The importance of floor screed design - part 1
Sand-cement and concrete screeds are often skimmed over or overlooked completely when it comes to design. Often, the guy on the tools on the day of screeding ends up designing the screed.
And guess what? He usually doesn’t give two shits about whether the screed will curl, debond, crack, lift or crumble.
..Or whether the heights of the screed still allow waterproofing membrane termination heights and balustrade heights to meet Australian Standards.
..Or whether the correct threshold at the door entry is maintained.
This is why it is important to give the humble floor screed the respect it deserves and why designers should play a more proactive role in screed design to prevent defects.
Risks of Inadequate Screed Design
According to the Concrete Centre (UK), the main risks of failure for a bonded screed is debonding from the substrate. Bonded screeds generally rely on the substrate for strength, so debonding will likely lead to cracking and possibly displacement of the screed.
The main risk for unbonded screeds (also according to the Concrete Centre) is curling and lifting. Curling and lifting can disrupt the waterproofing and floor coverings, adversely affect falls, induce cracking in adjoining building elements and cause unevenness in floor surfaces.
Other risks or types of failure due to neglected screed design include:
- Excessive shrinkage cracking
- Expansion cracking and delamination of floor coverings
- Damage to waterproofing layers
- Abrasion and wear
- Insufficient falls and ponding water
- Non-compliant heights/levels of surrounding elements such as balustrades, door thresholds, etc.
- Overloading the supporting slab
- Excessive efflorescence leaching
- Differential movement cracking – such as rendered slab edges
Design-Related Causes of Screed Failure
Most screed failures are borne out of poor installation practices, according to Flowcrete. However, it is important that the specifier plays their role in undertaking the necessary due diligence to avoid design-related defects.
It is easy for the specifier to copy and paste a previous specification or to place the onus on the contractor to design the particulars of the screed. Unfortunately, the installer is not qualified to identify all the important design considerations unique to the project, nor does he probably care.
For example, it shouldn’t be up to the screed installer to determine the anticipated shrinkage in the substrate, deflection in a cantilevered concrete slab or if the screed happens to be in the only location that a 20-tonne crane can get access to replace the air-conditioning unit in 20 years. All of which could cause screed failure in the short, medium or long term respectively.
And thus, input is required from the professional designers as to how the screed should be designed to accommodate the anticipated service conditions of the specific conditions of the building.
So, let’s go through some common design-related causes of screed failure and how they can be avoided.
Inadequate strength is, in my experience, most commonly a result of excessive sand to cement and/or poor on-site mixing practices. It can lead to failure or crumbling under load once in service. I have personally been involved in projects where the finished screed was akin to a children’s sandpit.
It is prudent to clearly stipulate both the required mix design and a minimum nominal strength requirement in the specification.
Specifying a mix ratio will reduce confusion on site by not putting the onus on the installer to determine the mix ratio – he is not going to call his local concrete technologist to work out what the mix design should look like to achieve that strength. He will just make a best guess of what the mix ratio should look like and proceed undelayed. This can lead to problems as his best guess is most likely wrong.
Where tiles are to be used for standard residential applications, the mix ratio is quite often in the range of 1:3 to 1:5 (cement:sand). However, this range is often too weak and the final ratio is going to depend on the application, service conditions and floor coverings. For example, such a mix design is too weak and therefore not suitable for resilient flooring systems, as ARDEX points out in their technical bulletin Issues with Sand-Cement Screeds as Substrates for Non-Ceramic Tile Flooring Systems.
The strength requirement should be derived from anticipated worst-case service conditions. For example, is the screed going to have any plant traffic during maintenance for example? Or, is truck access needed over the screeded area during construction?
The specifier should start by identifying all use cases including during all phases of the building’s lifecycle: construction, operation and maintenance. Working backwards, the required screed strength can be identified.
Whilst there is clearly increased liability on the designer associated with stipulating a mix design, by doing the calculations properly upfront and not putting the onus on the contractor, it will save much heartache down the track.
If the finished screed bed is to act as a wear surface, then strength needs to consider abrasion resistance. Table 4.6 in AS 3600 provides the minimum strength required for abrasion resistance for a range of traffic types.
Remember that the top of screed bed is usually the weakest as the water and fine aggregate rises to the top. This should be considered in conjunction with water-cement ratio to determine the expected strength of the surface, not just the average strength of the whole screed.
Ensure any stipulations in the specification can be enforced on site during construction as part of the QA process. It is best to think the QA process through from the beginning as it will guide what needs to be included in the specification
For example, if a required mix design is stipulated, you may include a witness point in the specification to ensure you’re present on site at the time of mixing.
If a nominal strength is stipulated, you may include a requirement in the specifications that samples be taken randomly from the batch for independent testing. Or request that no floor coverings be applied until insitu hardness testing is undertaken such as Schmidt hammer testing.
It is also a good idea to specify the requirement for mixing to be undertaken by an electric mixer. This will minimise the chances of localised weak spots due to high sand concentrations, a common problem with sandy screeds.
Failure to control shrinkage
Shrinkage control is particularly important when dealing with a large area such as a rooftop deck, airport or factory floor.
Shrinkage occurs as water in the mix bleeds out of the cement paste and evaporates. Long-term drying shrinkage can continue for up to 3 or 4 years which can crack the screed and any floor coverings well after installation.
Lifting and curling of the screed occurs due to shrinkage differentials through the thickness of the screed. This is usually due to the top surface drying out faster than the layers below from evaporation. Consequently, the top surface layer contracts and lifts the edges of the screed.
One of the main ways a specifier can minimise the risk of shrinkage-induced damage is to specify a maximum water-cement ratio. This is because the higher the water-cement ratio – the more water that leaves the mix and the greater the volumetric change that takes place.
The maximum water-cement ratio will be dependent on a number of factors such as the aggregates used and the ultimate strength needed to be achieved.
Other methods of reducing the effects of shrinkage include:
- Specifying specific locations of contraction/control joints: AS 3958.1 recommends movement joints be placed at approximately 4.5 m intervals. However, this is quite general and in reality, contraction/control joints should be designed with some thought. For example, where there is an irregular shape or an opening in the screed, a control joint or several control joints may be required to relieve the stresses that develop at the corners.
The placement of joints is too important to be left to the installer. As the Tile Council of North America’s (TCNA) Handbook for Ceramic, Glass, and Stone Tile Installation states:
“The design professional or engineer shall show the specific locations and details of movement joints on project drawings.”
Which is further reinforced by Scott Carothers from the Ceramic Tile Education Foundation:
“… it is not the installer’s responsibility to design and locate these joints. That is to be done by the design professional or engineer.”
- Stipulating the use of reinforcement in the screed: Reinforcing can absorb tensile stresses associated with shrinkage to mitigate the effects of shrinkage and reduce shrinkage crack size. In fact, AS 4654.2 stipulates that any unbonded screed should be reinforced with mesh. Reinforcement may consist of fibreglass mesh or small gauge steel mesh.
- Fibreglass mesh, such as Domcrete’s AR Glass Scrim, or galvanised steel mesh is recommended so that it can be continued across joints and to the outer edges of the screed to mitigate curling, without the risk of corrosion in future.
Note: it is generally accepted that the use of fibre reinforcement may help with plastic shrinkage cracking but evidence suggests that is not effective for controlling long-term drying shrinkage.
- Stipulating minimum screed thickness: A thicker screed is more resistant to curling and lifting. Not to mention, generally more resistant to cracking and crumbling when put into service.
AS 3958.1 stipulates a minimum unbonded screed thickness of 40 mm, however, AS 4654.2 recommends a 50 mm minimum. A 40-50 mm screed is still prone to curling so, where possible, keep the minimum screed thickness well above this to assist in preventing curling and lifting - The Concrete Centre recommends 100 mm thickness for unreinforced unbonded screeds.
- Specify minimum curing requirements: Explicitly state that screeds are to cure (continuously wet) for a minimum of 7 days followed by 2 weeks of air-drying before tiling, as stipulated in AS 3958.
This does two things: (1) prevent rapid drying out, and (2) allow controlled shrinkage of the screed to take place prior to floor covering application, thereby reducing the risk of debonding.
- In some scenarios, a liquid-applied curing compound, such as Fosroc Concure WB30, may be applied to lock in moisture within the screed and reduce shrinkage. This may present problems if delayed shrinkage occurs due to the inability to dry out, so this approach must be assessed case-by-case.
- Specify shrinkage-reducing or water-reducing admixtures: In some circumstances, the risk of shrinkage cracking or curling may have unacceptable consequences on performance and durability. Consideration may be given to the use of admixtures, such as GCP Applied Technologies’ Eclipse 4500, which, for example, can more than half the amount of shrinkage experienced after 90 days.
Failure to control expansion
According to the U.S. Department of Transportation Federal Highway Administration concrete has a coefficient of thermal expansion in the range of 8-12 millionths per degree Celsius or 0.012 mm/m/ oC. Across a 30-metre deck, with a 38 degree temperature change, this would result in expansion of 13.68 mm.
Expansion-induced popping, cracking and lifting is a common problem particularly where porous ceramic or terracotta tiles are used on a sun-drenched deck which is also exposed to rain. Whilst concrete expands as a result of heat, tiles expand in two ways – from heat and moisture absorption.
According to Latham Australia’s article on expansion joint design, ceramic tiles can expand in the order of 0.004-0.008 mm/m/oC from thermal expansion. Across a 30-metre deck, with a 30 oC temperature change, this would result in expansion of 7.2 mm. However, it’s the moisture absorption that can actually induce the greatest amount of expansion in some cases.
AS 4459.10 nominates a limit of 0.06% moisture expansion for ceramic tiles. At this limit, the effect of moisture expansion is almost three times as much as thermal expansion during a 30oC temperature variance.
Nonetheless, AS ISO 13006 ambiguously states:
“The majority of glazed and unglazed [ceramic] tiles have negligible moisture expansion, which does not contribute to tiling problems if tiles are correctly fixed (installed). However, with unsatisfactory fixing practices or in certain climatic conditions, moisture expansion in excess of 0.06% (0.6 mm/m) can contribute to problems.”
This doesn’t make sense to me. Whilst, a lot of this moisture expansion may occur during storage and transportation, if a locally produced tile is used in a low humidity environment, it may not get the chance to fully expand in storage and transport. And, therefore, will expand whilst in service. This is particularly true for more porous tiles such as terracotta which can expand up to twice as much as ceramic.
In his paper “The Need for Establishing a Mositure Expansion Convention for the Analysis of Tiling System Failures”, Richard Bowman from the CSRIO suggests that a major cause of tiling pop-up or lifting failures is the combination of tile expansion and shrinkage of immature concrete slabs/screeds when tiling is expedited.
Let’s face it, most construction jobs are going to try expedite tiling as soon as possible after the screed goes down. So, for this reason, moisture expansion therefore not be shrugged off so easily as AS ISO 13006 is trying to do.
Other measures to reduce the risk of expansion-related defects include:
- Calculations should be undertaken based on the specific type of material being used and its ‘worst-case’ properties of moisture and thermal expansion. Consider the anticipated temperature variations in service and add an additional margin of error.
The TCNA’s Handbook for Ceramic, Glass, and Stone Tile Installation recommends joint width of four times expected movement.
At the very least, follow the recommendations of AS 4654.2 which recommends 10 mm expansion joints at 4.5 m intervals.
- It important that the joints are filled with “permanently deformable material”. There is no point having expansion joints if they are filled with non-compressible material. So, as the specifier, it is prudent to specify a filler or sealant material that has the compressibility to accommodate the calculated expansion.
- Furthermore, there is no point specifying expansion joints and joint filler material if the contractor doesn’t install them. Include a hold point in the specification to inspect the formed-up joints prior to screed placement.
More to come in Part 2..