Stainless Steel Welding & Fabrication

Karl Gillihan

⚙️ STAINLESS STEEL

Welding & Fabricating Stainless Steel: Distortion, Sensitization, and Corrosion You Can Weld Away

A stainless weld can look perfect and still corrode in six months if you sensitize the grain boundaries or skip back-purging the root. The whole game in stainless fabrication is protecting the chromium that makes it stainless, while managing heat input on a metal that warps like a tin roof. Here is what changes between the austenitic, precipitation-hardening, and duplex grades, and where the corrosion failures actually start.

ISO 9001ISO 13485AS9100

When austenitic stainless like 304 sits in the 800-1500 F range during welding, chromium near the grain boundaries combines with carbon to form chromium carbides. That locally strips the chromium below the ~11% needed for passivity, leaving a sensitized band that corrodes preferentially in service. The part passes inspection, then fails at the weld months later. This is the single most important concept in stainless fabrication. The fix is in the grade. 316L and 304L carry carbon capped at 0.03% instead of 0.08%, starving the carbide reaction so the HAZ stays corrosion-resistant in the as-welded condition. Stabilized grades 321 and 347 add titanium or niobium that grab the carbon preferentially. For anything that will see a corrosive environment, specify the L grade up front; paying for 316L over 316 is cheap insurance against sensitization failures that are nearly impossible to repair in the field.

Back-Purging and the Sugared Root Nobody Wants

Stainless oxidizes aggressively at welding temperature, and an unshielded weld root exposed to air turns into a black, granular oxide called sugaring or coking. A sugared root is a corrosion initiation site and, in sanitary or high-purity work, an outright rejection. Pipe, tube, and vessel welds get back-purged with argon on the inside of the joint to hold oxygen below roughly 100 ppm, sometimes much lower for ultra-high-purity semiconductor and pharma lines. This matters for sourcing. A shop set up for sanitary stainless will have orbital welding, purge dams, and oxygen monitoring; a general fab shop may not. For food, pharma, semiconductor, and medical fluid-path work, confirm the vendor purges and can document weld oxygen levels and discoloration limits (the AWS D18.2 weld discoloration chart is the common acceptance reference). For ordinary structural or cosmetic stainless, back-purging may be skippable, which lowers cost.

Distortion Control and Heat Input on a Metal That Hates Heat

Austenitic stainless has about 50% higher thermal expansion than carbon steel and only a third of the thermal conductivity, so heat concentrates at the joint and the part warps dramatically. Thin sheet weldments buckle and pull without aggressive fixturing, chill bars, intermittent welding, and balanced welding sequences. Plan tack spacing and a back-step or skip-weld pattern from the start. Low heat input is the lever for both distortion and metallurgy: pulsed TIG and pulsed MIG let you fuse the joint with less total energy, narrowing the HAZ, reducing sensitization risk, and cutting warp. After welding, corrosion-critical parts get pickled and passivated (nitric or citric acid treatment) to remove free iron, heat tint, and weld scale and rebuild the passive chromium-oxide layer. Skipping passivation on medical or pharma parts is a common and expensive miss.

17-4PH and Duplex 2205: When Standard Procedures Don't Apply

17-4PH is a martensitic precipitation-hardening grade that behaves more like an alloy steel than a stainless. It is weldable, typically with 17-4 or 630 filler, but the heat-affected zone forms martensite and the part usually needs a post-weld solution treatment and aging to restore the H900/H1025/H1150 condition uniformly. Welding it in the hardened condition and skipping heat treatment leaves a soft, mismatched HAZ. Budget for the furnace cycle. Duplex 2205 has a roughly 50/50 austenite-ferrite microstructure that gives it twice the yield strength of 316 and superior chloride stress-corrosion-cracking resistance, which is why it dominates offshore and chemical processing. But that balance is heat-sensitive: too little heat input and the weld goes ferrite-heavy and loses toughness; too much and you precipitate sigma phase and intermetallics that embrittle it. Duplex welding requires controlled heat input (typically 0.5-2.5 kJ/mm), nitrogen-enriched shielding gas to preserve austenite, and often ferrite-number checks on the finished weld. It is not a job for a shop without duplex experience.

Frequently Asked Questions

Specify 316L for anything welded that will see moisture or a corrosive environment. The only difference is carbon: 316 allows up to 0.08% carbon, 316L caps it at 0.03%. During welding, that extra carbon in standard 316 combines with chromium at the grain boundaries in the 800-1500 F sensitization range, forming chromium carbides that strip local corrosion resistance and set up intergranular attack at the weld. 316L's low carbon starves that reaction, so the weld and HAZ stay corrosion-resistant in the as-welded condition without a post-weld solution anneal. The strength penalty for going L is minor (a few ksi lower yield), and for thin-section welded fabrication it is usually irrelevant. The cost premium is small. The one case where standard 316 makes sense is a heavy section that will be solution-annealed after welding anyway, or a part with no welding and a need for higher room-temperature strength. For welded tanks, piping, brackets, and enclosures exposed to chlorides or washdown, 316L is the default and 316 is a false economy.

Almost always one of three causes, all preventable. First, carbon-steel contamination: if the shop ground or brushed your stainless with tools previously used on carbon steel, embedded iron particles rust on the surface and look like the stainless itself failed. Dedicated stainless-only grinding wheels and brushes prevent this. Second, sensitization: standard high-carbon grades welded without an L or stabilized grade form chromium carbides at the weld, creating a chromium-depleted band that corrodes; the fix is grade selection. Third, missing passivation: welding leaves heat tint (the blue/straw discoloration) and free iron on the surface, both of which are less corrosion-resistant than the base metal. Pickling and passivating with nitric or citric acid removes the heat tint and free iron and regrows the protective chromium-oxide passive layer. A sugared, unpurged weld root is a fourth cause for tube and pipe. For corrosion-critical work, insist the shop uses stainless-dedicated abrasives, an L grade, and post-weld passivation, and the rusting stops.

Yes, but the part almost always needs heat treatment after welding to recover uniform properties. 17-4PH (also called 630) is a martensitic precipitation-hardening stainless that gets its high strength (up to ~190 ksi in the H900 condition) from aging, not from cold work. It welds readily with matching 17-4 filler, but the heat of welding forms a martensitic HAZ and overages or re-solutionizes the precipitate locally, leaving a soft, non-uniform band. The standard recovery is to weld in the solution-annealed (Condition A) state, then solution treat and age the whole assembly to the target condition (H900, H1025, H1075, or H1150) so strength and hardness are uniform. Welding a part already in a peak-aged condition and skipping the furnace leaves a mismatched joint. Budget the furnace cycle, the distortion risk from heat treatment, and possible re-machining of critical dimensions afterward. For lightly-loaded or non-critical joints, some shops weld and use as-is, but you lose the strength guarantee.

Expect duplex 2205 to run meaningfully more than 316L on both material and labor. Material is typically 1.5-2x the per-pound cost of 316L depending on market nickel and molybdenum pricing, though duplex's higher strength sometimes lets you down-gauge and recover some of that. The bigger cost is process control: duplex welding requires managed heat input (commonly 0.5-2.5 kJ/mm), nitrogen-added shielding gas, controlled interpass temperatures, and often ferrite-number or impact testing on the completed weld to confirm the austenite-ferrite balance and absence of embrittling sigma phase. That means a shop with documented duplex procedures and qualified welders, not a general fabricator, so labor rates and qualification overhead are higher. Lead times typically run a few days longer than equivalent 316L work because of the added testing and the narrower process window. The payoff is roughly double the yield strength and far better chloride stress-corrosion-cracking resistance, which is why offshore oil and gas, desalination, and aggressive chemical service pay for it.

Manage heat input and restrain the part, because austenitic stainless expands about 50% more than carbon steel and conducts heat only about a third as well, so energy piles up at the joint and the metal pulls hard. The practical toolkit: fixture or clamp the assembly in a jig that holds geometry while it cools; use chill bars or copper backing to draw heat away from thin sections; weld with a skip, back-step, or balanced sequence rather than running one long continuous bead; and keep total energy down with pulsed TIG or pulsed MIG so the HAZ stays narrow. Thinner gauges warp worst, so for sheet enclosures expect tack welds first, a planned weld sequence, and sometimes intermittent stitch welds where a continuous seam is not required. After welding, badly distorted parts can be straightened mechanically or thermally, but that adds cost and risk, so distortion is far cheaper to prevent than to fix. Low heat input also doubles as metallurgical insurance, narrowing the sensitization-prone band.

Last updated: July 2026

Find Stainless Steel Welding & Fabrication Suppliers

Search verified shops that handle Stainless Steel welding & fabrication.

Search Manufacturers →Get Your Shop Listed (Free)

No logins. No email gates. Just results.