Article on fiberglass’s IR translucency
Dana has mentioned fiberglass’s IR-translucency several times. In a recent thread, it was identified as contributing to a particular heat gain issue. And I’ve experienced it myself after having blown fiberglass added to my attic and experiencing a positive difference in winter but a nonexistent to negative one in summer. If I’d known about this before having this work done, I wouldn’t have had it done it, either doing it myself or searching farther and wider for someone willing to use cellulose instead of fiberglass. Given the ubiquity of fiberglass insulation in North American markets, this seems like a really important thing to know about and I think it deserves its own article to highlight it. Maybe Dana could do a guest post!
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
I have wondered about this issue as well, my "house" (known as the unmitigatable energy efficiency disaster) has 2x6 ceiling joists with fiberglass between them (and none on top). When i can someday get repair work and electrical done i plan on nuking it and installing cellulose to save energy and see if it reduces the high A/C use in the summer
Nate, if you remember I tried to get you to go with cellulose on general principles when you first brought it up. I remember grinding my gears to get you to go in that direction. It's always good to remember when a major decision is imminent that human beings are pack animals and that their consensus opinion can be wrong.
Yes, I remember, Eric. I'm sorry I didn't take your advice, but I just couldn't find anyone willing to do cellulose and I really didn't want to do the work myself. If I'd known then what I know now, I wouldn't have pulled the trigger.
It would also be good to know what the options are in these kinds of situations. Adding non-fiberglass bulk insulation materials like cellulose (and mineral wool, probably) is one option, but what about simply laying some kind of IR-opaque layer on it, like kraft paper or construction paper or foil? Or faced fiberglass batts with the facer facing up?
Nate,
Fiberglass insulation isn't really "translucent" to infrared radiation, at least at the thicknesses usually seen in attics. Although manufacturers of radiant barriers sometimes claim that radiant heat can "go right through a thick layer of conventional insulation," that isn't so. When radiant heat hits one side of a deep layer of insulation, only a tiny percentage of that heat is “shine-through” radiation that manages to miss all of the fibers in the insulation blanket and emerge unscathed on the other side.
There are three reasons that blown-in fiberglass insulation performs worse that cellulose on an attic floor. The first reason is that blown-in fiberglass is easy to fluff when it is installed, so it often gets installed at a low density.
The second reason is that blown-in fiberglass, even when properly installed, tends to have a lower R-value per inch than blown-in cellulose. Some types of blown-in fiberglass have an R-value as low as R-2.2 to R-2.6 per inch.
The third reason is that fiberglass insulation is much more air-permeable than cellulose. This undermines its performance two ways: by allowing more air to escape through ceiling air leaks, and by allowing convective loops to establish themselves in the insulation layer, especially in very cold weather, even in the absence of ceiling leaks.
If you're worried about the performance of blown-in fiberglass installed in your attic, the best solution is to install a layer of cellulose insulation (at least 3 inches thick) on top of the existing fiberglass.
It's translucent, not transparent. At some thickness & density it's effectively blocked from passing completely through, but it's still absorbed in the initial inch or two. The absorption at an internal layer is the problem, since it can't be directly convection cooled by the attic air through R3 or more of insulation. The manufacturers have been adding materials to the fiber to further limit the IR translucency for at least 3-4 decades, and higher density definitely helps.
The solution for blown fiberglass performance in recent years has been moving toward higher density products (and making it thicker as the IRC bumps up attic R-values also makes it less of a problem.) But R30 batts are a mid-density product, and more susceptible, though not nearly as problematic as the ubiquitous low-density R19 batt insulation used in attics in the pre-1990 era.
Researchers at Texas A & M University published multiple papers published on this phenomenon back in the early 1980s, many of which were available online as of 10 years ago.
The radiant barrier folks would have you installing reflective materials from the underside of the rafters from soffit to ridge. This works, but in terms of overall bang/buck it's is ridiculously expensive compared to the performance boost you get with an R10-R20 overblow of cellulose. If you go the radiant barrier route, use a perforated aluminized fabric type product, not foil or aluminized mylar or bubble pack. The perforated fabric products run about 5 perms, and thus can't create a moisture trap. If you're not sure about the product, search out it's ASTM E96 specifications- if it has none, don't use it. Aluminized polyester or similar that does NOT have a fine grid of perforations is nearly as vapor-tight as foil or 6 mil polyethylene. On the underside of rafters with vapor-tight goods you'd have to leave a gap at both the bottom top end so the rafter bays can convection-dry into the attic, but that also defeats some of it's thermal performance.
Note that this isn't only a summertime heat gain effect--it's also a winter heat loss effect. The fibers in the layer that absorb radiation in the summer can radiate heat into the attic in the winter as well. Radiation always works equally in both directions. The sales pitches for radiant barriers often imply otherwise, but that's just one of the many misleading things about radiant barrier ads.
What doesn't work equally in both directions is convection. So sometimes the relative important of radiation vs. convection is different according to heat flow direction. But it's the convention that is changing, not the radiation. An attic is an example of that, where convection is stronger in the winter for heat flowing upwards. Thus, addressing radiation isn't as beneficial in the winter as it is in the summer, only because convection is so much more important in the winter.
Paper or cardboard over the top of the insulation would work to stop this effect, and make the fiberglass behave pretty much the same as cellulose in this respect, and would also protect it some from wind washing. Incoming radiation would hit the paper and warm it, instead of penetrating into the insulation. A bit of the heat in the paper would then conduct down through the insulation, but most of it would get released to the air in the attic by convection. That's better than having the heat deposited deeper into the insulation where a larger fraction ends up traveling down into the house rather than going back up to the attic air.
But if you are going to bother with laying something over the top of the insulation, and you don't have room for adding cellulose insulation, a radiant barrier would do more than the paper does--it would reflect the radiation back up to the roof, instead of absorbing it. The only problem is that it only does that well only until it gets dusty on top, at which point it reverts to being not much better than the paper.
Note that deeper in the insulation, or in insulation in a wall, there is always some heat transfer by radiation, over small distances within the insulation. That is part of what determines its R-value. So normally you don't have to worry about it as a separate thing from the heat transfer through the insulation that you calculate with the R-value--that effect is part of what makes some insulations better than others, most prominently in graphite infused EPS (trade name Neopor), which has a higher R-value than regular EPS because the diffusion length for IR within the insulation is reduced. So in most applications, this isn't a new and different concern--the R-value gives the overall performance and there's no need to worry about what mix of heat transfer mechanisms occur within the insulation. But the exposed insulation surface in the attic is a special case where this can make a bit of difference.
I found a paper that I had access to through my library that has some data. It's a chapter from "Thermal Conductivity 20," titled
"Apparent Thermal Conductivity of High Density and Low Density Fiberglass Insulations,"
by S. Yajnik and J. A. Roux of the University of Mississippi .
It uses a model including radiation transfer within the insulation to predict the R-value and shows that the predictions work. It doesn't address the phenomena that Dana mentions that happen with open-top insulation in an attic, but it does have numbers on the extinction length--roughly the distance that the IR diffuses into the insulation. They considered two cases, very low density (0.58 lb/cu ft) and very high density (8 pcf). The 0.58 lb has a 1/4 cm extinction length, and the 8 lb/cu foot about 10 times shorter (1/4 mm). That's shorter than I expected, and if that's right, it means that this is a pretty tiny effect once you have a few inches of insulation. But the paper didn't directly address the question at hand and it's possible I misinterpreted some aspect of it.
They cite this paper that sounds like it is more on topic, but I don't have easy access to it:
Rish, III, J .W., and Roux, J .A., "Heat Transfer Analysis of Fibrous
Insulations With and Without Radiant Barriers for Summer
Conditions," Journal of Thermophysics and Heat Transfer, Vol. I, No.
I, January 1987, pp. 43-49.