Vapour control

For the fourth lesson, Professor Ruth Abernathy, who taught the environmental systems course the previous spring, joins Lavoie in the studio. On the screen is the temperature profile Abernathy drew for the community hub's wall: a line falling steeply across the insulation, and a dashed line for the dew point of the indoor air. 'Last term we found that the sheathing in this wall sits below the dew point on a design night,' Abernathy says. 'At minus 33 °C, even with exterior insulation, the sheathing is warmer than it would otherwise be, but not warm enough. We concluded that the warm-side air barrier and vapour retarder must keep indoor moisture away from it.' Lavoie adds the question this lesson answers: 'How tight must that retarder be, where should it go, and what happens if we put a vapour-tight layer in the wrong place?' Vapour control is the most misunderstood of the four control functions. It is also the one where well-meant rules of thumb, applied in the wrong climate or the wrong assembly, cause the most damage.

Vapour, humidity and diffusion

Air always contains some water vapour. Its amount can be described by the vapour pressure, the part of the air pressure due to the water molecules, or by the humidity ratio, the mass of water per mass of dry air. Warm air can hold more vapour than cold air before it becomes saturated; the saturation vapour pressure roughly doubles with every 11 kelvins rise in temperature. Relative humidity is the ratio of the actual vapour pressure to the saturation pressure at the same temperature. When moist air touches a surface colder than its dew point, the temperature at which it would be saturated, vapour condenses as liquid or, below freezing, as frost.

In a heated building in winter, the indoor air at 21 °C and 35 of every 100 parts of saturation has a vapour pressure of about 870 pascals, while the outdoor air on a January day might have a vapour pressure of a hundred or so, even near saturation, because it is so cold. Vapour therefore moves outward through the envelope in winter, from high vapour pressure to low, by two mechanisms. The first, air leakage, was the subject of the previous lesson and is by far the larger. The second, vapour diffusion, is the slow movement of water molecules through solid materials, independent of any airflow. Diffusion depends on the vapour pressure difference and on the vapour permeance of the materials, the rate at which they let vapour through.

Permeance and vapour retarders

Vapour permeance is measured in nanograms per pascal per second per square metre, written ng/(Pa·s·m²). The National Building Code calls a material a vapour barrier if its permeance is no more than 60 of these units. Materials span an enormous range. Polyethylene sheet and foil are almost impermeable. Some materials, such as plywood, oriented strand board and closed-cell foam, are fairly vapour-tight when dry but become more permeable as they get damp. Others, such as gypsum board, mineral wool, cellulose and house wraps, are very open. Paint can change a surface's permeance greatly, from the open finish of an ordinary latex paint to vapour-retarder primers that bring gypsum board close to the code's limit.

Where to put the retarder in a cold climate

The principle is simple: a cold-climate assembly should be more vapour-tight on the warm side than on the cold side, so that any vapour that enters the wall from inside can leave it toward the outside faster than it arrived. A low-permeance layer on the warm side slows vapour entering in winter; open layers toward the outside let the assembly dry outward. Problems arise when a vapour-tight layer sits on the cold side of most of the insulation, because vapour that reaches the back of that layer cannot pass and condenses there. The classic example is a wall with polystyrene or foil-faced boards of only modest thickness on the outside of the sheathing: the boards are too thin to keep the sheathing warm, and tight enough to trap moisture behind them.

The way out of this trap is to make the cold-side layer warm. If enough of the insulation is outside the vapour-tight layer, the layer stays above the dew point of the indoor air for most of the winter and condensation cannot form on it. The National Building Code allows a low-permeance material on the outside of a wall if the ratio of the thermal resistance outboard of it to the resistance inboard meets a minimum that rises with the heating degree-days; in Winnipeg's climate, roughly a third or more of the total resistance must be outside the membrane. The perfect wall takes this to its logical end: all the insulation outside the membrane, so that the membrane and the structure stay at room temperature all winter.

The Glaser method

Designers check an assembly for condensation risk with a simple steady-state calculation, called the Glaser method after the German engineer who published it in the late 1950s. It uses the same series calculation as heat flow. The temperature drop across each layer is proportional to its thermal resistance, which gives the temperature at each interface and therefore the saturation vapour pressure there. The vapour pressure drop across each layer is proportional to its vapour resistance, the reciprocal of its permeance, which gives the actual vapour pressure at each interface if no condensation occurs. Wherever the actual vapour pressure exceeds the saturation pressure, condensation is possible. The calculation below applies the method to four versions of the hub's wall in an average January, with the same studs, sheathing and 75 mm of exterior mineral wool, varying only the interior retarder, the membrane on the sheathing and the position of the insulation.

# Glaser method: temperature and partial pressure of water through a wall in January
import math
t_in, t_out = 21, -16              # indoor and mean January outdoor temperature (C)
rh_in, rh_out = 0.35, 0.80
def to_c(t): return t
def psat(t):                              # saturation pressure, Pa
    c = to_c(t)
    return 611.2 * math.exp(22.46 * c / (272.62 + c)) if c < 0 else 611.2 * math.exp(17.62 * c / (243.12 + c))
def wall(name, layers):
    rt = sum(r for _, r, _ in layers)
    zt = sum(1 / m for _, _, m in layers)
    pv_in, pv_out = rh_in * psat(t_in), rh_out * psat(t_out)
    print(name)
    r_sum = z_sum = 0
    for layer, r, m in layers[:-1]:
        r_sum += r; z_sum += 1 / m
        t = t_in - (t_in - t_out) * r_sum / rt
        pv = pv_in - (pv_in - pv_out) * z_sum / zt
        ps = psat(t)
        flag = "  <-- condensation risk" if pv > ps else ""
        print(f"  outer face of {layer:18s} {t:6.1f} C  pv {pv:4.0f} Pa  psat {ps:4.0f} Pa{flag}")
# layer, thermal resistance (RSI), permeance (ng/(Pa s m2))
base = [("inside film", 0.12, 1e6), ("gypsum board", 0.08, 0), ("stud cavity", 3.34, 1200),
        ("sheathing", 0.11, 55), ("membrane", 0.0, 0), ("exterior board", 2.1, 2000),
        ("cladding and air", 0.15, 10000), ("outside film", 0.03, 1e6)]
def build(inner_perm, membrane_perm, cavity=3.34, board=2.1):
    out = [list(x) for x in base]
    out[1][2] = inner_perm; out[4][2] = membrane_perm
    out[2][1] = cavity; out[5][1] = board
    return [tuple(x) for x in out]
wall("A: polyethylene inside, adhered membrane outside", build(3.4, 3))
wall("B: paint only inside, adhered membrane outside", build(350, 3))
wall("C: paint only inside, permeable membrane outside", build(350, 1500))
wall("D: paint only, adhered membrane, all insulation outside it", build(350, 3, 0.15, 5.44))

A: polyethylene inside, adhered membrane outside
  outer face of inside film          20.3 C  pv  868 Pa  psat 2369 Pa
  outer face of gypsum board         19.8 C  pv  528 Pa  psat 2297 Pa
  outer face of stud cavity          -1.1 C  pv  527 Pa  psat  559 Pa
  outer face of sheathing            -1.8 C  pv  506 Pa  psat  528 Pa
  outer face of membrane             -1.8 C  pv  121 Pa  psat  528 Pa
  outer face of exterior board      -14.9 C  pv  121 Pa  psat  167 Pa
  outer face of cladding and air    -15.8 C  pv  121 Pa  psat  153 Pa
B: paint only inside, adhered membrane outside
  outer face of inside film          20.3 C  pv  868 Pa  psat 2369 Pa
  outer face of gypsum board         19.8 C  pv  862 Pa  psat 2297 Pa
  outer face of stud cavity          -1.1 C  pv  861 Pa  psat  559 Pa  <-- condensation risk
  outer face of sheathing            -1.8 C  pv  822 Pa  psat  528 Pa  <-- condensation risk
  outer face of membrane             -1.8 C  pv  122 Pa  psat  528 Pa
  outer face of exterior board      -14.9 C  pv  121 Pa  psat  167 Pa
  outer face of cladding and air    -15.8 C  pv  121 Pa  psat  153 Pa
C: paint only inside, permeable membrane outside
  outer face of inside film          20.3 C  pv  868 Pa  psat 2369 Pa
  outer face of gypsum board         19.8 C  pv  776 Pa  psat 2297 Pa
  outer face of stud cavity          -1.1 C  pv  749 Pa  psat  559 Pa  <-- condensation risk
  outer face of sheathing            -1.8 C  pv  161 Pa  psat  528 Pa
  outer face of membrane             -1.8 C  pv  140 Pa  psat  528 Pa
  outer face of exterior board      -14.9 C  pv  124 Pa  psat  167 Pa
  outer face of cladding and air    -15.8 C  pv  121 Pa  psat  153 Pa
D: paint only, adhered membrane, all insulation outside it
  outer face of inside film          20.3 C  pv  868 Pa  psat 2372 Pa
  outer face of gypsum board         19.8 C  pv  862 Pa  psat 2301 Pa
  outer face of stud cavity          18.9 C  pv  861 Pa  psat 2174 Pa
  outer face of sheathing            18.2 C  pv  822 Pa  psat 2085 Pa
  outer face of membrane             18.2 C  pv  122 Pa  psat 2085 Pa
  outer face of exterior board      -14.9 C  pv  121 Pa  psat  167 Pa
  outer face of cladding and air    -15.8 C  pv  121 Pa  psat  153 Pa

The four walls tell the story of this lesson. Wall A, with a sealed polyethylene retarder on the warm side, stays just below saturation at the back of the sheathing in an average January; on a design night it would cross the line briefly, as Abernathy's profile showed, but the amount of vapour reaching the sheathing through polyethylene is tiny and dries in the spring. Wall B, with only latex paint inside and an impermeable membrane on the sheathing, accumulates moisture at the back of the sheathing all winter, because vapour enters easily and cannot leave. Wall C replaces the adhered membrane with a vapour-permeable one; the outside of the sheathing is now safe, but the sheathing itself is fairly tight when cold and dry, and this conservative model still shows moisture building up behind it. In reality the sheathing becomes more permeable as it absorbs moisture, and walls like C usually work if their interiors have at least a vapour-retarder primer, but they have little margin. Wall D moves all the insulation outside the membrane; the sheathing and studs sit near room temperature, and nothing can condense inside the wall, whatever the interior finish. It is the perfect wall of Lesson 1, and it is the safest choice for a building with high indoor humidity.

The limits of the Glaser method

The Glaser method is a useful first check but a crude model. It assumes steady conditions, while real walls see temperatures and humidities that change by the hour. It ignores the moisture that materials store: wood, gypsum and cellulose absorb and release large amounts of water and buffer short periods of condensation. It ignores liquid water moving by capillarity, rain absorbed by claddings and solar heating. And it ignores air leakage, which, as the previous lesson showed, usually carries far more moisture than diffusion. For important assemblies, such as walls of buildings with high humidity, unusual combinations of materials or novel products, designers use hygrothermal simulation, computer programs that model heat, air and moisture transport hour by hour over several years with real weather data. The results are judged against criteria such as those in the ASHRAE standard for moisture control design analysis, which ask whether surfaces stay wet long enough to grow mould or decay.

Summer, sun and inward vapour drive

In summer the vapour drive reverses in air-conditioned buildings: hot humid outdoor air has a higher vapour pressure than cooled indoor air, and vapour moves inward. A polyethylene sheet behind the drywall then becomes a cold surface for vapour coming from outside, and condensation can form on its outer face, wetting the stud cavity. In Winnipeg, summers are short and moderately humid, and the problem is limited, but it becomes serious when absorptive claddings are involved. Brick, stone, stucco and some fibre cement boards soak up rain; when the sun comes out, it heats the wet cladding, raising the vapour pressure in it far above the indoor level, and drives vapour inward through the wall. This inward vapour drive explains many failures behind brick veneers and stucco: wet sheathing, mould on drywall and stained wallpaper in buildings that never had a winter problem. The remedies are a well-ventilated cavity behind the cladding, which carries the vapour away, a water-resistive barrier that is not too vapour-open, and, in some cases, a variable-permeance retarder on the inside instead of polyethylene.

Variable-permeance membranes, often called smart vapour retarders, were developed for this kind of problem. Their permeance depends on the humidity around them: in winter, when the air on their warm side is dry, they are tight and limit vapour entering the wall; in summer, or whenever the wall behind them becomes damp, they open up and let it dry inward. They give assemblies two drying directions instead of one, which adds margin against both winter condensation and summer drive.

Indoor humidity

The vapour load depends on the indoor humidity, which the architect and the mechanical engineer set together. Houses and offices in a cold climate run dry in winter without humidification, often between 20 and 35 of every 100 parts of saturation, which is why winter condensation in their walls is uncommon if the air barrier is good. Apartments with many people, cooking and showers run more humid. Some buildings are humidified deliberately: hospitals, museums, archives, swimming pools, laboratories and some manufacturing. A museum that keeps its galleries at about 50 parts in 100 all winter, as many collections need, puts several times the vapour load of an office on its envelope, and needs walls, roofs and windows designed for it, usually with the perfect wall approach, very tight air barriers and windows whose frames stay above the indoor dew point. Swimming pools are the most demanding buildings of all; their envelopes are typically designed with hygrothermal simulation and reviewed by a specialist.

Exercise 4: condensation check

This exercise takes about one week.

• List the layers of each wall and roof of your mixed-use building with their thermal resistance and vapour permeance.
• Run the Glaser method for an average January and for a design night; mark every interface at risk.
• Check whether any vapour-tight layer sits on the cold side of the insulation, and if so whether enough insulation lies outside it.
• Identify any absorptive claddings and describe how inward vapour drive is managed.
• Set the indoor humidity for each occupancy with your mechanical consultant and adjust the assemblies where needed.