The Impact of Vapor Retarders on Basement Foundation Moisture Transport

INTRODUCTION

The question frequently is asked whether there is a net benefit to including a vapor retarder in basement envelope systems, since, subjectively, upon entering a basement that is unconditioned, the moisture levels often appear to the senses (in terms of odor, clamminess, etc) to be significant regardless of whether a vapor retarder is installed or not.  Over 10 years of testing at the FTF, a variety of vapor retarders have been evaluated in order to measure their quantitative impact on whole basement envelope water vapor transmission.


EXPERIMENTAL METHOD


 A more detailed review of the experimental method is presented elsewhere on this web site (Description of the FTF - section 3).  In summary, a constant dehumidification rate approach was chosen and implemented with identical, continuously operating dehumidifiers all in 4 basement test modules.  Since the interior module temperatures are held equal and constant, the above-grade module superstructures are identical, the geotechnical conditions around the modules are the same at least to a first order of magnitude, and, there are no interior sources of water vapor, the condensate collected is a direct measure of the vapor transport across the foundation envelope.  Data was collected at weekly intervals over the heating season (roughly December through April).  All the data is normalized with respect to the north reference module (module ARN with poured concrete walls and slab) which remained unchanged for all test heating seasons.  Thus data gathered over different heating seasons may be compared on a normalized basis.


RESULTS

These results also are reported elsewhere on this web site, but they are presented here in a summary fashion that seeks to demonstrate the time dependent impacts of the various insulation and/or vapor retarder configurations tested in the sequence that they were installed.  Two configurations are discussed for masonry block walls, namely, interior and exterior insulation placement (modules ATE and ATS, respectively).

The results for the exterior insulation placement (module ATS) are presented in figure 1 and table 1.


Figure 1

Table 1  Normalized seasonal condensate removal values for figure 1

HEATING SEASON NORMALIZED CONDENSATE REMOVAL
1989/90 .563
1990/91 .519
1991/92 .548
1992/93 non-continuous dehumidification
1993/94 .886
1994/95 .561
1995/96 .362

During the first five heating seasons (89/90 through 93/94) the envelope insulation system remained unchanged.  During the first three seasons, the exterior R-10 extruded polystyrene insulation acted as an effective vapor retarder, reducing the whole envelope vapor transport by an average of 0.46 compared with the reference.  However, two heating seasons later, the vapor retarding capacity of the the exterior insulation had diminished very significantly, yielding just a 0.11 reduction in vapor transport compared with the reference.  A physical model that may be used to explain this behavior is based on the concept of "diffusion (or molecular) tunneling".  In this model, vapor is transported through the insulation along tortuous paths driven by the imposed soil/module humidity ratio gradient.  Given a fairly uniform, and more importantly, unidirectional diffusion path of water vapor from the soil into the module, during at least the first three heating seasons, the vapor tunnel "face" had not progressed through the thickness of the insulation, that is, the diffusion tunnels had not become "filled" (reached vapor equilibrium) and so the insulation as a whole was still able to absorb vapor.  By the fifth heating season, the tunnels had become filled (that is, the amount of vapor in the tunnels had reached equilibrium with the imposed humidity ratio gradient) yielding the maximum equilibrium diffusion rate.  Hence, from this, it appears that under FTF geotechnical and temperature conditions, the "tunneling rate" for extruded polystyrene is of the order of 0.4 to 0.7 inches/year.

Adding a sealed 4-mil. polyethylene vapor retarder to the interior module wall surface prior to the onset of the 1994/95 heating season, reduced the normalized vapor transport to 0.56, a significant reduction of 0.32 .  Prior to the 1995/96 heating season, the floor of the module was covered with a 6-mil. polyethylene vapor retarder (which was well sealed to the wall retarder) further decreasing the normalized vapor transport by 0.20 .  Thus in the case of exterior insulation, a combination of slab and floor retarders yields a substantial 0.52 normalized reduction in vapor transport, the bulk of which comes from the wall retarder alone.

 

The results for the interior insulation placement (module ATE) are presented in figure 2 and table 2.


Figure 2

Table 2  Normalized seasonal condensate removal values for figure 2

HEATING SEASON NORMALIZED CONDENSATE REMOVAL
1989/90 .705
1990/91 .703
1991/92 .810
1992/93 non-continuous dehumidification
1993/94 1.155
1994/95 .604
1995/96 .687
1996/97 .404

During the first three heating seasons, the masonry block cores were unfilled and no insulation or vapor retarders were installed on the walls or floor, yielding an average normalized reduction of 0.26  with respect to the poured concrete reference module (ARN).  The noticeable jump of 0.11 in collected condensate between the second and third heating seasons is still under investigation (pending the deployment of an adequate foundation envelope moisture transport computer simulation code), but possible systematic errors have been excluded.  The average 0.26 reduction in envelope vapor transport compared with a poured concrete (solid) block wall is perhaps counterintuitive in view of the nominally more porous nature of masonry blocks.  The difference is explained well by the occurrence of buoyancy driven convective loops within the cores which transport vapor from the bottom to the top of the wall where it condenses and freezes.  This effectively reduces the ground moisture source strength (see reference GL94, page 50 for a complete explanation).  So while poured concrete indeed is less permeable than masonry block, the apparent permeability is greater owing to the convective loops in hollow masonry block walls.  This is confirmed by the increase in normalized vapor transport to 1.16 in 1993/94 when the cores of the masonry blocks were filled with a lightweight artificial aggregate.  That is, largely eliminating the convection loops (since the porosity of the filled cores is significant), increased the vapor transport to the intuitively "correct" relative magnitude.

Prior to 1994/95, interior R-8 molded polystyrene was installed on the walls, reducing the normalized vapor transport to 0.60, again a significant reduction of 0.55 (compared with 0.46 for exterior insulation).  However, adding an interior 4-mil. vapor retarder prior to 1995/96, yields an apparently anomalous increase in vapor transport of 0.08.  However, this really is a combination of two effects, an increase in permeability of the molded polystyrene owing to molecular tunneling plus the decrease in wall system permeability produced by the addition of the vapor retarder.  As might be expected, because of  its lower density (and higher porosity), molded polystyrene has a larger tunneling rate than extruded polystyrene, evidently of the order of 2 inches per year.  So assuming that the molded polystyrene approached vapor equilibrium prior to the 1995/96 heating season, then the net reduction in envelope transport accruing to the vapor retarder is 0.47 (compared with 0.32 for the exterior extruded insulation case).  Finally, addition of R-4 molded polystyrene on the floor for 1996/97 decreased the overall envelope moisture transport by an additional 0.20, identical to the result obtained for the 6-mil. polyethylene in the exterior insulation module.  This is reasonable, since given the 0.43 year extent of the 1996/97 heating season, at a nominal 2 in./year tunneling rate (which probably is an over-estimate), the tunnel face would have extended 0.86 inches into the insulation, less than the 1 inch thickness available.  So again, as in the exterior insulation case, a combination of slab and floor retarders yields a substantial 0.75 normalized reduction in vapor transport, the bulk of which comes from the wall retarder alone.

Finally, it can be argued that if the fiduciary condition is taken as the unfilled core case (rather than the filled case as argued above), then the impact of adding a wall retarder is only 0.05 (0.18 for the exterior insulation case) with an additional 0.28 provided by the floor retarder.  This can be used to argue that a floor retarder alone is not of much value.  However, this is an artifact of the convective wall cavity flows and would not apply to any wall without cavities or to a hollow masonry block wall where the convective flows are blocked at grade level, for example.  Still, in cases where unrestricted convective wall cavity flows do occur, the impact of a wall retarder alone is of much reduced significance.


REFERENCES

GL94  Goldberg LF, Langenfeld DT and Lively RS.  "Foundation Test Facility Experimental Results Part 1: 1993/94 Test Period System Data", Underground Space Center Report, 1994.

return to CONTENTS