The Sealant That Never Cured

Why this matters

A caulk bead that washes down the siding is the kind of callback that costs more than the original work, because the redo includes cleaning someone's house. It also tends to land on a shop that did everything a reasonable person would check: the product was right, the joint was prepped, the air was above the number on the cartridge when the bead went on. This case is about why that last fact was true and still not enough, and about the afternoon cutoff the shop wrote afterward so the next crew does not have to reason it out on a ladder at 4:30.

The call

A painting and trim crew of two spent a late-October Tuesday recaulking exterior trim on a two-story house ahead of a repaint scheduled for the following week. They caulked the east and south sides from 9 a.m. to 1 p.m., broke for lunch, and did the west and north sides from 2:30 to 5:15 p.m. Every joint got the same exterior acrylic latex caulk, from cartridges out of one case with a single lot number.

Thursday afternoon the homeowner called. It had rained since mid-morning, and there were milky white streaks running down the siding below most of the windows on the west and north sides. Where she touched a bead, it smeared. The east and south sides looked fine.

Four suspects, and what removed each

The lead went out Thursday at 3 p.m. with the job notes, the empty cartridges from the truck and a putty knife. He worked the possibilities in order of how cheaply each could be checked.

A bad batch. Caulk that has frozen in storage or aged past its shelf life can fail to cure anywhere it is used. But the east and south beads, from the same case and the same lot, were firm, slightly springy and paintable-looking. A batch fault does not respect the compass. Removed.

The wrong product. An interior-grade caulk on an exterior joint will soften and wash in rain. The empty cartridges in the truck all read exterior grade, and they were the same cartridges used in the morning. Removed.

Joint geometry. A bead laid too deep, with no backer rod, can skin over and stay wet in the middle for days, and a rain can open the skin and wash the core out. He cut two beads out on the north side. Both were about a quarter inch wide and roughly as deep, which is a normal trim bead, and they were soft all the way through, skin included. A deep-core problem leaves a firm skin over soft material; this had no skin worth the name. Removed.

Paint or wash on top too soon. Nothing had touched the beads. The repaint was a week out and nobody had pressure-washed. Removed.

What was left was the one thing that differed between the firm beads and the soft ones: the time of day they went on, and the weather the bead lived through afterward.

What the cartridge said and what the night did

The cartridge's directions, typical of water-based exterior caulks, said to apply with air and surface above 40 F and to protect the bead from rain and from freezing for 24 hours. It also said, in smaller print, that cure times are given at moderate temperature and humidity and lengthen in cool or damp conditions.

The lead pulled the hourly record for Tuesday and Wednesday from the same weather service the shop reads each morning:

Day  Hour   Air  Dew   Sky      Wind
           (F)  pt(F)           (mph)
Tue  2pm    52   36    sun      5
Tue  4pm    47   36    sun      3
Tue  5pm    44   36    clear    2
Tue  7pm    39   35    clear    calm
Tue  11pm   34   33    clear    calm
Wed  6am    30   29    clear    calm
Wed  2pm    50   31    sun      6
Thu  10am   46   44    rain     10

Read against the cartridge, the 5:15 p.m. bead went on at 44 F, inside the 40 F minimum, and a person standing on the ladder would have called it a fine afternoon. Here is what the bead then lived through.

• The evaporation slowed to almost nothing within two hours. Acrylic latex caulk cures by losing water, so its cure rate rides on temperature and on how dry the air is. By 7 p.m. the air was 39 F, under the cartridge minimum, and within 4 F of the dew point. A bead in those conditions is not curing; it is waiting. See related: Humidity and Cure: Moisture-Cure, Evaporative and Reactive.
• Dew settled on it. Clear sky and calm air are the setup for radiational cooling: the trim radiates heat to the open sky and runs several degrees colder than the air. With the air at 34 F and the dew point at 33 F by 11 p.m., any surface even a degree below the air was at or under the dew point, so the beads were wetted from outside while they were still trying to lose water from inside.
• Then it froze. By 6 a.m. the air was 30 F, and the trim, running colder than that on a clear calm night, would have been well below freezing. Water in an uncured latex bead that freezes disrupts the film before it can form, which is why cartridges ask for freeze protection and not only a minimum temperature at application. A surface can frost with the air still above 32 F on nights like this one. See related: Radiational Frost: Why a Surface Freezes Above 32.
• Wednesday did not rescue it. A 50 F afternoon in sun gave the west beads a few hours of slow drying. The north beads, shaded all day, got less.
• Thursday's rain found a bead that was still soft. The rain arrived about 41 hours after the last bead, well past the cartridge's 24 hours. That 24 hours is a figure at moderate conditions, and this bead had spent most of its first night below the cartridge minimum and part of it frozen. Its clock had barely run.

The pattern on the house confirmed the reasoning without needing a lab. North-side beads, laid last and in shade, were the softest. The west beads laid at 2:30 were softer than the morning work but firmer than the north. The morning beads on the east and south had roughly five to nine hours above the 40 F minimum, much of it in sun, before the evening drop, and they were fine. The damage tracked how long each bead had above the minimum before evening, which is the variable the cartridge describes and the crew did not read.

Why the crew missed it

Nobody on the crew was careless. They read the minimum on the cartridge, saw it was 44 F at their last bead, and treated the minimum as a gate at the moment of application. It is not. The application minimum and the protection period are two separate limits, and on a falling afternoon the second one is the binding one. The forecast that showed 30 F by morning was on the lead's phone at 7 a.m. that day. It was read for whether the crew could work, not for what the work would sit through. See related: Sealant Joints and the Temperature You Caulk At, which owns the general rule.

The redo

The repair cost more labor than the original, and the order mattered.

1. Clean the siding first, the same day, from a ladder footed on firm, dry ground. Uncured latex residue washes with water and a soft brush while it is still fresh; once it dries on painted siding in streaks it takes scrubbing and sometimes touch-up. The lead and one tech washed the west and north sides Thursday evening under the eaves, out of the rain, with a ladder standoff so the ladder bore on the wall and not on the gutter. Wet siding and wet rungs are slick, so they worked with the ladder at the correct angle and someone footing it, and stopped when the light went.
2. Strip every afternoon bead, not just the ones that smeared. A bead that looks firm on the surface but went on after 2:30 had the same night. They cut out all west and north beads with a utility knife and a plastic scraper, which does not gouge the trim paint the way a metal blade can.
3. Let the joints dry before recaulking. A joint that took a day of rain holds water in the wood and the old paint. They waited for a dry forecast and confirmed the joint surfaces were dry to the touch and above the cartridge's surface minimum with an infrared thermometer.
4. Recaulk inside a window that fits the protection period. The following Monday showed highs near 60 F, lows in the low 40s and no rain for three days. They caulked west and north from 9 a.m. to 12:30 p.m., which left the beads roughly six hours in the 50s before evening.

The original west and north work took about 2.75 crew-hours. Cleaning, stripping and recaulking took about 6.5, roughly 2.4 times the original, with none of it billable.

The other option the lead considered was switching products. Some solvent-based and hybrid-polymer sealants list application minimums lower than a typical latex caulk, and some cure by reacting with humidity rather than by drying, which changes how cold, damp air affects them. Those figures are product-specific, and a sealant that will take paint has to be confirmed compatible with the paint system on the trim, so the choice is made from the two data sheets, not from the shelf. Where the schedule cannot move, that is the right fork; here it could, and the original product was what the paint system was specified for.

The afternoon cutoff the shop now uses

The shop wrote one rule for water-based sealants and caulks from October through March, and the lead checks it against the hourly when he writes the day's plan.

• Each bead gets at least 4 hours with the air forecast above the cartridge minimum plus 5 F. Find the hour the hourly drops below that line and count back 4 hours. That is the last bead for the day.
• If the overnight low is forecast below 32 F, or rain is forecast within 48 hours, water-based caulk stops at 1 p.m. or the job switches to a product whose sheet covers those conditions.

Run against the Tuesday in this case: the cartridge minimum was 40 F, so the line is 45 F. The hourly crossed it between 4 p.m. (47 F) and 5 p.m. (44 F), call it about 4:40, and 4 hours earlier is about 12:40. The second rule fired twice, with a 30 F low forecast and rain about 43 hours out from a 3 p.m. bead. Both rules put the last bead between 12:40 and 1 p.m. on that day, which would have kept the morning's east and south work, short of its last twenty minutes, and moved every bead that washed.

The numbers are a starting point. A shop in a dry climate where evenings stay well above freezing in fall will find the 4 hours conservative; a shop on a shaded northern street with damp air may need more. The rule's real job is to make the crew read the forecast for the bead's first night, not just for the crew's afternoon. Reading the hourly that way is covered in Reading a Forecast for the Work, Not the Weekend, and the record that would have shown this on Tuesday is The Coating and Sealant Conditions Log SOP.

References

• Manufacturer cartridge directions and product data sheets for the sealant in use, which set the application minimum and the protection period.
• See related: Sealant Joints and the Temperature You Caulk At
• See related: Radiational Frost: Why a Surface Freezes Above 32
• See related: Humidity and Cure: Moisture-Cure, Evaporative and Reactive
• See related: The Coating and Sealant Conditions Log SOP