# Sensitization of stainless steel

Chromium-depleted microstructures, time–temperature dependence and why susceptibility is different from observed corrosion.

Reviewed: 2026-11-02

Tags: production, stainless-steel, corrosion

Two illustrative specimens cool linearly from **900 to 500 °C** at **20 or 5 °C/s**. In Outokumpu’s **850–550 °C** comparison interval they spend **15 or 60 seconds**, respectively. Residence time alone leaves sensitization unresolved. Alloy, carbon content, prior thermal history and precipitation kinetics matter. [Outokumpu corrosion-resistance guidance](https://www.outokumpu.com/en/expertise/stainless-basics/corrosion-resistance).

In susceptible austenitic stainless steel, chromium-rich grain-boundary carbides can leave adjacent regions depleted in chromium and vulnerable to intergranular attack. [Outokumpu corrosion-resistance guidance](https://www.outokumpu.com/en/expertise/stainless-basics/corrosion-resistance)

Distinguish thermal exposure, sensitized material and subsequent corrosion. Heat tint alone does not demonstrate attack.

## Why a stainless alloy can become locally vulnerable

The corrosion resistance of stainless steel depends on its ability to maintain a passive surface. Chromium depletion near a grain boundary can create a local region less resistant than the surrounding material, even though the bulk chemical analysis still matches the grade. A bulk chromium percentage therefore cannot establish that every microscopic region retained the same corrosion behavior after thermal processing.

Outokumpu identifies welding and inappropriate hot-forming exposure as possible sources of sensitization, and notes that the effect can also increase susceptibility to other localized corrosion forms. Its illustrative **550–850 °C** range describes a metallurgical risk region, not a universal instantaneous threshold for every stainless grade. [Outokumpu, intergranular-corrosion section](https://www.outokumpu.com/en/expertise/stainless-basics/corrosion-resistance)

Actual corrosion also depends on environment. Nickel Institute's *Design Guidelines for the Selection and Use of Stainless Steels* explicitly distinguishes the degree of sensitization from the aggressiveness of the medium and notes that many environments do not cause intergranular attack in sensitized austenitic material. This supports a susceptibility-versus-damage distinction, not a service approval for an unspecified fluid or temperature. [Nickel Institute publication 9014, intergranular corrosion](https://www.nickelinstitute.org/media/1667/designguidelinesfortheselectionanduseofstainlesssteels_9014_.pdf)

## Time, temperature and carbon content interact

Sensitization is a kinetic process. Peak temperature alone does not describe the time available for carbide nucleation, growth and redistribution of chromium. A time–temperature-sensitization diagram applies to the material condition and test basis behind it; it is not a generic calendar of safe furnace holds.

ASM International's article dated **June 17, 2016** illustrates the strong effect of carbon content using austenitic-steel sensitization curves. Lower carbon delays carbide precipitation. The article's illustrated exposure times are source-specific examples, not maximum permissible service times for every 304L product. [ASM, time–temperature sensitization curves](https://www.asminternational.org/news/26579541/)

A thermal path through the HAZ can differ from a furnace hold, and successive weld passes can create multiple exposures. Material composition, prior processing and the initial carbide state affect the response. Adding elapsed seconds from different temperatures does not create a validated sensitization index without a model that accounts for those differences.

## Low-carbon and stabilized grades reduce particular risks

Low-carbon austenitic grades such as 304L and 316L reduce the carbon available for chromium-carbide formation. Nickel Institute's selection guide also cautions that low-carbon grades can become sensitized after extremely long exposures in the relevant range. The “L” suffix is therefore not a claim of immunity to every thermal history or corrosion mechanism. [Nickel Institute selection guide](https://www.nickelinstitute.org/media/1667/designguidelinesfortheselectionanduseofstainlesssteels_9014_.pdf)

Stabilized grades use additions such as titanium or niobium to bind carbon preferentially. This is a different metallurgical strategy from lowering carbon alone. Stabilization also has limits: The IMOA high-performance-austenitic fabrication guide, co-published by Nickel Institute and ICDA, **first edition 2010 (© IMOA 2010)**, describes knife-line attack in Types 321 and 347 when welding dissolves stabilizing carbides and a subsequent thermal exposure permits chromium carbide formation in a narrow adjacent region. The guide distinguishes this from intermetallic-phase problems in high-performance austenitics. [Nickel Institute fabrication guide, §6.3.2](https://nickelinstitute.org/media/3856/fabrication-of-high-performance-austenitic-stainless-steels.pdf)

These mechanisms explain why grade family matters. A statement about chromium-carbide sensitization in a conventional austenitic alloy does not establish the behavior of every ferritic, duplex or precipitation-hardening stainless steel. Nor does preventing this mechanism eliminate pitting, crevice corrosion or all welding defects.

## Sensitization can evolve rather than increase monotonically

Original research by Kolli and coauthors, published **November 15, 2019** in *ISIJ International* **59(11), 2090–2097**, compared double-loop electrochemical reactivation measurements with precipitation/diffusion predictions. Its model treats carbide nucleation and growth separately and describes both sensitization and self-healing through chromium-depletion behavior. This is evidence that the microstructural response can evolve with exposure, rather than a rule that every additional second always increases susceptibility. [Kolli et al., DOI 10.2355/isijinternational.ISIJINT-2019-264](https://www.jstage.jst.go.jp/article/isijinternational/59/11/59_ISIJINT-2019-264/_article/-char/en)

“Self-healing” in this metallurgical context does not mean that an already corroded component regrows lost metal. Similarly, a treatment capable of redissolving carbides does not reverse existing section loss. Surface cleaning and passivation address different surface conditions; removal of contamination alone is not evidence that a chromium-depleted microstructure has been eliminated.

## Original bounded thermal-history example

Assume two invented specimens cool linearly from **900 to 500 °C**, one at **20 °C/s**, the other at **5 °C/s**. For illustration only, calculate time spent between **850 and 550 °C**, using Outokumpu's stated range as a bookkeeping interval.

The temperature span is 300 °C. The first specimen spends 300/20 = **15 seconds** inside it; the second spends 300/5 = **60 seconds**. Total time from 900 to 500 °C is respectively **20 and 80 seconds**. Thus the slower path has four times the residence time within the chosen interval.

Neither 15 nor 60 seconds proves sensitization, absence of sensitization or a corrosion rate. The calculation contains no alloy-specific precipitation kinetics, carbon content, grain-boundary condition, heating history or exposure environment. A second pass would add another thermal path, not automatically an equivalent isothermal hold. The arithmetic illustrates the difference between a temperature range and a time-dependent transformation assessment.

## What testing demonstrates

ISO 3651-2:1998, **second edition, May 1998**, concerns resistance to intergranular corrosion of ferritic, austenitic and duplex stainless steels in sulfuric-acid-containing media. ISO lists the publication as confirmed in 2021 and currently to be revised; a revision project is not itself a replacement published edition. Its stated media and material scope limit what the result represents. [ISO 3651-2:1998](https://www.iso.org/standard/9098.html)

A micrograph, electrochemical susceptibility result and corrosion-exposure test measure different aspects of the problem. Their interpretation depends on specimen condition and test method. A passing result under a defined accelerated test is not a universal lifetime guarantee, while susceptibility does not prove that a component has experienced attack in its actual service.

The online producer guidance is undated. The cited fabrication guide identifies its first edition as 2010 and carries © IMOA 2010. The publication date of the cited Nickel Institute 9014 PDF was not independently verified here and is not inferred from its upload path. The ASM article, research paper and ISO edition dates above are explicitly stated by their publishers. Sources were reviewed on **2026-11-02 PST**.

## Related reference

[Heat-affected zones](./heat-affected-zones.md) · [Stainless-steel passivation](../../processes/finishing/stainless-steel-passivation.md)