Indent Fracturing of Stone Tile: What is it… and how can it be prevented?

by Katie Daniel | July 11, 2017 2:55 pm

[1]

by Daren S. Kneezel, RA, and Michael J. Scheffler, PE
Soon after a stone-tile floor is installed for a hotel lobby, an architect hears from the client: “In the early morning, the tile exhibits surface depressions and has a wavy appearance—not necessarily at joints between tiles, but randomly throughout the tile. When we review the tile close up, the waviness disappears. It’s also less apparent in the afternoon. How do we fix this? Is it because of your design or is it an installation issue?”

The condition is likely indent fracturing—permanent distress manifested as slight depressions in the tile, often accompanied by very fine, barely perceptible hairline fractures. An example of a tile installation exhibiting indent fractures is shown in Figure 1. Typically developed within weeks or months of installation (i.e. through the curing and drying phase of the mortar materials), this distress is most apparent under oblique reflected light conditions, such as in the early morning—it may not be immediately noticed unless these specific conditions are present.

A stone-tile installation exhibiting indent fracturing, where evident, is often found to be undesirable by owners and can often be rejected or be subjected to a warranty claim. Where the indents are extensive throughout an installation, the most economical method of repair typically involves full removal of the tile installation, including the thin-set setting mortar, as large-scale replacement of individual tiles is rarely cost-effective. However, where distress is limited, individual tile replacement is possible and may be economical, provided matching tiles are available.

Removal and replacement of stone tile can be labor-intensive, disruptive to building occupancy, and the cause of significant material waste. Knowledge of a few key factors can help avoid this potentially costly, irreversible issue, which is not attributed to lippage or exceeding setting tolerances.

[2]

Factors leading to indent fracture
Several factors have been shown to directly contribute to the development of indent fractures in stone-tile installations. They include:

• shrinkage of the mortar setting bed and crack development;
• stone tile water absorption;
• tile dimensional changes upon wetting;
• tile strength;
• construction timing (e.g. allowing mortar to be set in lifts and cure before setting the tile, and providing sufficient time to construct and evaluate mockups); and
• construction detailing—particularly lack of shrinkage restraint when the mortar is set on a flexible substrate.

These critical factors have been identified by this article’s authors, through examination of individual elements of stone-tile assemblies. The tested system and components included several types of nominally 9.5-mm (3/8-in.) thick stone tile set in varying mortar thicknesses over a sound-attenuation mat on a concrete substrate. They represent various conditions noted in a floor system comprising several thousand rectangular stone floor tiles. The frequency of indent fractures varied by stone type and mortar thickness.

Absorption and rate of absorption of stone tile
The extent of indent fractures increased with increased tile water absorption (and rate thereof), as shown in Figure 2. Of the four stone types exhibiting the most frequent occurrence of indent fractures, two had approximately 14 times greater absorption—when tested in accordance with ASTM C97, Standard Test Methods for Absorption and Bulk Specific Gravity of Dimension Stone—than the others. The stone types with greater absorption had approximately two times greater frequency of indent fractures.

Stone tiles with higher total absorption were also found to have higher rates when partially immersed in water. The more-absorbent stone types had moisture first observed on the top surface of tiles in less than 20 minutes, while the less-absorbent ones did not have moisture after several days of monitoring.

Warping and elongation of stone tile due to water exposure
Testing showed stone types exhibiting a higher frequency of indent fractures in-situ generally display greater warp and elongation when immersed in water than stone types experiencing less-frequent indent fractures. The four stone types with the most frequent occurrence of indent fractures showed approximately twice the measured warp and elongation of the three stone types with less-frequent occurrence. This strongly indicates stone-tile types with greater moisture expansion are likely to be subjected to greater internal stresses and potential for cracking after mortar bond develops and the stone attempts to return to its pre-wet and preset size. However, the measured warp and elongation do not correlate precisely with observed indent fracture occurrence, as other factors also contribute.

All other factors being equal, an assembly with stone tile that is more moisture-sensitive (i.e. greater warp and elongation upon wetting) has greater potential for cracking that may lead to indent fractures after bond with the setting mortar has developed. Figure 3 shows a partially immersed specimen being monitored for warp and elongation.

Mortar shrinkage
After the stone tile and mortar bond, a greater magnitude of mortar shrinkage will result in greater applied forces and stress to the tile, increasing the potential for tile cracking and indent fractures. Testing (in accordance with ASTM C596, Standard Test Method for Drying Shrinkage of Mortar Containing Hydraulic Cement) showed there are differences in the rate and magnitude of mortar drying shrinkage of three different types of commonly used thin-set mortar during curing at 50 percent relative humidity (RH). Depending on the mortar type, approximately 80 to 90 percent of the 28-day shrinkage occurred within the first week of curing. While most mortar shrinkage occurs in the first seven days, continued shrinkage may sometimes result in the later appearance of indent fractures.

Further, shrinkage cracks are less likely to develop in mortar prevented from curing rapidly, as shown by comparative evaluation of sealed and unsealed thin-set mortar applications. Based on this, the rate of mortar shrinkage and potential for shrinkage cracks may be significantly reduced by the presence of the tile, particularly impervious ones that absorb little water from the setting mortar. Testing performed on mortar with two different water contents showed initial moisture content does not necessarily affect mortar shrinkage, though it may affect the moisture expansion of the tile, as more water would be available for tiles to absorb.

Flexible substrate materials
Where mortar is set on a flexible substrate, the flexible material may provide only minimal restraint against shrinkage of the setting mortar, leaving the job primarily to the stone tile. Where restraint is insufficient, the setting mortar will crack upon curing. Testing showed the tensile modulus of a sound-attenuation mat was insufficient to significantly restrain shrinkage of the setting mortar, and was also significantly lower (i.e. more elastic) than elements such as the stone tile, mortar, and concrete substrate (Figure 4).

Close-up examination of indent fractured stone tiles
A characteristic of indent fractures is there are only cracks in the stone tile where they align with cracks in the thin-set mortar below. A greater number of cracks is exhibited by the thin-set mortar than the stone tiles, as not all cracks extend into or through the stone.

Cracks in the assembly are typically wider at the bottom and narrower at the top, as shown in Figure 5. These features are a strong indication mortar shrinkage cracks initiate at its minimally restrained bottom surface, at the sound-attenuation mat. These features are characteristic of not only indent fractures, but also structural overload of the tile. However, the presence of a coincident indent and curvature at the top surface of the stone tile at the crack is a defining characteristic of an indent fracture resulting from mortar shrinkage, rather than an overload condition.

Evaluating critical factors through mockup assemblies
To evaluate the effect of several factors on development of indent fractures, mockup assemblies were fabricated with various combinations of conditions (i.e. stone type, mortar type and thickness, presence of sound-attenuation mat). This enabled replication of the occurrence of indent fractures in a laboratory setting. There are numerous variables in adhered stone tile systems of which only a few were included in the mockups. However, this mockup testing further established the key contributing factors to development of indent fractures.

Mockup assemblies with stone types that had the greatest frequency of indent fractures in the actual installation also exhibited indent fractures without exposure to external loading. These were typically first observed between two and four weeks after assembly, and were nearly identical in appearance to those observed in-situ, as typically shown in Figures 6 and 7.

Laboratory mockups also showed indent fractures were far more frequent in assemblies with at least 13-mm (½-in.) thick setting mortar than companion assemblies with 6.4-mm (¼-in.) thick mortar. For example, all eight of the mockup assemblies set on 13-mm (½-in.) thick mortar and sound-attenuation mat for a tile with frequent in-situ distress exhibited indent fracture, while none of the five companion assemblies set using 6.4-mm (¼-in.) thick mortar did. These observations are consistent with reduced stress applied to the tile from reduced mortar thickness.

Indent fractures were also observed at a reduced frequency in mockups without sound-attenuation mat. This indicates a concrete substrate provides greater restraint of the setting mortar—and therefore resistance to mortar shrinkage crack development—than does a more-elastic sound-attenuation mat or any similarly flexible membrane or bond-inhibiting membrane that does not restrain mortar shrinkage.

As was observed in-situ, indent fractures in the mockup assembly tiles aligned with cracks in the mortar, as shown in Figure 8; cracking was also more frequently seen in the mortar than in the stone. The cracks in the mortar appeared to grow wider over time due to continued drying shrinkage of the mortar and stone tile, as shown in Figure 9. Based on these observations, and because no external loading was applied to the mockups, formation of indent fractures can result from mortar shrinkage alone.

Elongation of mockup assemblies without external loading
Elongation of the stone in mockup assemblies exhibiting indent fractures typically increased rapidly to an initial maximum positive elongation, followed by a more gradual descent to a negative elongation (shrinkage) and another ascent to positive elongation. The initial positive elongation of the stone is likely caused by moisture absorption of water into the stone from the mortar during curing.

The shrinkage suggests the stone is in compression resulting from mortar shrinkage after bond development. The subsequent positive elongation indicates tile cracking at mortar shrinkage crack locations.

The elongation for the assemblies that did not exhibit indent fractures typically had a similar initial increase, followed by gradual descent. However, the negative elongation tended to stabilize, indicating the mortar shrinkage stresses did not exceed the strength of the stone tile. Several of the tile types exhibited indent fractures in-situ, but not in mockups when observations were halted after four weeks. However, the stone tile in these mockups was continuing to shrink or had begun to elongate again—this indicates indent fractures may occur even beyond four weeks.

Essential factors leading to indent fracture
Based on testing and observations of indent fracture development in mockup assemblies, several factors result in indent fracturing of stone-tile applications. Appropriate modification of the following essential factors can significantly mitigate or eliminate the potential to develop indent fractures in thin-set stone-tile applications.

Mortar thickness
The stress applied to the stone is proportional to the thickness of the mortar. Additionally, a thicker mortar setting bed yields a greater amount of free moisture. This causes elongation of the tile prior to restraint from the mortar and greater subsequent stresses due to stone shrinkage as the moisture evaporates. As a result of both these factors, greater mortar thickness will contribute to a greater likelihood of indent fracturing. This was confirmed through the greater frequency of crack development in mockup assemblies with greater mortar thicknesses.

Stone tensile strength
The use of a stone type with a greater tensile strength or greater thickness (if an inherently weak stone is to be used) will provide greater resistance to cracking from the applied mortar shrinkage stresses.

Contributing factors that can lead to indent fracture
Appropriate modification of the following contributing factors can minimize the potential for indent fracture formation. (However, the aforementioned essential factors must also be sufficiently addressed.)

Elastic or low-permeability membrane below mortar
A primary contributing factor is the use of an elastic substrate such as a sound-attenuation mat, or bond-inhibiting surface, which will not restrain shrinkage of the mortar bed, as would be provided by a more rigid and bond-developing substrate such as concrete. As a result, tile applications with an elastic or bond-inhibiting substrate have a greater potential for mortar shrinkage cracks and associated indent fractures. Cracks originate in the setting mortar at the contact with the flexible sound-attenuation mat before extending upward through the mortar and into the more-rigid stone.

Additionally, low-permeability substrates will result in a greater amount of free moisture availability for absorption into stone tile. This means greater moisture expansion of the tile prior to restraint from the mortar, increasing potential for indent fractures.

Stone moisture sensitivity
All other factors being equal, stone tile with greater total absorption (and/or rate thereof) is more likely to exhibit indent fracturing than tile with less absorption or a reduced rate. This is probably due to more-rapid mortar curing adjacent to a more-absorptive tile, resulting in a greater potential for shrinkage cracks to develop in the mortar. However, low stone-tile absorption alone does not preclude development of indent fractures, as stone types with very low absorption can also exhibit indent fractures in-situ.

Stone types with greater moisture expansion are also likely more vulnerable to indent fractures due to the greater stresses after bond is developed and the tile recovers to its original pre-installation dimensions. Where two or more stone types are being considered for a project, the less-moisture-sensitive stone is less likely to exhibit indent fractures.

Mortar shrinkage magnitude
Mortar exhibiting greater total shrinkage likely contributes to a greater frequency of indent fractures, as increased shrinkage after bond development leads to greater applied stresses to the stone tile.

Indent fracture mechanism
Based on these findings, the physical/mechanical mechanisms resulting in the occurrence of indent fractures can be best described as follows:

1. The stone tile elongates and can slightly warp upon wetting from moisture in the setting mortar. A greater amount of available water in the mortar results in more wetting of the stone tile.

2. The stone tile then becomes restrained in its wet and lengthened condition as the mortar develops a bond with the stone, altering from a plastic state to a partially cured and rigid one.

3. The drying shrinkage of the mortar setting bed results in shrinkage cracks in the mortar, particularly for stone types with high absorption or where restraint is not provided by the substrate below (such as where there is a flexible sound-attenuation mat or crack-isolation membrane).

4. Mortar shrinkage results in tensile stress being imparted to the underside of the tile to which the mortar is bonded at the mortar crack.

5. Tile cracks result where the applied tensile stress from mortar shrinkage exceeds the tensile strength of the stone tile. The greater width of cracks at the tile’s bottom surface—and resulting downward curvature—is likely due to greater applied stress (and induced shortened length of tile) at the bond surface between the stone and mortar than is present at the top surface.

The mechanism of indent fracture distress is inherent in the stone-tile assembly materials, and no external applied forces (e.g. service or construction loading) are required. Of course, external loading of a tile can result in cracks, but these are not likely to exhibit indents coincident with the crack, as mortar shrinkage shortening the bottom of the tile is needed to locally indent or curve the tiles at cracks. It may be difficult to differentiate the exact cause of cracks between loading-induced types and mortar shrinkage-induced indent fractures where both are suspected; however, loading-induced cracks are typically observed only locally to where loads were applied. Indent fractures are very likely to occur throughout the installed assembly as long as the aforementioned factors are present either singly (for essential factors) or found in combination.

Limiting potential for indent fractures
While most stone-tile installations do not exhibit indent fracturing, there are a few key contributing factors that can, when acting in combination, result in this distress. These factors can be modified to minimize the potential for indent fracturing of a stone tile installation. Each of the contributing factors alone is not likely to result in indent fracturing, provided the other factors are appropriately accommodated.

Mortar
Mortar should be installed in lifts where necessary, and its thickness minimized. When setting beds thicker than 6.4 mm (¼ in.) are required, consideration should be given to installing an initial thick bed of mortar, allowing it to fully cure, then installing a second thin-set bed for adhering the tile. This limits the mortar drying shrinkage stresses applied to the tile, and also the amount of tile moisture expansion prior to bond formation and the subsequent stresses from tile shrinkage.

Stone tile
The stone-tile thickness should be increased and a less-moisture-sensitive stone selected. If a particular stone type with high absorption or low tensile strength is desired for its aesthetic, a greater stone-tile thickness may be appropriate. Increased thickness provides greater stone tensile strength, as well as greater dimensional stability and less potential for curling where a hairline fracture occurs—this makes any distress that occurs far less noticeable.

Additionally, increased thickness may slow the rate of free moisture evaporation from the mortar, resulting in reduced mortar cracking from drying shrinkage. Application of an impervious tensile reinforcement mesh to the back surface of the stone may be an alternative means to increase the tensile strength of the tile while reducing the rate of mortar shrinkage.

Membrane
If possible, one should avoid using elastic membranes below the mortar. If an elastic substrate is desired (for reasons such as soundproofing or for isolation of substrate cracks), the stone tile properties and mortar thickness must be carefully considered.

Mockups
As the potential for indent fractures can be compounded by a number of contributing factors, precise failure criteria for each factor cannot be established. For example, a highly moisture-sensitive thin stone tile may perform well when set in a thin mortar application. Consequently, when schedule permits for large tile installation projects or those including the factors noted in this article, the authors recommend construction of mockups and monitoring for at least four weeks prior to installation to evaluate the potential for indent fracturing. However, monitoring should take place for as long as the construction schedule allows, as the cracks may take longer to develop in some applications. Indent fractures are most easily identified on mockup assemblies in
a dimly lit room with oblique lighting conditions, and also using a straightedge and flashlight.

Conclusion
A few essential factors are necessary for indent fractures to occur, but this distress may be aggravated when the other contributing factors act in combination. Modification of factors related to material selection may be impractical where certain aesthetic or performance characteristics are desired, assembly depth is predetermined, or budget/schedule do not allow for extensive comparative testing between different materials. Addressing factors related to construction timing, such as proper schedule allowance for application of mortar in lifts and creation of mockups, is likely a more practical approach to avoiding the development of indent fractures.

Each of the factors that result in indent fractures can be avoided through proper preparation and specification, and at minimal additional cost to the owner relative to the significant expense of replacement. Armed with an understanding of how to mitigate the various factors that are described in this article, an informed specifier should be able to confidently design a stone tile floor without fear of indent fractures.

Daren S. Kneezel, RA, is a senior associate with Wiss, Janney, Elstner Associates (WJE). He has been involved in projects ranging from investigation of high-rise building façades to laboratory testing of stone and other construction materials. Kneezel has extensive experience performing laboratory testing of various materials with a concentration on stone and stone-faced honeycomb composite to evaluate structural performance, durability, and to determine compliance with project specifications and industry standards. He is a member of ASTM technical committee C18 (dimension stone). Kneezel can be reached at dkneezel@wje.com[9].

Michael J. Scheffler, PE, is a principal with WJE. Over more than 35 years, he has been involved in thousands of investigations of deterioration and distress in buildings and other structures. Scheffler also has extensive experience performing laboratory and in-situ testing of building assemblies related to stone material performance, durability, and structural performance, and has performed long-term monitoring and instrumentation of stone distress and construction. He is a member of ASTM technical committees C18, C24 (building seals and sealants), and E06 (performance of buildings). Scheffler can be e-mailed via mscheffler@wje.com[10].

Source URL: https://www.constructionspecifier.com/indent-fracturing-of-stone-tile-what-is-it-and-how-can-it-be-prevented/

---