What the Surface Defect Pattern Tells You: Dusting vs Scaling vs Crazing Single-Observable Decision Tree

Why this matters

Three of the most common surface complaints on flatwork (dusting, scaling, crazing) get blamed on each other constantly, and the wrong diagnosis sends you down the wrong repair path. A crazed surface that someone treats as scaling gets needlessly ground and overlaid; a dusting surface that someone seals over keeps dusting under the sealer and fails it. The good news is that each defect has a single dominant observable you can read in seconds, and each points to a distinct cause in the finishing, bleeding, or curing window. This tree keys off that single observable so you can name the defect on the first look, then trace it back to the process failure that caused it so the repair holds and the next pour does not repeat it.

Symptom presentation

• Dusting: a fine, loose powder rubs off the surface under foot traffic or a dragged hand. The surface looks intact but never hardens at the top.
• Scaling: the top layer flakes or peels away in thin sheets or patches, exposing the aggregate or a rougher layer beneath. Often starts at edges or high traffic.
• Crazing: a network of fine, shallow, interconnected cracks (like cracked glaze or a dried mud flat), visible especially when the surface is damp and drying. The surface is sound and does not come off.

Per PCA IS177 and ASTM C672 (scaling), these are distinct mechanisms with distinct fixes.

Quick check (the single observable)

Drag a coin or hard plastic edge across the surface and observe what happens:

• Powder comes up, surface stays flat: dusting.
• A flake or sheet lifts, leaving a step: scaling.
• Nothing comes up, but you see a fine crack web: crazing.

That one test branches the whole tree.

Isolation tree

Branch A: powder lifts (dusting)

The top paste is weak and underhydrated. Causes, in order of likelihood:

1. Finishing while bleed water was on the surface. Troweling bleed water back into the surface raises local water cement ratio at the top, leaving a weak skin. The classic and most common cause.
2. No curing, so the surface paste dried before it could hydrate.
3. Carbonation of fresh surfaces from unvented heaters in cold weather enclosures (CO2 reacts with surface paste).
4. Too much fines or dirty aggregate.

Confirm: the dusting layer is thin and uniform; a scratch test (per a hardness probe) shows the surface scratches far easier than the body. Per ACI 302.1R, never finish water back in and never start finishing while bleed water stands.

Branch B: flake or sheet lifts (scaling)

The surface bonded layer is delaminating, usually from freeze thaw with or without deicers acting on a non durable surface. Causes:

1. Inadequate air entrainment for a freeze thaw exposure (target around 5 to 7 percent for severe exposure per ACI 318 durability tables).
2. Deicer salts on a surface that was finished too wet or not air entrained (ASTM C672 simulates exactly this).
3. Premature finishing trapping bleed water, then freeze thaw exploiting the weak skin.
4. No cure, low strength at first freeze.

Confirm: scaling concentrates where water and salt pool, and worsens after winters; ASTM C672 visual rating describes the progression. Distinguish from spalling (which is deeper, often reinforcement or pop out driven).

Branch C: fine crack web, nothing lifts (crazing)

Surface only. Caused by rapid surface drying or a high surface paste content shrinking faster than the body. Causes:

1. Rapid early evaporation (sun, wind, low humidity) before cure.
2. Overworking the surface, bringing up fines/water.
3. Fog/water sprayed during finishing then troweled.
4. Premature or uneven curing.

Confirm: cracks are shallow (do not penetrate), the surface is structurally sound, and the pattern is most visible when wetted and drying. Per PCA, crazing is cosmetic and does not progress structurally.

Confirming diagnosis

Cross check with the placement record: weather and evaporation rate (ACI 305R 0.2 lb/sq ft/hr line for plastic surface effects), air content from the batch ticket (scaling), and the finishing log (was bleed water present at finishing). A core or scratch hardness on the dusting/scaling surface confirms a weak skin versus a sound body.

Two confusions are worth naming because they drive wrong repairs. First, crazing versus mapping/structural cracks: crazing is shallow and surface only and the pattern cells are small (often well under an inch or two across); a structural map crack penetrates and the cells are larger, so probe depth before calling it cosmetic. Second, scaling versus spalling and pop outs: scaling is a thin surface flake driven by freeze thaw on a weak skin, while spalling is deeper material loss often tied to reinforcement corrosion or a hard reactive aggregate particle (a pop out). If you see a single conical crater with an aggregate at the bottom, that is a pop out, not scaling, and the fix is a patch, not an air entrainment lecture. Matching the right defect to the right mechanism is the entire value of reading the single observable first.

Remediation

• Dusting: remove loose material, then apply a penetrating silicate densifier to harden the surface; severe cases need grinding to sound concrete and a thin overlay. Do not seal over a dusting surface (the sealer delaminates with the dust).
• Scaling: grind/clean to sound concrete and place a bonded repair or overlay rated for freeze thaw; fix the cause with proper air entrainment and cure on the next pour. Keep deicers off new concrete the first winter.
• Crazing: cosmetic only. Clean and, if appearance matters, apply a sealer or stain to blend; no structural repair needed.

References

• PCA IS177, Concrete Slab Surface Defects: Causes, Prevention, Repair.
• ASTM C672, Standard Test Method for Scaling Resistance of Concrete Surfaces Exposed to Deicing Chemicals.
• ACI 302.1R, Guide to Concrete Floor and Slab Construction (finishing, bleed water).
• ACI 305R, Guide to Hot Weather Concreting (evaporation and surface drying).
• ACI 318, durability air content requirements for freeze thaw exposure.