Building Envelope -- Waterproof Design: Construction Materials, Supplies part 2

Both membranes working together

As we have mentioned previously, liquid water does most of the damage to buildings. Historically, designers and builders did not try very hard to stop air or vapor penetration in the building envelope. As recently as 40 years ago, you were lucky to see a window installer fill the voids around his or her products with a can of open-cell foam. It was rare to see weather stripping seals, and nobody worried much about infiltration. Asbestos shingles were installed over impregnated sheathing. Interior walls were made of wood lath and plaster. All these materials (except for the siding) were breathable. Most buildings were operated at negative pressure, out of ignorance and for economic reasons. It was the way it had always been done! As labor costs and the demand for buildings increased, builders started using paper-faced gypsum boars in lieu of lath and plaster. This was just the beginning.

Changes in building envelope design continued with the energy crisis and rising costs for electricity. People began to care about wasting energy. Sealing the building envelope was seen as a must; it had to be done in hot or cold climates. As a result, designers and builders came to rely more on building air-conditioning systems than ever before. They introduce outside air, don't they? How much? Is it enough to be cool? Few occupants open windows except for those few days in the fall or spring (assuming that they have operable windows). As building envelopes got tighter, it seems more and more people complained about the air in buildings.

Well, surprise! The inside smells worse than the outside in most small cities.

People started reporting illnesses and blamed them on the air in their buildings.

The term indoor air quality (IAQ) started to take on negative connotations.

It seems that buildings were designed and /or being operated after modification by owners and operators in a negative-pressure condition. Even if the architects, engineers, and builders provided enough outside air in their designs, the professionals began to get named in lawsuits at astoundingly increasing rates. Negative-pressure buildings tend to suck in outside air (often hot and moist in those climates). This air gets drawn in at the points of least resistance-at windward windows and doors or at intersections of systems. This moist air seems to always be at a different temperature than the indoor set points for HVAC systems. Pressure and temperature gradients caused this air to move through walls and corridors, condensing in places that occupants could not see or clean. The cycle continued until such a time as a trigger event occurred, such as a hurricane dumping rain for days at a time coupled with a loss of power-and boom!-mold growth increases exponentially.

The best way to prevent mold growth is to stop it before it grows. Minimize the number of spores, and then prevent moisture from dissolving food sources that fuel its growth. Thus the AIB plays a very important role. It permits you to effectively keep out humid air that even may contain new spores and at the same time allows you to cost-effectively maintain positive pressure on the envelope. In some systems, you will see that there is a separate, dedicated sheet material that acts as the AIB. In others, the MRB is used as the AIB. In still others (see ill. 72), you might see a layer of paint over CMU or applied directly over sheathing for the AIB. If a wall sheathing is to be used as an air barrier, the joints must be properly sealed. There are several products on the market, most of which must be protected from UV exposure for long-lasting adhesion. There are many reasons to choose one over the others, but it all comes back to the three big C's-climate, clients, and cash. How much rain is expected? What are the temperatures of the outside air compared with the inside set points? How much does it cost? These are the issues that really matter.

Ill. 72 Paint as AIB. Paint as Air Barrier; Stucco Finish; Vapor Retarder; Concrete Wall; SAM Strip

Ill. 73 Image showing AIB outside MRB.

In mixed arid climates, there may not be a need for moisture barriers. In these situations, you might wish to have the entire wall system breathe. This entire discussion is based on the assumption of air-conditioned interior spaces. In a warm, dry scenario, an AIB that breathes might be a great solution. Locating this AIB inside or outside the insulation would be determined based on heating and cooling day considerations, as well as individual occupant comfort (for single-family buildings). High-thermal-mass masonry on block wall systems could provide a good solution, if coupled with the right sun exposure, for cool, dry climates, where temperatures vary greatly over a 24-hour cycle. In other areas, where you don't want to store the heat energy for long periods of time, low-thermal-mass wood frame wall systems would be better. You would want a low-permeability AIB for when the difference between inside and outside temperatures or humidity is greatest. There are some variable membranes that allow less air and moisture through them when the humidity is high than when it's low.

These "smart membranes" are also manufactured in conjunction with insulation as the facing material. We don't recommend their use in hot, humid areas.

There are many differing views on where the AIB gets installed. This is particularly applicable to multi-component wall systems such as the metal or wood frame examples with sheathing in the rain screen system. In warm climates, AIB's are installed outside the insulation, just like the MRB; we know that. But where exactly does the AIB go? As a general statement, the AIB should be installed wherever the installers have the best opportunity to complexly seal the envelope. In the framed wall examples shown in ill. 73, the AIB goes on over the SAM. This permits the SAM (acting as the MRB) to have reduced pressure on it. The AIB effectively reduces ambient air pressure on the MRB. Another reason is that the SAM adheres better to raw sheathing, such as DensGlas or plywood. You don't really want the SAM over the top of the AIB, especially if a sheet AIB is used.

Wrinkles in the AIB make it difficult to install the SAM without potential leaks where the wrinkles are below. Finally, the fasteners from the AIB will seal them selves as they penetrate the SAM. In this scenario, we have created two effective air barriers because the SAM also functions to prohibit air movement through it.

For the AIB to seal against air penetration, all seams, fasteners, laps, and penetrations need to be sealed with the manufacturer's approved sealing tape.

We need to detail roof and floor wall joint geometries for sealing the AIB and SAM to slab and walls. Roof membranes need to relieve moisture and pressure by venting out at relief vents or under flashings at parapets (see ill. 74). Note in the image that we cantilevered the ladder supports over the face of the roof coping. This prevented roof sidewall membrane penetrations as well as holes in the coping itself. Hot gases can escape under the coping and cleat system through the small voids in the membrane. All our membranes and air vapor and moisture barriers need to be designed to work effectively in concert with the wall and roof skin systems to control pressure and limit infiltration. In the section pro vided (see ill. 4.75), we show a roof edge detail with ventilation designed for relief under the metal flashing. This lets the hot gases out, and keeps wind-driven water out along the edges. Good ventilation leads to a longer lasting roof since the sheathing and membrane stay drier beneath the roof.

Sealants

Ill. 74 Image of roof membrane that will vent at coping.

Ill. 75 Roof membrane will vent at wall receiver, detail. Sidewall, Parapet, etc.

Top of Flashing Tapered Shims (Sealant and B. R. Optional) Hot Gases Escape Under Counterflashing Vented Base Sheet up Wall Cap Sheet Plies Roof Membranes Coverboard Insulation Thermal Insulation Roof Deck

Ill. 76 Three lines of defense. Primary Means of Defense; Stucco, Paint, and Sheathings; Second Line of Defense; Membranes and Flashings; Third Line of Defense; Backer Rod and Sealant

We don't use caulking on the exterior of our buildings very often. Caulking is limited to use where joints don't move. There may be an occasional case where caulking is appropriate on the building envelope, such as … well! we really can't think of a good one. Caulking is for painters and generally is limited to the interiors of buildings where temperatures are relatively stable. As an experienced designer, we consider sealants to be an important part of the building envelope system. We like to think of the skilled worker applying sealants as waterproofing experts. We think of sealants as the third line of defense against water and vapor intrusion. The building skin, with coatings, veneers, and sheathings, is the first line of defense. Membranes and flashings are the second (see ill. 76). Sealants generally are used to reduce the sizes of openings exposed to the weather and to fill voids. We urge you to never rely on sealants by themselves to stop water intrusion into a wall or roof cavity. They are only a part of the entire system.

Sealants are temporary. If selected and applied properly, a good sealant can last for 5 to 10 years without significant reduction in performance. Sealants should be inspected annually and removed and replaced if they are seen to be cracking or separating from one surface or another for (rule of thumb) more than an 1” per 10 feet of length (2.5 centimeters per 3 meters). If more has separated than this, the entire area needs closer inspection.

If sealants are seen to be coming loose from a surface, this often can be the result of several possible root causes. If both edges are still fully adhered, the sealant did not have enough elasticity. It could be the wrong product for the job. You must calculate the expansion and contraction of the materials being sealed, looking at total movement as well as differences in adjacent materials. Then you should select a sealant product that far exceeds the calculated movement. A factor of 2 is minimum; 4 is not uncommon. If the materials are going to move 1/16” (0.15 centimeter), the sealant should be able to stretch (and compress) 1/8” (0.3 centimeter) without coming loose from either of its two adhered surfaces. The higher percentage the material elongates, the less it fatigues and cracks or pulls loose from its bond.

Thermal coefficients of common materials range from 0.000003” per degree Fahrenheit for pine to 0.000007 for Portland cement and 0.0000128 for soft metals such as aluminum. Glass has a coefficient equal to 0.0000047. To get the total expansion, multiply the span in inches times the coefficient for thermal expansion times the _T (difference from low to high temperature the material may see in a year) or, as a rule of thumb, 120ºF. Note that this is not air temperature, but material temperature exposed to the sun.

Let's work a sample problem. For a 100-foot-long (30 meter) masonry wall (see ill. 77), this would be 100 _ 12, (to get total inches) _ 0.000034 _ 120ºF = 0.49” (1.27 centimeters). Let's call this 1/2”. Let us imagine that we are trying to minimize the appearance of the sealant joints, so we might choose to limit them to 3/8” (1 centimeter) in size so that they would be the same size as the mortar joints in our typical masonry wall.

We prefer to design the joints to move about one-quarter of their width so that we can't see them bulging out noticeably in the summer months when materials are at their largest. In the example, the structural system is steel columns and beams on a 25-foot (8-meter) grid. We could introduce three joints along the entire length, giving us vertical sealant joints at the center of interior column lines B, C, and D. We recommend keeping the masonry unit corners laced, without a movement or sealant joint, for both aesthetic reasons and for structural integrity. In this scenario, each joint would move about a third of 1/2”, or 1/6” (0.4 centimeter).

Ill. 77 Plan showing 100' wall.

We would need to specify a sealant whose material properties exceed 1/6” times our factor of 2, or 1/3” (0.85 centimeter). This requirement is easily exceeded by most sealants. A good silicone sealant will be able to stretch five to seven times its installed length (500 to 700 percent). If our joint is 3/8” (1 centimeter) wide, the product could stretch more than 1 1/2” (4 centimeters).

This leads us to the second possible cause of sealant separation-bad design.

Sealant joints are not to be confused with expansion joints; rather, they should work in concert with them. Expansion joints are designed where movement is excessive. Long buildings require expansion joints. The conditions that require expansion joints generally include: (1) new building meeting existing building, (2) different systems coming together, (3) long wings meet main structures, and (4) tall buildings transition to shorter roof areas or any geometric configuration where excessive differential settling is likely to occur. This book does not address expansion joints; however, in many projects, expansion joints are not required, and cracking must be minimized through the use of control joints (typically with sealant), to be referred to hereafter as sealant joints or joints.

There are three principal areas where design decisions must be made relative to joints: frequency, location, and geometry. Frequency can depend on the primary structure, soil-bearing capacity, temperature differential, etc. Joints should be located so as to minimize cracks in exterior walls. A rule of thumb for stucco is to place joints every 20 to 25 linear feet (7 meters) or 250 square feet (20 square meters). Manufacturers and industry associations will publish their recommendations for maximum distances between joints. Location depends on the type of material, support conditions, and many other factors. In brick veneer, we typically place vertical control joints where support angles are fastened to the primary structure, at the center of column lines, for example. In a tall wall, we often locate horizontal joints at places where ledger angles are required or where bond beams exist. The idea is to minimize cracking in the finish materials that could result from movement related to thermal energy, vibration, or differential settling. It is also a good idea to place joints where forces are concentrated, such as in line with window or door heads and sills. Another recommended location is where materials or systems change, such as CMU walls to floor or roof slabs (see ill. 78). We recommend joints in the 1/2- to 5/8-inch-wide range (1.25 to 1.5 centimeter) in most materials. You can work this backwards to determine how far apart the joints need to be. We try not to ever design sealant joints over 1 1/2” wide (3.75 centimeters).

Joint geometry is also important, as discussed in the preceding paragraph.

There are three common kinds of sealant applications. A typical butt joint is perhaps 5/8” (1.59 centimeters) wide. The depth of the sealant should range between half and two-thirds of the width (see ill. 79).

It is common to use foam backer rod material at the back side of the joint. The backer rod also should be clean and dry when installed. It should be com pressed to about two-thirds or three-quarters its full size in the void. Backer rods should be installed so that the remaining depth for the sealant is about half the width at the center. You should have more depth for bonding at both ends where the two materials are joined, for example, in a butt joint (see ill. 80).

Use a 5/8-inch or maximum 3/4-inch (1.5 to 1.75 centimeters) backer rod for a 1/2-inch-wide (1.27 centimeter) joint. The depth should be constant. If the depth of the joint is insufficient for the backer rod, an alternative is bond-breaker tape.

This tape adheres even better than most sealants when applied properly. It can be used in shallow joints to ensure that only two-sided adhesion takes place. In most cases, the tape is applied to the backside to prevent three-sided adhesion.

In cases where the movement exceeds half the elongation, an expansion joint should be used, or a different sealant should be selected.

Poor preparation can be another cause of joint failures. The most common cause of sealant adhesion problems is insufficient surface preparation of bonding surfaces. Most sealants adhere best to dry surfaces, and all surfaces to be in contact with the sealant need to be clean. By this, we mean free of all loose contaminants such as dirt and dust, oxidation, oils, and most paints. Some painted surfaces, when the paint is adhered sufficiently to the metal, such as shop-applied kynar-coated aluminum or powder-coated or baked-on enamel, may not need the paint removed prior to application of the sealant. The sealant selected must be reviewed carefully to make sure that it's right for the materials with which it will be in contact. Some sealants contain materials that will react with solvent-based paints.

We don't recommend applying a sealant over another trade's liquid- or fluid applied material. If, for example, the vertical joint in a tilt wall panel comes in contact with a roofing detail involving sealant, it's best to make the transition with a metal cover plate, counterflashing (see ill. 81), or similar separation/ isolation of trades. Another solution might be to make certain that the two sealant products that come in contact are made by the same manufacturer. It is always a good idea to try to get all sealant products in a single project from a single source (manufacturer, if possible) so as to not reduce or void warranty issues.

Sealants should not be used as adhesives, as a general rule. Materials being joined generally should use mechanical fasteners to resolve forces acting on the joint. For a sliding lap joint connection, one side is fastened fixed, and the other side of the joint is permitted to move as forces are applied. Sealants should be used in either a fillet joint, lap joint, or butt joint configuration (see ill. 82). In no case should there be three-sided adhesion; the two sides being sealed preferably are parallel to each other. While some of the better sealants have great adhesive properties, up to 25 or 30 pounds per inch, they should be using those properties to stay in place, not to hold window jambs in place or copings to the parapet wall.

Ill. 78 Sealant at roof to wall joint. Locate Sealant Joint at Change in Materials and Wherever Likely to Move.

Ill. 79 Joint depth-to-width ratio.

Ill. 80 Hourglass shaped sealant at butt joint.

Ill. 81 Avoid applying two sealants in contact.

Ill. 82 Three kinds of sealant joints.

Ill. 83 Photo, mock-up of roof crown element.

Ill. 84 Sealant pull testing on a mock-up.

Joints need to be tested in order to ensure anticipated performance. There are two steps to sealant testing. The first step is in full-scale mockups. Wall and roof segments should be built to mock test the most challenging conditions.

These typically include window and door openings and wall-to-grade, wall to-soffit, and roof details (see ill. 83). We can't overstress the importance of full-size mockups. Even the best-intentioned designers with time-tested details should make certain that the contractors build the mockups well in advance of wall framing and sheathing. There may be a quarter of an inch one way or the other that could make the details work better. Or a sequence-of construction tip could be learned. You can have pull tests done on sealed joints (see ill. 84) and penetration tests done on the representative systems as mocked up.

In nearly every instance, there are tweaks that can be done to improve the details (at least one of them). Assuming that everything finally passes the tests, the contractor now has a great idea of your expectations. After the final details are worked out and construction is complete, additional tests should be per formed on a portion of the completed building to make sure that the subcontractors who performed the work in place did so at the same level of competence as the mockup. Final thought: Insist on mockups!

Flashings

Another of the very important skin components we rely on to minimize water intrusion is the family known as flashings. Flashings come in a wide variety of shapes, sizes, and materials. They play a very important role. We believe that the first flashings were used in roofing, such as valley flashings where two roof planes slope toward a common intersection (see ill. 85). Valley flashings were used in shake and shingle roof systems to minimize water flowing down one roof plane from getting behind the membrane at laps. Metal valley flashing helps to bridge the gaps between sheathing materials on both sides of the valley. This can function as a conductor for water that gets behind the membranes, as well as a moisture barrier.

Valley flashings typically are light-gauge sheet metal made of galvanized steel, lead, or copper. Exposed metal valley flashings also function to reduce friction of organic material such as leaves and branches that tend to build up in valleys. A buildup tends to make small dams in the valley that can cause water to flow up and behind the membranes placed in the valleys, thereby causing leaks. It is important to maintain barrier-free valleys so that owners can clean the valleys when material accumulates.

Flashings also were common at roof-to-wall conditions, where a piece of L or Z metal (see ill. 86) was placed behind wall finishes and on top of roofing materials to prevent water from coming down the wall after having gotten under the roof membranes. This flashing also makes it more difficult for wind driven rain to get behind the wall membranes. Roof-to-wall flashings are used at the tops of the roof planes, as well as at any roof-to-sidewall intersection (see ill. 87). Chimney flashings fit into a similar condition, where vertical surfaces meet sloping roof planes. The level top-of-roof-plane flashings are often simpler to install than sloped-roof sidewall flashings. Sidewall flashings need to be coordinated with finish materials. In our example (ill. 4.87), the kickout flashing which is required in many jurisdictions is not yet installed.

Ill. 85 Valley flashing.

Ill. 86 Metal flashing at wall to roof. Concrete Tilt Wall Formed Reveal or Reglet Drive Wedges, to Secure in Slot Backer Rod and Sealant Counter Flashing Receiver Metal Base Ply Membrane Transition Flashing "Z'' Closure (Fastened and Set in Sealant, or Soldered as Material Permits) Architectural Metal Roofing Panel

Ill. 87 Roof to sidewall ready for membrane.

Ill. 88 Chimney flashing to roof.

Kickout flashings reduce potential for leaks when water coming down a roof surface terminates into a gutter or similar condition at the low end. The kick out flashing forces water away from the face of the wall reducing amount and force of water striking the wall-to-roof flashing.

Bricks are small, rectangular masonry units that are laid on a level line. The flashing for brick walls to a sloped roof are usually cut to horizontal lengths that coincide with the size of the bricks. This is referred to as step flashing. The horizontal wall must be sheathed and membrane applied, as well as the roof.

Short segments of horizontal flashing are coursed in with each brick from the bottom up. The step flashing is fastened in such a way as to prevent water that might get through or behind the brick from getting behind the wall membrane.

If this is a brick-on-CMU (or brick) backup wall, the flashing should be laid up with the backup wall, installed approximately 8” (20 centimeters) above the elevation of the finished brick. Each successive piece of step flashing is bent tight to the face of the finished brick. The low side of the flashing then is bent to fit with the sloping roof and fastened above the roof membrane (see ill. 88). Many installers recommend installing another piece of roof membrane over the step flashing, continuous from bottom to top.

There is one thing about the preceding that does not protect the joint in wind driven rain. Do you know what that's ? What about the rain or snow that's blowing up the roof? What keeps wind-driven precipitation from getting behind the laps in the step flashing? As usual, it's a question of how far you want to go with your protection.

Ill. 89 Ridge cap.

There are several other flashings used in conjunction with sloped-roof systems. Starting at the high point, there is ridge flashing (see ill. 89), also called a ridge cap. A ridge cap is designed specific to the roofing material and function it plays. In a ventilated ridge cap, air is expelled under the metal. This requires a special water trap so that wind-driven rain doesn't enter the roof system or ceiling cavity below. At the bottom edge of a roof plane, drip metal and perhaps fascia trim flashing pieces are used. For a look at all common flashing shapes, please refer to ill. 90.

Ill. 90 Common flashing shapes. Drip Edge Rake and Corner Base Angle Fascia J-Channel Double Angle Sidewall Drip Cap Endwall F and J Overhead Door Trim Inside Corner Drip Z Counter-Flashing End wall Valley Rake Ridge Caps

Window flashings

Window flashings are also common. The most important window flashing is probably the head flashing, also referred to as a drip cap. It functions to keep water coming down the wall from getting behind the window head. Head flashing gets installed behind the wall membrane above the window opening.

Window sill flashings are becoming more common. The best sill flashings are three-sided welded or molded sill pans. Sill pans keep water that may defeat sealants from getting into the wall cavity. These are commonly made from stainless steel. Stainless steel is weldable and does not react with most other metals. Stainless steel also resists rust and corrosion well, especially 304 and 316 series. Ill. 91 shows a sill pan for a window. This image shows it being checked for fit, however the air and vapor barriers would be installed prior to pan installation. Then the sides were to be stripped in with SAM, over the vertical sill pan lips at the jambs.

One of the challenges of sill pan installation is how to fasten the window system through the sill pan without allowing it to leak. Many window systems, especially curtain wall systems, have large-diameter fasteners through the head and sill that transfer wind loads to the primary structure. Fastening a sill pan without allowing it to leak can be done by fastening it at the sides into jam framing. The fasteners can be located more than 1” (2.5 centimeters) above the bottom edge of the sill pan. Water seldom, if ever, will accumulate in the sill pan to this depth. The front (outer) edge of the sill pan should be the low point of the pan, allowing water to drain out owing to the force of gravity.

Ill. 91 sill pan.

Ill. 92 large bolt holes on curtain wall sill.

Ill. 93 Isometric showing sealant at rear of sill pan.

Ill. 94 SAM strip over sill pan lip, plan view. Sheathing Exterior SAM Membrane Future Finish Mat'L Sill Pan SAM Strip Over Lip SAM Jamb Return Rear Lip on Sill Pan

The more difficult challenge is sealing the large bolt holes (see ill. 92). We recommend a combination of SAM to the top side of the sill pan and application of a sealant under the head of the bolt. Another solution places rubber washers, or O-rings, under the T-clips and F-clips.

Many manufacturers have begun to develop their own sill flashings and will warrant their performance with their window systems. The difference between a sill pan flashing and a sill flashing is the side and rear vertical lips on sill pans. Which should we trust? Would you rather mix concrete on a sheet of ply wood or in a wheel barrow? Think about it! Through years of evaluation, we have come to recommend a minimum of 2” (5 centimeters) of vertical leg height on all sill pans. In a holistic solution, we would design the details so that installers put a three-sided bead of sealant in the bottom inside corner of the sill pan prior to installing the window (see ill. 93). This sealant will make it more difficult for wind-driven water to defeat the sill pan system and blow into the wall cavity, where it can provide moisture for mold growth.

Sill pans need to be well fitted to the opening they are going to protect. On most projects, we try to limit the number of window opening sizes so that the fabricator makes a few standard sill pans and uses them throughout. This places importance on the accuracy of layout and framing. If the openings vary by more than 1/2” (1.27 centimeters), the integrity of the detail can be compromised.

It is important for the trades to be well coordinated and sequenced properly. We prefer the AIB to be complete, outside the exterior wall sheathing first. Then the MRB, in this case, a SAM, goes on the surfaces. Again, we like the self-sealing aspect of a nice SAM to seal fastener penetrations, etc. Then the sill pans are installed. Ideally, there is a very small space between the vertical lips of the sill pans and the SAM. This void should be about 1/8” (0.25 centimeters) or less.

Then SAM stripping is applied to close off the void, making a pathway for any water that may reach the jamb to drain vertically into the sill pan, where it should drain out past the finished face of the wall system (see ill. 94). If the sill pan and related membranes are installed as directed, you will have effectively water proofed the bottom edge of the window opening. It should not need face sealants to perform well. Once the jambs are flashed, the final application of finish materials then can go forward.

Jamb flashings are not yet common. Except where we specify fin-type windows, we seldom see jamb flashings. In our opinion, far too many designers rely on sealants to close off the jamb openings. In the old days, we had wood windows and mostly wood frame construction. After the windows were installed, some times after the finish skin materials were installed (such as brick or stucco), a piece of wood trim was installed to close off the void. This was done to reduce drafts through the void, as well as for improved appearance. This piece of wood trim, often quarter round, or casing would reduce the pressure and volume of wind-driven rain. If you had to replace a window, you simply would set the finish nails using a nail set (driving their heads through the depth of the trim) and pry off the trim. The windows then could be removed by unscrewing their fasteners. This was the typical condition; some builders applied stucco or brick or wood sheathing right over the window jambs, making replacement much more difficult.

Ill. 95 Section through sill at jamb.

In today's construction methods, wood and applied trim are not as common. A large percentage of the commercial, industrial, and government jobs we see use sealants for closure between walls and window jambs. This goes against our first rule of envelope design and detailing: Use gravity, geometry, and then technology. In our opinion, you need a piece of something solid, such as wood or metal, to close off the potential pathway at the jamb. In modern metal windows, many manufacturers use two-piece sill and jamb construction. It is incumbent on our designers to draw the extrusions as they really are. Show all the fasteners, all materials being fastened into, the shims, and shim space to scale. If windows are being fastened to the walls, many of these two-piece window jambs use a pressure plate that gets screwed into the vertical wall framing members.

In this way, jamb fasteners are not exposed. In order to seal the jamb, you must close of the void between the window and wall. The location of the window system in the depth of the wall is very important. Ill. 95 illustrates one way to close off this void, improving the joint geometry. In it, we use a small piece of closure metal to close the gap between the two primary systems.

You will note that we show the shim [usually about 1” (2.5 centimeters) in total width, one half on each side] with backer rod and sealant behind the closure. Depending on the ultimate geometry of the window system, this may have to be a Z-closure piece to make up for nonaligned surfaces. In our detail, the closure metal would be installed prior to a window set from the inside or after one set from the outside. The embedded portion of the closure metal should be outside the MRB and AIB. It should fit tightly to the face of the window jamb. To make it function as a wall opening air barrier, the closure metal should be installed over a continuous bead of flexible sealant.

After the skin is finished, a final fillet sealant should be applied between the closure piece and wall material. The closure metal can be fastened to the window jamb with rivets or machine screws, manufacturer permitting. Head flashings should go on last. Make sure not to void warranties if attaching to a window.

Other wall flashings

Ill. 96 Closure metal at jamb. Stucco or Finish System; Membrane; Sheathing; Z-Closure Metal; Backer Rod and Sealant; SAM Jamb Wrap; Rear Lip on Sill Pan; Window Sill; Window Stool

We recommend that wall louvers be treated similar to window openings. Louvers that are part of an outside air makeup system need special treatment, not at the exterior face of the finish wall, but inside. Owing to the negative pressure (and velocity) of an outside air louver, it can bring in more moisture than a passive louver.

Depending on where in the section these louvers exist, you must plan for the increased amount of potential moisture. In the best case, these louvers are protected by long, overhanging roof planes. This will effectively reduce water droplets and the volume of water that will need to drain out.

The worst case is where a gable louver is high up on a building facade and unprotected. Water separator blades alone are insufficient in even a moderate rain storm. Imagine that you are bringing in 100,000 cfm of outside air makeup into a large building. First, you have to get the water past the wall cavity without it entering the cavity, and then you have to remove the moisture from the air and drain it back out again. A good solution is to start by creating a deep sill pan, sump, or basin just inside the louver section.

This sump needs to be treated like a shower pan or small swimming pool.

Depending on the amount of air being introduced, this pan could measure more than 50 feet wide by 10 feet (15 meters by 3 meters) deep. In this scenario, the vertical flanges would need to be more than 6 feet high (2 meters).

Now you need to use gravity and applied physics to remove the large droplets from the air prior to it coming in contact with filter or heat-exchange media.

Ill. 96 illustrates how you trap the moisture-by making it go down and then up. As it goes up, you increase the cross-sectional area of the plenum, thereby slowing the velocity. This aids gravity in pulling the heavier droplets from the airstream. After the air climbs vertically more than 6 or 8 feet (1.8 meter or 2 1/2 meters), most of the water droplets will have been removed.

This water drains down into the sump, where it's collected and discharged out past the face or, even better, into the storm drainage piping system. In our section, you will see that the discharge piping is about 8” (20 centimeters) higher than the drains, acting as an overflow drain for intense rain events or if the drains get clogged. Another component of the detailed system is a fiber glass screen at the face that reduces the size and energy of droplets reaching the louvers. Note: This also adds significantly to the required surface area of the opening owing to increased static pressure loss of the screening and , as such, may not be aesthetically pleasing or desirable. The good news is that it also reduces bugs and large particles that could clog drains and filters.

Flashings have proven beneficial in a variety of other skin conditions. One of the most common is through wall flashings. Such flashings are used in many places. In single-level buildings, flashings are used near grade. Masonry walls have flashings installed so as to drain water that gets behind the outer wythe.

Wood exterior sheathing relies on flashings, such as Z flashing (also referred to as double-angle flashing), to keep water from getting behind the sheathing.

These flashings are used every 8 feet (2 1/2 meters) vertically or where sheets are butted. During the hurricanes of the 2004 season, many Florida buildings failed where poorly conceived wall flashings between block ground floor walls and frame second-floor walls allowed water to get behind the membranes.

Ill. 97 flashing up the rake.

Roof flashing

On simple V-crimp roofing, one of the most important flashings is called a rake and corner flashing. It very effectively prevents wind and rain from get ting behind the metal roofing or membranes at the roof edges. In addition, it provides for closure between the roof and fascia framing (see image and section view), and because of its tensile strength, helps keep the roof and fascia tied together in hurricane-force winds. We don't think that any other roofing system (nonmetal) uses such an outstanding piece of closure metal flashing.

Inside corners are another good place to use flashing. Most wall finish systems would benefit from the addition of an inside-corner flashing. Wall finish systems such as stucco, brick, block, wood siding, or sheathing or any other non transparent system will likely perform better in high-wind-driven rain if metal corner flashings are used. These flashings would be integrated with all other roof and wall flashings to act as a conveyance system that also protects the softer, more flexible products that may be inside of them, such as felts, AIBs, or MRBs that should not be exposed to the sun.

Low-slope roof systems also rely on flashings for their success. Designers have a variety of shapes and sizes of metal products at their disposal. The most important of them all, copings and receivers, help to keep rain water from getting behind the roofing plies. Most low-slope roofs are designed with parapet walls to prevent observers from seeing the "stuff" on the roof, as well as to protect them from falling rainwater. Interior roof drains are used commonly to conduct the rain that falls on a low-slope roof to the underground rain leader or storm piping.

Metal copings are used to protect the roof edges (see isometric sketch, ill. 99).

These copings serve two very important purposes: (1) to keep water from getting behind the wall membranes and (2) to prevent water from getting behind the vertical roof membranes. Parapet walls vary in height and width. Sometimes you might wish to block the view of roof-mounted mechanical equipment and the like. If the vertical face of the parapet extends more than 12” (0.3 meter) or at most a few feet (0.6 meter) above the roof surface, some manufacturers will not warrant their materials. We would like to have the warranty extended to the out side face of the parapet if we could so that the owners are protected to the fullest extent possible.

Ill. 98 Rake flashing section; V-Crimp Roofing; Membrane; Rake and Corner Flashing; Sheathing; Membrane; P.T. Drip; Fascia

Ill. 99 Coping isometric.

Metal copings can be done in a variety of ways. Depending on the process of construction and coordination of trades involved, the details may vary. The principle is the same. Use metal to protect the top of the wall and to keep water out of both roof and wall systems. Since the roofer may not be on the job site until months after the exterior walls are started, many walls get built before the roofer is on the job. There are occasions when the building needs to be weather resistant or dried in before the copings have been fabricated. In these instances, the walls need to have a temporary (or permanent) membrane installed over the top of the wall. It laps over the wall and possibly roof wall membranes on parapets, preventing rainwater from getting behind them.

Once the metal copings are available, they can be installed over the dry-in membrane in some systems. Saddle flashings can be supplied by the roofer for the wall installer to place behind the wall membranes after the top-of-wall membrane placement and before the top-of-wall coping. This prevents reliance on sealants in the future, and the metal protects the membranes from the sun's harmful rays. It also resists the force of wind-driven rain, birds pecking, etc.

Metal is good.

Copings are not normally provided in lengths exceeding 20 feet (6 meters). The laps or joints often are placed aesthetically but need to be installed to permit movement as the metals expand and contract. Since we like welded metal corners on our copings, not riveted (see Image of riveted corner, ill. 100), we plan for our joints to be located no closer to the corners than 4 feet (1.2 meters) (see ill. 101). Provide details for the joint covers as you would like them built. Some designers like underplate and coverplates with the void between lengths of coping materials about 1/4” apart. The cover plates extend about 2 to 4” (1.5 centimeters) in both directions from the joint (see Section A, ill. 102). We most often choose to fix one side of the cover plate and permit the other side to slide. We prefer to have no penetrations through the horizontal surfaces of flashings or cover plates. We place slotted screw fasteners, or rivets on vertical faces, since they are less prone to have standing water (see Section B, ill. 103).

A slight over-break (also referred to as a hem edge) in the coverplate reduces wind-driven rain intrusion past the cover plate. Double rows of sealant improve the performance of this detail. The under plate stops whatever defeats the cover plate, preventing it from getting behind the membranes.

Ill. 100 riveted coping.

Ill. 101 Plan View. Cover Plate; Welded Corner; Cover Plate; Coping Ill. 102 Section A.

Ill. 103 Joint lap coverplate, section B.

Ill. 104 Receiver metal and termination bar.

Ill. 105 Roof edge flashing detail, similar to standard NRCA. When Sealant Fails, Water can Easily get Behind Roofing Below and Enter the Envelope. This Detail Relies Upon Sealant Being Applied Properly. This Shape is not Often Done Right in the Field.

Install Polyurethane Sealant and Tool to Facilitate Water Run-Off Expanding Shank Fasteners (Approx. 12" [305 mm] O.C., Depending Upon Wind Zone and Local Conditions) Sheet Metal Counterflashing (See Table 2) Fasteners Approx. 8" [203 mm] O.C.

Seal Top of Flashing with a 3-Course of Vertical Grade Roof Cement and Reinforcement Fabric Compressible Elastomeric Tape to Span Irregularities

Depending upon site conditions and building orientation, you might not wish for others to see sidewall roofing membranes on the roof side of the parapet.

Vertical metal panels and stucco are two common materials used on the back sides of parapet walls. Whatever the choice, there needs to be a transition between vertical roofing membrane and vertical finish material. You need to terminate the roof plies so that they stay in place, prevent water from getting behind them, and allow for repair and replacement over time. This is where we often rely on termination bars and receiver flashings (see ill. 104).

Termination bars take many shapes, but are used to evenly distribute forces from fasteners to sheet membranes in order to keep water from getting behind them as water may flow down vertical surfaces above.

There are several acceptable means for terminating roof membranes, and each has pros and cons. The National Roofing Contractors Association (NRCA) has published a wide range of details covering a majority of roof-related conditions. While they do provide a variety of time-tested means that most roofing installers can follow easily and that can be warranted, you will find they offer a range of performance options. Some approved details that are adequate for a warehouse may not be suitable for a condominium. You may find ways they can be improved for your project's specific needs.

Ill. 105 shows a standard NRCA roof edge detail that relies upon sealant at the top. We don't recommend that for every type of project, depending upon the occupancy, maintenance, and other factors. Ill. 104 goes against the first two of our three rules. It breaks gravity and geometry (rules number 1 and 2). That detail (once sealant fails) makes a funnel on top so that any rain water falling down from above has a bigger target. Then it allows any water that hits the funnel to get behind the roofing membrane, where gravity and vapor drive will make sure that it finds an occupied space or cavity to get wet. The third thing it does wrong is to rely on sealant. By now, you should know that this is not a good practice for long-lasting protection. We inherited the detail shown in ill. 104 on a recent project as it was shown on approved shop drawings (reviewed by others).

We met with the installer in a pre-construction meeting. We expressed our concerns. When pressed for an improved detail, we were told, "We can install another funnel above the one we have and caulk it too!" While we admitted that redundancy did reduce the weathering on the lower sealant, thereby increasing its lifespan, we did not jump at the chance to rely on two funnels.

Rather than embrace two funnels as a means for reducing water intrusion over time, we went out on a limb. "Isn't there a better way," we asked? "One that cuts the top edge of the flashing into the surface of the tilt concrete wall, thereby avoiding the funnel top altogether?" This was at a preconstruction meeting with the construction manager, owner's project manager, two representatives from the roofing installation subcontractor, and several other tradesmen. We began to draw a quick sketch on the back of our notebook. As soon as we had finished a few lines, the roofing superintendent acknowledged that he had done something like it once before. We agreed that it was far better than the initial submittal indicated and got the owner to agree. Then the construction manager chimed in with his support. Of course, he was making a commitment the general superintendent had to live with and complained about later, but we all knew it was being done for the right reason. For most commercial, government and similar projects we recommend something like Ill. 106.

Ill. 106 Reglet detail, from WLW. Form or Cut Reveal into Face of Wall, 1" Deep × 5/8" Install Silicone Sealant to Fill Void Utilize Removable Screw Fastener to Permit Removal in Case of Re-Roofing or Repairs in Future This Detail Takes Advantage of Gravity and Geometry to Prevent Water Intrusion When Sealant Fails

After the meeting, we had the contractor send in a revised detail as part of the revised roofing submittal. Since it was part of a concrete tilt wall project, we began to take it a step further. We drew up how it affected the panel joints and both inside and outside corners.

This was about the time the clients decided that they could no longer afford to maintain painted stucco walls located more than 8 feet above the finish grade (where they could pressure wash them). We changed from stucco on lath over sheathing to a metal panel system. This caused us to rethink how we were doing our intersections between parapet walls and vertical wall surfaces. Ill. 107 shows a image taken a few weeks later, after the metal panels had been installed and before the coping went on. We drew up a saddle flashing detail (see one of the sections, ill. 108) that was later installed, after selective removal of a portion of the metal panels. In this way, the saddle went back under the wall membrane as it should be, (see ill. 4.109). This ensures that any water that might later defeat the sealants at the coping to wall joint, will be conveyed out side a protected wall system and not get behind the membranes.

Ill. 107 CMU parapet wall intersecting metal wall system.

Ill. 108 Saddle flashing detail. Vertical Saddle Face Top of Saddle Bent Corner Wood on Brick Front Face of Saddle Finish Coping

Ill. 109 Section at wall. Plywood Substrate SAB Strip Over Saddle Metal Panel, Starter Concealed Cleat Welded Corner Saddle Flashing Finish Coping Wood Plate

Table 5 Insulating values of materials

Number Product Thickness R Value 1 Gypsum board 1/2" 0.45 2 Impreg. Sht'g 1/2" 2 3 Felt building paper 15 lb. 0.06 4 Fiberglass (unfaced) 3.5”. 11 5 Expanded Polyurethane 1”. 5.5 6 Extruded Polystyrene 1”. 4 7 Stucco 3/4" 0.15 8 Brick 4" Nom. 0.7 9 Concrete Block 8" Nom. 1.1 10 Stone 4" Nom. 0.32 11 Cement Plaster interior, w/gypsum 3/4" 0.4 12 Asphalt Shingles 240 lb. 0.44 13 Built-up roof 3 ply 0.3 14 Wood Siding 15/8"2

Insulation

Thermal insulation does a lot more than keep us from getting too hot or too cold. It affects how much it costs us to heat or cool our buildings and plays a significant role in preventing condensation. Insulating materials retard the movement of heat energy though a wall, roof, or any other part of the building envelope. The higher the insulation R value, the slower is the transfer of heat energy through the system. Refer to Table 5 for approximate R values for common insulating materials. There are several effective means for calculating the required minimum amount of insulation for anticipated conditions, but you must start out with good information. Historical data for average monthly temperatures can be used to see what the climate was in the past; however, we should be building to last more than 100 years. What if global climatic changes do occur? How will R10 walls perform in temperatures that exceed 20ºF beyond historical data? Many HVAC engineers only design mechanical equipment that maintains our set points when ambient temperatures are about 90 or 95 percent of the aver age high and low temperatures from historical data anyway. Even if it never gets much hotter, what happens if it doesn't cool off at night? The total aver age insulation required by most minimal codes will not keep occupants comfortable if the world sees wholesale changes. What we are recommending is designing and building future buildings with a greater average R value. Run scenarios with 10 percent higher and lower values. If your envelope performs well, then you will have reached a good minimum R value. Use a 30-year pay back when calculating the marginal return on insulation, and use a realistic, conservative inflation rate of 6 to 10 percent per year for energy costs.

Having said this, we would add that most engineers don't design envelopes;

architects do. Computer software is available, but it can be as simple as drawing lines on graph paper. We created our own spreadsheets using standard software.

We began by placing a list of materials in the wall section in rows, and adding R values for each. We list the inside and outside temperatures and calculate the difference. The _T equation is used to calculate each temperature as heat passes through the wall section. In an R20 wall, for example, a material with an R value of 5 would resist one-fourth the total resistance in the wall. If _T = 28ºF, that material would reduce the temperature one-fourth of that, or 7ºF. We then draw a simple graph in Excel and add that same information to a computer-aided design (CAD) line drawing (see ill. 110).

Insulation design to prevent condensation

Ill. 110 Temperatures in wall section.

Minimize Condensation in or on Wall

1. Keep Temperatures above Dew Point

2. Keep Moisture Levels (RH or Grains) Below Dew Point

3. Increase Permeability of Materials from Moisture Barrier Both Ways

4. Provide Drainage Pathway for Condensation that may Form

It is no longer enough to have a wall or roof system with minimal insulation for creature comfort. As designers, we should do more than look at a map of the region with a blanket recommendation for average R values. We should consider and calculate insulation in order to control temperature relative to dew point in the building envelope. By now, we should have designed air and moisture barriers that will work together to limit the amount of moisture in the cavity. This will reduce the amount of water formed by condensation. By reducing the volume of water, we minimize the potential growth of mold spores.

It is not difficult to calculate temperature gradients through roof and wall sections. The most difficult part is determining which values to use for indoor and outdoor temperatures. There are several means for finding average and record high and -low ambient outside temperatures. The question is what values to use. You are never going to know what inside temperatures to design for. Be conservative. We look at both winter and summer temperatures, since one wall has to perform well in both extremes. In summer, vapors move from the outside (warm side) toward the inside as represented in the dashed lines in Ill. 111. In the winter, vapors move from in to out. We can calculate dew points in walls (also floors and roofs) based upon temperature and relative humidity. This example uses a straight line for humidity, such as if there were no vapor reduction in the walls. The point where temperature intersects dew point is marked with a big circle and other lines to signify the importance of this value. We need to keep the temperature above dew point to prevent condensation.

Recent lawsuits against an architect have come as a result of occupants not being able to afford to use the building heating system in cold winter weeks in Florida. The interior spaces got down below 45ºF (7ºC). During the day, the outside temperature climbed into the low 70ºF range (25ºC). The resulting condensation on the inside face of the gypsum sheathing led to claims against the designer. In this case, the moisture was finding its way in through substituted windows and poor detailing, but it was condensing on the inside face of the drywall. It appeared to form on the metal studs because they were the coldest impermeable surface in the path of temperature driven moisture. You could see where each stud was located through the wet stains on the wall.

Ill. 111 Calculating Temperatures in Wall. Summer 95 Degrees 95% RH; Winter 40 Degrees 80% RH

Considering the recent number of record-high and -low temperatures we have seen and unknown consequences of global environmental changes, we recommend that you use very conservative temperatures for determining the amount of insulation in your designs. The incremental cost for added insulation thickness is very low. To be conservative, we would design for 5 percent below the record low and 10 percent above the record high. We prefer not to use the monthly average figures, using instead the daily minimum and maximum for winter and summer. This establishes the values for outdoor temperatures.

From there, you must determine the inside temperatures. We know people are comfortable in the summer within 5 to 10ºF of 70ºF (25ºC). But people are not always home, and people are different, and some people can't afford air conditioning or heating in the peak-load times. What values should you use? To protect yourself from litigation, use extreme values. Do not settle for low R values because there was only a small space for insulation in the void created by furring strips. Change the wall section to accommodate whatever insulation is required to do the job well. Ill. 112 shows a typical wall section with a table of values that we use to illustrate the process of predicting thermal performance for a frame wall. We list all the materials and airspaces in the first column and the R values next to that. In the third column we indicate the resulting temperature that will occur based on that R value. To get the _T, we first have to calculate the total temperature difference on both sides of the wall, in this example 23ºF. We then sum the total R value for the composite wall, in this case 15.41. We then divide the total R by the R value for that line to get the effect of that material on the temperature and multiply that value times the

_T. The R value for the interior still air film is 0.68. When we multiply 0.68 times 23ºF, the product is about 1.3ºF. We continue the same process until we have calculated the temperature for every point in the wall. Computer spread sheets make this and subsequent comparisons very easy to do.

There are several ways to approach the dew point calculation. You may choose to use a straight line based on wall permeability. This is not very accurate. The best way is to calculate absolute humidity in grains of water per unit volume of air at a certain pressure. As you move through the wall section, you can calculate the grains of water that pass through each successive material based on perm rating.

By calculating a moisture gradient, you can show the dew point values for each portion of the wall. Then we change values for operating and /or ambient, run the calculations again, and draw the charts again (see ill. 113). By doing many variations, you can model likely dew point scenarios. Then you change insulation values and locations within the wall. The goal is to prevent condensation from forming inside the vapor or MRB in any temperature likely to occur.

There are several ways to calculate dew points. The old way was to use a psycho metric chart. There are several Web-site calculators that do the math for you. Type in "dew point calculators," and use the one you choose. Once we establish the predicted temperatures and grains of moisture, we can find dew points. We then look to see where in the wall that value is reached and place our target (circle with cross hairs) at that point on the temperature-gradient line, (see ill. 114). You will note that there is not enough insulation in this wall section to prevent condensation for the sample temperatures, resulting in condensation taking place in the wall. This location is inside the CMU and likely would be creating the right environment for the growth of mold spores. If the target were to occur outside the MRB, you would succeed in preventing water from forming in the wall. There are additives for block that reduce potential bacterial growth in CMU, but the best prevention is prevention of condensation from forming in the block.

Ill. 112 Data and graphed temperature in wall.

Ill. 113 Temperature and dew points.

Ill. 114 Temperature gradients in CMU walls.

Ill. 115 Temperature and dew point.

In the next example we have used a far better insulated wall system, batt insulated metal frame walls with exterior rigid insulation and finish system. We also varied the grains of water in the wall to illustrate how that impacts dew point. Since this wall has a vapor barrier outside the insulation, it's not likely that the moisture level would ever reach more than about 25% in the wall. It would take major leaks in the air or moisture barrier to introduce enough moisture for condensation to reach more than 40 percent in the wall. The three humidities shown are 40, 60 and 80 percent relative humidity, (see ill. 115).

The predicted temperatures are 79 degrees dew point for 60% and 67 degrees for 40% RH. With an R24 wall as shown, and with a vapor barrier it would take a major introduction of moisture from the interior, such as 60 percent humidity or greater to permit conditions in the wall to reach dew point. The designer then should look at the performance of the same wall in winter conditions expected for the location.

You can see in ill. 115 that the same wall would have condensation when temperatures reached in the wall are only a few degrees F lower than outside temperatures (around 8ºF). If condensation in the insulation is to be prevented, we need to reduce moisture and dew point by placing the vapor barrier outboard of the insulation. This wall also performs well in winter conditions, with the wall sheathing reducing moisture from the conditioned (interior) side of the wall as it moves toward the cooler exterior face. By calculating temperature and dew points in the wall in both summer and winter conditions, we can prevent condensation year round. While it may sound like a lot of work, after you have set up the spreadsheets the first time, it's fast and easy to change one value and see the results. Whatever the method you use, the results should be pretty much the same. What is important is performance over time. Insulation is cheap insurance against condensation-related moisture problems, that's , moisture on piping and ducts, as well as within the building envelope. You must keep in mind thermal bridging such as metal framing provides and voids in insulation, as may exist in wood frame construction. Both affect localized temperatures. The preceding and following discussions assume uniform values.

Ill. 116 Temperature gradients, CMU wall.

Masonry walls are not necessarily good for all climates either. While there are advantages to increased thermal mass in dry climates, and moderately cool climates, they can be problematic in hot, humid and cold wet climates, unless properly designed with insulation and vapor retarders in the right locations and with the right values. In a cool, dry climate, in the winter time, the wall section indicated in the following example might be adequate, (see ill. 116). In the summer, however condensation will always take place interior of the vapor retarder.

In the example provided, exterior latex paint is the vapor retarder. It is located on the exterior face of the CMU. Regardless of the humidity, dew points are reached inside the vapor retarder. Depending on the perm rating of the paint, it may sufficiently reduce grains of moisture when new to maintain dew point below temperatures reached in the wall. What happens when the wall gets old is that it may crack, and paint may begin to loose some of its ability to prevent moisture trans mission. That is when localized and /or global condensation may occur. This is why barrier walls need frequent maintenance checks. Paints and sealants must be reapplied when degraded. There is no secondary line of defense on a barrier wall.

Types of insulation

There are many different materials that have insulating properties. Many of them also have structural properties, such as concrete and wood. For the sake of clarity, this section considers only materials that you can count on primarily for their thermal properties. These include loose-fill materials, sheet products, expanding or sprayed foams, and roll goods. Each has a different insulation value per unit thickness, as well as different permeabilities, etc. We use insulation below floors, in walls, and in roof or ceiling systems. Since glass has a different functional role than just thermal, we will not include it in this discussion, other than to say that every component that contributes to the building envelope, contributes to the average R value, and therefore affects air and water movement in and around the skin. Similarly, we will not address air films, wind velocities, and such because they are not really a part of the envelope.

Generally speaking, the more dense a material, the lower is its R value because it's the airspace or voids in the cells that act to reduce the transmission of heat energy. The most dense of all insulating materials is insulating concrete. It differs from normal-weight concrete owing to the addition of other, light materials such as perlite or vermiculite (like kitty litter). It is placed on horizontal bases, most often metal deck or normal-weight concrete. Some lightweight concrete is placed by pumping long reaches using a boom and hopper system.

Workers slope it to or from the drain sump areas as they place it. Once the concrete is fully cured, the first layer of roofing membrane can be installed over it. The cure time reduces the amount of trapped water under the membrane.

When we speak of rigid or semi-rigid insulation, we are commonly talking about rectangular sheets of a chemical derivative of the petroleum industry.

Among the many names you might recognize, are isocyanurate, urethane, and styrene. Semi-rigid boards are adhered or mechanically fastened using impalement pins, typically on vertical surfaces. Rigid boards are used in nearly every application imaginable. On the roof, you have a multitude of possible uses, including tapered insulation systems or constant-thickness boards on sloped decking. Roofing products are installed over the foam. Foam boards are also used under floors and in walls and foundation systems. Foam boards have many benefits, including resistance to water transmission. R values for foam insulation range from 1 to more than 7R, but they average about 3.5 per inch (2.5 centimeters). Batt insulation is typically fiberglass or polyolefin that comes rolled up in different lengths and widths, with choices in backing. Kraft paper, "smart membranes," and aluminum foil are common backings; plastic is not quite so common.

Cellulose-based fill and spun glass fiber can be applied loosely by blowing it from a hopper through a long, large hose.

Spray-applied foam products gained popularity during the energy crisis of the 1980s and early 1990s. It is typically polyurethane foam with R values in the 4 to 6 per inch range. As it cures, it tends to skin over, which gives it more dimensional stability and allows it to be used as an air and moisture barrier as well. It is closed-cell foam, meaning that the cells don't absorb water easily.

These properties help to increase the use of sprayed-on polyurethane foam (SPF) in attics, basements, and crawlspaces. It is also fast and easy to install, making it a low-cost solution.