What the Heat-Affected Zone Is and Why It Governs

Why this matters

Three documents travel with a welded joint: a mill test report for the plate, a classification for the filler, and a procedure with a qualification record behind it. Two of the three state properties measured on material that does not exist at your joint. The mill report's yield came from unaffected wrought plate in a stated orientation, in the as-supplied condition. The filler's strength came from all-weld metal in a test assembly built the way the filler metal specification says to build it, in a stated heat treatment condition. Between those two materials sits a band of parent metal that was heated to somewhere between just short of melting and barely warm, cooled at whatever rate the joint allowed, and now holds properties nobody ordered. The joint's governing property is usually in that band, and the qualification record is the only document in the pile that ever put a test in it.

Before any check in this card: preparing a surface for a hardness traverse means grinding and polishing parent metal and any coating on it, which is an inhalation route, so use capture at the tool or a respirator issued under a written program meeting 29 CFR 1910.134 rather than eye protection alone. Stainless and chromate primers release hexavalent chromium under 29 CFR 1910.1026 in general industry or 29 CFR 1926.1126 in construction, and old paint can release lead under 29 CFR 1910.1025 or 29 CFR 1926.62. Etching a prepared face uses an acid reagent that burns skin and eyes and gives off vapour, so work it in a ventilated area with the chemical-resistant gloves and eye protection named on the product's safety data sheet, add acid to water and never the reverse. Any weld deposited to make a coupon carries the arc's ultraviolet at the minimum protective shade in OSHA's filter-lens table at 29 CFR 1910.133(a)(5), with fire prevention and the fire watch owned by the hot work permit under 29 CFR 1910.252(a) or 29 CFR 1926.352.

Where the band is, and how wide

The heat-affected zone starts at the fusion line, where the parent metal reached melting, and ends where the parent metal never got hot enough to change anything. In a single arc pass on carbon steel it is commonly somewhere between a sixteenth and a quarter of an inch wide, set by how much energy went in per unit length and how quickly the surrounding metal drew it away. The related card on heat input owns that control.

   section through one fillet, weld at left

   weld | HAZ                        | unaffected
   metal|                            | plate
   -----+----+------+--------+-------+-----------
        |CG  | FG   | inter- | sub-  |
        |    |      | crit   | crit  |
        ^    ^      ^        ^       ^
     fusion grain  A3       A1     no change
      line refine

Two things follow from the picture. The band is graded, not uniform: every point saw a different peak temperature, so a single hardness reading describes one location and nothing else. And a multipass joint is not this picture, because each later pass re-heats the earlier one's zone, which is why a cap pass over a hard root can improve the root and why an unlucky bead placement can leave a coarse, untempered region buried in the middle.

The four regions of one pass on carbon steel

• Coarse-grained. Nearest the fusion line, above roughly 2000 degrees F. Grain grows, toughness falls, and on a hardenable steel this is where the hardest structure forms.
• Grain-refined. Cooled from just above the upper transformation temperature. Usually the best material in the joint.
• Intercritical. Between the lower and upper transformation temperatures, about 1340 to 1600 degrees F for plain carbon steel. Only part of the structure transformed, so the region ends up mixed.
• Subcritical. Below about 1340 degrees F for plain carbon steel. Nothing transforms; what happens instead is tempering, and on material that arrived hardened or cold-worked, tempering is a loss.

The two directions, and which material tells you which

The sibling card on what a welded joint changes about the parent metal owns the metallurgical changes themselves. What this card adds is that the direction of the change is set by the condition the plate arrived in, and the direction decides which document misled you.

Plate that arrived soft gets a harder zone. Hot-rolled carbon and low-alloy structural steel is not hardened when it reaches you. Weld it and the coarse-grained region can transform to a hard structure, and hard structure plus dissolved hydrogen plus tensile stress is the delayed cracking the hydrogen card owns. The mill certificate's yield is then conservative for the joint, and the risk is cracking rather than strength.

Plate that arrived hard gets a softer zone. Quenched-and-tempered steel, thermomechanically processed steel, cold-worked material and precipitation-hardened aluminum all arrive holding strength from a treatment the weld undoes. The subcritical region over-tempers, and the joint's weakest tensile section becomes parent metal outside the weld entirely. The mill certificate is then optimistic, and the risk is that somebody sized a member from it.

The aluminum case is where this is written down plainly: for common heat-treatable structural alloys, the Aluminum Design Manual, in the edition your project specification adopts, reduces allowable strength within about one inch of a weld, precisely because the material there is no longer in the temper the certificate names. Steel practice handles the same effect by controlling the procedure rather than by publishing a derated zone, which is why it is easier to miss.

What each document actually measured, and where

A shop that answers an inspector with "the plate is 50 ksi and the wire is 70 ksi" has answered about two materials that are not carrying the load at the location in question.

What a hardness number is a number of

Hardness is the cheap proxy for what happened in the band, and it is only usable with its basis stated. Print four things every time: the scale and load (Vickers at 10 kgf reads HV10, and a Rockwell C figure converted from it is a converted number, not a measured one), the location as a distance from the fusion line, the condition when tested (as-welded, or after a stated post-weld heat treatment), and whether it is a traverse on a prepared cross-section or a portable reading on a surface. Those last two are not interchangeable, and the field figure carries the larger spread.

Worked example: one gate, two plates, opposite failures

The gate: the joint's governing property is the extreme value across the traverse at the location the specification names, not the average and not the base plate figure. Both plates below are half-inch material, both get the same single-pass fillet, both are read as-welded on a prepared cross-section with HV10. Say the project specification sets 350 HV10 as the maximum permitted in the heat-affected zone, a limit commonly specified where hydrogen cracking is the concern; that number is owned by the specification or the engineer, never by the shop.

Plate A, ordinary hot-rolled carbon structural plate. Base metal reads 150 HV10. Welded in a 40 degree F shop with no preheat, short beads, low energy per unit length, so the joint cooled fast.

• Reading at the coarse-grained region, 0.02 in from the fusion line, as-welded: 395 HV10.
• Against the 350 HV10 limit: over by 45 HV10, which on the 350 limit is about 13 percent over. Fail.
• Against the base plate's 150 HV10: the band is 2.6 times the base reading, and it is the hardest metal in the assembly.
• Consequence: the correction is cooling rate, which means preheat and interpass control and hydrogen control, owned by those cards. Nothing about the filler changes.

Plate B, quenched-and-tempered plate of the same thickness. Base metal reads 310 HV10 because it arrived hardened and tempered. Same welder, generous energy per unit length, long time at temperature.

• Reading at the coarse-grained region, 0.02 in from the fusion line, as-welded: 330 HV10. Against the 350 HV10 limit, under by 20 HV10. Pass.
• Reading at the subcritical region, 0.15 in from the fusion line, as-welded: 250 HV10.
• Against the base plate's 310 HV10: down 60 HV10, which on the 310 base is about 19 percent below. Nothing in the specification's hardness limit catches this, because the limit is a ceiling.
• Consequence: the weakest tensile material in the joint is now parent plate the welder never melted, about an eighth of an inch outside the visible weld. If the member was sized from the mill certificate's yield and the connection loads that band in tension, the design value was never delivered. Ask the engineer of record whether the design used welded-condition properties.

Same procedure change, opposite sign. Adding energy per unit length fixes Plate A and worsens Plate B, which is why the procedure, not the welder, owns the number.

Check the example against the rules the siblings state

• Basis printed on every hardness figure. All five readings (150 and 395 on Plate A, 310, 330 and 250 on Plate B) are HV10 on a prepared cross-section in the as-welded condition, and the three heat-affected-zone readings also carry a distance from the fusion line: 0.02 in for the 395 and the 330, 0.15 in for the 250. No converted Rockwell figure appears.
• The ceiling caught one case and not the other. 395 against 350 fails by 45 HV10. 330 against 350 passes by 20 HV10, and the 250 HV10 softening on that same plate is invisible to a ceiling, by construction.
• Arithmetic re-derived, not paraphrased. 45 over 350 is 12.9 percent, stated as about 13 percent over. 60 over 310 is 19.4 percent, stated as about 19 percent below. 395 over 150 is 2.63, stated as 2.6 times.
• Direction check against the siblings. The preheat and heat input cards both have more energy per unit length slowing cooling and softening the zone on hardenable steel. Plate A agrees at 395 HV10 fast-cooled. Plate B runs the same relationship to the opposite consequence at 250 HV10 in the subcritical region, because the base started at 310 HV10 rather than 150.

References

• AWS D1.1, structural welding code for steel, in the edition invoked by your project specification or the adopted building code, for procedure qualification and the location of impact test notches when toughness testing is required
• ASME Boiler and Pressure Vessel Code Section IX, as invoked by the applicable piping or vessel code and adopted by your jurisdiction
• Aluminum Design Manual, in the edition your project specification adopts, for the reduced allowable strength zone adjacent to a weld in heat-treatable alloys
• 29 CFR 1910.1026 and 29 CFR 1926.1126 (hexavalent chromium), 29 CFR 1910.134 (respiratory protection), 29 CFR 1910.133(a)(5) (filter shades)
• See related: What a Welded Joint Changes About the Parent Metal; What Heat Input Actually Controls; Why Preheat Exists and What It Is Controlling