# Cryogenic treatment

Subzero processing of metals: retained-austenite transformation, dimensional stabilization, deep-cryogenic claims, treatment sequence, and application-specific qualification.

Reviewed: 2026-11-01

Tags: processes, heat-treatment, cryogenic-treatment

BÖHLER's **K340 ECOSTAR** brochure, **K340EN – 03.2020**, describes vacuum hardening at **1050 °C for 30 minutes**, **5 bar nitrogen quench**, **−70 °C for two hours** and **three two-hour tempers**. Confirm the applicable tempering temperatures and section response for the part. Measure retained austenite and any dimensional change separately. [BÖHLER K340, pages 8–9 and final-page edition](https://www.bohler-edelstahl.com/app/uploads/sites/248/productdb/api/k340en_ecostar.pdf)

For hardened steels, subzero treatment can reduce retained austenite. Studies of carbide precipitation and wear depend on the alloy and full treatment route.

Define the required outcome before specifying a minimum temperature. A tool-life claim, a retained-austenite limit, and a dimensional-stability requirement are different acceptance questions.

## Separate the mechanisms and terminology

Linde groups subzero processing into **cold treatment** associated with austenite-to-martensite transformation and **cryogenic treatment**, described on its technical page as below **−120 °C**, associated with conditions for subsequent fine-carbide formation in higher-alloy steels. This is the supplier's terminology, not a universal boundary shared by every producer. The same page distinguishes shrink fitting, which uses temporary contraction to assemble parts, from metallurgical treatment. It also identifies liquid-nitrogen-cooled chambers as equipment, rather than requiring direct immersion of the workpiece. The page has no verified issue date. [Linde: subzero treatments](https://www.linde-gas.com/processes/heat-treatment/sub-zero-treatments)

| Objective | Relevant evidence | What the evidence does not establish |
| --- | --- | --- |
| Reduce retained austenite | Phase measurements for the specified alloy and complete thermal route | An automatic improvement in every fatigue, wear, or toughness property |
| Improve dimensional stability | Dimensions before/after treatment and after representative exposure | Zero dimensional change during treatment itself |
| Improve tool life | Comparable tools tested under controlled application conditions | The same improvement for another cutting material or wear mechanism |
| Shrink-fit assembly | Temporary temperature-dependent dimensional change and assembly fit | A permanent metallurgical improvement after warming |

Record actual temperatures and sequence, not only a label such as “deep cryo” or “cryogenic tempering.” A subzero step does not replace the specified elevated-temperature tempering operation.

## Alloy, starting condition, and treatment order

Uddeholm's *Heat Treatment of Uddeholm Tool Steels*, **Edition 9, October 2017**, explains that retained austenite can transform during service, causing local distortion and fresh untempered martensite. It describes subzero treatment and high-temperature tempering as routes to low retained-austenite content. The guide notes that high-temperature-tempered tools may show little or no additional hardness increase from subzero treatment. For its repeated cold/temper stabilization route, it explicitly finishes with tempering to avoid leaving untempered martensite. These are tool-steel guidelines; the target retained-austenite content and permissible property changes remain application-specific. [Uddeholm guide, page 12](https://www.uddeholm.com/app/uploads/sites/216/2024/05/Tech-Uddeholm-Heat-treatment-EN.pdf)

Linde's white paper discusses both stabilization of austenite before cold treatment and cracking risks when susceptible steels are left untempered. It describes an optional preliminary temper in some sequences, followed by cold treatment and final tempering. Its final page prints **©2017** and document code **3332_0917**. Thus, “always cold-treat immediately before any temper” and “always temper first” are both inadequate general instructions. The qualified alloy-specific route must resolve the tradeoff. Its discussion of unsuccessful applications also prevents treating reported wear gains as universal. [Linde white paper, sections 2.1.5 and 2.2.4](https://cdn2.hubspot.net/hubfs/189660/3332_0917_Sub%20Zero%20Treatment%20of%20Steel%20WP.pdf)

Before selecting a route, identify the producer grade, product form, austenitizing and quenching history, existing temper, and elapsed time since quench. Also identify case-hardening, welds, coatings, dissimilar inserts, and previous service. A finished assembly may contain materials or fits that cannot tolerate the proposed thermal excursion even when its steel component can.

## Bounded producer example

BÖHLER's **K340 ECOSTAR** brochure, printed **K340EN – 03.2020**, provides a sequence under the heading “Cryogenic treatment”: vacuum hardening at **1050 °C for 30 minutes**, nitrogen quenching at **5 bar**, subzero cooling at **−70 °C for two hours**, and **three two-hour tempering operations**. This is a producer-specific example, not a recipe for D2, high-speed steel, bearings, or an assembled machine.

Its terminology also demonstrates why the word “cryogenic” alone cannot establish a temperature: the K340 example is warmer than Linde's deep-treatment category. The adjacent tempering graph identifies **20 mm square specimens** and its own test conditions; it does not establish the through-section response of an arbitrary large tool. Obtain the applicable producer recommendation and qualified tempering temperatures rather than combining isolated values from different graphs. [BÖHLER K340, pages 8–9 and final-page edition](https://www.bohler-edelstahl.com/app/uploads/sites/248/productdb/api/k340en_ecostar.pdf)

## Benefits need complete-condition evidence

An original **21 November 2019** study by Kusý and colleagues examined **Vanadis 6** treated at **−75 °C**, with different durations and subsequent tempers. It found changes in retained austenite, carbides, and hardness, but fracture-toughness results depended on the tempering condition and were often worse than the comparison condition. The paper is useful evidence that improved hardness or phase conversion does not automatically mean improved resistance to cracking. Its measured specimens and thermal schedules are the scope of those findings, not a general tool-life guarantee. [Original study, Materials 12(23), 3827](https://pmc.ncbi.nlm.nih.gov/articles/PMC6926819/)

For production trials, compare otherwise equivalent tools and document material lot, complete heat treatment, geometry, finish/coating, workpiece, speed/feed, lubrication, and retirement criterion. Separate abrasive wear, chipping, fracture, and loss of dimensional accuracy. A change in the dominant failure mode can make a single “life improvement percentage” misleading. Include replicate results and scatter before making a purchasing assumption.

## Measurement, dimensions, and an original example

ASTM **E975-22**, listed as active with a catalog update of **15 February 2023**, measures retained austenite by integrated X-ray diffraction intensities. Its scope assumes near-random crystallographic orientations and applies from **1 volume %** retained austenite upward. It flags carbide-peak interference and cautions for total alloy content above **15 mass %**. Consequently, a laboratory should confirm method applicability and specimen preparation rather than reporting an unexplained phase percentage. The standard measures phase content; it is not a cryogenic-processing recipe or a universal acceptance limit. [ASTM E975-22](https://store.astm.org/standards/e975)

Consider a **hypothetical qualification result**, not measured producer data: comparable measurements give retained austenite of **12 volume % before** and **3 volume % after** the complete route. The change is **9 percentage points**, or **75% of the initial measured fraction**:

`(12 − 3) / 12 × 100 = 75%`

This does not mean a part changes length by 9% or 75%. Suppose a separately measured length changes from **100.000 to 100.012 mm** at the same reference temperature. That is **+0.012 mm**, **+0.012%**, or **+120 µm/m**. Those dimensional observations require their own uncertainty and repeatability assessment. Neither result alone predicts future drift, complete transformation throughout the part, or tool life. Measurements made while the part is cold would additionally mix thermal contraction with any permanent change.

## Process handoff and acceptance

Specify controlled cooling, soaking, rewarming, and tempering together, including how representative load temperature is established. Review geometry and support with the processor: transformation and thermal gradients can create distortion or cracking rather than simply remove it. Reserve appropriate finishing allowance and inspect dimensions after the complete route and thermal equilibration.

A useful order includes four groups of information:

1. **Material and input state:** producer grade, lot, form, dimensions, hardening/quench/temper history, and prior service or rework.
2. **Qualified sequence:** temperatures, rates, holds, delays, load arrangement, rewarming, and final tempering requirements.
3. **Output requirements:** phase fraction where applicable, hardness, dimensions, crack inspection, toughness or application testing, and approved limits.
4. **Traceability and comparison:** load records, measurement conditions, sampling, reference tools, and disposition of deviations.

## Related reference

[Grades, tempers, and conditions](../../materials/grades-tempers-and-conditions.md)