Shrinkage Compensation in Dry Mix Mortar
Thin Sections, Bonded Substrates, and Why Mortar Is Not Small Concrete
Most guidance on shrinkage compensation is written for reinforced concrete, where steel provides the restraint and sections are thick. A dry mix mortar is a different problem: the layer is thin, it dries quickly, and the restraint comes from the substrate it is bonded to. That changes what can go wrong, what to dose, and how to test it.
The Substrate Is the Restraint
Shrinkage compensation only works when expansion is restrained; unrestrained expansion does little useful work. In reinforced concrete the restraint is the steel, and the designer can specify a reinforcement ratio to obtain it.
In a bonded mortar there is usually no steel. The restraint is the substrate: the concrete floor, the wall, the base slab that the mortar is bonded to. This has three consequences that shape the whole problem.
- Restraint is one-sided and total. The bonded face cannot move at all while the free face can. That gradient is what produces curling and edge lifting when shrinkage wins, and it is why a mortar can fail at the bond line rather than by cracking through the section.
- You do not control the restraint. A rough, sound, saturated concrete substrate restrains very differently from a smooth, dusty or dry one. The same mortar can perform differently on two floors in the same building.
- Bond strength becomes part of the shrinkage system. If shrinkage stress exceeds bond strength, the layer debonds. Compensating shrinkage protects the bond as much as it protects against cracking.
Thin Sections Change the Timing
The expansive reaction needs water. A thin layer loses water faster than anything in concrete practice.
Drying starts almost immediately
A 5 to 30 mm layer has a very high surface-to-volume ratio. Water leaves through the exposed face and, on a dry substrate, into the substrate as well. The moisture needed for ettringite growth can be gone before the expansive reaction has done its work.
The useful window is shorter than in concrete
Expansion has to occur after set but while the matrix is still compliant. In a thin mortar that window is compressed at both ends, so curing during the first hours matters more here than in a mass pour.
Substrate suction competes for water
A dry, absorbent substrate pulls mixing water out of the fresh mortar. This starves both the binder hydration and the expansive reaction, and it is one of the most common site causes of underperformance. Substrate preparation is part of the formulation’s success.
Which Shrinkage Are You Compensating?
Dosing an expansive agent against the wrong shrinkage mechanism produces disappointment that looks like product failure. Three act in mortars, on different timescales.
Plastic shrinkage – before setting
Surface water evaporates faster than bleed water arrives, and the still-plastic surface cracks. An expansive agent does nothing about this, because it acts after setting. The answers are wind protection, evaporation control and mix design, not dosage.
Drying shrinkage – days to months
Water leaves the hardened pore structure and the material contracts. This is the mechanism that CSA expansive agents address most effectively, and it is the dominant one in most bonded mortars.
Autogenous shrinkage – early, in low water-cement systems
Self-desiccation as hydration consumes water internally. It matters in high-strength, low water-cement products such as some repair mortars and grouts. Expansive agents have a weaker effect here than on drying shrinkage, so a high-strength mortar dominated by autogenous shrinkage needs internal curing or shrinkage-reducing admixtures alongside, not simply more expansive agent.
By Product Type
Starting directions from supplied product information. Dosage is by trial; see the dosage guide for the reasoning behind the ranges.
Repair mortar and bonded overlays
The classic case for compensation, because the substrate restrains fully and any shrinkage goes straight into bond stress. Compensation protects the bond line. Watch substrate preparation and pre-wetting as closely as dosage, and test on the substrate type the product will actually meet.
Non-shrink grout
Here dimensional stability is the specification rather than a nice-to-have, and the confinement is severe: grout under a baseplate is restrained on nearly every face. Our CSA Binder starting direction for non-shrink grout with an expansive agent is 10 to 15% of the OPC blend. Flow and dimensional stability must be assessed together, since adding water to recover flow costs stability.
Self-leveling compound and screeds
Large plan area, thin section, one bonded face: the geometry that curls. Compensation reduces edge lifting, but it cannot rescue a product whose water content is too high or whose curing is neglected. Flatness measured at 24 hours and again at 28 days tells you more than shrinkage bars alone.
Waterproof and crack-control mortar
Crack risk is the whole point of the product, so compensation is central. Check the interaction with any redispersible polymer and waterproofing admixture, since polymer films and the ettringite reaction develop on different timescales and can interfere.
Tile adhesive
Usually not a compensation application. The layer is very thin and the tile itself restrains the surface; the governing properties are open time, deformability and bond. Adding an expansive agent here is more likely to disturb setting than to solve a shrinkage problem.
Testing at Mortar Scale
Concrete practice measures restrained expansion on specimens with a restraining bar. That method is still the reference for classifying an expansive agent, and our product information references restrained expansion in the CSA-I, CSA-II and CSA-III direction. But a bar-restrained prism does not represent a thin bonded layer, so it should not be your only test.
For a bonded mortar, add tests that reproduce the real restraint:
- Bonded overlay or ring-type restrained tests, which impose restraint in a geometry closer to the application.
- Curling and edge-lift measurement on a slab specimen for screeds and self-leveling compounds, since curling is the failure mode users actually see.
- Bond strength after drying, not only at early age, because shrinkage stress accumulates against the bond over weeks.
- Testing on the real substrate, prepared and pre-wetted the way site practice will prepare it, rather than on a laboratory ideal.
- Curing that matches reality. If site curing will be minimal, evaluate under minimal curing. A result obtained under 14 days of moist curing is not a prediction for a floor that gets none.
Frequently Asked Questions
Does a shrinkage-compensated mortar stop cracking completely?
No. It reduces the tensile stress that drives cracking by storing a compression first. Cracking can still occur if restraint, curing, thickness or substrate conditions fall outside what the design assumed.
How much expansive agent should a mortar contain?
Our supplied product information gives 6 to 8% for shrinkage-compensating concrete and 9 to 10% for post-poured strips, subject to trial. Mortars differ from concrete in restraint and drying, so the concrete figures are a reference point rather than a transferable dosage. See the dosage guide.
Why did my repair mortar debond instead of cracking?
Because in a bonded thin layer the substrate restrains the bonded face completely. If shrinkage stress exceeds bond strength, the layer releases at the interface rather than cracking through. Compensation protects the bond line as much as the section.
My screed curled at the edges. Will an expansive agent fix it?
It can reduce curling, which is caused by the moisture gradient between the drying top face and the restrained bonded face. But excessive water content and inadequate curing are usually the larger contributors, and dosage will not rescue either.
Do I still need to cure if the mortar is shrinkage-compensated?
More than ever. The expansive reaction consumes water, and a thin section loses water quickly. Without moisture during the expansive window the designed expansion never develops, so a compensated mortar that is not cured can perform worse than expected.
Can I use the same dosage as my concrete mix?
Not reliably. Restraint conditions, section thickness, drying rate and binder content all differ. Re-establish the dosage for the mortar system by trial.
Does an expansive agent help with plastic shrinkage cracking?
No. Plastic shrinkage occurs before setting, and the expansive reaction acts after setting. Control evaporation instead.
Related Products
CSA Expansive Agent
Learn MoreCSA Binder
Learn MoreCSA Cement
Learn MoreDry Mix Mortar Premix
Learn MoreRelated Reading and Solutions
How CSA Expansive Agent Works
Read ArticleCSA Expansive Agent Dosage Guide
Read ArticleShrinkage Compensation Solution
View SolutionTechnical Note
This article is intended for technical communication and product selection discussion. Dosage directions quoted are typical starting points from supplied product information for formulation testing, not a specification. Actual performance depends on formulation design, raw materials, sulfate source, water amount, additives, substrate condition, restraint, curing regime, temperature and application method. Trial mixing and local verification are required before commercial use.
Designing a Shrinkage-Compensated Mortar?
Tell us the product type, layer thickness, substrate and how it will be prepared, the curing you can realistically expect on site, and your base cement. We can suggest a starting direction and support trial mixing.
Send Your Inquiry