I have owned an 1893 house in Edgewood for eleven years and I have made every matching mistake available. The worst was a rear wall repair where I matched the brick beautifully and got the mortar wrong, and the repair is visible from the yard to this day. Nobody looks at the brick. They see the pattern of joints.
Mortar color does most of the work
In a typical wall, mortar is 15 to 20 percent of the visible surface, and it is continuous. Brick is discrete. The eye reads a continuous grid far more strongly than it reads individual units, so a mortar that is two shades too light will draw attention to a patch where a slightly-off brick would not.
Mortar color comes from three things, in order of importance:
- The sand. Sixty-five to seventy-five percent of the mix by volume. Its color and grain size are the color of the mortar. Getting the sand right is 80 percent of matching.
- The binder. Lime is chalky white, gray Portland is gray, white Portland is available and useful.
- Pigment. A last resort for fine adjustment, never the primary tool. Pigmented mortars fade unevenly and the patch that matched perfectly in year one stands out in year five.
The professional route is a mortar analysis on a sample of original joint — $150 to $400, and it comes back with the binder ratio and a sieve grading and color description of the original sand. The homeowner route is to take a lump of original mortar to a few sand suppliers and compare in daylight.
Making test panels, which is the only method that works
Mortar dries several shades lighter than it looks wet, and lime-based mixes keep lightening for weeks as they carbonate. Judging a match from a wet trowel is guaranteed failure.
What works: mix three or four candidate batches, point a few joints of each in an inconspicuous area, label them, and look at them after two weeks in dry weather. Then choose. It costs a Saturday and it is the difference between an invisible repair and a permanent one.
Do the same for joint profile. A concave joint on a wall originally struck flush changes the shadow line down the entire elevation, and shadow reads at a hundred feet where color does not.
One thing I got wrong for years: judging a match by holding a sample brick up against the wall at arm’s length. Stand back thirty feet, which is where anyone actually sees the elevation from, and half the differences that looked serious disappear while the joint pattern becomes the only thing you notice.
Matching brick
Priority order, which is not what people expect:
- Size. Old brick varies but averages differ from modern modular by enough that a few courses will not line up. This is the one thing you cannot fudge.
- Texture and face treatment. Sand-struck, water-struck, wirecut and pressed brick all catch light differently.
- Hardness and absorption. A modern brick set into a soft wall will not weather at the same rate and can cause damage to its neighbors.
- Color. Last, because new brick weathers toward the wall over about five years, and a slightly-off color usually resolves itself.
For the sourcing question — salvage versus new — see when salvaged brick is worth it. The Brick Industry Association at gobrick.com publishes the technical notes on unit sizes and properties that make the size comparison straightforward once you know what you are measuring.
Frequently asked questions
Can I tint mortar to match? Yes, with mineral oxide pigments at up to about 10 percent of cement weight, but treat it as fine tuning after the sand is right. Pigment-heavy mixes are weaker, and they fade at a different rate from the surrounding wall.
How long before a repair blends in? Three to eight years for brick in an exposed location, longer on a sheltered elevation. Mortar takes one to three years to lose its fresh look. A repair that is invisible on day one was probably tinted, and it may look wrong later.
Where do I get the original sand? Local pits, which is usually where it came from originally. Take a sample of original mortar with you and compare washed sands in daylight. Bagged silica from a big-box store is uniform, pale and almost never right for a nineteenth-century wall.