Engineering engineered stone installations to prevent failures

There have recently been a number of differential movement failures of engineered stone floor tiling systems. Since many factors can contribute to such failures, each failure must be separately investigated in order to establish the predominant causal factors in that particular circumstance. This paper does not seek to analyse such failures individually or generically. It reflects upon useful data that manufacturers have published, detailing two engineered stone characteristics that must be respected. A consideration of the published data indicates why some engineered stone tile fixing practices, such as the butt jointing of large tiles, are extremely inadvisable.

This paper reports a few facts in order to better educate retailers and consumers about using engineered stone tiles on floors. It contains past and current material that has been extracted from manufacturers' catalogues and websites, with minimal editing. Proprietary names have been deliberately suppressed, partly since consumers should always obtain current product literature. Manufacturers must also provide retailers and consumers with comprehensive educational material that promotes sensible product use. While consumers may want a seemingly seamless, absolutely level polished floor, they must be made fully aware of any risks that are associated with the choice of cheap setting materials, the butt-jointing of tiles, the omission of movement joints, and other design and installation issues.

This paper focuses upon the two engineered stone characteristics most relevant to differential movement failures: dimensional stability and linear thermal expansion.

What is engineered stone?

Engineered stone (also known as reconstituted, re-composed, agglomerated and synthetic stone) is manufactured from a mix of stone aggregate chips (most commonly quartz or marble, but also igneous rocks such as granite and basalt); mineral fillers (generally the ground aggregate); a resin binder (typically an unsaturated polyester); pigments and additives. Tiles are typically formed using a vibrocompaction process, where applied vacuum minimises any porosity, usually reducing the water absorption to much less than 0.4%. There are many different kinds of polyester resins, with different viscosity, colour and hardening time. Tile curing may be accelerated by using ovens or steam. Some tiles are given post-curing heating treatments to increase the degree of cure.

Since there are no standards for adhesive fixing of engineered stone tiles, they tend to be fixed according to ceramic tiling standards, using adhesives designed for fixing ceramic tiles. Tiling codes of practice have evolved from an empirical basis, where what was found to work was adopted, and attempts were made to eliminate those practices that were associated with failures.

What is a differential movement failure?

Tiling systems are comprised of several elements that may be assembled in a number of ways, taking into account the characteristics of the materials that make up each layer and how they will interact when the complete composite system has to perform functionally. Differential movement can occur where two or more differing materials are joined, where the movements in the materials may be in different directions, of differing magnitudes, and may occur at different times and rates in response to a variety of exposure conditions. Expansion and contraction of any tiling system component will have the potential to cause problems if not accommodated in the design, or if sufficient time is not allowed for movements and reactions to occur within the project scheduling. Other types of movement such as structural deflections, creep and foundation movements sometimes need to be considered.

Thermal movement - most materials are subject to reversible thermal movements, expanding as temperature increases and shrinking as temperature decreases. Engineered stone tiles have a linear thermal expansion coefficient in the range of 5 to 40 x 10-6 mm/mm C-1. Most engineered stones have a higher thermal expansion than the concrete substrates that they are fixed to (8 to 12 x 10-6 mm/mm C-1), and a much higher thermal expansion than natural stone (3 to 12 x 10-6 mm/mm C-1) and ceramic tiles (4 to 9 x 10-6 mm/mm C-1). The magnitude of the linear thermal expansion has very important implications for the tiling system design.

Moisture movements - porous materials capable of absorbing or releasing water or water vapour will generally be subject to some increase in size as their moisture content increases but not all types of movement are fully reversible. They are divided into two categories: reversible and irreversible. Irreversible (permanent) changes in size principally occurring during the early life of some manufactured materials. For example, the moisture expansion of some ceramic tiles and the irreversible drying and carbonation shrinkage of cement-based products.

Reversible size changes occur in service due to moisture content changes. Engineered stone can have an appreciable reversible moisture movement. However, since there is no recognised test, the extent of such movement may not have been widely determined.

Dimensional stability refers to the ability of an engineered stone tile to resist curling or warping, when exposed to the water contained in most adhesives. All materials, even glass, will absorb some water on its surface. As the surface absorbs water it expands. This expansion makes the wet side of the tile larger than the dry side and the tile will attempt to curve or warp in order to relieve the stress. How much the tile cups will depend on the amount of water absorbed per unit of time. The amount of warpage also depends on the size and the thickness of the tile. "Double the size of the tile and the apparent warpage goes up by four". "Double the thickness and the strength to resist warping doubles".

Mapei designed a computerised apparatus and test procedure to measure the dynamic warpage of various natural stone and engineered tiles. Certain minerals in natural stones are sensitive to and react with water (eg. serpentine in 'green' marbles). This causes tiles to expand unevenly and warp when installation with water-containing adhesives. Also, certain resins used in the manufacture of agglomerate tiles can expand.

The Mapei dimensional stability test uses 'simulated' conditions to measure the maximum amount of linear movement and warpage occurring during initial contact with wet 'mortar' on the back of the sample. A sample of each size and thickness is laid face down, and is supported on 3 corners, leaving one corner to move in all directions. A cut-to-size wet felt is applied over the entire back of the sample. A computer records values from strategically positioned sensors at 10 minute intervals over a 12 hour period. However, since the most detrimental movement is considered to occur during the initial 6 hours of installation, the test result is based on the maximum movements recorded during that period.

When a tile is fixed, it must remain stable without moving for that period of time necessary for the adhesive to react, so as to become partially hardened. In the process, available water should be consumed (so that no further moisture is available to promote movement by the tile) whilst sufficient adhesive strength is developed to restrain the tile against any continuing tendency to curl or move.

Tiles are classified according to the Mapei test results. Adhesive manufacturers specify adhesives according to the tile's classification. Class 1 (or A) tiles generally have no movement or warpage and can be fixed with standard latex-modified cementitious adhesives. Class 2 (or B) tiles require fast setting adhesives that produce sufficient adhesion and consume excess water before the tile can move. Class 3 (or C) tiles can only be fixed with epoxy or polyurethane reaction resin adhesives. Some manufacturers discuss dimensional stability and indicate the 'Mapei' classification for some product ranges. However, remember that tiles may move to a lower class with size increases. Some manufacturers provide no explicit information.

One manufacturer advises, "Too much water in the adhesive will keep tiles from sticking and make them come loose". Another advises, "When the water cannot escape through the underlying screed, the grooves between the tiles or the perimeter border, it results in detachment and tension, which inevitably leads the tiles to warp in an uncontrolled manner". This highlights one function of the grout joints. Another function is to accommodate small amounts of differential movement due to reversible thermal movements.

Dealing with expected movements

One manufacturer advises "Installation of products without spacing (and thus without joints) is generally not advised; if this option is chosen, the following limitations must be followed: small surfaces; small formats; low-traffic areas; and products must retain stable dimensions". The same literature notes, "It is necessary to make generous use of expansion joints and these must be consistent with the architectural project".

While little guidance is provided on grouting materials, one manufacturer states, "Most cement, resin or latex based grouting products are suitable". The guidance on the recommended width of grout joints for large tiles is generally for a minimum of 3 mm, although some manufacturers specify a minimum of 4 or 5 mm, particularly for darker coloured tiles. "The width of the groove depends on the size of the tiles and must conform with European regulations (with minimum values from 2 to 5 mm); and also consider the size and thermal strains".

Some manufacturers infer that tiling system should be engineered: "Structural joints should follow those already present in the building and (expansion joints) should be of a size consistent with the expected movement and coefficient of linear thermal expansion of the tile to be installed". This requires that the architect be able to identify and productively use the linear thermal expansion coefficient of the product.

With regard to the "generous use of expansion joints", one manufacturer recommends dividing the tiling into bays of 3 x 3 m. Other manufacturers recommend bays of between 4 x 4 m and 7 x 7 m. "In places where a tile will be in direct sunlight, especially if it is of a dark colour, we recommend larger expansion joints".

Manufacturers also focus on the need for perimeter joints, perhaps recognising a general aversion to the visual presence of movement joints. "Other joints must be planned around the floor perimeter and structures such as columns. It is a good idea to plan an open joint of 10 mm because it will be influenced by the foundation, the adhesive and the tile". Another states, "It is advisable to leave joints wide enough to allow filling with one-component silicone sealant around the perimeter and at walls, pillars and steps". Such joints need to be properly designed if they are to accommodate the expected amounts of movement.

What might happen if insufficient provision is made for thermal movement? One manufacturer, when discussing storage, states, "Additionally, it is possible for larger size 10 mm panels to "bend", or assume the profile of whatever it is stacked against". If a 600mm long tile with a thermal expansion coefficient of 25 x 10-6 mm/mm C-1 increases by 30C, it will expand by 0.45 mm. If this movement cannot be transferred into functioning movement joints, or be absorbed by the grout joints, it might be relieved by the tile bending upwards. Shrinkage of cementitious screeds and concrete substrates could result in similar stress relief at the grout joints, where the ends of the tile bend upwards, mimicking the warping associated with dimensional instability.

Given that substrates can contribute to differential movement, there are strict requirements for floors that are to be tiled: "Concrete surfaces must be sufficiently mature. The maturity of screeds is calculated as about 6-8 days per cm of thickness, unless special quick-drying cements are used". Furthermore, given the potential instability of some tiles if water is present, "The supporting surfaces or foundations must be absolutely dry", and "the residual humidity of the foundations must be less than or equal to 3-5%". This requirement has recently been changed to a maximum residual humidity of 2.5%, without any guidance on how this should be determined. Different moisture test procedures yield different results.

There are naturally other requirements that must be complied with. "The supporting surfaces must also be stable, solid, resistant to compression, sufficiently level and free of dust, grease, wax, paint, dismantling agents and anything else that could prejudice the efficacy of the adhesion".

One manufacturer recommends (for 300 x 300 mm tiles) "the use of a notched trowel to ensure at least 80% adhesive coverage of the tiles", but most manufacturers require the use of the double spreading system for the laying of larger sizes and for areas subjected to heavy traffic (as this guarantees 100% coverage).

We are reminded, "floors are complex construction systems with multiple layers, and as such, they are complete and integral systems in themselves, and not just surface coverings. They must be conceived to resist mechanical loads and withstand the temperature and chemical conditions of the local environment".

One manufacturer notes that a higher frequency of expansion joints is necessary where engineered stone tiles are laid over in-floor heating systems. However, there does not seem to be any specific guidance on how to install tiles over acoustic membranes. An engineered stone tiling system that incorporates both an acoustic membrane and in-floor heating is a complex system. As one tile manufacturer advises, "For large or special projects, consult with the setting material manufacturer directly for specific recommendations and installation methods". It is best to consult broadly until all manufacturers approve a system that can then be specified.

Tile testing - there are no standards defining requirements for engineered stone tiles. Manufacturers use a range of test methods to characterise their products. Some manufacturers publish two values for the water absorption of their products, where the ASTM C 97 cold water absorption value is always less than any value obtained from boiling tests. This indicates the difficulty of comparing manufacturers' test data where several different test methods have been used. Several different test methods have been used to determine thermal expansion. Some manufacturers publish two results for individual products, where a lower value is obtained in the range of -20 to 20C, than between 20 and 70C. The thermal expansion coefficients of these materials may vary continuously, increasing very slightly as the temperature range increases. Some manufacturers indicate the temperature range over which the mean linear thermal expansion was determined. This further demonstrates the difficulty in comparing and using manufacturers' data. It also emphasises the need for specific test methods for engineered stone tiles.

In summary, while it may sometimes be hard to determine or compare the specific physical characteristics of individual products, there are generally some hard and fast design rules that must be observed, particularly with larger sized products. Engineered stone tiles are more likely to warp as they become larger or thinner. While the extent of any such reaction may depend on the type of binder and its degree of cure, the scientific literature does not seem to contain appropriate studies. The Mapei test can identify which natural stones react readily, but it does not indicate the long-term potential of tiles to warp. There is a need for further studies that might lead to the development of a reliable accelerated test method.

Although some manufacturers' websites are now less accessible, architects must ensure they are adequately informed when selecting engineered stones and specifying tiling systems. Engineered stone manufacturers' product literature places a significant design responsibility on the specifier. There is still an unfilled need to provide architects with simple means of calculating the size of grout joints and the spacing of movement joints, given the other relevant project data.

When consumers demand butt-jointed 600 mm engineered stone tiles free of bisecting movement joints, they may have to accept the need for project scheduling delays; premium quality fixing materials; automated shading systems to protect the tiles from thermal variations; and other interventions to ensure the integrity of the system. Their other option is to absolve the specifier and installer of any associated liability.