All grade guides — one page, every alloy we cover
Browsing for a grade? The full library of per-grade technical guides is on one list.
Open the list321 and 347 exist for one reason: to survive the 425-900 °C band where ordinary 304 precipitates chromium carbides and starts to fall apart from the inside. 321 ties up carbon with titanium, 347 ties it up with niobium plus tantalum, and both keep chromium in solution so that welded equipment can be put into hot service without a post-weld anneal. The two grades are chemically almost twins and commercially very different — 347 is the better welder, 321 is the cheaper and more formable one, and choosing wrongly costs either money or a weld. This guide sets out the numbers, the stabilisation ratios that actually govern acceptance, and the decision rules we apply to refinery, power and exhaust-system enquiries.

When an austenitic stainless steel is held or slowly cooled through roughly 425-900 °C, chromium and carbon combine as chromium-rich carbides (M23C6) along the grain boundaries. The chromium is consumed locally, the adjacent zone drops below the ~12 % needed for a passive film, and the material becomes sensitive to intergranular corrosion. This is sensitisation, and in welded equipment it happens in the heat-affected zone during the weld thermal cycle itself.
There are only three industrial answers. Reduce the carbon (the "L" grades, 304L/316L). Anneal after welding (rarely practical on a large fabrication). Or add a strong carbide former that grabs the carbon before chromium does — titanium in 321, niobium in 347. These are the stabilised grades: UNS S32100 and S34700, W.Nr 1.4541 and 1.4550, with high-carbon variants 321H (S32109, 1.4878) and 347H (S34709, 1.4912) for creep duty.
The consequence is a grade family that behaves like 304 in ambient corrosion but like a different material at temperature — and, importantly, that can be welded and put straight into 500-900 °C service without heat treatment. That combination is why 321 and 347 remain in every refinery and power-plant material specification written since the 1950s.
The single fact that justifies the grade: between 425 and 900 °C, unstabilised 304/304L will sensitise and can fail by intergranular attack; 321 and 347 will not. Below 425 °C, 304L does the same job for less money.
| Element | 321 (S32100) | 321H (S32109) | 347 (S34700) | 347H (S34709) | Role |
|---|---|---|---|---|---|
| Carbon | 0.08 max | 0.04-0.10 | 0.08 max | 0.04-0.10 | Higher C in the H grades raises creep strength; it is also why the stabiliser ratio tightens |
| Chromium | 17.0-19.0 | 17.0-19.0 | 17.0-19.0 | 17.0-19.0 | Passive film and oxidation resistance |
| Nickel | 9.0-12.0 | 9.0-12.0 | 9.0-13.0 | 9.0-13.0 | Austenite stability, toughness, delayed SCC |
| Titanium | 5×(C+N) min, 0.70 max | 4×(C+N) min, 0.70 max | — | — | Carbide former; the ratio, not the absolute value, is the acceptance criterion |
| Niobium + Tantalum | — | — | 10×C min, 1.00 max | 8×C min, 1.00 max | Stronger carbide former than Ti and it survives the welding arc |
| Manganese | 2.00 max | 2.00 max | 2.00 max | 2.00 max | Deoxidation, austenite balance |
| Silicon | 0.75 max | 0.75 max | 0.75 max | 0.75 max | Deoxidation |
| Phosphorus / Sulphur | 0.045 / 0.030 max | 0.045 / 0.030 max | 0.045 / 0.030 max | 0.045 / 0.030 max | Residuals; low S for weldability |
| Iron | Balance | Balance | Balance | Balance | Matrix |
Two specification details are routinely missed. First, the stabilisation requirement is a ratio: a 321 heat at 0.06 % C needs at least 0.30 % Ti (5 × C, ignoring the small nitrogen term), and a 347 heat at 0.06 % C needs at least 0.60 % Nb+Ta. A certificate showing "Ti 0.25 %" against 0.06 % C fails that ratio even though it is full of titanium. Second, 321H carries an additional requirement of ASTM grain size No. 7 or coarser, which is what delivers the elevated-temperature strength; a fine-grained 321H should be queried.
Never accept "Ti: present" or "Nb: present" as evidence of stabilisation. Recalculate the ratio from the certificate: Ti ≥ 5 × C for 321, ≥ 4 × C for 321H, Nb+Ta ≥ 10 × C for 347, ≥ 8 × C for 347H.
Room-temperature minima in the annealed condition, with typical mill ranges in square brackets. All four grades share the same 515/205 MPa floor — the H grades earn their keep above 540 °C, not at room temperature.
| Grade / form | Yield 0.2 % (MPa) | Tensile (MPa) | Elongation | Hardness |
|---|---|---|---|---|
| 321 plate & sheet, annealed — A240 | ≥ 205 [230-290] | ≥ 515 [550-640] | ≥ 40 % [45-58] | ≤ 95 HRB (≈ 217 HB) |
| 321H plate & sheet, annealed — A240 | ≥ 205 [240-300] | ≥ 515 [550-650] | ≥ 40 % [40-50] | ≤ 95 HRB |
| 347 plate & sheet, annealed — A240 | ≥ 205 [230-290] | ≥ 515 [550-650] | ≥ 40 % [45-55] | ≤ 92 HRB (≈ 201 HB) |
| 347H plate & sheet, annealed — A240 | ≥ 205 [240-310] | ≥ 515 [555-660] | ≥ 40 % [40-50] | ≤ 92 HRB |
| Bar, annealed — A276 / A479 | ≥ 205 | ≥ 515 | ≥ 40 % | ≤ 95 HRB |
| Seamless pipe / tube — A312, A213 | ≥ 205 | ≥ 515 | ≥ 35 % | — |
| Cold-worked (drawn tube, formed strip) | Not standardised; rises with cold reduction | Rises with cold reduction | Falls with cold reduction | Rises |
At temperature the picture changes. The stabilised grades retain higher strength than 304 above roughly 540 °C, and the difference between 347 and 347H becomes significant in the creep regime — ASME allowable stresses for 347H reflect that higher strength, which is why 347H is the default for superheater and reformer duty. Do not extrapolate room-temperature values: for design above 425 °C use the ASME BPVC Section II Part D tables for the specific grade, product form and temperature.
The two grades are near-identical physically; the small differences come from titanium versus niobium atomic weight. Values are typical for annealed material.
| Property | 321 | 347 | Note |
|---|---|---|---|
| Density | 7.92 g/cm3 (some tables 8.03) | 7.96 g/cm3 | Grade-data sheets in circulation often round both to 7.90 — acceptable for estimating, not for weight-critical fabrication |
| Melting range | ≈ 1398-1446 °C | ≈ 1398-1446 °C | Practically identical; drives the 1150-1200 °C hot-working window |
| Thermal conductivity | ≈ 16.1 W/m·K at 100 °C; ≈ 22.2 at 500 °C | ≈ 16.1 W/m·K at 100 °C; ≈ 22.2 at 500 °C | About one third of carbon steel — thermal gradients across walls are steep |
| Specific heat (0-100 °C) | ≈ 500 J/kg·K | ≈ 500 J/kg·K | Same basis for transient calculations |
| Mean expansion 20-100 °C | ≈ 16.6 ×10-6/K | ≈ 16.6 ×10-6/K | Use the band that matches the duty. Expansion roughly doubles the carbon-steel value — design expansion joints accordingly |
| Mean expansion 20-315 °C | ≈ 17.2 ×10-6/K | ≈ 17.2 ×10-6/K | |
| Mean expansion 20-538 °C | ≈ 18.6 ×10-6/K | ≈ 18.6 ×10-6/K | |
| Modulus of elasticity | ≈ 193 GPa | ≈ 193 GPa | Falls steeply with temperature; check the value at design temperature |
| Electrical resistivity (20 °C) | ≈ 0.72 µΩ·m | ≈ 0.73 µΩ·m | Relevant for resistance welding and electrical heating elements |
| Magnetic behaviour | Non-magnetic annealed; µ < 1.02 at 200 H | Non-magnetic annealed; µ < 1.02 at 200 H | Weld ferrite, if any, raises permeability locally |
Because the physical properties are identical, 321 and 347 are interchangeable in thermal-stress and expansion-joint calculations. The choice between them is driven by welding, corrosion in specific media, and availability — not by physics.
Three separate phenomena are often lumped together as "high-temperature behaviour". They need to be checked separately.
The 425-900 °C band is a materials-selection gate, not a nuance. Below 425 °C use 304L/316L. Between 425 and 900 °C use 321 or 347 (or the H grades for creep duty). Above ~900 °C move to 310S or a nickel alloy. Equipment that crosses the band during welding is exactly why stabilised grades were invented.
In the annealed condition, 321 and 347 resist general corrosion about as well as 304. Stabilisation is not a corrosion upgrade in the way molybdenum is; it protects one specific failure mode, intergranular attack, and mostly after welding. Both grades contain no molybdenum, which gives them a PREN of only about 18.
| Environment | 321 | 347 | Comment |
|---|---|---|---|
| Intergranular corrosion after welding (e.g. ASTM A262 Practice E) | Good | Better | 347 is more tolerant of off-ratio heats and of heavy sections |
| Polythionic acid (refinery shutdowns, sulphidic scale + moisture + oxygen) | Acceptable | Excellent | 347 is the refinery standard for this failure mode; soda-ash neutralisation is still required |
| Nitric acid, oxidising media | Excellent | Excellent | Low carbon plus stabilisation keeps welds sound |
| Dilute sulphuric / phosphoric acid | Fair | Fair | Use 904L or Alloy 20 for anything beyond trace concentration |
| Chloride pitting and crevice | Poor (PREN ≈ 18) | Poor (PREN ≈ 18) | Neutral chlorides are the wrong duty for these grades — go to 2205, 317L or a 6 % Mo grade |
| Chloride stress-corrosion cracking | Poor above ≈ 60 °C | Poor above ≈ 60 °C | Same limitation as 304; duplex or high-nickel alloys are the answer |
| High-temperature sulphidation / carburisation | Moderate | Moderate | Better than 304, far below nickel alloys |
One practical point that catches fabricators: 321 does not polish well. Titanium forms stringers and inclusions during melting, and these tear at the surface during polishing, leaving a speckled finish. For decorative or hygienic polished surfaces, 304/316L or 347 are preferred over 321.
| Criterion | 321 wins | 347 wins |
|---|---|---|
| Welding extent | Light welding, or formed components joined at a few seams | Heavy welded assemblies, thick sections, no post-weld anneal possible |
| Weld filler availability | 321 filler is effectively unavailable — 321 base metal is welded with ER347 anyway | ER347 is a stock item worldwide; the practical default filler for both grades |
| Service temperature | Up to about 900 °C peak, non-creep-critical | Conservatively rated to about 816 °C continuous, with 347H extending creep capability above that |
| Refinery / polythionic acid | Acceptable with neutralisation | Standard choice; niobium gives markedly better resistance |
| Formability and cost | Cheaper and easier to form; titanium is the lower-cost stabiliser and 321 is the higher-volume grade | Dearer; niobium-bearing heats cost more and availability is narrower in thin sheet |
| Surface finish | Poor polishing behaviour — avoid decorative use | Accepts a normal polished finish |
| Cyclic thermal fatigue (exhaust systems) | Often preferred: better formability, slightly lower density, adequate at 400-750 °C | Better for steady high-temperature service than for vibration-fatigue duty |
Heuristic we apply to enquiries: exhaust and thermal-oxidiser work at 400-800 °C with light welding → 321. Refinery, petrochemical and heavy welded pressure equipment at 427-816 °C → 347 (347H for creep duty). If the item is a pressure vessel and the temperature exceeds 540 °C, specify the H grade and verify code acceptance first.
Code caveat worth checking before you commit: some mill datasheets note that 321H is not listed for ASME Section VIII Div. 1 construction (347H is), and that 321H is limited to 427 °C in Section III nuclear applications. If the vessel is code-stamped, confirm the grade is in the applicable ASME table before releasing the order.
| Industry | Typical components | Grade usually specified |
|---|---|---|
| Aerospace | Exhaust manifolds and cones, afterburner and tailpipe components, bellows, heat shields | 321 (formability, cyclic 400-750 °C) |
| Oil refining / petrochemical | Catalytic reformer internals, exchanger tubes, transfer line piping, reactor internals, flare tips | 347 / 347H (polythionic acid, creep) |
| Power generation | Superheater and reheater tube supports, boiler tube hangers, soot-blower parts, expansion joints | 347H / 321H |
| Chemical processing | Expansion joints, bellows, thermal-oxidiser shells, nitric acid equipment, catalyst beds | 321 (light welding), 347 (welded) |
| Automotive | Exhaust systems, catalytic converter shells, flex couplings, turbo downpipes | 321 (also 409/441 at lower duty) |
| Heat treatment & furnaces | Retorts, muffles, fans, radiant tubes, conveyor parts below ~900 °C | 321 / 347, 310S above 900 °C |
| Nuclear | Piping, surge lines, core support structures (grade-specific code limits apply) | 347 |
| Pulp & paper / food | Digester piping in elevated-temperature zones, steam-side equipment | 321 / 347 |
The most common welding error with these grades is ordering ER321 filler. There is essentially no commercial ER321 supply, and using it would be wrong anyway: 321 base metal is welded with ER347, and that is what the procedure qualification should say.
Hangbo supplies 321, 321H, 347 and 347H as plate, sheet, bar, seamless and welded pipe, heat-exchanger tube, forgings, flanges and fittings, cut and machined to drawing. Certificates are EN 10204 3.1 with heat-number traceability, PMI on every item and stabiliser-ratio verification before dispatch; third-party inspection (SGS, BV, TUV) on request. Stock sizes usually ship in 7-15 days, mill-roll and forged items in 25-45 days.
If the equipment is heavily welded, thick-sectioned, or in refinery/polythionic-acid service, choose 347 (347H for creep duty). If it is formed and lightly welded, running to about 800-900 °C, and cost or formability matter — exhaust systems, thermal oxidisers, bellows — 321 is the better value. Note that 347 is the filler for both, so 347 base metal never creates a consumable problem.
Titanium is a strong oxide former and is largely lost to the slag and atmosphere in the welding arc, so a 321 weld deposit would not be reliably stabilised. Niobium transfers across the arc much more efficiently, which is why ER347 is the AWS-standard match for 321 base metal. ER321 is essentially not commercially available.
No, not to restore corrosion resistance — that is the whole point of stabilisation. PWHT may still be required for other reasons: ASME VIII Div. 1 for thick-wall vessels above 38 mm with certain exceptions, caustic service, polythionic-acid service, or dimensional stability. A stabilising anneal at 870-900 °C is sometimes specified for severe high-temperature duty.
High carbon: 0.04-0.10 % versus 0.08 % max for the standard grades. The higher carbon raises creep and stress-rupture strength above roughly 540 °C, which is why superheater tubing and reformer internals are specified as 347H. The trade-off is that the stabiliser ratio requirement tightens (4 × C for 321H, 8 × C for 347H) and carbon cannot be verified by handheld PMI.
No. Both contain no molybdenum, giving a PREN of only about 18, and they pit in warm chloride solutions much like 304. They also stress-corrosion crack above roughly 60 °C in concentrated chlorides. For chloride duty use 316L, 317L, 2205 duplex or a 6 % Mo grade, depending on severity.
When the welded equipment will operate between 425 and 900 °C, or when it will be stress-relieved after welding. 304L can sensitise in that band and, being low carbon, has lower hot strength above about 500 °C. Below 425 °C in the annealed condition, 304L performs the same job for less money and is more readily available in every product form.
Send us your drawing or specification for a quotation — we reply within one business day.
Buy this grade as: Plate · Round bar · Seamless tube · Flanges