Open Cell vrs. Closed, where to draw the line
Here in NC we generally spray 8″ to 12″ of open cell insulation, generally Demelec or Icenene, on the underside of our roof deck. I am constantly asked about closed cell vrs. open cell and what are the considerations that would make one or the other preferable.
Last year at the IBS I asked Joe Lstiburek this question and his response was that open cell was fine in the south where you have good drying potential and short winters but that north of New York it was better to use closed cell because open cell could take on moisture by diffusion and it could freeze within the structure of the foam and accumulate throughout an extended cold spell reducing the effectiveness of the insulation and leading to moisture problems in the spring when it dried out.
Joe’s word is good enough for me, but when I quote him to the skeptical masses they want more in the way of citations and footnotes than just “Joe sez so.” Does anybody have any such hard data about problems using open cell in areas with extended winters especially in roofs?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Michael,
Articles on moisture problems in cold-climate homes insulated with open-cell foam appeared in the April 2005 and July 2005 issues of Energy Design Update. Briefly, the following cases were documented:
1. A "building science consultant who requests anonymity told EDU that he has seen a few houses with saturated roof sheathing behind Icynene."
2. Clyde Potts, a builder in Big Fork, Montana, had moisture problems in a cathedral ceiling insulated with Icynene.
3. Homeowners Elizabeth and Matt Moffitt of Warren, Vermont has extensive moisture problems in the walls and ceiling of a small house insulated with Icynene.
Further details can be found in the two articles. These cases had a few things in common:
1. The homes had high levels of interior moisture.
2. The homes had inadequate mechanical ventilation systems.
3. The Icynene installers forgot to include a vapor retarder.
There are several possible morals to be drawn. Of course, Icynene can work fine in a cold climate, as long as you get the details right. But closed-cell foam is certainly more foolproof.
Michael,
I've had the same debate in our office and with clients. I've found about 80-90% of the market to be open cell. Costs to acheive the same R-Value is roughly the same, closed cell being offering twice the thermal resistance per inch, at double the cost per board foot, being sprayed half as thick.
Of note, all open cell sprayers swear that closed cell will trap moisture from a roof leak and it will rot your framing and ruin your sheathing before anyone notices. All say they know of real life examples of such ruin. I've personally asked sprayers of Demilac, NCFI and Iceynene products for a single case study to prove the theory. They all have my card, they all said they'd send something. I'm still waiting.
We prefer closed cell for a few reasons. I work on old historic homes, and install typically after a gut job. Closed cell acts as a vapor barrier, air seals, doesn't absorb mositure, and, often forgotten, is a structural product, adding strength to underframed structures. I rarely get more than 6" of rafter depth, which isn't enough for open cell to meet code. Balloon framers didn't have sealed attics in mind. As Martin noted, it's just the simplest solution.
Aaron
I'm with you except that I'm seeing higher cost per R-value for the closed cell than you are and I'm generally only putting foam in my roofs and band joists so the thickness is less of an issue than in your renovations. If cost per R were truly equal as you say then I might be tempted to go to the closed cell for the structural hurricane resistant properties alone. I'll price it both ways for a few houses and see where we end up.
I too have never heard of an undetected leak causing problems due to closed cell, but I don't know anyone but you who is spraying closed cell around here so there is not much of a data set to draw from.
m