Sealant Joints and the Temperature You Caulk At
Why this matters
A sealant joint is not a gap filler. It is a strip of rubber that stretches and squeezes every day as the materials on either side grow and shrink with temperature. Its rated movement is measured from the width it had when you installed it. Caulk on the coldest day of the year, when the joint is at its widest, and the bead spends every summer squeezed toward its limit. Caulk on the hottest afternoon, when the joint is narrowest, and it spends every winter stretched. The bead that tears off one side of a coping joint the second winter was usually doomed by the day it was put in, not by the product.
Choosing the sealant itself (chemistry, substrates, primers, paintability) is in Sealant + Caulk Selection Quick Reference. This card is about the one thing every product has in common: it is installed at one temperature and lives at all the others.
Movement capability, and what it is measured from
Most construction sealants sold for moving joints are rated by class under ASTM C920, the standard specification for elastomeric joint sealants. The class number is the movement the sealant has been tested to take, as a percentage of the joint width at installation, in both directions unless the class says otherwise. The common classes are 12.5, 25, 35, 50 and 100/50, where 100/50 means it can extend 100 percent and compress 50 percent. A Class 25 sealant installed in a 1/2 inch joint is rated to handle that joint opening to 5/8 inch and closing to 3/8 inch, plus or minus 25 percent of 1/2 inch, which is 1/8 inch either way.
The key words are "at installation". The rating does not know what season it is. It only knows the width the bead had when it cured. Whatever happens after that is measured from that starting point.
Many non-rated or low-rated products, such as general-purpose acrylic latex caulk, are meant for joints that barely move: interior trim, small static gaps. In a joint that cycles every day in the sun, they will fail regardless of the day you install them. The temperature question matters most for the rated products in the joints that move.
Why joints move
Building materials grow when they warm and shrink when they cool, by an amount set by the material's coefficient of thermal expansion, the length of the piece, and the temperature change of the piece itself.
It is the surface temperature, not the air. A dark metal coping in full summer sun runs far hotter than the air, and the same coping on a clear winter night can run below it. The range that matters for the joint is the range of the material, which is wider than the range of the weather report.
Long pieces of high-expansion material move most. Aluminum moves about 12.8 x 10^-6 inch per inch per degree F, roughly twice as much as steel and several times as much as masonry. Vinyl and other plastics move more than aluminum. A long aluminum coping, a vinyl window frame or a run of plastic trim puts real movement into the joint at each end.
Wood moves with moisture more than with temperature. Wood siding and trim swell when wet and shrink when dry, across the grain. A wood joint caulked while the wood is wet from a week of rain opens as the wood dries.
Worked example: one coping joint, three installation days
A shop is sealing the joints in a dark aluminum parapet coping. The coping comes in 10 foot sections with a 1/2 inch joint between sections, and each joint takes the movement of one 10 foot section. The sealant is Class 25. For this example, assume the coping's surface runs from 10 F on the coldest winter night to 160 F in full summer sun; the real range for a given building comes from its exposure and color, and ASTM C1472 is the standard guide for estimating it.
The total movement. One section is 10 feet, or 120 inches. The full temperature swing of the metal is 160 F minus 10 F, or 150 F.
120 in x 12.8 x 10^-6 per F x 150 F = 0.2304 inch.
Check it another way: 0.2304 inch over 150 F is 0.001536 inch per degree for the section, and 0.001536 divided by 120 inches is 12.8 x 10^-6, the coefficient we started with. Over the year, the joint's width swings through about 0.23 inch.
Installed on a mild day, mid-range. The middle of the 10 F to 160 F range is 85 F. Install the 1/2 inch joint with the metal at 85 F, and the metal has 75 F to warm (160 minus 85) and 75 F to cool (85 minus 10).
- Each 75 F swing moves the joint 0.001536 x 75 = 0.1152 inch.
- Summer: 0.50 minus 0.1152 leaves 0.3848 inch, a compression of 0.1152 / 0.50 = 23 percent.
- Winter: 0.50 plus 0.1152 gives 0.6152 inch, an extension of 23 percent.
Both are inside the Class 25 rating of 25 percent. The bead lives in the middle of its range.
INSTALLED, METAL AT 85 F: joint 0.50 in, bead at rest
[ coping A ]|<----- 0.50 ----->|[ coping B ]
WINTER, METAL AT 10 F: metal shrinks, joint opens
[ coping A ]|<------ 0.62 ------>|[ coping B ]
bead stretched 23 percent
SUMMER, METAL AT 160 F: metal grows, joint closes
[ coping A ]|<--- 0.38 --->|[ coping B ]
bead squeezed 23 percent
Installed on the coldest day. Install with the metal at 10 F and the joint at its widest, 0.50 inch. Every degree from there is warming, so the entire 0.2304 inch swing goes into compression. In summer the joint closes to 0.50 minus 0.2304, or 0.2696 inch, a compression of 0.2304 / 0.50 = 46 percent. A Class 25 sealant is rated for 25. The bead is squeezed out of the joint, bulges, and over a few seasons loses its shape and bond.
Installed on the hottest afternoon. Install with the metal at 160 F and the joint at its narrowest, 0.50 inch. Now the whole swing goes into extension. In winter the joint opens to 0.7304 inch, a 46 percent extension. The bead stretches past its rating every winter, and the first cold snap tears it off one side or splits it down the middle.
Same joint, same product, same total movement of 0.2304 inch. Installed at mid-range, it takes 23 percent each way. Installed at either extreme, it takes 46 percent one way.
What would fix the cold-day install. If the work genuinely has to happen in January, there are three honest fixes, and each comes straight out of the numbers.
- A higher class. A Class 50 sealant in the same 1/2 inch joint is rated for 50 percent. The 46 percent compression is inside it.
- A wider joint. For Class 25 to take all 0.2304 inch in one direction, the installed width needs to be 0.2304 / 0.25 = 0.92 inch. Widening a coping joint is a design change, not a field choice, but it shows why a sealant spec ties width and class together.
- Install at the warmest part of the day, on the side that sees sun. Moving the install temperature from 10 F to 40 F shrinks the one-way compression to 0.001536 x 120 F (160 minus 40) = 0.1843 inch, or 37 percent of 0.50 inch. That is still over Class 25, so this one helps only alongside fix 1.
Picking the day in the field
Nobody runs the expansion arithmetic for a run of window perimeters on a house, and they do not need to. The working rule is to install when the material's surface sits near the middle of its own yearly range, which for most sun-exposed exterior joints in most of the US means a mild spring or fall day, not a January morning or a July afternoon. Read the joint surface with an infrared thermometer rather than trusting the air temperature: on the same mild morning, an east-facing dark frame in the sun and a north-facing one in shade can be 20 F apart, and the shaded one may be near the low end of its range.
Two cases move the rule. Wood trim and siding that have been soaked by a week of rain are swollen across the grain, so the joints between them are at their narrowest. Caulking that day puts the bead in the same position as the hot-afternoon coping, with all of its movement still to come as extension when the wood dries. Let the wood dry first. And joints that live indoors, or on a shaded wall that never sees sun, have a narrow range, so the install day hardly matters for them.
The other limit on a cold day
Every sealant data sheet also sets an application temperature, often around 40 F minimum for the surface and the product, along with a requirement that the surfaces be clean, dry and frost-free. On a cold morning, a surface can be below the air temperature and carry a film of frost you cannot see. That limit is separate from the movement question and is covered in How to Caulk and Seal Outside in Cold Weather. A joint can pass the application limit and still be installed at the wrong point in its movement range, and the reverse.
Movement during cure
A sealant does not reach its rated properties the moment it is tooled. Many cure from the surface inward over days to weeks. A joint that moves a lot while the sealant cures, such as a dark coping caulked in the morning and baking by afternoon, is being worked before the rubber is fully formed, and the bead can wrinkle, crack on the surface or lose bond. Where you have a choice, install in conditions where the material's temperature will not swing far in the next day or two: an overcast mild day beats a clear one with a large daily range, and a shaded side beats a sunny one.
What else the joint needs to survive
The temperature choice only pays off if the joint is built to move. Two details matter here, and the selection card covers them in full: the sealant must stick to the two sides and not to the back of the joint (two-sided adhesion, usually with a backer rod or bond-breaker tape), and its depth must be sized to its width, because a bead that is too deep for its width cannot stretch without tearing. A three-sided bond turns a Class 50 sealant into a crack.
References
- ASTM C920, Standard Specification for Elastomeric Joint Sealants, for movement classes, as called up by the project spec or the product's listing
- ASTM C1472, Standard Guide for Calculating Movement and Other Effects When Establishing Sealant Joint Width
- Sealant manufacturers' data sheets for movement class, application temperature and cure time
- See related: Sealant + Caulk Selection Quick Reference
- See related: The Sealant That Never Cured