Sensitization is the single biggest metallurgical problem in stainless steel welding. It strips the corrosion resistance from the heat-affected zone by depleting chromium at grain boundaries. The mechanism is straightforward: carbon in the steel combines with chromium at temperatures between 800 and 1500F to form chromium carbide (Cr23C6) particles along grain boundaries. Those carbides pull chromium out of the surrounding matrix, dropping the local chromium content below the 10.5% minimum needed for stainless behavior. The result is a narrow band of material that rusts like carbon steel in a corrosive environment.
In service, sensitized stainless fails by intergranular corrosion (IGC), sometimes called “weld decay.” Corrosion attacks preferentially along the chromium-depleted grain boundaries adjacent to the weld, causing the grains to literally fall out of the surface. The weld itself looks fine. The base metal away from the HAZ looks fine. But a narrow band 1-3mm from the fusion line dissolves.
How Sensitization Happens During Welding
When you weld austenitic stainless (304, 316, 321, 347), the base metal adjacent to the weld heats through a thermal cycle. Different zones reach different peak temperatures:
- Fusion zone: Melts completely, solidifies with weld metal chemistry
- Near HAZ: Heats above 2000F, dissolves any existing carbides, then cools through the sensitization range
- Mid HAZ (sensitization zone): Reaches 800-1500F, where chromium carbides form if carbon is available
- Far HAZ: Below 800F, unaffected
The mid-HAZ is where sensitization occurs. This zone heats into the danger range and stays there long enough for carbon atoms to diffuse to grain boundaries and combine with chromium. The amount of sensitization depends on three factors:
- Carbon content of the steel: Higher carbon = faster and more severe sensitization
- Time at temperature: Longer time in the 800-1500F range = more carbide formation
- Grain size: Larger grains have less grain boundary area per unit volume, concentrating the carbide effect
| Factor | Increases Sensitization Risk | Decreases Sensitization Risk |
|---|---|---|
| Carbon content | Standard grade (0.08% C) | L-grade (0.03% C) |
| Heat input | High (weaving, slow travel) | Low (stringers, fast travel) |
| Interpass temp | High (over 350F) | Low (under 350F) |
| Number of passes | Many (repeated thermal cycles) | Few (single pass when possible) |
| Material thickness | Thick (slower cooling) | Thin (faster cooling) |
| Stabilizing elements | None (304, 316) | Ti (321) or Nb (347) |
The Chromium Depletion Mechanism
Stainless steel’s corrosion resistance comes from a passive chromium oxide film that forms spontaneously on the surface when chromium content exceeds about 10.5%. In properly processed 304 (18% Cr), there’s plenty of chromium distributed uniformly through the matrix.
During sensitization, chromium carbide (Cr23C6) precipitates at grain boundaries because grain boundaries are high-energy sites where diffusion happens fastest. Each carbide particle contains 23 chromium atoms for every 6 carbon atoms. That’s a lot of chromium being pulled from the surrounding matrix.
The critical point: chromium diffuses much more slowly than carbon at these temperatures. Carbon atoms travel quickly to grain boundaries, but the chromium needed to replace what the carbides consumed can’t replenish fast enough. The result is a chromium-depleted zone (typically 0.5-2 micrometers wide) on either side of each grain boundary where chromium drops below 10.5%.
That narrow depleted zone is no longer “stainless.” It corrodes preferentially in any environment that attacks carbon steel, creating a network of corrosion paths along every grain boundary in the sensitized region.
Prevention Method 1: Use L-Grade Materials and Filler
The most common and practical prevention method is controlling carbon content. L-grade stainless (304L, 316L) has a maximum carbon content of 0.03%, compared to 0.08% in standard grades.
At 0.03% carbon, the precipitation kinetics slow dramatically. There simply aren’t enough carbon atoms to form significant carbide volumes during the brief time the HAZ spends in the sensitization range during normal welding. The material can still sensitize with extended time at temperature (hours in a furnace at 1200F), but normal welding thermal cycles don’t produce meaningful sensitization.
Always use L-grade filler metal (ER308L, ER316L, E308L-16, E316L-16) regardless of whether the base metal is standard or L-grade. The weld deposit should have the lowest carbon content possible because it undergoes the most severe thermal cycling from multiple passes.
| Base Metal | Correct Filler | Wrong Filler | Risk of Wrong Filler |
|---|---|---|---|
| 304L | ER308L | ER308 (standard) | Sensitization in HAZ and weld |
| 304 (standard) | ER308L | ER308 (standard) | Sensitization risk from dilution |
| 316L | ER316L | ER316 (standard) | Sensitization in HAZ and weld |
| 316 (standard) | ER316L | ER316 (standard) | Sensitization risk from dilution |
Prevention Method 2: Stabilized Grades (321, 347)
Stabilized stainless steels contain elements that preferentially combine with carbon before chromium gets a chance.
321 stainless contains titanium (Ti). Titanium carbide (TiC) is more thermodynamically stable than chromium carbide. The titanium ties up the carbon as TiC, leaving chromium in solution. 321 is commonly used for high-temperature service where extended time in the sensitization range is expected.
347 stainless contains niobium (Nb, also called columbium). Niobium carbide (NbC) serves the same function as TiC in 321. Some engineers prefer 347 because niobium carbides are more stable at higher temperatures than titanium carbides.
Welding stabilized grades: Use matching filler (ER347 for 347 base, ER321 for 321 base). Note that ER321 filler rod is difficult to manufacture and not widely available. Most shops use ER347 filler for both 321 and 347 base metal, which is an acceptable substitution per most codes.
Prevention Method 3: Control Heat Input
Even with L-grade material, minimizing heat input is good practice. Less heat input means faster cooling through the sensitization range and less time for any carbide precipitation to occur.
Practical heat input controls for stainless welding:
- Stringer beads only. No weaving. Weaving dramatically increases heat input.
- Interpass temperature below 350F. Measure between passes. Don’t weld until it cools.
- Minimum amperage for proper fusion. Don’t run hot “just to be safe.”
- Fast travel speed. Move briskly. A smaller, faster weld beats a big, slow one.
- Minimum number of passes. Each pass re-heats the previous HAZ. Fewer passes = less cumulative sensitization.
- Use TIG for thin material. TIG’s lower heat input per unit length reduces the sensitized zone width.
Solution Annealing: Fixing Sensitized Material
If stainless has already been sensitized, solution annealing restores full corrosion resistance. The procedure:
- Heat the part to 1900-2100F (1040-1150C) in a furnace
- Hold at temperature long enough for the chromium carbides to dissolve back into the austenite matrix (typically 1 hour per inch of thickness)
- Rapid cool (water quench for heavy sections, air cool for thin material) through the sensitization range to prevent carbides from reforming
Solution annealing is effective but has practical limitations:
- The part must fit in a furnace
- Rapid cooling may cause distortion on complex weldments
- Scale formation requires post-anneal pickling
- Cost is significant for large assemblies
For most welded assemblies, preventing sensitization through material selection and welding practice is far more practical than fixing it after the fact.
Testing for Sensitization
ASTM A262 Practice A (Oxalic Acid Etch)
A quick screening test. Etch a polished cross-section in 10% oxalic acid at 1 amp/cm2 for 1.5 minutes. Examine under a microscope at 250-500x magnification. Three structure classifications:
- Step structure: No continuous ditches at grain boundaries. Material is not sensitized. Passes.
- Dual structure: Some ditched grain boundaries. Borderline. Requires further testing.
- Ditch structure: Continuous ditches surrounding individual grains. Material is sensitized. Fails.
Practice A is a screening test only. A “step” result means the material passes without further testing. A “dual” or “ditch” result requires a more definitive test (Practices B through E).
ASTM A262 Practice E (Strauss Test)
The definitive test. Boil samples in an acidified copper sulfate solution (copper-copper sulfate test) for 24-120 hours, then bend. If the sample shows intergranular cracking on the bend, the material is sensitized. This test is time-consuming but gives a clear pass/fail result.
Electrochemical Tests (EPR, DL-EPR)
Electrochemical potentiokinetic reactivation testing quantifies the degree of sensitization numerically. DL-EPR (double-loop EPR) is the most common variant. It’s faster than boiling tests and provides a continuous scale rather than pass/fail. Results correlate to the severity of chromium depletion. Used primarily in research and for qualifying welding procedures.
Real-World Examples of Sensitization Failure
Chemical plant piping: 304 stainless piping welded with standard (non-L) filler in a chemical plant handling dilute sulfuric acid. Within 18 months, the HAZ adjacent to every weld developed knife-line attack, with through-wall corrosion penetrating the pipe. The entire piping run was replaced with 304L base metal and 308L filler.
Food processing equipment: A brewery installed 304 stainless vessels welded with high heat input (multiple wide weave passes). The caustic cleaning solutions used for sanitization attacked the sensitized HAZ, producing pitting and eventual contamination of the product. Re-welding with L-grade filler and proper heat control solved the problem.
Exhaust systems: Sensitization in standard 304 exhaust components exposed to road salt and condensate. The combination of chlorides and sensitized grain boundaries caused rapid intergranular stress corrosion cracking. 321 stabilized stainless eliminated the sensitization issue.
The lesson is consistent: sensitization failures almost always come from one of two mistakes. Either the wrong filler (standard instead of L-grade) or excessive heat input (high interpass temperature, weaving, too many passes). Both are completely preventable with proper procedure. The metallurgy is well understood. Apply it.