When 304 stainless is the wrong material for a part

Most 304 specifications are not chosen. They are copied. A bracket that ran for a decade inside a dry plant room inherits the same callout when the next machine goes onto a coastal site, and the drawing is released without anyone asking what the environment does to the passive film. By the time the parts are pitted, the material decision is a warranty conversation rather than a drawing change.

304 is a good default and a poor universal. It resists atmospheric corrosion, food contact and most industrial indoor environments, and it forms and welds predictably. It fails in a narrow but repeating set of conditions: chloride above about 200 mg/L, chloride with heat above 60 °C, welded joints left in the sensitisation range, and sliding contact against itself.

The 304 stainless alternatives below are ranked by what they solve, not by price. Each one gives something up, and the crossover numbers are what the procurement decision needs.

The failures that mean the grade is wrong

Chloride pitting. The passive film breaks down at inclusions and at manganese sulphide stringers, and a pit grows through the thickness. It shows as a brown stain and a dark pit, usually on a downward-facing surface where chloride deposits and moisture sits. A 304 panel at 1,000 mg/L chloride will pit; that is not a defective heat.

Crevice corrosion. The same mechanism in a gap: under a gasket, between two bolted sheets, under a weld spatter that was never removed. The crevice excludes oxygen, the film cannot rebuild, and the attack is faster than on the open surface because the chloride concentrates.

Chloride stress-corrosion cracking. Above about 60 °C, chloride plus residual stress from a weld or a tight bend produces a brittle crack with no visible metal loss. It propagates through the section, so a part that passes a visual inspection fails without warning. This is the failure that makes a grade change urgent rather than scheduled.

Weld decay. Chromium carbides precipitate at the grain boundaries between 450 °C and 850 °C, and the chromium-depleted band beside the weld corrodes in preference to everything else. 304 at 0.08% carbon is exposed to it; 304L at 0.03% is not. ISO 3651-2 is the test method for intergranular corrosion resistance, and it is the acceptance criterion when a weld will see an aggressive service.

Galling and seizure. Austenitic stainless against austenitic stainless in sliding or threaded contact cold-welds. A 304 bolt into a 304 nut tightened dry will seize at roughly the torque a carbon steel fastener takes comfortably. The material is not defective; the friction pair is wrong.

The 304 stainless alternatives, and where each one crosses

Read the fourth column with the third. Each alternative trades a property for a cost, and the trade only pays if the property it gives up is not the one the part depends on. A 430 panel in a dry server room is a reasonable substitution for a painted 304 cover; the same panel on a coastal genset canopy is a corrosion test.

The decision rule in two numbers

Chloride concentration and service temperature settle most cases before cost is discussed.

Below 200 mg/L chloride and below 60 °C, 304 is adequate. That covers indoor industrial air, most potable water, food contact without concentrated sanitisers, and clean inland outdoor exposure.

Between 200 and 1,000 mg/L, or within about 500 m of a coast, 316L is the working grade. The molybdenum raises the pitting resistance equivalent to about 24 against 18, and the critical pitting temperature in a ferric chloride test moves up by roughly 10 °C. ASTM G48 is the method, and the scatter between heats is large enough that the test is a screening tool rather than a guarantee.

Above 1,000 mg/L, or in continuous immersion in chloride, neither austenitic grade is right. A splash zone, a brine line, a pool plant room and a de-icing salt tunnel all sit above the line. Duplex, a super-austenitic, a coated carbon steel or a non-metallic part is the answer, and the material decision belongs with the end user rather than the fabricator.

Above 60 °C with chloride present, the mechanism changes from pitting to stress-corrosion cracking, and the chloride limit collapses. A 304 part at 80 °C in 300 mg/L chloride can crack where the same part at 30 °C would only stain.

The same canopy panel on two sites

Take an acoustic canopy side panel, 1,200 × 800 mm at 1.5 mm, eight panels per unit. The panel area is 0.96 m², so the set is 7.68 m² and weighs 92 kg at 8.0 kg per m² per mm. The material callout is the only thing that changes between the two orders.

Inland, in a plant room at 40 °C with condensation but no chloride source, 304 is correct and stays correct for the life of the genset. Coastal, 500 m from the shore, the same panel sees airborne chloride deposited on the leeward face and held there by dust. Pitting begins at the cut edges and the fixing holes within two to three years. 316L roughly doubles the time to first pit, and a powder coat over a blasted surface buys more than the grade change does, because it keeps the chloride off the surface altogether.

The coated option adds about 0.17 kg per panel at a 60 µm film on both faces, which is 1.4 kg across the set. It also hides the surface, so a pit under the coating is invisible until it perforates. For a canopy the customer will not clean, coating plus 316L is the combination that survives; for a canopy that is washed twice a year, 316L bare is enough.

Galling, threads and the assembly decision

A stainless fastener in a stainless thread is the most common avoidable failure in the whole category, and it is usually specified by the same person who chose the grade. Austenitic stainless work-hardens at the contact patch, the oxide does not act as a solid lubricant, and the surfaces cold-weld. Once welded, the torque needed to move the fastener exceeds its torsional strength and the head shears.

Three controls work. Use a lubricated fastener — A2 or A4 stainless with a molybdenum-disulphide or PTFE coating — and specify a torque rather than "tight". Use a dissimilar pair where the design allows, such as a stainless bolt into a plated steel insert. And keep the engagement length within reason: 1.5 times diameter in stainless is enough for most sheet assemblies and every extra thread adds contact area that can gall.

The torque number is the one to put on the drawing, because it is what separates a joint that holds from a joint that seizes. For an M6 A2 bolt at a 6 kN clamp load, the tightening torque is about 12.6 N·m dry at a friction coefficient of 0.35 and about 5.4 N·m with a lubricated coating at 0.15. Both give the same clamp load. The dry value is the one a fitter will exceed by feel, and the joint that fails is the one that felt tightest.

Where a stainless fastener must sit against galvanised steel, the couple is the issue rather than the thread. Stainless sits near −0.05 V and zinc near −1.0 V, so the zinc around the fastener becomes the anode and disappears. Isolating washers are the cheap fix.

Where the cost actually sits

Procurement sees one number per part and the alloy content is buried inside it. Three lines are worth separating at quotation.

The alloy surcharge. Nickel at 8–10.5% in 304 and 10–14% in 316L, plus 2–3% molybdenum, are priced on a metal index that moves independently of the conversion work. Ask for the surcharge as its own line so a price rise can be traced to the metal rather than negotiated as a whole.

The conversion route. 304 welds and forms much like CRCA, so the conversion cost is close. 316L is slower: more galling in forming means more tool cleaning, and the same bend takes the same time but the die needs attention more often.

The finish. A coated part carries the coating cost twice over on a box, because both faces and the cut edges are treated. On a 2 m² panel the coated area is closer to 4 m², and that arithmetic often closes the gap between CRCA with a full coating system and bare 304.

The cheapest wrong decision is down-gauging to save weight. Dropping from 2.0 mm to 1.5 mm removes 25% of the mass and 58% of the bending stiffness, because stiffness scales with the cube of thickness: (1.5 ÷ 2.0)³ = 0.42. A panel that deflected 4 mm now deflects about 9.5 mm. If the part is stiffness-driven, the thickness is not a cost lever.

The mirror-image error is paying for corrosion resistance the environment does not need. A 316L cover on a machine in a dry, heated hall is protected by the building, not by the molybdenum, and the premium is paid on every unit for a decade. The check takes one line on the RFQ: name the chloride source and the maximum temperature, and the grade follows.

Where 304 is exactly the right answer

Three cases argue for 304 rather than against it. Dry indoor enclosures with a competent coating do not need the molybdenum, and bare 304 in a clean room or a food-handling area is easier to clean than a coated steel panel because there is no film to chip. Potable water below 200 mg/L chloride, running cold, is a 304 application with decades of service behind it. And low-temperature service is where austenitic stainless outperforms carbon steel outright: 304 keeps its impact toughness down to about −196 °C, while a carbon steel part that is fine at 20 °C can behave like glass at −40 °C.

There is also a maintenance argument that cost comparisons usually miss. A bare 304 panel can be cleaned, inspected and left in service. A coated panel that is scratched through to the steel has to be repaired, and the repair is a site visit. On a part that people will lean on, hose down or drag equipment across, the uncoated surface is often the lower whole-life cost even at a higher material price.

What this changes on a drawing

1. The environment, written out: chloride in mg/L, maximum service temperature, indoor or outdoor, cleaning chemicals.
2. The grade with its standard, and a note that substitution requires approval.
3. Whether the part is welded, and the post-weld treatment if any.
4. The fastener specification and its lubricant, not just "stainless".
5. The finish and the faces it covers, including cut edges.
6. Any contact with a dissimilar metal, so the couple can be designed out.
7. The expected service life, because a ten-year part and a two-year part do not need the same grade.

What Ganesh will and will not run

Ganesh cuts 304, 304L, 316L, 430 and 201 on the 12 kW fibre laser up to 20 mm, forms to 10 mm on the 200 T brake, and welds through three CLOOS robotic cells and more than 30 manual stations. Seven-tank pretreatment, the 16 ft × 16 ft blast chamber and the 18 ft × 14 ft powder booth handle the coating. The custom parts range covers the material mix.

We do not stock or process duplex 2205, we do not roll-form profiles, and we do not operate a passivation line. Where a drawing calls for any of those, say so at enquiry rather than at first article, and the route will be agreed before the quotation is issued. The limits on each grade are set out in 304 stainless sheet metal and in the materials and grades hub.

Send the drawing with the service environment attached, and the review will confirm whether 304 is right or name the grade that is. Request a quote with the annual quantity and the assembly it fits.

Related reading: the 1050 aluminium corrosion and finish compatibility and the 1050 aluminium for sheet metal parts: grade and formability take the same problem further.