Insulating a Steel Box Without Trapping Water In It

A shipping container will outlast almost anything you build next to it, on one condition: that water is never given somewhere cold and dark to collect. Get the insulation wrong and you will grow rust on the inside face of a wall you cannot see. Get it right and the frame will outlive you.

Why steel is unforgiving

Steel conducts heat roughly four hundred times better than the timber a conventional house is framed in. Any point where the inside surface touches the outside surface without insulation between them is not merely a heat loss, it is a cold spot, and warm interior air meeting a cold spot deposits its moisture there. In a timber wall that dries out. In a sealed steel wall it sits against bare metal.

This is the whole problem, and it is why the two obvious approaches both fail.

What does not work

Batt insulation in a stud wall built inside the container. This is the cheapest route and the one most self-builders take. You have now created a cavity between the batt and the corrugation, that cavity is connected to the interior air by every screw penetration and service hole, and the corrugation is the coldest surface in the building. Water condenses on the steel, runs down the ribs and collects in the bottom rail where you will not see it for four years.

Rigid board cut to fit between the ribs. Better, but the corrugation profile means every board has an air gap behind it somewhere, and cutting rigid board to a perfect corrugated profile at scale is not realistic. You end up with intermittent voids, which is the same problem distributed more evenly.

What we do

Closed-cell spray polyurethane foam, applied directly to the inside face of the corrugation at a minimum of two inches, before any framing goes in. It works because it is three things at once:

• Insulation, at roughly R-6 to R-7 per inch, which is why two inches in a corrugation profile is worth more than four inches of batt in a stud bay
• An air barrier, because it is continuous and bonded, so there is no cavity for air to move through and nowhere for it to deposit moisture
• A vapour retarder, at the thicknesses we apply, so warm interior vapour never reaches the steel in the first place

Framing goes on after the foam, not before, and it is held off the steel so that no stud ever touches the container wall. Services run in that framing zone, not through the foam.

The details that actually leak

The wall build-up is the easy part. The failures we are called to look at are almost always at four places:

1. The floor. The original marine ply deck sits directly on steel cross members that are exposed to outside air underneath. Insulating above the deck alone leaves those members cold. We insulate below the deck between the cross members and above it under the finished floor.
2. Openings. A window jamb welded into the side panel is a continuous piece of steel from outside to inside. The reinforcement has to be thermally broken or wrapped, and the reveal detailed so the foam runs into the frame.
3. Corner castings and the roof rail. These are heavy sections that stay cold a long time. They need more foam, not the same amount.
4. Where two containers meet. A welded joint between units is a straight steel path. It is insulated as a single continuous line across the joint, which means planning the sequence before the two boxes are welded together.

Coatings and the outside face

Cor-Ten weathering steel forms a stable oxide layer that protects the metal underneath, but only if it gets alternating wet and dry cycles. Where a container is buried, permanently shaded or in constant contact with soil or standing water, that layer never stabilises and it corrodes like ordinary steel. We shot-blast every unit back to bare metal, prime with a zinc-rich primer and finish with a low-VOC topcoat, and we never set a container directly on the ground.

How you know it worked

You test it. Every unit gets a blower-door test in the yard before it ships, and we are looking for 0.6 air changes per hour at 50 pascals or better. That number is not a marketing figure, it is the evidence that the foam is continuous. A unit that leaks air will leak vapour, and a unit that leaks vapour will eventually rust from the inside. Testing before delivery is the only point at which fixing it is still cheap.