10, 15, 20, 30, or 40 kg/m³: When discussing blown-in insulation, density is generally considered an installation specification. But can it also affect how the insulation actually performs once it is installed in an attic?
In France, density is not an overlooked factor. It is taken into account in technical documents and in monitoring the installation of loose-fill insulation. However, are the densities specified for the various products always sufficient to guarantee the expected thermal performance under the actual conditions of an attic?
Density is already taken into account in France
The certifications and technical documents applicable to loose-fill insulation include parameters related to their installation. For insulation blown into unused attic spaces, the density measured on-site is one of the factors used to verify that the product has been installed in accordance with the manufacturer’s instructions.
DTU 45.11, which focuses in particular on the thermal insulation of unused attic spaces using blown-in loose-fill insulation, thus specifies inspections covering various installation parameters: thickness, distribution of the insulation, quantity of material used, and resulting density.
But do the specified and verified densities ensure the expected thermal performance of the insulation under all real-world conditions of use?
Why can a layer of insulation allow air to circulate?
Loose-fill insulation is not a compact material. It consists of fibers separated by numerous air-filled spaces.
This structure is precisely what provides good thermal resistance: the air trapped between the fibers limits heat transfer.
But when the layer is thick enough, has a very open structure, and is subjected to a significant temperature difference between its hot side and its cold side, the air trapped within the insulation can begin to move.
This is the phenomenon of natural internal convection.
In practical terms, warm air tends to rise while cooler air sinks. Insulation then no longer works solely through conduction: internal air movements can contribute to heat transfer.
In other words: the more easily a loose-fill insulation material allows air to circulate, the more its thermal behavior may deviate from a simple calculation based on conduction.
A phenomenon that has been studied for more than 30 years
This is not a new issue. Natural convection in loose-fill insulation has been the subject of research for several decades.
Beginning in the early 1990s, research conducted at the Oak Ridge National Laboratory (ORNL), under the auspices of the U.S. Department of Energy, examined the thermal behavior of loose-fill fiberglass insulation in attic configurations.
A study published in 1991, based on tests conducted in a setup replicating a residential attic, observed a significant decrease in thermal resistance as the temperature difference increased. Under the most severe conditions tested, the measured resistance could be up to half the estimated nominal thermal resistance of the insulation. The authors attributed this change to natural convection within the insulation.
They noted, however, that the tests involved only one type of loose-fill fiberglass insulation and that the results could not be automatically generalized to other types of insulation with different properties.
In 1992, Delmas and Wilkes further investigated the phenomenon through numerical modeling of conduction and natural convection in loose-fill fiberglass insulation. Their analysis varied, in particular, the density, air permeability, and thickness of the insulation. Their general conclusion was that as convection increases, the total heat transfer increases, and the thermal resistance decreases as the temperature difference across the insulation increases.
This research is particularly relevant to our discussion because it shows that the thermal performance of loose-fill insulation does not depend solely on its thermal conductivity as measured under standardized conditions. The performance of installed insulation also depends on its physical properties and the conditions to which it is subjected.
Sources: Wilkes et al., Thermal Performance of One Loose-Fill Fiberglass Attic Insulation, ASTM STP 1116, 1991; Delmas & Wilkes, Numerical Analysis of Heat Transfer by Conduction and Natural Convection in Loose-Fill Fiberglass Insulation, ORNL/CON-338, 1992.
So, what role does density play?
Density is one of the parameters that influence the behavior of a loose-fill insulation material, along with its structure, its air permeability, its thickness, and the temperature gradient to which it is subjected.
But it has one obvious physical consequence: for structures of comparable size, a lower density generally means a more open layer, with more volume available for air.
Yet it is precisely the possibility of internal air movements that can promote the occurrence of natural convection.
Density alone is therefore no guarantee of performance, but it is one of the parameters that determine the structure of the layer and its thermal behavior.
What does the 2026 experimental study tell us?
A study published in 2026 in the *Journal of Building Physics* provides a new experimental contribution to this issue.
Johra, Léard, Veit, and Jensen conducted large-scale tests on horizontal layers of blown glass wool, with a thickness of 30 or 60 cm and a density ranging from 11.9 to 19.7 kg/m³. In the tests, the temperature difference between the two sides of the layer was varied from 5.2 to 59.5 K.
In this experimental setup, the authors identify a critical Rayleigh number ranging from 13 to 15, beyond which internal natural convection becomes detectable through changes in heat transfer.
The results are of interest for attic applications, since the conditions studied correspond to a horizontal layer of blown glass wool, with an open top surface and a structure consisting of joists.
The authors themselves point out that the results must be viewed within the context of their experimental setup and that the properties of the materials as well as the boundary conditions can influence the threshold at which convection occurs.
The study thus adds another piece to an already extensive body of scientific literature.
And what about the " cellulose insulation " in all of this?
Another experimental study provides particularly interesting insights. In 2020, Kivioja and Vinha studied internal convection in heavily insulated roof structures using various types of loose-fill insulation.
In particular, the tests compared blown glass wool and cellulose insulation, in various roof configurations, with thicknesses of 300 and 600 mm.
In the configurations studied, the glass wool exhibited an increase in heat flux attributable to internal convection, whereas the effect observed with the cellulose insulation was significantly more limited.
The authors emphasize, in particular, the influence of the material’s structure and its air-permeability properties. Density is therefore a factor in the analysis, but it cannot be considered in isolation from the material’s bulk form and permeability.
cellulose insulation The published results indicate an increase in heat flux of the order of 30 to 40 percent for glass wool in the configurations tested, compared with 0 to 10 percent for the blown-in insulation studied in this article.
Obviously, these results do not allow for a direct application of these percentages to all products on the market. Nevertheless, they show that the same line of reasoning—based solely on nominal thermal resistance—can mask differences in behavior when materials are subjected to conditions conducive to internal convection.
Source: Kivioja, H., Vinha, J. (2020), Hot-box measurements to investigate the internal convection of highly insulated loose-fill roof structures, Energy and Buildings, 216, 109934.
The Real Challenge: Moving from Theoretical Performance to Actual Performance
Thermal resistance is, of course, a key factor in characterizing an insulation material. But it does not necessarily tell the whole story about the loose-fill insulation installed in a building.
Between the laboratory and actual attic conditions, several factors can come into play: thickness, density, product distribution, layer structure, air permeability, the presence of joists, and temperature variations.
That is why the issue of density may be worth looking at from a different perspective.
It is not simply a parameter to be monitored on the job site. It helps define the physical structure of the insulation and, therefore, the conditions under which it will operate.
What if we finally asked the question differently?
Today, density is already controlled during the installation of blown-in insulation.
But verifying that a product meets the specified density does not necessarily answer another question: Is this density sufficient to ensure the expected thermal performance under all conditions to which the layer will be exposed?
The available scientific research does not allow for a universal answer. However, it does show that internal convection is a real phenomenon that has been studied for several decades, and that its significance depends on several physical parameters of the insulation and its environment.
So perhaps the question is no longer just about what thermal resistance to list on a technical data sheet. We must also ask under what conditions this performance is actually maintained once the insulation has been blown into the building.
And in this context, density certainly deserves to be viewed as more than just a processing parameter.
FAQ – Density and Blown-In Insulation
Is density taken into account for blown-in insulation?
Yes. In France, certification documents and installation specifications include parameters related to the density of loose-fill insulation. In particular, NF DTU 45.11 specifies inspections related to the quantity of product, the thickness, and the resulting density.
Does a low density mean that insulation is less effective?
Not necessarily. Performance depends on several material properties: form (loose or in panels), thermal conductivity, air permeability, density, thickness, and thermal conditions. A low density, however, can promote certain conditions conducive to internal convection.
What is internal convection in blown-in insulation?
These are natural air movements within the insulating layer, caused primarily by temperature differences. These movements can increase heat transfer and thus reduce the effective thermal resistance of the layer under certain conditions.
Is there a universal minimum density for all blown-in insulation materials?
The data examined here do not allow for the establishment of a universal threshold applicable to all materials. The behavior also depends on the structure of the insulation, its air permeability, its thickness, the geometry of the attic, and the temperature gradient.
Why are studies on convection important for unused attic spaces?
Unused attic spaces can accommodate thick layers of blown-in insulation. During cold weather, a significant temperature difference may occur between the warm part of the building and the cold part located beneath the roof. These conditions can promote air movement within certain layers of loose-fill insulation.
Scientific Sources
- Delmas, A. A. & Wilkes, K. E. (1992). Numerical Analysis of Heat Transfer by Conduction and Natural Convection in Loose-Fill Fiberglass Insulation — Effects of Convection on Thermal Performance. Oak Ridge National Laboratory, ORNL/CON-338.
- Wilkes, K. E., Wendt, R. L., Delmas, A., Childs, P. W. (1991). Thermal Performance of One Loose-Fill Fiberglass Attic Insulation. ASTM STP 1116, pp. 275–291.
- Kivioja, H. & Vinha, J. (2020). Hot-box measurements to investigate the internal convection of highly insulated loose-fill roof structures. Energy and Buildings, 216, 109934.
- Johra, H., Léard, P., Veit, M., & Jensen, R. L. (2026). Full-scale experimental study on the critical Rayleigh number for the onset of internal natural convection inside a horizontal insulation layer of blown glass wool. Journal of Building Physics, 50(2), 374–437.
Notes and Reading Guidelines
- The results presented in this article do not allow for the definition of a universal minimum density applicable to all loose-fill insulation materials. Thermal performance depends on several parameters.
- The ORNL's historical tests involved specific products and configurations. Their results should not be automatically generalized to all blown-in insulation materials.
- Comparing materials cannot be reduced to density alone. Fiber structure, air permeability, thermal conductivity, thickness, and boundary conditions also play a role.
