IS 4031 Part 8 : 1988Methods of Physical Tests for Hydraulic Cement - Part 8: Determination of soundness by autoclave method

IS 4031 Part 8:1988 is the method for determining the soundness of hydraulic cement by the autoclave-expansion method. Soundness — freedom from large delayed expansion after setting — is a *pass/fail safety property*: an unsound cement expands and disrupts hardened concrete from within, and the autoclave method is the severe test that detects the slow-acting causes (free lime and especially magnesia / periclase).

It sits in the cement-testing stack:

• IS 4031 Part 1 — physical tests for hydraulic cement (soundness among them; Le-Chatelier is the companion soundness test) · IS 4845 — terminology (soundness defined)
• IS 269 — OPC (soundness is a mandatory requirement) · IS 456 — concrete · IS 4032 — chemical analysis (free lime/MgO)

Unsoundness is delayed, destructive expansion *after* the concrete has hardened — there is no remedy once it is in the structure, so cement must be proven sound *before* use. The autoclave method matters because of *what* it catches:

• Free lime (CaO) and, critically, magnesia (MgO / periclase) hydrate very slowly and expansively — slow enough that the milder Le-Chatelier test (sensitive mainly to free lime) can miss MgO-driven unsoundness
• The autoclave subjects a specimen bar to high-pressure steam, dramatically accelerating these slow reactions, and the % expansion must be within a strict limit
• It is a mandatory acceptance property in IS 269 and the cement specs — not optional

The engineering point: soundness is binary and unforgiving — an unsound cement is a structure that cracks and disintegrates from within with no fix. The autoclave test is the severe screen for the slow expansive constituents (especially MgO) that a milder test underestimates; failing it is an absolute reject, not a borderline negotiation.

Scenario: acceptance testing of a cement consignment.

Step 1 — representative sample: sample per the cement-sampling standard (IS 3535) — soundness on a non-representative sample certifies nothing.

Step 2 — prepare the bar: mould the specimen at normal consistency (IS 4031 Part 3) with reference gauge length.

Step 3 — autoclave: subject to the prescribed high-pressure steam cycle (accelerates the slow free-lime/MgO reactions).

Step 4 — measure expansion: compute the % length expansion; compare to the IS 269/spec limit.

Step 5 — accept/reject: within limit → sound; over limit → absolute reject (do not use — there is no site remedy for unsoundness). Run the Le-Chatelier companion too, but rely on the autoclave for MgO-type unsoundness.

This is a safety hold-point: a failed soundness test ends the consignment's use, full stop.

1. Relying only on Le-Chatelier. It is sensitive mainly to free lime and can miss MgO-driven unsoundness — the autoclave is the severe screen for that.

2. Treating soundness as negotiable. It is a binary safety property — an over-limit result is an absolute reject, not a borderline pass.

3. Non-representative sample. Soundness of a convenience sample certifies nothing — sample per IS 3535.

4. Wrong specimen prep / gauge length. Expansion is a small precise measurement — sloppy moulding/measurement invalidates it.

5. Assuming 'it set fine, so it's sound'. Unsoundness is *delayed* — normal early behaviour says nothing about later expansive disruption.

• IS 4031 Part 1 — physical tests (soundness; Le-Chatelier companion) · IS 4031 Part 3 — consistency (specimen water)
• IS 4845 — terminology (soundness defined) · IS 4032 — chemical analysis (free lime / MgO)
• IS 269 — OPC (soundness is mandatory) · IS 3535 — cement sampling
• IS 456 — concrete (an unsound cement destroys the structure)

IS 4031 Part 8 is reaffirmed and is a safety gate, not a quality metric: soundness is freedom from delayed expansive self-destruction, there is no remedy once unsound cement is cast, and so it must be proven *before* use. The practitioner essential is the why the autoclave: the milder Le-Chatelier test catches free-lime unsoundness but can underestimate the slow, expansive magnesia (periclase) problem — the autoclave's high-pressure steam accelerates exactly those slow reactions and is the severe screen for them. Treat a soundness failure as an absolute reject (no borderline acceptance, no site fix), always test a representative sample, and never infer soundness from normal early setting — unsoundness is delayed by definition. It is one of the few cement results where the only safe response to a fail is to refuse the consignment.