Mapei Keraflex Maxi S1 vs Ultralite S2 for Bonding 60x120 Porcelain on Underfloor Heating

A 60x120 cm porcelain tile has a face area of 0.72 m2 and, in a 9 mm body, weighs roughly 15 to 17 kg. Bonded over a heated cementitious screed that swings from around 15C at plant shutdown to 40 to 45C at flow-temperature peak, that tile expands and contracts against an adhesive bed on every cycle. The metric that decides whether the bond survives is the deformability class under EN 12004 and EN 12002: S1 (deformable, transverse deflection 2.5 to 5 mm) or S2 (highly deformable, above 5 mm). Keraflex Maxi S1 is a C2TE S1 cementitious mortar. The Ultralite variant in Mapei’s range carries an S1 rating as well in most published data, so the honest comparison for a genuine S2 requirement is Keraflex Maxi S1 against a true S2 product such as Mapei Ultralite S2 or Keraflex Easy S2, not against the S1 Ultralite.

Where the S1 versus S2 line actually sits

Heated screeds move. A sand-cement screed at 45 mm thickness over 100 m2 undergoing a 25C to 30C differential shifts several millimetres across its length, and every restraint (perimeter, doorway, column) concentrates that strain. An S1 adhesive absorbs transverse deflection up to 5 mm before it transfers load into the tile edge as shear. For 60x120 format the failure mode is rarely the tile cracking; it is the bond letting go at a corner, producing a hollow that telegraphs as a drummy sound underfoot.

S2 buys headroom above 5 mm deflection. That headroom matters on anhydrite (calcium sulfate) screeds, on screeds where the heating was commissioned late, and on assemblies stacking multiple restrained edges. The trade is real: S2 mortars cost more per bag, cure with a softer bond line, and on a rigid, well-cured, fully commissioned sand-cement screed the extra deformability sits unused. For a standard C2TE S1 bond over a properly heat-cycled sand-cement base, Keraflex Maxi S1 carries the load the assembly will realistically impose.

The uncoupling membrane changes the calculation

Install a polyethylene uncoupling membrane, Schluter DITRA-HEAT or Dural Durabase, and the adhesive question shifts. The membrane’s cavity structure absorbs in-plane movement between screed and tile, so the mechanical demand on the bonding mortar above the membrane drops. On a DITRA-HEAT assembly the substrate movement is decoupled at the membrane, and an S1 mortar bonding the porcelain to the membrane fleece is the specified and sufficient choice.

The critical detail on any uncoupling assembly is the anchoring layer beneath the membrane, which is where trowel selection earns its keep. Schluter specifies full contact between membrane fleece and its setting mortar, verified by lifting a corner during the first square metre and reading the transfer pattern. A skinned-over ridge that a tile presses onto but never wets out leaves a void, and that void, not the mortar’s deformability class, is where a heated assembly fails first. Membrane or no membrane, the coverage on the underside of a 60x120 tile is the number that predicts the drummy corner two heating seasons later.

Yield, trowel, and why Ultralite exists

Ultralite’s selling point is mass, not chemistry. A 20 kg bag of Keraflex Maxi S1 Ultralite covers roughly the same area as a 25 kg bag of standard Keraflex Maxi S1, because the lightweight fillers cut density by around 25 percent. On a large job the back-and-knee arithmetic is concrete: fewer kilos carried up a stairwell, more square metres per bag lifted. The mortar also buttes up more readily under a large tile, which matters when you are back-buttering a 0.72 m2 face and combing the substrate in the same pass.

Both products need a large-notch trowel for this format. A 12 mm or 15 mm half-round or a 20 mm slant-notch, combed in one direction, then the ridges knocked down by laying the tile with a firm sliding motion perpendicular to the ridges, is the method that achieves the 90 to 95 percent contact a heated large-format bond requires. Under-combing at 6 mm is the single most common cause of hollows under 60x120 tile, and no deformability class rescues a bed that only touched 60 percent of the porcelain.

Open time is the other variable. Keraflex Maxi S1 gives around 30 minutes at 23C and 50 percent humidity, shrinking sharply on a warm screed. Comb no more than one tile ahead on a floor already carrying residual heat from commissioning, and skin-test every ridge with a fingertip before laying.

Levelling clips do not fix a soft bed

MLT Lash or Raimondi levelling systems pull adjacent 60x120 edges flush and hold them through initial set, which is close to mandatory at this format where a 1 mm lip reads as a shadow line under raking light. The clips manage lippage. They do not compress air out of an under-combed bed, and they do not compensate for a deformability class the assembly needed and did not get.

A worked coverage check on 40 square metres

Take 40 m2 of 60x120 porcelain over a DITRA-HEAT membrane. At 90 percent target coverage with a 15 mm trowel, bed consumption runs about 4.5 to 5.5 kg/m2 of mixed mortar, so budget 180 to 220 kg of mixed Keraflex Maxi S1 for the bonding layer alone, before the membrane’s own anchoring coat. That is nine 25 kg bags standard, or roughly the same coverage from eleven or twelve 20 kg Ultralite bags, the Ultralite trading bag count for lift weight.

Back-butter every tile with a thin skim regardless of trowel size on the floor. The skim wets the porcelain’s low-porosity face and lets the combed ridges key into it, and skipping it on a dust-glazed factory surface is why tiles that felt solid at lay-up sound hollow by the second heating cycle. Verify by lifting the third and thirtieth tile and reading whether the transfer covers the corners, because the corners debond first and the middle almost never does.

Grout haze on a matt-textured face

Matt and structured 60x120 porcelain traps cementitious grout in its micro-texture, and a haze that wiped clean off a polished tile at 20 minutes will set into a matt surface as a dull grey film. Sponge the initial clean-up while the grout is still plastic, water changed every few metres, then return after full cure with an acid-based haze remover such as Fila Deterdek or Mapei Keranet diluted per the surface, worked with a white nylon pad, never a green scour. Rinse twice. On very fine structured porcelain a first grout pass in a shade close to the tile hides any residual film that a darker contrast grout would advertise across the whole floor.

The deformability decision and the haze decision are unconnected, but they fail in the same place: the assembly reads as finished at handover and reveals its shortcuts one heating season in, one at the drummy corner and one as a grey bloom that no longer sponges away. Whether a genuine S2 mortar was worth its premium over Keraflex Maxi S1 is answerable only against the specific screed’s commissioning record and restraint map, and that record is the document most large-format installations never actually consult before the first bag is opened.