Type 321 exists because of a competition inside the steel. When an 18-8 austenitic stainless is held in the 425–860°C range - during welding, during service, or during slow cooling - carbon and chromium combine at the grain boundaries to form chromium carbides. The boundaries lose chromium, and the steel becomes vulnerable to intergranular corrosion in any aqueous electrolyte that later reaches them.
The standard solution is to reduce carbon. The 321 solution is different and more elegant: add titanium, an element that competes for the same carbon and wins. This article explains that competition, the 5×C rule that governs it, the distinction between 321 and 321H, where the grade earns its premium, and the procurement details that decide whether the titanium is actually doing its job.
Two Elements Competing for One Carbon Atom
Titanium has a stronger affinity for carbon than chromium does, and titanium carbide (TiC) is more stable at high temperature than chromium carbide (Cr₂₃C₆). In a titanium-stabilized austenitic, the titanium grabs the available carbon first, forming TiC particles inside the grains.
Chromium is left in solid solution where it belongs, and the grain boundaries keep their chromium content. The result is a steel that can be welded and held in the sensitization range without developing the chromium-depleted band that would otherwise make it fail an intergranular corrosion test.
That is the whole mechanism, and it is worth stating precisely because the marketing language around "stabilized" grades tends to blur it. 321 is not more corrosion resistant than 304 in general service - it is specifically resistant to sensitization, and therefore to intergranular attack after welding or high-temperature exposure. In a low-carbon 304L application at ambient temperature, 321 offers nothing that 304L does not; its value appears where the thermal history would otherwise create a sensitized structure and where a post-weld solution anneal is impractical.
The 5×C Rule and Why It Is Not a Formula
ASTM A312 expresses the titanium requirement as a ratio, not a fixed addition: titanium shall be not less than five times the carbon content, and not more than 0.70%. For 321H the rule is slightly different - titanium not less than four times the carbon content minus 0.02, with the titanium held between 0.40 and 0.70%.
| Requirement (A312) | TP321 (S32100) | TP321H (S32109) |
| Carbon | 0.08 max | 0.04–0.10 |
| Titanium | ≥ 5 × C, ≤ 0.70 | ≥ 4 × (C − 0.02), 0.40–0.70 |
| Tensile, min | 515 MPa | 515 MPa |
| Yield, min | 205 MPa | 205 MPa |
| Elongation, min | 35% | 35% |
The ratio exists because stoichiometry is not enough: some titanium is lost to titanium nitride and to oxide inclusions, and the 5×C multiplier is the industry's empirical allowance for that loss plus a margin. The upper limit of 0.70% matters just as much as the lower one: excess titanium forms coarse TiN and Ti(C,N) particles that hurt formability, weldability and surface finish, and a heat with titanium at 1.2% is not "more stabilized" - it is a different, harder-to-process steel.
The practical consequence for a buyer is that "321" on a certificate is not sufficient evidence. The certificate must show the actual carbon and the actual titanium, and the ratio has to be checked: a heat with carbon 0.07% and titanium 0.30% fails the 5×C rule even though it is stamped 321. That is a real failure mode in low-cost supply, and it is invisible unless someone reads the numbers.
321 vs 321H: The Carbon Split at 0.04–0.10
The two grades share the same chemistry family and the same room-temperature minimums, and the difference is the same one that separates 310S from 310H: carbon range, and what it does to creep behavior.
321 caps carbon at 0.08% and is the general-purpose stabilized grade. 321H requires carbon between 0.04 and 0.10% - a window deliberately set high enough that the steel carries the carbon needed for creep resistance, and low enough that the titanium can still handle it. The short-time tensile values are identical, which is why datasheets make them look interchangeable. They are not, above roughly 540°C: 321H carries higher allowable stress in the ASME Section II-D tables because the carbon level supports creep rupture resistance that 321 cannot match.
The rule of thumb that follows: below about 540°C, specify 321 and stop paying for 321H; above roughly 650°C under sustained load, 321 is the wrong order and the design should be rechecked against 321H allowables. In between, the calculation decides, not the datasheet.
Where the Titanium Earns Its Keep
321's commercial home is high-temperature equipment that also has to survive a welding operation without a solution anneal. The classic applications:
- Automotive and diesel exhaust systems. Manifolds, downpipes, expansion joints and muffler internals that run 600–800°C and are welded in high volume with no possibility of post-weld heat treatment. The titanium is what keeps the weld heat-affected zone from sensitizing.
- Heat exchangers and superheater tubing. Where the tube is welded into a tubesheet or a header and the joint will see elevated temperature in service; 321 and 321H are standard boiler and heat-exchanger materials under A213.
- Steam and hot-water piping above the 304L comfort zone. Saturated steam lines, condensate systems and refinery furnace piping where the design temperature is high enough that 304L's low carbon becomes a creep penalty but not so high that a 310-grade is needed.
- Expansion bellows and flexible components. Formed and welded parts where the annealing furnace is not part of the production route.
The boundary is as important as the list. 321 is not the grade for aqueous chloride service - its molybdenum content is zero, and its pitting resistance is essentially 304's. It is not a substitute for 316L in chemical process duty, and it is not a substitute for 347 in the highest-temperature cyclic service where niobium stabilization is preferred.
321 vs 347: Titanium or Niobium
The other stabilized 18-8 is 347, which uses niobium plus tantalum instead of titanium, at not less than ten times the carbon content and not more than 1.00%. Both grades solve the same sensitization problem, and the choice between them is narrower than the catalogs suggest:
| Point | 321 (Ti) | 347 (Nb + Ta) |
| Stabilizer | Ti ≥ 5 × C, ≤ 0.70 | Nb + Ta ≥ 10 × C, ≤ 1.00 |
| High-temperature behavior | Good to roughly 800 °C | Preferred for higher and more cyclic service |
| Weldability | Ti can oxidize in the arc; 347 filler is often used | Nb is more forgiving in the weld pool |
| Typical use | Exhaust, heat exchangers, general high-temp piping | Heavy cyclic thermal service, refinery, power |
Two practical notes. First, 347 welding wire is routinely used to weld 321 - the niobium does the stabilizing job in the weld pool where titanium would tend to burn out, and the resulting joint is at least as resistant to sensitization as the parent. Second, if the service is severe cyclic thermal duty at 800°C and above, 347 (or a higher alloy) is usually the better answer; 321's titanium carbide can coarsen over long exposure, and the stabilization benefit degrades with time at the top of its range.
Welding and the Stabilizer That Does Not Travel
One advantage of titanium stabilization is that it is less sensitive to weld dilution than niobium-based grades: the titanium sits in the parent metal and the weld pool does not need to carry it, provided the filler is chosen correctly. The practical rules for 321 fabrication:
- Use 347 or 321 matching filler. ER347 is the usual choice for 321 pipe because it eliminates titanium burn-out in the arc and gives a weld with equivalent sensitization resistance. Matching ER321 is available but less common.
- Control heat input. 321 welds without difficulty, but excessive heat input still coarsens the heat-affected zone and can dissolve the TiC that was doing the protecting. Stringer beads and controlled interpass are standard.
- Root protection for thin wall. On thin-wall 321 tube, an oxidized root is a corrosion and fatigue initiation site; argon purge is cheap insurance.
- No post-weld anneal needed in most cases. That is the point of the grade - but if the fabrication does go through a solution anneal at 1050–1100°C, the titanium carbide re-dissolves and the stabilization is re-established on cooling, which is why the anneal is not harmful, just unnecessary.
The Procurement Detail: Ti/C Ratio on the Certificate
Three checks close the loop on a 321 or 321H order, and all three are done on the mill certificate rather than the datasheet:
- The ratio. Multiply the certified carbon by five and compare with the certified titanium. If the titanium is below that number, the heat is not 321 in any meaningful sense, whatever the stamp says.
- The upper limit. Titanium above 0.70% is out of specification and signals a heat that will be difficult to form and weld.
- The grade designation. 321 and 321H are different products; a certificate for 321 does not satisfy a 321H design, and vice versa, because the carbon range decides the creep allowable.
Seamless 321 pipe under A312 is available from 1/2" to 12" NB in schedules 5S through 80S, with welded pipe to larger diameters and tube for heat exchangers down to light wall; A213 covers the boiler and superheater tube product. The grade is widely stocked in standard schedules and is one of the few stabilized austenitics with genuine mill availability at short notice.
Ours quotes 321 and 321H by UNS number with the carbon and titanium values called out on the certificate, and we will tell you when 304L, 316L or 347 is the more honest answer for the duty.
FAQ
Q1: Is 321 better than 304L?
Not in general corrosion resistance - 321 has no molybdenum and its pitting resistance is essentially 304's. It is better in one specific way: it resists sensitization after welding or high-temperature exposure, so it does not suffer intergranular attack where 304L's thermal history would leave it vulnerable. If the service is ambient and low-chloride, 304L is the cheaper and equally correct choice.
Q2: Why is titanium limited to 0.70% if more stabilization is better?
Because excess titanium forms coarse titanium nitrides and carbonitrides that hurt formability, weldability and surface finish. The 5×C ratio with a 0.70% cap is the balanced point: enough titanium to consume the carbon, not so much that the steel becomes difficult to process.
Q3: Can 321 be used in seawater?
No. 321 contains no molybdenum and has pitting resistance comparable to 304; it will pit and crevice-corrode in seawater. Seawater service belongs to the duplex, super-austenitic or nickel alloy families, not to a stabilized 18-8.
Q4: What is the maximum service temperature for 321?
Continuous service is generally quoted to about 800°C, with 321H preferred for sustained load above roughly 540–650°C because of its higher creep allowable. Above 800°C, oxidation and scale become limiting and a 25Cr-20Ni grade (310S/310H) or a nickel alloy is the appropriate step up.
Q5: Why do fabricators weld 321 with 347 filler?
Because titanium tends to oxidize and burn out in the arc, while niobium survives the weld pool and provides the same stabilization in the deposit. An ER347 weld on 321 pipe gives a joint at least as resistant to sensitization as the parent metal, with a more forgiving procedure.
Q6: Does 321 need post-weld heat treatment?
In most cases no - that is the commercial reason the grade exists. Where the fabrication involves a solution anneal for other reasons, the titanium carbide re-dissolves and re-forms on cooling, so the anneal does not damage the stabilization. What 321 does need is a clean, oxide-free weld surface and a filler that carries a stabilizer.
CTA
Send the design temperature, the welding route and the certificate you are being offered - we will check the Ti/C ratio, confirm whether 321 or 321H is the grade your allowable stress actually requires, and tell you honestly when 304L or 347 is the better answer. 📧 mail to:market@htpipe.com | WhatsApp +86-19339900201
