Mapei Keraflex Maxi S1: Why Large Format Porcelain Stops Lipping on a Heated Screed

A 12mm porcelain slab measuring 1200x600mm weighs roughly 20kg and covers 0.72 square metres. Bond it over a cement screed carrying underfloor heating pipes and you are gluing a rigid plate to a substrate that expands and contracts every time the thermostat swings. Mapei rates Keraflex Maxi S1 as C2TE S1 under EN 12004, where S1 means transverse deformation between 2.5 and 5mm measured on the EN 12002 test rig. That flexibility band is what absorbs the differential movement instead of transferring it to the tile edge as a lip.

Lipping, the industry term for one tile edge standing proud of the next, is usually blamed on bad screeding or a careless fitter. On a heated slab the more common cause is a rigid adhesive that will not yield when the screed grows a fraction of a millimetre per linear metre. The slab tries to stay flat, the screed pulls underneath it, and the stress concentrates at the joint. An S1 mortar takes up that stress in the bond line.

Why the S1 deformation figure matters more than the bond strength number

Most porcelain adhesives on a builders merchant shelf already exceed the C2 bond threshold of 1 N per square millimetre. Keraflex Maxi S1 hits that comfortably. The number that separates it from a standard C2TE mortar is the S1 classification, and on underfloor heating that single letter code decides whether the installation survives thermal cycling.

EN 12002 loads a hardened mortar prism until it bends, then records the deflection at the point of transverse fracture. A rigid C2 mortar deflects less than 2.5mm and is classed as no deformability. S1 sits in the 2.5 to 5mm band. S2 exceeds 5mm. For a 6 metre run of screed heated from 18 to 45 degrees, cement expansion of around 10 microstrain per degree produces movement in the region of a millimetre across that length. A rigid bed passes almost all of that into the ceramic. An S1 bed elastically shears within its own thickness and the tile stays put.

The consequence shows up at the perimeter first. Slabs near a large window or an external door see the steepest temperature gradient and the biggest movement. Fitters who report lipping appearing only along the south wall three months after handover are usually describing an adhesive that had adequate grab but no give.

The uncoupling mat argument, and where it stops helping

Schluter Ditra and similar polyethylene uncoupling mats promise to break the bond between screed and tile so movement in one layer does not reach the other. On a heated screed with large format porcelain the mat introduces a second variable that many specifications ignore.

Ditra is 3mm thick and demands its own thinset in the dimpled cavities plus a bed on top. Two bond lines now sit under a 20kg slab. If the fitter beds the mat with a rigid mortar underneath and an S1 above, the weak layer governs. Worse, a 1200mm slab spans dozens of the mat dimples, and any air trapped in an unfilled cavity becomes a hollow spot that a heavy foot or a dropped pan can crack straight through. The uncoupling principle works well for 300x300 mosaics and awkwardly for oversized panels.

There is a specific case where the mat earns its place: a screed with known residual moisture or a new build slab that has not fully completed its drying shrinkage. Here the mat buys time by decoupling early movement. On a mature, correctly dried screed the S1 adhesive alone handles the thermal component, and the extra 3mm of buildup plus its own failure modes is difficult to justify. The decision hinges on the age and moisture state of the screed, not on the tile size.

Getting the screed flat before any adhesive opens

A 1200mm edge is unforgiving. A 2mm hollow under a 300mm tile disappears into the adhesive comb; the same hollow under a 1200mm slab becomes a visible ramp at the joint and a drum-sounding void underneath.

Ardex A45 is a rapid-setting repair and levelling mortar that takes foot traffic in around 60 minutes and can be feathered from a hard edge to around 30mm in a single pass. On a heated screed it fills local depressions before the main levelling compound goes down. For the finished plane, a self-smoothing compound poured to at least 3mm and checked with a 2 metre straightedge should show no gap over 3mm under the rule. Anything worse and the large format tiles will telegraph it.

The order of work matters on underfloor heating. The screed must complete its commissioning heat cycle, ramping up over several days and back down, before levelling. Pour a smoothing compound onto a screed that has never been heated and the first commissioning cycle after tiling can crack both the compound and the bond above it. Fitters who skip the pre-heat to save a week are the ones who return to lifted tiles.

Back-buttering is not optional at this size. The manufacturer solid-bed requirement for exterior and heated floors is effectively 100 per cent coverage. That means combing the floor with a 10 or 12mm notch trowel and skimming a thin coat onto the tile back, then laying and beating in with a rubber mallet or a vibrating levelling tool until the ribs collapse and fill. A slab laid on ribbed adhesive with 60 per cent coverage has voids that flex, and flex at a joint is lipping in waiting.

Open time and the 20 minute mistake

Keraflex Maxi S1 has a rated open time that shortens sharply on a warm slab. Comb out more than about a square metre ahead of yourself on a floor already at commissioning temperature and the ribs skin over before the tile lands. A skinned rib will not wet the back of the porcelain and you get a dead spot. Test by dragging a finger across the combed bed; if adhesive no longer transfers to your skin, scrape it off and recomb.

Levelling clips versus a good eye

Tile levelling systems from Raimondi or the cheaper generic wedge-and-clip kits hold adjacent edges in the same plane while the adhesive cures. On 1200x600 porcelain they are close to mandatory, because no amount of manual adjustment holds a 20kg slab dead flat against a neighbour for the hours an S1 mortar takes to firm up.

The clip works by clamping under both tile edges and forcing them to a common surface as the wedge drives home. Space them at roughly 200 to 300mm along every joint, more densely on the long edge where a slight bow in the slab does the most damage. Snap the clips off after 24 hours by kicking parallel to the joint, never across it, or you risk lifting a corner that has not reached full grab.

Clips correct for tile bow and minor bedding differences. They do not correct for a hollow under the centre of a slab, which is why solid bedding comes first and levelling clips second. Fitters who rely on clips to pull down a slab floating on ribbed adhesive create a joint that is flush on day one and drummy by month three, because the unsupported centre keeps working against the clamped edge until something gives.

There is also a calibration issue the clip cannot fix. Rectified porcelain is ground to a tight dimensional tolerance, but slabs from different production batches can vary by up to a millimetre in thickness. Mix batches across a floor and the levelling system fights a size difference it was never meant to absorb. Checking the batch number on every box before the first tile goes down costs nothing and prevents a plane that no clip can rescue.

Efflorescence on the grout after the first heating season

A white bloom appearing in the grout joints after the underfloor heating runs for a few weeks is efflorescence, soluble salts drawn to the surface as moisture migrates up through the screed and evaporates. On a cement grout over a heated slab it is common in the first season and largely cosmetic. A dry brush lifts the loose deposit; a proprietary efflorescence remover based on a mild acid dissolves the bound crystals, though acid on a limestone or marble insert will etch the stone and must be kept off it entirely.

The deeper question the white bloom raises is whether the screed was fully dry before tiling. A residual moisture reading above the accepted threshold for heated cement screeds, commonly cited around 1.8 CM per cent by carbide hygrometer, means the salt supply keeps coming and the bloom returns after every cleaning. If that is the source, the fix is not the acid but time and heat to finish the drying that should have happened before the adhesive opened. Which leaves the awkward possibility that a floor which looks finished, grouted, and flush is still telling you it went down too soon.