Bubbles or Blisters Across Your Flat Roof Membrane? What's Trapped Underneath

Quick Answer: Those raised bubbles on a flat or low-slope commercial roof are almost always pockets of trapped air or moisture vapor sealed under the membrane during installation, or pulled in later through a weak seam. Sun-driven heat makes that trapped gas expand, and it pushes the membrane up wherever the bond underneath is thin or missing. A small blister that stays the same size season after season is usually cosmetic — a bubble of air with nowhere to go. One that grows, spreads, or feels soft and spongy under your boot is sitting over wet insulation, and that moisture doesn't sit still. It spreads sideways under the membrane, and Connecticut winters make it worse every single year it goes unaddressed.

A facilities director at a distribution center off I-95 walked her roof in early June with the same clipboard she carries every spring. Drains clear. Flashing tight. Then she stopped at the northeast corner, where a cluster of raised bumps had shown up across a six-foot stretch of TPO that had been dead flat during last year's walkthrough. Nothing was dripping below. No stain on the ceiling tile. Just bubbles — some the size of a silver dollar, one nearly the size of a dinner plate.

She called her roofer half expecting to hear it was nothing. It wasn't nothing. It also wasn't the five-alarm fire she'd braced for. It sat somewhere in between, and figuring out exactly where takes more than a look from the parking lot.

That gap between "harmless" and "call in the crew" is where almost every blister inspection lands. And it's not a judgment call you want to make from the ground.

A membrane blister forms when something gets trapped between layers — the membrane and the insulation below it, or between plies within the membrane itself on a built-up or modified bitumen system. Three things put it there in the first place.

The first is moisture sealed in during the original install. If a roof gets covered over a substrate that's even slightly damp, or the crew works through humid conditions without giving materials time to dry, that moisture has no way out once the membrane goes down. It sits there for years, waiting for enough heat to move.

The second is poor adhesion. Adhesive applied too thin, too cold, or over a surface with dust and debris on it never bonds fully. That leaves microscopic voids — pockets where the membrane is technically attached but not really attached. Air finds its way into those voids over time, whether through a hairline seam gap or just atmospheric pressure working its way in.

The third is the sun. On a July afternoon, a dark TPO or EPDM surface can run 40 to 60 degrees hotter than the air temperature around it. That heat doesn't just warm the membrane — it heats whatever's trapped beneath it, and trapped air or moisture vapor expands when it's heated. With nowhere to release that pressure, the membrane bulges upward at every weak point. Cool the roof down overnight and some blisters shrink back. Others don't, because there's actual liquid water in there now instead of vapor, and liquid water doesn't behave the same way air does.

I've cut open enough of these to know the pattern by feel before I ever get the knife in. A blister with give to it, one that flattens under thumb pressure and pops back, is usually just air. One that feels dense, almost like pressing on a wet sponge through a rubber glove, means there's water sitting under that membrane right now.

Not every blister is a problem, and treating all of them the same way wastes money and tears into membrane that didn't need to be touched.

A cosmetic blister stays roughly the same size across seasons. It might swell a little on a hot afternoon and settle back down by evening. Press on it and it feels firm, almost hollow — because it is hollow. There's air in there and not much else, and the membrane above it is still doing its job of keeping water out.

A moisture-driven blister behaves differently. It grows between one inspection and the next. It shows up alongside other blisters nearby, because wet insulation spreads laterally under the membrane the way a coffee stain spreads through a paper towel. Press on it and it feels soft, sometimes almost mushy, with none of that hollow bounce-back. Sometimes the membrane surface around a wet blister looks slightly discolored or has a faint sheen, which is condensation working its way toward any weak point it can find.

Here's the part a lot of property owners don't think about until it's too late. Water expands by roughly nine percent when it freezes. If there's moisture trapped under your membrane going into December, every freeze cycle that winter pushes outward against whatever void that water is sitting in, widening it just a little. Then it thaws, the pressure releases, and the cycle waits for the next cold snap to do it again.

A blister that measured four inches across in October can be six or seven inches across by March, not because anything new happened but because the same water froze and thawed a dozen times between the two dates. Connecticut doesn't do one hard freeze and call it winter — it does weeks of temperatures swinging back and forth across 32 degrees, which is close to the worst-case scenario for anything holding trapped moisture. Every one of those swings does a little more damage to the membrane's bond and the insulation board underneath it.

This is different from what ponding water does to a roof, and the distinction matters. Ponding water sits on top of the membrane in a low spot, adding weight and slowly working its way into any seam or fastener it can find from the surface down. Trapped moisture under a blister is already inside the system, doing its damage from underneath through repeated expansion rather than surface infiltration. A roof can have one problem, the other, or both at once, and the fix for each looks different.

Guessing isn't part of the process, or shouldn't be. When we get a blister call, the first step is a hands-on walk of the roof — pressing on every blister, noting size, firmness, and location relative to drains, flashing, and rooftop equipment, since those are the spots where installation seams and penetrations tend to be weakest.

For anything ambiguous, the next step is a test cut. That means slicing a small section of the membrane open right at the blister, folding it back, and looking directly at what's underneath. If the insulation board is dry and the substrate looks intact, that section gets patched and sealed back up in under an hour. If the insulation is dark, soft, or visibly saturated, that tells us the problem extends beyond what we can see from the surface.

On larger roofs, or when blisters are scattered across multiple areas, an infrared moisture scan comes before any cutting starts. That survey has to happen at night, usually one to two hours after sunset on a day with good sun exposure, because dry insulation releases the day's heat quickly while wet insulation holds onto it longer. Through an infrared camera, the wet zones show up warmer than everything around them, which lets us map the full extent of a moisture problem before opening a single test cut. The scan tells us where to look. The test cut confirms what we're looking at.

Once we know what we're dealing with, the repair matches the scope. A localized dry-air blister gets a patch — cut it open, let it flatten, reseal it with compatible membrane material and the right adhesive for that system, whether that's TPO, EPDM, modified bitumen, or a built-up roof. A confirmed wet section gets cut back to dry, sound insulation, which sometimes means removing a few square feet and sometimes means a much larger area once you follow the moisture to where it stops. Trying to patch over wet insulation without removing it just seals the moisture in for another winter of freeze-thaw cycles, and that's a call we won't make regardless of how much smaller the patch job would look on paper.