Last updated: September 2026

Short answer: A steam room vapor barrier is the continuous low-permeance layer that sits on the warm side of the enclosure — behind the tile bed, in front of the insulation — and its only job is to stop water vapor diffusing into the construction where it will condense. It almost never fails in the middle of a sheet. It fails at 6 crossings: fittings, light penetrations, the steam inlet, the extract duct, the door frame and the floor junction.

Every one of those 6 is a hole somebody had a good reason to make. A steam room vapor barrier is therefore not a material decision; it is a sequencing decision, and it is lost on the day a second trade arrives after the membrane is signed off. This guide sets out where the layer gets cut, how each crossing is detailed, how continuity is inspected before the mosaic goes on, and what the failure looks like months later from both sides of the wall. We build commercial steam rooms as a turnkey contractor, so these are crossings we detail, seal and sign off ourselves.

What is a steam room vapor barrier, and what does it have to do?

It is 1 continuous layer of very low water vapor permeance, fixed to the warm face of the enclosure before any insulation, screed or tile goes on. Its job is diffusion, not water. It stops vapor migrating through the build-up and condensing on the first cold surface it reaches inside the wall.

That distinction gets lost on site constantly. A tanking membrane keeps liquid water out of the substrate. A steam room vapor barrier keeps vapor out of the construction. The two layers can be the same product in a well-chosen system, but they are two separate performance claims, and only one of them is tested against vapor.

The layer also has to survive the rest of the programme. It is installed early, then walked on, drilled through, leaned against and worked over by the tiler, the electrician and the mechanical fitter. A barrier that is perfect on the day it is installed and unprotected for the next 6 weeks is not a barrier.

Why does a steam room drive vapor harder than any other wet room?

Because the pressure difference is roughly 7 times a normal bathroom. At 45 °C and saturation, the internal vapor pressure is about 9.6 kPa. A corridor at 22 °C and 50% relative humidity sits at about 1.3 kPa. That leaves about 8.3 kPa pushing outward through every square metre of the enclosure.

A domestic shower produces a comparable peak for perhaps 10 minutes a day and then the room dries. A hotel steam room runs a saturated atmosphere for 12 to 16 hours a day, every day, for the life of the facility. The load is not the peak; it is the integral of the peak over years.

That is why detailing that is perfectly adequate in a bathroom is not adequate here. A shower enclosure forgives a pinhole because the drive is intermittent and the wall dries back between uses. A steam room never dries back. Whatever gets through in the morning is still in the insulation at night, and the next morning adds to it.

Where does the steam room vapor barrier actually get cut?

In 6 places, in roughly this order of frequency: mechanical fixings and pipework for showers and benches, light fittings, the steam inlet, the extract duct, the door frame, and the junction between wall barrier and floor tanking at the gully. Every one is a deliberate hole made by a trade with a legitimate reason.

Notice what is not on that list: the flat field of a wall. Sheet membranes fail at laps and terminations, not in the middle. That is the single most useful thing to know when you are standing in a room deciding where to look, and it is why a penetration register is worth more than a moisture meter.

The pattern repeats across rooms that look nothing alike. It is the same in a compact residential cabin, in a hotel wet floor with a Vichy shower next door, and in a full Turkish hammam where the heated marble adds its own set of crossings. The geometry changes; the list of 6 does not.

How should a pipe or fitting penetration be detailed?

With a bonded collar from the same system as the sheet, not with sealant. The pipe passes through a sleeve, the sleeve is bonded to the membrane with a preformed collar, and the collar is dressed a minimum of 50 mm onto the flat of the sheet in all 4 directions before anything else covers it.

Silicone applied on the face of the finished tile is the commonest substitute and the commonest failure. It is a surface treatment on the cold side of the problem. Vapor does not care what the tile face looks like; it moves through the bed and around the fixing, and the sealant is nowhere near the plane it would have to seal.

The rule that catches most of it on site is simple. Nothing is fixed through a completed steam room vapor barrier after sign-off. If a bench bracket, a rail or a hand shower is added later, the barrier is opened again and re-collared, and that work is inspected again. There is no such thing as 1 small extra hole.

What goes wrong at the light fitting and the fibre-optic ceiling?

A starlight ceiling is treated as 1 crossing when it is detailed properly and as 200 crossings when it is not. The fibre tails should be gathered and taken through a single sealed gland plate bonded to the barrier, with the fibres distributed on the warm side of the plate, inside the room build-up.

Drilling the membrane once per point is the failure mode, and it is almost invisible afterwards. Each hole is a couple of millimetres. Two hundred of them, unsealed, add up to more open area than the extract duct, and they sit at the highest and hottest point in the room where the vapor drive is strongest.

Recessed downlights have the same problem in a coarser form. A fitting that needs a 90 mm cut-out needs a bonded collar around the cut-out and a housing rated for the room, and the cable entry into the housing is its own separate crossing. Cove lighting is easier only because the barrier can usually run behind the whole cove uncut.

How do the steam inlet and the extract duct cross the barrier?

Through 2 collars with different requirements. The steam inlet carries a pipe well above 100 °C, so its collar has to be rated for that temperature and the pipe has to fall back towards the generator so condensate drains rather than pooling. The extract is the biggest single hole in the room, commonly 100 to 150 mm.

The extract crossing deserves a flanged sleeve bonded to the membrane, not a duct pushed through an oversized hole and foamed. Foam is not a vapor control layer, it shrinks, and it hides the gap it leaves behind. We size the sleeve so the duct is a close fit and the flange has a clean 50 mm bonding face all round.

The steam generator itself sits outside the room, and it is not our product — Condair and its peers make those units and any certification on them belongs to the manufacturer. What is ours is the crossing: the sleeve, the collar, the fall on the pipe and the sign-off that says the membrane was intact when the pipe went through it.

Why is the door frame the hardest junction in the room?

Because the barrier has to be lapped onto the frame on 3 sides before the frame is set, and once the frame is in, that lap can never be made. The reveal is also the coldest part of the enclosure, which means it is where condensation appears first if the barrier stops short of the frame.

The sequence that works is: barrier dressed into the reveal and around the corner onto the face of the opening, frame set into the finished opening and bedded against the dressed barrier, collar or tape applied to the frame-to-barrier junction, then tiling. The sequence that fails is frame first, because it feels faster on the programme.

Glass steam room doors add a second issue. The frame is usually fixed with mechanical anchors through the tile bed into the substrate, which is 6 to 10 fixings straight through the plane you have just sealed. Each one gets a collar or a sealed anchor, and each one gets recorded. This is the junction we photograph most.

How does the barrier meet the floor tanking and the gully?

With a lap of at least 100 mm, wall over floor, so gravity works with the detail rather than against it. The floor tanking runs up the wall 150 mm or more, the wall barrier comes down over it, and the gully gets its own bonded flange that is clamped, not just taped, into the floor layer.

Getting the overlap the wrong way round is a real and common error. Floor tanking dressed up over the wall barrier gives you an upward-facing lap edge that any water running down the wall can find. Nothing fails immediately; the screed simply stays damp and the substrate never dries.

The gully flange is the part that gets rushed, because it is the last thing done before the screed and the tile setter is waiting. A clamped flange is the only version of this detail we accept. Adhesive alone under a drain body that thermally cycles from 20 °C to 46 °C twice a day does not stay bonded.

Which membrane class does a steam room vapor barrier need?

Class I, which means a water vapor permeance of 0.1 US perm or less measured to the ASTM E96 dry cup method. Converted into the metric form used across Europe, 0.1 perm is roughly an equivalent air layer thickness of 34 m. Steam-grade sheets are commonly specified far tighter than that.

The conversion is worth doing once and remembering. Permeance of 0.1 perm is about 5.7 ng/(s·m²·Pa); dividing the water vapor permeability of still air, about 1.94 × 10⁻¹⁰ kg/(m·s·Pa), by that figure gives roughly 34 m of equivalent still air. A sheet sold as sd 100 m is therefore about 3 times tighter than the Class I threshold.

We specify by the metric value because the products we buy in Europe are declared that way, and because a single number makes substitution arguments short. If a proposed alternative does not declare an sd value, it has not been tested for what we are asking it to do, and the answer is no.

Is tile waterproofing the same thing as a vapor barrier?

No, and assuming otherwise is behind a large share of the failures we are asked to look at. A bonded waterproof membrane for tile is tested for liquid water performance under tile, to standards such as ANSI A118.10. That test says nothing about how many nanograms of vapor pass per second per square metre.

Plenty of products do both jobs, and the good system suppliers publish both figures. The error is not in using a tile membrane; it is in assuming that a product with a waterproofing certificate automatically has a vapor figure. Ask for the declared permeance or sd value. If the data sheet has one, the question is settled in 30 seconds.

The same logic applies to boards. A foam construction board carries its own facing, and how the boards are jointed decides whether the assembly is continuous. The board is a substrate; the continuity is in the joint treatment, and that is a separate line in the specification.

How do you inspect continuity before the mosaic goes on?

With a written register, not a walk-round. Before tiling, every crossing in the room is numbered, photographed close up, and signed off individually. A room with 2 benches, a hand shower, a starlight ceiling, a steam head, an extract and a door typically produces 20 to 30 numbered entries.

The register does two things a visual check cannot. It forces somebody to count the crossings, which is how forgotten ones are found, and it produces a photographic record of a layer that will be invisible for the next 20 years. When a room misbehaves in year 3, the register tells you where to open the wall.

Water testing still has its place for the floor, and we flood-test to 24 hours. But a flood test is a liquid test. It will happily pass a floor whose vapor continuity is broken, which is precisely why the two checks are recorded separately in our handover file rather than folded into 1 line.

What does a failed steam room vapor barrier look like from inside?

The first sign is a room that stays wet after it has gone cold. Mosaic over a saturated bed holds moisture for hours, so the surface still reads damp 3 or 4 hours after shutdown while a sound room is dry. The grout in that area darkens permanently rather than drying to its original tone.

Then the mechanical symptoms arrive. Tapping the tile gives a hollow note where the bed has debonded, usually starting within 1 m of the crossing that is leaking vapor. Individual tesserae lift at the edges. On heated benches and seat noses the movement is faster because those areas cycle through a wider temperature range.

People often blame the tiler at this stage, and it is almost always wrong. A bed that has been wet continuously for 2 years will debond regardless of who laid it. The question worth asking is not who tiled it but which of the 6 crossings is feeding it, and the register answers that in minutes.

What does the same failure look like from the corridor outside?

Usually earlier, and usually somewhere nobody connects to the steam room. Paint blistering on a corridor wall 2 or 3 m from the door, a tide mark on plasterboard, a musty smell in the ceiling void, or insulation that has slumped in the cavity and left a cold strip you can see on a thermal image.

That distance is what makes the diagnosis hard. Vapor that gets past the barrier condenses at the first cold surface, runs down inside the construction, and shows itself wherever the construction lets it out. The visible damage can be metres from the hole, on the other side of a wall, or a floor below.

This is the failure mode that turns into a dispute, because by the time it shows, three or four trades have left site. It is also the reason the crossings are photographed. A dated photograph of every collar, taken before the tiling, converts an argument into a 10 minute file check.

What does a repair cost in programme, and what does it touch?

Assume a minimum of 2 m² of mosaic removed per crossing, and 3 to 4 weeks of room closure for anything more than a single fitting. The tile comes off, the bed comes off, the barrier is opened back to sound material, the crossing is re-collared, and the whole area is rebuilt and left to set before tiling.

Matching the mosaic is the part that surprises clients. A tessera from the same range, 5 years later, is rarely the same batch, and on a wall of graded colour the patch reads as a patch. On several jobs the honest answer has been to retile a whole elevation rather than accept a visible repair.

Compare that with the cost of the collar itself, which is a consumable and 10 minutes of labour. There is no other detail in a wet spa where the ratio between doing it right and putting it right is so lopsided, and that ratio is the whole argument for the register.

What belongs in the specification so the barrier survives the site?

5 clauses, and they take a page. Name the required permeance or sd value. Require all collars, tapes and boards from 1 system. Require a numbered penetration register with photographs. Make the barrier a hold point that is signed off before tiling. Forbid any fixing through the barrier after that sign-off without a re-inspection.

Those 5 clauses are cheap to write and they move the decision to the only time it can be made cheaply — before anyone is on site. They also give the main contractor something enforceable, which is worth more than a drawing note, because the trades that cut the barrier are usually not the trade that installed it.

If you are specifying a wet spa now, the same discipline carries into the sauna cabins and across the rest of the suite, and it is the first thing we set out in our turnkey spa contracting services. For projects in the Gulf, where the outside air is itself humid for much of the year, our spa contracting work in Dubai follows exactly this register. Send your steam room drawings to our design office before the membrane is ordered.

Who is writing this

Sauna Dekor has been building wellness and thermal facilities since 1987 and is now in its 40th year, manufacturing in its own Istanbul facility with a team of 19, working under TS EN ISO 9001:2015, with projects delivered in more than 35 countries. On steam rooms we work turnkey: substrate, vapor barrier, tanking, tiling, fit-out and commissioning.

What we do not make is equally clear. The steam generators, the control panels, the heaters and the lighting drivers are not our products; they come from the equipment manufacturers we buy from and work alongside, and any certification on that equipment belongs to its manufacturer rather than to us. The TS EN ISO 9001:2015 certificate is ours; the rest are theirs. We built the sauna, steam rooms, spa hot tub and pool areas at the Diplomatic Club in Doha, which is where the crossings photographed above were detailed.

Frequently asked questions about the steam room vapor barrier

Does a steam room vapor barrier go on the inside or the outside of the insulation?
On the inside, meaning the warm room side, in front of the insulation. Putting it on the cold side traps vapor in the insulation instead of excluding it, and the room will then run 2 or 3 degrees cooler than the setting for no obvious reason.

Can a vapor barrier be added to an existing steam room?
Only by stripping back to the substrate. The layer sits under the tile bed, so there is no version of this work that leaves the finish in place. On an existing room we quote the strip-out first and let the client decide before any further design work.

Is foil-faced insulation enough on its own?
Rarely. The foil facing can reach the right permeance, but the continuity is in the joints between boards, and a taped joint on foil is the weakest part of the assembly. Treat the facing as 1 component of a system, never as the whole barrier.

How many penetrations is too many for one steam room?
There is no fixed limit, but a room that needs more than 30 recorded crossings usually has a services layout that could be simplified. Grouping cable entries into a single gland plate often removes 10 or more crossings before anything is built.

Does a residential steam room need the same detailing as a hotel one?
The detail is identical; the duty cycle is not. A home room may run 4 hours a week and a hotel room 90, so a weak crossing takes far longer to show at home. The crossing is still wrong, and it still has to be collared.

Who should sign off the barrier on site?
Whoever carries the defect risk for the room, which on our projects is us. The hold point should sit with a single party that inspects, photographs and dates every crossing, because shared responsibility for this layer reliably produces no responsibility at all.

Sources

• Building Science Corporation, BSD-106: Understanding Vapor Barriers — defines Class I as 0.1 perm or less, Class II as 1.0 perm or less, Class III as 10 perm or less, measured by ASTM E96 Test Method A.
• ASTM E96/E96M-24, Standard Test Methods for Gravimetric Determination of Water Vapor Transmission Rate of Materials, the dry cup and wet cup methods behind every declared permeance figure.
• BS EN ISO 13788:2012, Hygrothermal performance of building components and building elements — internal surface temperature to avoid critical surface humidity and interstitial condensation, calculation methods, published 31 January 2013, 52 pages.
• BS EN 15026:2007, Hygrothermal performance of building components and building elements — assessment of moisture transfer by numerical simulation, published 31 August 2007, 28 pages, for cases the simplified method cannot carry.
• BS EN 13984:2013, Flexible sheets for waterproofing — plastic and rubber vapour control layers, definitions and characteristics, published 31 March 2013, the product standard for the sheet itself.
• Tile Council of North America, ANSI standards list, including A118.10 for load bearing bonded waterproof membranes (2023) and A108.13 for their installation.