Abstract:

The research reported in this paper is aimed at increasing the understanding of the hygrothermal performance of interior basement insulation systems by a combination of field monitoring of four assemblies and one-dimensional computer modeling. The work described here is part of a Canada Mortgage and Housing Corporation (CMHC) sponsored research program to determine the significance or insignificance of potential moisture problems due to an impermeable polyethylene layer in above- and below-grade walls (Wilkinson et al. 2007).

Introduction

Due to a combination of demand to lower energy use, stricter building energy codes, and the desire for maximizing finished space in a house, an increasing number of basements are being finished and insulated (often on the interior). However, this practice has resulted in more moisture-related failures of their wall insulation systems (Lstiburek 2006), including condensation and/or moisture accumulation within the vulnerable portions of the assembly.

Background

Heat and moisture flows in the basement environment have been addressed by many researchers, including Timusk & Pressnail (1997), Cheple and Huelman (2001), and Ueno and Townsend (2006). Some important points that inform this research include the following:

The soil adjacent to the basement walls is for all effective purposes at 100% relative humidity (saturation) throughout the year, except for the very topmost surface interface. As a result, the dewpoint (absolute moisture content) of the soil surrounding the basement is equal to its temperature.
The thermal mass and insulating value of the soil moderate temperatures seen by the belowgrade portions of the basement wall. As a result, the temperatures seen at the below-grade portion of basement walls are warmer in the winter and cooler in the summer than the abovegrade portion. This effect increases with depth in the soil. Furthermore, the thermal mass of the soil causes a time delay of the below-grade temperatures relative to air temperatures (also increasing with depth). One example of this phenomenon is that condensation can occur on the inner face of the lower portions of wall during the spring, if exterior dewpoints are rising but the deep soil is still cool from the winter.
Based on the two previous points, the vapor pressure and temperature conditions on both sides of the wall can be plotted on a psychrometric chart (for example, see Timusk, 1997). This presentation, which compares the soil boundary conditions with exterior air conditions (summer and winter), shows that the relatively constant temperature/dewpoint of the lower wall and floor slab will dehumidify the interior air in summer, and humidify in winter. In addition, if typical interior temperature and humidity conditions are plotted, it demonstrates that in most cases the vapor drive is inward (i.e., the wall is drying to the interior). The exception would typically be the above-grade portion of the wall during the wintertime; this is the case where interior vapor control is needed to prevent moisture accumulation and condensation at the concrete-insulation interface.
Moisture is the primary cause of failure in these assemblies: it occurs when wetting exceeds drying, resulting in net accumulation. The following moisture sources may be acting alone or in combination in these failures (Rose 2005). Understanding the relative magnitude of various moisture transport mechanisms is useful in setting priorities for the design of the assembly. The mechanisms, in rough order from highest transport rate to lowest, are:
Bulk liquid water transports water at the greatest rates, as seen in massive water events such as flooding due to exterior or interior sources. This fact stresses the importance of keeping liquid water away from the foundation, with such measures as directing runoff away with eavestroughs and downspouts, and proper grading. Adding a drainage system to the wall, for instance with a dimpled drainage mat, reduces bulk water exposure of the concrete significantly. In addition, control of rising groundwater with a footing drain system is vital.
Liquid capillarity is a significant moisture transport mechanism; it is the absorption and transport of liquid water through the pore spaces of a porous medium such as wood or concrete. It is commonly referred to as “wicking;” an example would be water movement from the concrete footing sitting on wet soil into the bottom of a basement wall.
Air transported water vapor can be an important transport mechanism; it can both act across layers of an assembly, as a leak (e.g., warm humid interior air leaking into a cavity and condensing on a cold surface in winter), or within a layer (e.g., convective looping of air from one side of an insulated cavity to another, transporting moisture).
Vapor diffusion is the slowest-acting mechanism, although it often receives the most attention in regulations and codes (i.e., requirements for a vapor barrier). In common building applications, air transport typically moves moisture at rates orders of magnitude faster than vapor diffusion; however, vapor diffusion can still cause failures if excess wetting or inadequate drying is available.
There is a significant amount of construction moisture in block or site-cast concrete basement walls. This moisture can cause damage if it accumulates in a vulnerable part of the assembly; a design goal is to let concrete dry out without causing harm to the remainder of the wall, or to safely contain the moisture.
The critical moisture level for assembly failure is tied to the onset of mold growth initiation and amplification. There are several threshold levels stated in literature; typically, surface relative humidity levels below 80% or 20% moisture content (by weight) in wood are considered safe. Recent research (Doll 2002, Black 2006) has indicated that strong mold growth is linked with the presence of liquid water, as opposed to high relative humidity levels.
The lowest-risk approach to basement insulation is to apply the insulation to the exterior; this is in line with concepts long known in the field of building science (Hutcheon and Handegord, 1983). This approach eliminates the need for interior vapor control (as there is no cold condensing surface), and strongly reduces the chances of seasonal condensation on the lower portions of the wall. More importantly, exterior insulation protects against both bulk water and capillarity, by providing a robust drainage plane and capillary break from liquid ground water. (Kesik et al. 2001). However, the building industry has proven to be reluctant to adopt this practice, due to construction sequencing issues, insect control, and difficulty in protecting the above-grade portion of the insulation.
The lowest risk approaches for interior insulation of basement walls use non-moisture sensitive semi-vapor permeable materials at the interface between the concrete and the insulation (Lstiburek 2006). Alternately, in foundations known to leak liquid water, insulation assemblies specifically designed to safely drain this water can be retrofitted to the interior.

Previous Research

A selection of field surveys, simulation, and experimental work focused on interior insulation of basements is presented here:

Robert W. Anderson and Associates (1989) performed a field survey of 42 houses in Minnesota, measuring moisture content of wood framing of interior insulated basement walls. They included walls both with and without a polyethylene vapor retarder; the installation of the polyethylene was classified as “excellent,” “good,” or “poor” (i.e., from “air sealed” to having “many tears and rips”). Measurements were taken in both spring and summer; higher moisture contents were seen in the summer, both above and below grade. Some condensation was noted on the exterior side of the polyethylene during the summer at the above-grade portion. There was no correlation between the quality and/or presence of a vapor retarder and framing moisture content; walls without polyethylene did not show excessive moisture content levels. In addition, the authors noted that the polyethylene inhibited the drying of incidental wetting (such as leakage due to improper drainage), and suggested that it might be better to omit this layer.

Swinton and Karagiozis (1995) examined the phenomenon of condensation on the exterior side of the polyethylene vapor barrier during spring and summertime, at the above-grade portion of the wall. This problem is caused by inward vapor drives from the damp concrete; when there is an inward thermal gradient, the moisture moves inwards and condenses on the polyethylene. They replicated this problem using two-dimensional hygrothermal modeling in a Montreal climate and demonstrated that using semi-permeable materials (building paper) on both sides of a fiberglass batt cavity had the best overall performance. Although removing the interior polyethylene layer eliminated the summertime problem, it resulted in moisture accumulation at the concrete-insulation interface during the winter.

Goldberg has tested a variety of interior basement insulation configurations at a Minnesota test facility. After testing frame walls with fiberglass insulation (Goldberg and Huelman 2001), she recommended an assembly with polyethylene on both sides of the stud bay, as adopted by the Minnesota building code. However, a 2002 addendum provided warnings against using this assembly in “superficially dry” basements, which would accumulate moisture behind the exterior polyethylene. Tests were also run using a variable permeability vapor control layer made of polyamide-6 (PA-6) (Goldberg and Gatland 2006). The PA-6 wall experienced minimal condensation during the summertime, and was able to dry inwards, unlike similar polyethylene walls. The PA-6 had similar monitored performance to Kraft-faced batts. In addition, some walls were constructed with a latex elastomeric waterproofing on the concrete surface. These walls had noticeably higher wintertime condensation than the uncoated concrete walls (which largely showed only surface dampness), due to the elimination of the storage capacity of the hygroscopic material, by the hydrophobic dampproofing layer.

Zuluaga et al. (2004) examined seven interior basement wall insulation assemblies in a test installation in the Chicago area; they included two rigid insulation systems (foil-faced polyisocyanurate and expanded polystyrene), two fiberglass blanket systems (perforated and unperforated), and three types of framed walls with fiberglass insulation (encapsulated batts, unfaced batts, and Kraft-faced batts). They were run for two years; the summer conditions were changed between low relative humidity, and high RH between years. The moisture levels behind the rigid insulation systems were dominated by their vapor permeability; the impermeable foil-faced polyisocyanurate showed high sustained moisture levels. The roll blanket walls, at their above-grade portion, had moisture levels matching their permeability: the unperforated blanket accumulated moisture during the summer. The framed walls showed responses that did not correspond to their permeability; in contrast, they were dominated by airflow, bypassing the remainder of the assembly. This was demonstrated by noting that the absolute moisture content levels in the assembly closely tracked interior levels. In this experiment, this airflow caused drying of the assembly.

Ueno and Townsend (2006) examined eight interior insulation assemblies in a Chicago-area basement. These assemblies included rigid foam (foil-faced polyisocyanurate and extruded polystyrene/XPS), fiberglass batt frame and polyethylene walls (polyethylene on one or both sides), composite walls (XPS with a stud frame to the interior; two versions), a perforated roll blanket, and rigid fiberglass board with a PA-6 facer. The assemblies were periodically wetted to measure their drying response. In non-wetted operation, some of the walls showed signs of air leakage from the interior behind the assemblies, as indicated by identical dewpoints to inside (e.g., foil-faced polyisocyanurate). The perforated blanket also showed similar dewpoints to interior, but this could be due to the permeability of the facer (13 perms or 720 ng/Pa•m•s²). The composite walls showed that the wood framing was completely protected from concrete-sourced moisture by the XPS insulation. In the wetting experiments, some walls very quickly (polyisocyanurate, due to air leakage, and perforated blanket). Others dried in a controlled manner, allowing drying through the vulnerable portions of the assembly without damage (composite walls). The XPS (2”/50 mm) wall, however,
showed extended periods of high humidity after the wetting event, showing that drying occurred more slowly through the 0.5 perm or 29 ng/Pa•m•s² material; the PA-6 wall had a similar response. The double polyethylene showed minimal wetting, which was a surprise given its low drying potential. However, it is likely that during wetting, drainage out of the wall cavity occurred (due to the hydrophobic materials lining both sides), so a low initial dose of water was retained.

This review of the literature indicates that a consistent failure seen in a wall with a single layer of polyethylene on the interior is condensation due to spring or summertime inward vapor drives. Wintertime condensation on the above-grade portion was occasionally seen; this was worsened by removing the hygric buffer capacity of the concrete. Monitored data showed that interior insulation has drying potential to the interior in most cases, especially in the below-grade portions of the wall.

Experimental Program

Four interior basement wall insulation assemblies were monitored in a field installation; following the first year of operation, some of the walls were disassembled and inspected.