Floor screeds and finishes

A floor screed is not intended to act as a wearing course, so a suitable floor finish, offering both decorative and wearing capabilities, is an essential part of a finished floor. Failures can of course occur in the finishes themselves (due to incorrect or inappropriate use, workmanship or product failure) but often the root cause of failure is with the substrate itself. While flexible finishes may have some measure of tolerance to cracking or even curling, rigid finishes (such as tiles) or in-situ finishes (such as terrazzo) are vulnerable to deterioration due to defects in the substrate.

Possible defects include:

• rippling of flexible floor finishes over crack lines;
• damage to flexible floor finishes over movement joints;
• blistering of flexible, impervious floor finishes;
• osmosis in applied seamless floor systems; and
• disruption and delamination of rigid flooring materials due to substrate shrinkage.

Rippling of flexible floor finishes

Curling of unbonded floor screeds is a prevalent problem. It can occur some time after initial laying and so it is quite possible that the flexible floor covering could have been laid over an area that was either prone to curling, had started to curl or was fully curled. In the latter case, the problem may have been masked by the application of a latex screed or levelling compound prior to floor laying.

Given the time needed for a screed to dry out fully, the provision of an impervious floor covering could seal residual moisture into the slab and screed, causing a redistribution of the moisture content throughout its thickness. Curling occurs because of a differential moisture content between different surfaces and so if moisture is redistributed, the curling effect may be reversed. If this happens, the floor covering will be slightly wider than is needed; it cannot be compressed back to make it narrower and so it will remain elevated as a ridge. Repeated trafficking will cause the ridge to wear and perforate, leading to more extensive deterioration.

Similarly, if the floor was flat prior to floor laying, the curling of the screed would stretch the floor covering (assuming it was bonded). Subsequent reversal may be accommodated depending on the qualities of the covering, but it may not return to its previous width, leading again to the formation of a ridge.

You might expect that thermal movements in a substrate could create similar rippling problems, although there seems to be slightly more uncertainty as to this. Early work by BRE suggested that thermal movements did not replicate the problem, although more recently there appears to be more of an acceptance that thermal or normal diurnal movements could have an effect.

Lifting of rigid floor finishes

Rigid floor surfaces, such as ceramic or stone tiles and terrazzo, are often bedded in adhesive with grouted joints that contain little, if any, flexibility. So, if the substrate expands, cracking may occur in the grouted joints. However, rather more disruption can occur where the substrate shrinks significantly after the tiled finish has been laid - perhaps where for reasons of programming the tiles were laid before full drying shrinkage had taken place.

Depending on the measure of adhesion between the tiled surface and the screed, shrinkage of the latter will exert compression forces on the edges of the tiles. If the compression forces can overcome the adhesion force, the pressure will be relieved by the tiles suddenly popping up - often along the line of a daywork joint or formed joint in the screed. This can happen suddenly, several years after completion, although the most usual period is about a year after completion. Long term expansion of the tiling can produce similar effects.

Blistering of thin sheet flooring

The appearance of bubbles or blisters in sheet flooring materials is often a sign of moisture related problems:

• lack of an effective damp-proof membrane (DPM);
• water trapped in the floor screed; or possibly
• leaks from service pipes buried in the screed.

Occasionally, blistering can be due to lack of sufficient or patchy adhesive application during laying, or failure of the adhesive due to incompatibility with the substrate or flooring.

A loss of adhesion in bonded floor coverings means that in trafficked areas the sheet gradually stretches and eventually works up into a raised area or blister, which then wears rapidly and perforates. The presence of moisture beneath the sheet can affect certain types of adhesive, gradually causing them to emulsify and lose their effectiveness.

A frequent cause of this type of problem is a failure to allow sufficient time for a screed and floor slab to dry out before the floor covering is laid. Usually coverings should not be laid until the relative humidity of the slab is at or below 80%. For a ground bearing floor slab, the existence of a dpm beneath the floor could restrict the drying out of the slab, with the result that an impervious floor covering effectively seals the moisture in the floor construction.

Case study: supermarket

The premises comprised a small supermarket about 2 years old. Within a year of opening the floor coverings in the vicinity of the refrigerated display cabinets began to lose adhesion and display excessive wear. The resultant perforated coverings were a health and safety risk and needed to be replaced. The client was concerned that the blistering was the result of a failed damp-proof membrane or leaking service pipe.

Following removal of the damaged floor coverings, the screed surface was exposed and a series of hygrometer probes were inserted into holes drilled into the screed to determine its relative humidity. Also, the floor was cored to determine its overall composition and the existence of a dpm. The floor was found to be a ground bearing slab 300mm thick with a 75mm cementitious screed. Initial relative humidity (RH) readings were in excess of 95%.

Over a period of 3 months the RH of the floor screed gradually fell as the water was free to evaporate from the surface. After 4 months the RH had dropped below 80% and the floor was successfully relaid.

In this case, the considerable thickness of floor slab plus screed compounded the problem of moisture retention after construction. The provision of the impervious floor covering meant that evaporation could not occur and the floor coverings had been laid too soon. However, the luxury of 4 months drying time was unusual. In most cases operational demands mean that remedial measures must be put in place far earlier.

In this example, alternative approaches might have been appropriate to restore the damaged floor in a much quicker time. A water resistant adhesive might have been simple, but a more effective approach would have been to have applied a surface damp-proof membrane to the entire floor, covering it with a smoothing compound and then re-laying the impervious floor covering.

Case study: retail unit

The premises comprised a retail unit with a ground floor staff room at the rear. Staff complained of musty smells. When the floor coverings were taken up they were found to be very wet. The staff room was in an extension constructed about 10 years previously. A conventional cementitious screed had been laid over a concrete ground bearing slab and polythene dpm.

A series of hygrometer probes were inserted and left to stabilise for 72 hours before RH readings were taken. They revealed an RH in excess of 95%. Repeat measurements were taken every 2 weeks but these revealed little improvement and in 2 cases a worsening of the situation.

Given the lack of improvement it became apparent that the problem was not simply due to trapped moisture dating from construction. Taking into account the thickness of the slab it was felt unlikely that the dpm was at fault. Comparison of RH readings in the floor with RH readings of the ambient air failed to establish a correlation suggesting that a further source of moisture was at fault.

As an experiment, copper flow and return pipes serving the radiator in the staff room were isolated. The pipes passed through the thickness of the screed and could possibly have been pinholed. Over the following weeks the floor RH readings progressively fell indicating that a failure of the pipe had indeed occurred. Once the defective pipe had been replaced and the RH had dropped to below 80% the floor covering was replaced.