If you read product literature long enough, you will see "freezer-thaw tested" used to cover everything from surface scaling to full internal durability. The tests behind those words are different, and knowing which one a claim rests on is the difference between a useful result and a marketing borrowing.
ASTM C666: internal freeze-thaw durability
C666 cycles concrete specimens through freezing and thawing, typically while measuring the fundamental transverse frequency of the specimen. As internal damage accumulates, stiffness drops; the test expresses this as a durability factor. C666 therefore measures the concrete's internal resistance to repeated freezing: the property that determines whether a saturated element survives years of winters.
It does not involve deicing chemicals. It is an internal durability test of the paste and air void system.
ASTM C672: surface scaling with deicing chemicals
C672 instead evaluates the surface: salted specimens are cycled through freezing and thawing, and the result is read as both cumulative mass loss and a visual scaling rating from 0 (no scaling) to 5 (severe scaling with coarse aggregate visible over the whole surface).
C672 is a surface-scaling test. It tells you how the finished surface stands up to saline freeze-thaw exposure. It does not establish internal freeze-thaw durability; a specimen set can look clean on the surface (low mass loss) while the deeper durability question remains C666's business.
Why the tests get conflated, and why it matters
Product summaries often compress both into "freeze-thaw tested." The conflation matters because:
- A deicer-scaling result does not predict internal durability in saturated, unsalted freeze-thaw.
- An internal durability result does not by itself answer how a salted surface scales.
- Field performance depends on both, and on exposure specifics (drainage, saturation, salt load).
The honest citation names the test, the specimen condition, the control, and the result, as our PRO Tek test evidence article does for one product's data.
What each test cannot show
- Neither is a service-life calendar. A durability factor or mass loss result does not convert into "years of life" without project-specific correlation that does not generally exist.
- Neither characterizes cracking from other mechanisms (structural movement, ASR, corrosion), which need their own assessment.
- Laboratory specimens are cast and cured under controlled conditions; field concrete is not. Test results indicate the material's potential under the method's conditions, not a prediction of performance for a specific structure.
The wider reading habit
This pair is one example of a general rule: name the exact test, describe the specimens, state the control and the result, and let the method's scope define the claim. FHWA's bridge durability research and ASTM's standards descriptions are consistent on this. Our C1202 explainer covers the same discipline for a different test, and the test-evidence map shows all of them side by side for one product.
How the two tests are run in practice
Understanding the mechanics helps explain why their results differ. In C666, specimens are typically submerged and cycled through freezing and thawing on a rapid schedule (roughly one cycle every few hours), and stiffness is tracked throughout, because the property of interest is internal damage accumulation. In C672, specimens sit in a shallow pond of salt solution, freeze, and thaw on a slower schedule, and the evaluation reads the surface: how much material has scaled off, and how bad it looks against the rating scale.
The consequence is that the two methods stress different parts of the concrete. C666 exercises the paste and the air void system throughout the specimen; C672 works on the finished surface and its near-surface paste, with the salt solution actively involved. A concrete with an excellent air void system but a weak, over-finished surface can look good in C666 and poor in C672. The reverse combination is also possible. Neither result is "wrong"; they are answering different questions, and both questions exist in the field.
Reading product claims that cite these tests
The practical checklist when you see either test cited:
- Which test, exactly? C666 and C672 are different standards; "freeze-thaw tested" without a designation answers nothing.
- What procedure and how many cycles? C672 results are typically reported at 50 cycles; C666 runs hundreds of cycles and reports durability factors at defined points.
- What were the specimens? Size, age, curing, and substrate condition all shape the result.
- Was there a control? A treated-versus-untreated comparison on identical specimens is the informative design; a single number with no comparison is not.
- What category or units? Durability factor for C666; mass loss and visual rating for C672. Vague language ("excellent freeze-thaw performance") without these fields is a flag.
This is the same discipline our technical data sheet article recommends for any performance claim: name the method, describe the specimens, state the control, and let the result mean only what it means. Where we discuss a specific product's data, we hold our own content to that standard, as the PRO Tek evidence walkthrough shows.
What this article does not tell you
It does not provide acceptance criteria (they belong to specifications) and does not evaluate any product's results. Its purpose is to make the two test names mean what they mean when you encounter them.
Where to go next
Freeze-Thaw Damage in Atlantic Canada covers the field mechanism, What a Petrographic Examination Can Reveal covers how air void systems are actually examined, and What Road Salt Does to Concrete connects the exposure to the chemistry.