Most newer manufactured homes in the Pacific Northwest, as well as many older mobile homes, have a vapor retarder on the inside of the wall cavity--typically right behind the gypsum board. However, many older mobile homes, especially those built before the 1980s, were manufactured with a vapor retarder on the outside of the wall cavity--generally right behind the metal (or sometimes wood) siding.

Vapor retarder materials applied in those days included polyethylene sheet, thin foam with kraft paper coatings on each side of the foam, and asphalt-impregnated kraft paper. The apparent purpose of the exterior vapor retarder (EVR) was to keep moisture that condensed on the back side of the metal (or wood) siding from wetting the insulation.

Unfortunately, the EVR traps moisture migrating as a vapor through the wall cavity from the inside to the outside. The vapor condenses during cold weather on the inside surface of the vapor retarder, runs down it, and collects on the bottom plate or elsewhere. When sufficient moisture builds up within wood members, the wood decays.

Research projects undertaken in the Pacific Northwest for the U.S. Department of Energy have shown that conventionally constructed walls without an EVR located in northern heating climates exhibit no evidence of wood decay, except for that caused by leaks or direct contact between wood members and earth. Thus, decay in walls will probably not occur under normal conditions, especially in older homes. In fact, there is almost no field evidence of decay occurring in conventionally constructed walls, whether they are site-built or manufactured. However, the presence of an exterior vapor barrier can cause wood decay, with its resultant structural damage.

The first known cases of extensive decay in the wall framing members of hundreds, if not thousands, of mobile homes involved Tri State Homes in Wisconsin, Minnesota, and Michigan (the company has since gone bankrupt). The decay was attributed to high levels of indoor moisture, but the primary cause was really the presence of an exterior vapor retarder that trapped moisture within the wall cavity. The low permeability retarder was on the outside of the plywood sheathing behind hardboard lap siding. The presence of the vapor retarder caused the plywood to get much wetter than normal during the winter and spring and reduced the rate at which it dried. Thus the wood was still quite wet in the early summer when temperatures were high enough to promote the growth of decay fungi. The result was severe and extensive rotting of the plywood sheathing, which occurred over a period of 20 years. Because the decay progressed slowly, it was first noticed only about 8 years ago. Moreover, the wet wall conditions led to substantial mold growth which seriously affected the health of many of the occupants.

Our inspection of 17 Tri State homes that had had all the siding removed revealed that 16 had plywood decay and 12 of the cases were severe enough that the plywood could be torn apart by hand. Many walls were unusually wet during winter and early spring, and plywood moisture contents well above 60% were measured during late June and early July, when the plywood in a conventionally constructed wall is considerably drier (typically below 10%). The plywood moisture contents measured in the Tri State homes during that early summer, as well as during the previous winter and early spring, were higher than the highest values measured in any of the Pacific Northwest wall moisture studies.

A comparison of sheathing and siding moisture levels for walls with and without an exterior vapor retarder was also undertaken using the MOIST computer model developed at the National Institute of Standards and Technology (NIST). The MOIST software predicts the moisture content of wall components. The modeling results further reinforce the field inspection finding that the EVR caused the structural damage.

As part of a study conducted by the Bonneville Power Administration (BPA), the exterior siding (mostly metal) of 12 older mobile homes in Butte, Montana; Shelley, Idaho; Shelton, Washington; and Redmond, Oregon was temporarily removed the winter before last to install moisture monitoring sensors. (These sensors monitor wood member moisture content, relative humidity, and temperature.) Siding was typically removed in two small sections of the wall. Some of the homes had an exterior vapor retarder in place just inside the siding.

In several cases where an EVR was present, extensive condensed moisture was noted on the inside surface of the retarders. Often the bottom plate was also wet. Measured maximum wall-cavity wood-member moisture contents were 33%, 44%, and 52% for three of those homes. Most importantly, decay was present in structural wood members of four homes with an EVR (three in Washington and one in Oregon). Decay was noted mainly in the bottom plate and the subfloor; some of it was isolated and relatively minor, but some of it was major. In one case, a whole corner was rotten and lacked structural integrity. More extensive decay might have been discovered if all the siding had been removed and the rest of the wall areas had been inspected. Notably, there was no decay found in any of the homes that did not have an exterior vapor retarder. Moreover, for those homes, the maximum wall-cavity wood-member moisture content was less than 16%, considerably drier than the maximum figures noted above for walls with an EVR.

During the 1993-1994 winter, the structural wood moisture content continuously measured in the wall of one of the Shelton, Washington houses with an EVR became excessively high, while the other seven homes without an EVR in Shelton and in Shelley, Idaho, did not get overly wet. After a year's worth of hourly data were taken, wall moisture as high as 65% was measured in the house with the EVR (see Figure 1). That is abnormally wet. However, in the seven other homes without an EVR, moisture levels monitored during the same period were no higher than 28%. The existence of an EVR thus appears to cause serious moisture problems.

Figure 1.A moisture data plot from BPA's Residential Construction Demonstration Program shows moisture content in two areas within one wall of a mobile home with an EVR. Moisture probes were placed in the wettest location in the wall for the worst case data, as well as in a more typical portion of the wall. The maximum moisture content is higher than 60%, well above the wood's fiber saturation point.

Many weatherization measures will either directly or indirectly make the house more airtight. These activities include adding ceiling, wall, duct, or floor insulation, and replacing windows, along with many air-sealing measures. For example, simply adding dense-pack cellulose wall insulation will often seal leaks and reduce the air leakage of a house substantially (50% reductions are not uncommon). Tightening the house reduces its natural ventilation rate; thus indoor relative humidity levels increase as moisture generated indoors is not flushed out of the house as easily as before.

Higher indoor air moisture levels in homes with EVRs can lead to more condensed water in the walls and a greater chance of even more extensive wood decay. Previous BPA research has shown that wall wood-member moisture contents increase when indoor relative humidity levels increase. The lifetimes of the affected mobile homes could be shortened. In addition, there could be serious health repercussions associated with air tightening homes with an EVR. Utility and low-income housing agencies need to be made aware of these preliminary findings.

A project now under way at BPA should provide insights into resolving this issue. The project involves monitoring the moisture conditions of 11 of the 12 homes mentioned earlier. The overall purpose is to investigate the thermal performance, moisture levels, and ventilation of older mobile homes in the Northwest before and after weatherization. One key goal is to determine how weatherization may cause or otherwise affect moisture problems within the structure or inside the homes. In fact, the existence of the decay problem in some older mobile homes was discovered unexpectedly during the early phases of this project.

Further research is also needed to provide a solution to this problem. Computer modeling to predict and compare moisture accumulation conditions in walls with and without exterior vapor retarders in a wide variety of climates needs to be undertaken. It may turn out that some locations are less susceptible to this moisture problem than others. In addition, more extensive field surveys of a large sample of older mobile homes should be initiated to determine the extent of the problem. This probably should be done to older mobiles located both east and west of the Cascades, since the climates in the two areas are very different.

Remedial action may include removing the entire exterior vapor barrier, or removing it and replacing it with a permeable barrier such as Tyvek, or simply cutting the retarder open at the bottom of the wall cavity. It may be possible to add small vents through the siding and the adjacent vapor retarder (although, based on previous experience, that is not advisable at present). All these possible solutions need to be explored and compared, preferably with side-by-side testing.

Until this issue is fully resolved, walls need to be opened and checked for the presence of an EVR before any weatherization that would air tighten the home is undertaken. The large number of older mobile homes that are inspected and found not to have an EVR should be weatherized as planned. But if an EVR is present weatherization on that home should be temporarily delayed until satisfactory remedial actions are available. Weatherizing these types of mobile homes may make a bad situation worse.

George Tsongas is a professor of mechanical engineering at Portland State University and an independent energy and moisture consultant.