated this deterioration. Value engineering, installer inexperience, and fast-tracking also played a role. Perhaps the calamitous omission (i.e. the proverbial nail) was not the lack of movement joints, but failure to read, understand, and employ tile installation methods readily available and clearly stated in the TCNA handbook and ANSI standards for tile installation.

Howard Jancy, CSI, CDT, is a commercial architectural services representative for Custom Building Products. He has 30 years of experience with tile, stone, and concrete flooring, as well as paving design, installation, and remediation. Jancy’s responsibilities include specification writing and review, technical service, and continuing education. He has written articles for numerous industry publications, including Landscape Contractor National, the Journal of Architectural Coatings, and The Construction Specifier. Jancy has also been a presenter at World of Concrete (WOC). He can be reached via e-mail at hjancy@cbpmail.net[11].

Endnotes:
  1. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Celedon-7-e1488382451334.jpg
  2. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-1.jpg
  3. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-2.jpg
  4. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-3.jpg
  5. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-4.jpg
  6. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-5.jpg
  7. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-6.jpg
  8. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-7.jpg
  9. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-8.jpg
  10. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2017/03/porcelain_Figure-9.jpg
  11. hjancy@cbpmail.net: mailto:hjancy@cbpmail.net

Source URL: https://www.constructionspecifier.com/for-the-want-of-a-horseshoe-nail-identifying-causes-of-tile-failure/