|
QUICK ANSWER An ASTM A923 report shows whether a duplex stainless steel product contains harmful intermetallic phases, which are brittle compounds that cut toughness and corrosion resistance. Before accepting it, check five things: the method (A, B or C), the grade and acceptance limit, the heat-treatment condition tested, the sample location, and the traceability from the report to your actual heat and lot. |

Duplex stainless steels such as S31803 and S32205 are popular because they are strong and resist chloride attack. But that performance depends on one fragile condition: the steel must be cooled quickly enough after solution annealing. If it is not, hard, brittle phases can form. A standard mill test certificate may not reveal them. ASTM A923 exists to close that gap.
This guide explains, in plain language, what an A923 report proves, what it does not prove, and how to review one line by line before you sign off on a delivery.
What Is ASTM A923, and What Does an A923 Report Prove?
|
QUICK ANSWER ASTM A923 is a test method standard, not a product specification. It detects intermetallic phases in duplex stainless steel that are bad enough to significantly reduce toughness or corrosion resistance. A passing report proves the tested sample was free of those phases. It does not prove anything else. |
Duplex steel is a mix of two crystal structures, ferrite and austenite. When it is held between roughly 600 and 1750°F (320 to 955°C), brittle compounds can form. A923 names this range in its scope. These compounds, often called sigma phase and similar phases, make the steel crack more easily and let it pit in salt water.
What the standard says about its own limits. A923 states that meeting the chemical and mechanical rules of a product specification does not necessarily show that harmful phases are absent. It also says the tests will not necessarily detect losses of toughness or corrosion resistance that come from other causes. In short, A923 checks for one specific problem. It is a targeted tool, not a general quality stamp.
Current edition. A923-25 was published in September 2025 and is the edition currently listed by standards sellers. Earlier editions (2003, 2008, 2014, 2022 and 2023) remain in circulation on older reports, so always read which edition a report cites.
What Is the Difference Between A923 Methods A, B and C?
|
QUICK ANSWER Method A is a fast etch test for screening. Method B is a Charpy impact test at −40°C (−40°F). Method C is a 24-hour ferric chloride corrosion test. A, B and C answer different questions, so a report must say which one was run. |
The logic. Method A is cheap and quick. A sample with a clean etch can pass without further testing. A sample with a doubtful etch needs Method B or C to settle the question. Method B looks at toughness. Method C looks at corrosion resistance.
|
Method |
What is done |
What it detects |
Key rule from the standard |
|
A |
Electrolytic etch in sodium hydroxide, then microscope check at 400–500× |
Visible signs of intermetallic phase: unaffected, possibly affected, affected, or centerline structure |
May be used to accept material but not to reject it. Only an unaffected structure counts as acceptable. |
|
B |
Charpy V-notch impact test at −40°C (−40°F) |
Loss of toughness linked to intermetallic phase |
S31803/S32205/J92205: 40 ft-lb (54 J) base metal and heat-affected zone; 25 ft-lb (34 J) weld metal (full-size specimen). S32750: value agreed between buyer and seller. |
|
C |
24-hour immersion in ferric chloride solution (100 g FeCl3·6H2O in 900 mL water, pH about 1.3), then weight-loss check |
Loss of corrosion resistance from chromium- and molybdenum-poor zones |
Maximum 10 mdd (mg per square decimeter per day). Base metal S31803/S32205 at 25°C, weld metal at 22°C, S32750 at 40°C. |
Source: ASTM A923-03 text and tables as reproduced by a public copy; Tables 1–3 in later editions add more grades. Verify against the edition on your order.
Edition note. Later editions extend the grade tables. Copies of the 2008 and 2014 editions list S32550 and S32520 for Methods A and C. The 2014 edition also lists S32760, and the 2008 edition adds the cast grade J93404 for Methods A and B. Always confirm your grade appears in the table of the edition cited on the report.
A useful detail on Method B. For sub-size Charpy bars, the required energy drops in proportion to the smaller cross-section. The standard also notes that transverse specimens from mill products typically give about one half to two thirds of the longitudinal value. A report that quotes a number without saying the specimen size and direction is hard to judge.
What Should Buyers Check First on an A923 Report?
|
QUICK ANSWER Check identity, method, condition, location, result and retest history, in that order. If any one of the six is missing or unclear, the report cannot support acceptance. |

Why a checklist works. Most disputes about A923 do not come from the test itself. They come from a report that cannot be tied to the delivered material, or that tested a different condition than the one shipped. Use the table below as a desk review.
|
Review item |
What to confirm |
Why it matters |
|
1. Identity |
UNS grade, heat number and lot number match the material test certificate and the marking on the product |
A perfect report on the wrong heat protects nothing |
|
2. Standard and edition |
Report cites ASTM A923 with a year, such as A923-25 or A923-23 |
Grade tables and criteria change between editions |
|
3. Method |
Method A, B or C is named, and it is the one on your purchase order |
Method A alone cannot reject; B and C measure different properties |
|
4. Condition tested |
Product tested in the final solution heat-treated condition, or another condition you agreed in writing |
The standard requires the final condition unless otherwise agreed. Later heat or forming can undo a good result |
|
5. Sample location |
Taken from the region that cooled slowest, and through mid-thickness for heavy sections |
The standard says tests must hit the region most likely to form phase; a convenient edge sample can hide a problem |
|
6. Result and limit |
Actual measured values (ft-lb, J, mdd) next to the acceptance value, with test temperature |
A bare word "pass" cannot be checked |
|
7. Retests and re-anneals |
Report states whether any retest or full re-anneal took place |
The standard lets the producer retest or re-anneal after a failure. You should know it happened |
|
8. Laboratory |
Lab name, date, signature, and accreditation if you require it |
Supports audit and later claims |
Source: Review framework by EETA, based on ASTM A923 requirements for specimen location, test condition and retests.
Does an A923 Pass Mean the Product Meets Its Specification?
|
QUICK ANSWER No. An A923 pass and a product-specification pass are two separate checks. You need both: the mill test certificate for chemistry, strength and dimensions, and the A923 report for phase-related toughness and corrosion behavior. |
The reason is simple. The product standard, for example one for duplex pipe, plate or bar, sets limits on composition and tensile properties. A923 asks a different question: did the heat treatment leave harmful phases behind? A steel can meet every chemical and tensile limit and still fail A923 if it cooled too slowly.
|
Question |
Standard mill test certificate (MTC) |
ASTM A923 report |
|
Is the chemistry in range? |
Yes, heat analysis |
No |
|
Is tensile strength at least the minimum? |
Yes |
No |
|
Were harmful intermetallic phases detected? |
Not tested unless added to the order |
Yes, the whole purpose |
|
Is low-temperature toughness checked for phase damage? |
Only if impact testing was ordered |
Yes, with Method B |
|
Does it predict pitting resistance in service? |
No |
No. The standard says Method C is not a basis for predicting other forms of corrosion |
|
Can it detect toughness loss from other causes? |
No |
Not necessarily; the standard says so |
Practical rule. Treat the two documents as a pair. If your order calls for A923, hold the shipment until both documents are reviewed and linked to the same heat and lot numbers.
A923 Method C vs ASTM G48 Method A: Which Test Is on Your Report?
|
QUICK ANSWER They look alike but are not interchangeable. Both use ferric chloride. A923 Method C checks for harmful phase, while G48 Method A is a pitting test. Some project rules, such as ISO 17781 and IOGP S-563, call for G48 Method A, so the purchase order must name the test you want. |
Why confusion happens. The A923 standard itself warns about this in Note 4. It says Method C uses some equipment and steps similar to G48, but it should not be confused with G48. Method C does not find a critical pitting temperature and does not judge fitness for a particular environment. It is designed only to detect harmful phase.
A 2023 paper by Cowe, Marchebois and Mendibide compared the two methods side by side. Its summary notes that both have been used for many years to find intermetallic phases and precipitates in duplex steel, and that ISO 17781 and IOGP S-563 now stipulate G48 Method A. It also notes that A923 uses a higher test temperature for 25% chromium material.
|
Point |
A923 Method C |
ASTM G48 Method A |
|
Main purpose |
Detect harmful intermetallic phase |
Rank pitting resistance; used by some project specifications as a phase check |
|
Test period |
24 hours (unless otherwise specified) |
Typically 24 hours; confirm in your standard |
|
Pass criterion |
Corrosion rate at or below 10 mdd |
Defined by the governing specification, often no pitting and a weight-loss limit |
|
Temperature |
Set by grade: 25°C, 22°C (weld) or 40°C |
Set by the project specification |
|
Who typically requires it |
Product specifications and purchase orders that cite A923 |
ISO 17781 and IOGP S-563 users, per the 2023 paper |
Source: ASTM A923 Note 4 and Table 3; Cowe, Marchebois and Mendibide, AMPP/NACE 2023 (abstract). G48 details vary by edition and project specification; confirm before ordering.
What to do. Write the exact test name, method letter and temperature on the purchase order. If a report shows a ferric chloride test but does not say A923 Method C or G48 Method A, ask the lab to clarify before accepting.
How Does Heat Treatment History Change A923 Results?
|
QUICK ANSWER A few minutes at the wrong temperature can ruin duplex steel. In the standard's own example for S31803 plate, five minutes at 1550°F (843°C) cut longitudinal impact energy from 185 ft-lb to 42 ft-lb, a drop of about 77%, and cut the critical pitting temperature from 35°C to 20°C. |

Duplex steel gets its balance of properties from a correct solution anneal followed by fast cooling. A923 Appendix X1 shows what happens when that balance is disturbed. The data come from a 1991 NACE paper by Davison and Redmond and use 3/8-inch plate with three-quarter-size Charpy bars.
|
S31803 plate condition |
Charpy, longitudinal (ft-lb at −40°F) |
Charpy, transverse (ft-lb at −40°F) |
Critical pitting temp. (G48 Method A) |
|
Mill annealed |
185 |
127 |
35°C |
|
1950°F + water quench + 5 min at 1550°F |
42 |
25 |
20°C |
|
1950°F + water quench + 10 min at 1550°F |
32 |
16 |
15°C |
|
1950°F + water quench + 20 min at 1550°F |
21 |
9 |
15°C |
|
1950°F + air cool |
198 |
133 |
Not tested |
|
1950°F + slow cool |
80 |
88 |
30°C |
Source: ASTM A923 Appendix X1, Table X1.1 (Davison and Redmond, NACE Corrosion 91, Paper 302). Values are for three-quarter-size Charpy specimens, so they are not directly comparable with the full-size 40 ft-lb limit.
Three lessons for buyers.
- Cooling speed matters as much as temperature. Air cooling from 1950°F kept toughness high in this example, while slow cooling cut longitudinal toughness from 185 to 80 ft-lb.
- Later heat can undo a good result. Hot forming, welding or stress relief after the test sample was taken can bring phases back. That is why the standard calls for testing in the final heat-treated condition.
- Thick sections are the high-risk case. The centre of a heavy forging or plate cools slowest, which is why the standard asks for samples from mid-thickness or the slowest-cooled region.
- Forging example. ASTM A1049 sets minimum solution temperatures and requires water quenching to below 200°F (95°C). For example, S32205 (F60) is listed at 1870°F (1020°C) minimum and S32750 (F53) at 1880°F (1025°C) minimum. A heat-treat record that does not show quench completion below 200°F is a reason to ask questions.
Which Product Standards Call for A923, and How Do You Add It?
|
QUICK ANSWER A923 is usually something the buyer asks for. Some product standards list it as an optional supplementary requirement, and others do not mention it. Never assume it is included. Check the product standard, then state it on the purchase order. |
Verified example. ASTM A1049, the specification for duplex forgings for pressure vessels, lists A923 among its referenced documents. Under its supplementary requirement S1, the purchaser must specify the acceptance criteria for Method B or Method C. If you do not select S1, no A923 test is required.
Lean duplex is different. ASTM A1084 covers detection of detrimental phases in lean duplex grades. The A923 standard itself points out that A1084 describes similar tests but with procedures that differ significantly from A923's Methods A, B and C. Do not apply A923 limits to a lean duplex grade.
|
Product form |
Typical duplex specifications |
How to handle A923 |
|
Bar and rod |
ASTM A276, A479 / ASME SA276, SA479 |
Check each standard's supplementary requirements; if absent, add A923 by purchase order text |
|
Plate, sheet, strip |
ASTM A240 / ASME SA240 |
Add A923 by purchase order text; sample from the slowest-cooled region of the plate |
|
Pipe and tube |
ASTM A789 / A790, A928 (welded with filler metal) |
Add A923 by purchase order text and state whether the weld is included |
|
Forgings |
ASTM A182, A1049 |
For A1049 select S1 and name Method B or C with limits |
|
Lean duplex grades |
Check A1084 |
Use A1084, not A923 |
How Should You Write A923 Into a Purchase Order?
|
QUICK ANSWER State six items in the purchase order: the A923 edition, the method, the acceptance criteria, the test frequency, the sample location, and the reporting format. Anything left out will be decided by the supplier. |
Why detail matters. A923 does not set how many tests per lot you need. For Charpy testing, it describes a single specimen taken from the product piece or lot to be represented. If you want one test per heat, per heat-treat load, or per piece, you must say so.
Sample purchase order wording (adapt to your project and grade):
|
"Each heat and each heat-treatment lot shall be tested for detrimental intermetallic phase in accordance with ASTM A923 (state edition), Test Method [B or C]. Acceptance criteria shall be per Table [2 or 3] of the cited edition. Specimens shall be taken from the final solution heat-treated product at mid-thickness in the slowest-cooled region. Actual test values, test temperature, specimen size and orientation, and any retests or re-anneals shall be reported and traceable to the heat and lot numbers on the material test certificate." |
Extra points for welded products. Where welds are included, the standard says specimens should include base metal, heat-affected zone and weld metal. For Charpy testing of a weld, the notch location matters because intermetallic formation can vary a lot over very short distances in the heat-affected zone. Ask for the notch location to be recorded.
What Are the Red Flags on an A923 Report?
|
QUICK ANSWER The biggest red flags are a missing method letter, no actual values, no heat or lot number, an edge-only sample, and an unexplained retest or re-anneal. Any one of them is enough to pause acceptance and ask the supplier for clarification. |
|
Red flag |
What to ask the supplier |
|
Report says only "A923 passed" |
Provide the method letter, test temperature and measured values |
|
No heat or lot number on the report |
Provide traceability linking this report to the delivered heat and lot |
|
Method A listed as the only test, and the etch was not clean |
Method A cannot reject, so ask for Method B or C results |
|
Specimen from an edge or thin offcut of a thick product |
Provide the sample location and explain why it represents the slowest-cooled region |
|
Test date earlier than the final heat treatment |
Confirm the specimen was tested in the final solution heat-treated condition |
|
Sub-size Charpy bars with the full-size limit applied |
Show the proportionally reduced limit and the specimen dimensions |
|
Ferric chloride test without a standard named |
Confirm if it is A923 Method C or G48 Method A, and the temperature used |
|
Retest or re-anneal mentioned in passing |
Provide original and final results and the heat-treatment records |
|
Grade not found in the cited edition's table |
Ask which acceptance criteria were used and why |
Frequently Asked Questions
Q: What does ASTM A923 test for?
A: It tests for detrimental intermetallic phase in duplex stainless steel, meaning phases that significantly reduce toughness or corrosion resistance. It does not test for every kind of defect.
Q: Is Method A enough to accept a duplex product?
A: Only if the etch shows an unaffected structure and your purchase order allows Method A as the acceptance test. Method A can accept material but cannot reject it. A doubtful etch needs Method B or Method C.
Q: What is the Method B Charpy requirement for S32205?
A: For S31803, S32205 and J92205, the standard lists 40 ft-lb (54 J) at −40°C (−40°F) for base metal and heat-affected zone, and 25 ft-lb (34 J) for weld metal, using a full-size specimen.
Q: What is the Method C limit?
A: The calculated corrosion rate must not exceed 10 mdd after 24 hours in the ferric chloride test. Test temperatures are set by grade, for example 25°C for S31803 and S32205 base metal and 40°C for S32750.
Q: Can a product fail A923 and be tested again?
A: Yes. The standard permits a retest on two new specimens for Methods B and C, and the producer may give a failed product a full anneal and retest. Buyers should require the report to disclose it.
Q: Is an A923 report the same as a PREN or pitting rating?
A: No. A923 detects harmful phase. PREN estimates pitting resistance from chemistry, and G48 tests rank pitting behavior. The standard states that Method C does not predict resistance to other forms of corrosion.
Q: Does A923 apply to lean duplex grades?
A: No. Lean duplex grades are covered by ASTM A1084, which uses procedures that differ significantly from A923.

