[ Duplex ]

Duplex stainless steel quality -

ASTM A923 vs ISO 17781
Two standards are available for users when they are considering applying supplementary
testing to duplex stainless steels: ASTM A923 and ISO 17781. ASTM A923 was a strong
first attempt to formalise supplementary testing, while ISO 17781 is a newer contribution.

By Roger Francis, RF Materials & Glenn Byrne, Rolled Alloys

History processing etc. As the market for DSS concept of using Method A alone as a
Modern duplex stainless steels were expanded, more manufacturers around “rapid screening test” is seriously flawed.
developed in the 1970s and really came the world began offering duplex alloys This is because of how intermetallic
into their own in the 1990s and onwards. as part of their portfolio. Similarly, more phases form during conventional heat
They quickly became a corrosion fabrication shops were offering to weld treatment, and how failures in heat
resistant alloy (CRA) of convenience DSS too. A number of these new users treatment processes can cause localised
in the oil and gas industry for seawater either do not realise the importance of precipitation of intermetallic phases
cooling/firewater systems and mildly supplementary testing, or they consider rather than widespread formation. It
sour process fluids. it an unnecessary cost. So they relied on is also the case that metallographic
Duplex stainless steels (DSS) are roughly ASTM and ASME requirements alone and examination can be subjective, especially
50/50 austenite and ferrite and it is manufacturers and fabricators supplied for welds, and it is influenced by sample
necessary to have good control of the accordingly. preparation, etching and metallographer
composition and the heat treatment to The result has been that poor quality interpretation. For these reasons taking
obtain satisfactory properties in both duplex has been supplied to some the results of microstructure checks,
phases. Poor control of these can lead projects, either as parent metal or impact tests and corrosion tests as a
to the precipitation of third phases, as fabrications, which have failed collective is the best way forward.
such as nitrides, sigma, chi and alpha prematurely, usually by corrosion. Test method B is a Charpy impact
prime. These are all deleterious and can There have been numerous reports in toughness test, at minus 40°C. The
reduce both toughness and/or corrosion the literature of failures of both 22%Cr standard is clear that the toughness
resistance. Similarly, poor control during duplex and superduplex due to poor acceptance criterion for each grade
welding of DSS can also result in the quality material. Some of the failures is only used to detect the presence
precipitation of third phases and poor due to poor quality have cost millions of
properties of the joint. dollars.
ASTM ‘product’ specifications and There are two standards available to About the authors
the user that can be considered and Roger Francis has been
ASME fabrication codes do not include
decided upon when trying to apply a corrosion engineer for
microstructure, impact and corrosion
supplementary testing to DSS: ASTM 45 years, with 30 years spent
tests that individually and collectively
A923 and ISO 17781. ASTM A923 was largely on duplex stainless
indicate the presence of deleterious
a strong first attempt to formalise steels. He has published over
phases in these steels or their welded
supplementary testing and ISO 17781 is a 90 technical papers, many
joints. Since the mid 1980s, more
much more recent contribution. on corrosion of duplex stainless steels. He has
sophisticated users have developed
written 6 books and co-edited two more. He is
their own ‘material’ and ‘fabrication’
ASTM A923 currently helping to write a guide to avoiding
specifications that call up these tests
ASTM A923, is designed to detect and solving corrosion problems for desalination
as supplementary requirements to
sigma phase in 22%Cr duplex and plant engineers.
the product specifications and weld
procedure qualification codes in order superduplex1. It does not address nitrides
or alpha prime. (Later a second standard Glenn Byrne is a physical
to assess the quality of the steel and the
was written, ASTM 1084, for lean duplex). metallurgist with 35 years
procedures used to weld it. This was
It quickly became apparent that ASTM of experience in the metals
mostly satisfactory in the 1990s when the
A923 had some serious drawbacks. Test industry. He has worked
major users, such as oil companies, and
A is an etched microsection, but the for multiple markets and
the major engineering design houses
sample is only etched in NaOH, which industries around the world
were writing testing specifications for
will not show nitrides. In our opinion, on the application, development and sales of
these alloys and insisting on compliance.
some of the comparative micrographs in duplex and super duplex in all product forms.
However, the use of DSS spread to
A923 described as “possibly affected” He is Director of Technology & Projects at
other industries, such as chemical
are in fact definitely “affected” and the Rolled Alloys.
process, desalination, power, mineral

www.stainless-steel-world.net Stainless Steel World January/February 2022 1

[ Duplex ]
of unacceptable third phases, rather
Precipitate at a triple point
than the minimum toughness to suit
the actual application. We find that the
Charpy impact toughness test had a
rather low pass/fail criterion2, which
Precipitate at the
would not necessarily reject material
austenite / ferrite
with low levels of intermetallic particles
boundary growing into
in the microstructure. Indeed, when the
the ferrite phase
standard was developed, Davidson3
showed that 2205 plates meeting the
54J at minus 40°C acceptance level
had already suffered a considerable
loss in toughness and also some
loss in corrosion resistance. The
54J acceptance criteria was applied
because it was “regularly used for a
wide range of process applications
other than cryogenic applications”3,
and not because it was found to be Figure 1. Showing typical locations of precipitates in a duplex stainless steel wrought
related to some significant presence microstructure.
of intermetallic precipitates (IMP) in
the microstructure. Our own work for 40°C test temperature. In saying this, it the TWI method has a maximum weight
superduplex stainless steels2 showed is our experience that these grades are loss of 4g/m2. The lower weight loss
that an acceptance level of 70J at minus always deployed in arduous applications, in ASTM A923 is justified in terms of it
46°C was rather more discerning with so we find the concept of a commodity being “discernable, measurable and
respect to the presence of intermetallic application difficult to rationalise. not reliant on the subjectivity of visual
phases. The problem is that low levels determination of pitting”3, but not
of acceptance such as in ASTM A923 ASTM A923 and welds in terms of presence of IMP’s. When
Method B and other specifications, can In addition, ASTM A923 does not using ASTM A923 it is possible that all
allow materials that are predisposed address welds well; in particular, how the as-manufactured faces have been
to further precipitation of IMP to be to take samples, and this impacts on removed by grinding and only the
fabricated by welding. In such cases suitable pass/fail criteria. ASTM A923 bulk metal is being tested. Whereas,
they could suffer further and rapid IMP allows testing of both machined flat, when as-manufactured, or as-welded,
precipitation in the low temperature rectangular samples cut from welds, surfaces are tested higher weight loss
HAZ4 and then be deployed in service. with polishing of all faces and also limits are appropriate6. It has been
In such cases, if pre-existing rather testing of samples with surfaces in the found that for weight losses exceeding
than new weld procedure qualifications “as fabricated” condition and only cut 4g/m2 the rate of weight loss increases
(performed on the susceptible material) edges polished. Both NORSOK5 and rapidly, indicating stable pitting. As
are used it is unlikely that the problem TWI/ International Institute of Welding such a weight loss of 4g/m2 and lower is
will be discovered until it is too late. (IIW)6 only require a cut out of the weld considered acceptable. Further, the TWI
Finally, the corrosion test in ferric and polishing of cut faces, which means method was developed on the basis of
chloride solution, at 40°C was judged that both the as-manufactured and evaluating round robin testing of welds
by some to be at too low a temperature as-welded surfaces including the weld in a number of different laboratories.
for wrought and cast superduplex root run are tested. These areas are As far as the authors are aware, no such
in the solution treated and water what a user wants to know about. The testing has been done to justify the use
quenched condition, such that it could NORSOK requirement for a brief pickle of ASTM A923 Method C test for welds.
also pass material that might contain prior to testing is supposed to remove Indeed, users have reported problems
low levels of intermetallic phase. This test to test variability because it gave a when trying to apply A923 corrosion test
was rectified to some extent by the distinct transition between the passive requirements as part of weld procedure
inclusion of Supplementary Requirement state and active pitting7, but this might qualification8. For this reason, the
S1 which gave the option of corrosion also remove some surface defects like authors prefer the TWI method and its
testing at 50°C. Again our own work poor pickling or areas of nitrogen loss in 4g/m2 weight loss limit.
on superduplex in the solution treated the root of welds where corrosion could ASTM A923 and its acceptance levels
and water quenched condition shows initiate2. For this reason, the authors are based more on what a manufacturer
a test temperature of 50°C to be much prefer to test the mill finished and as would wish to supply rather than what
more discerning in terms of detection welded and cleaned condition (without an end user may need. When criticised
of intermetallic particles than testing at pickling of the sample, unless pickling because the test methods failed to
40°C2. However, other contributors to is going to be applied post welding). detect problems associated with
the standard argued that the 50°C test These two test methods have different nitrides, ineffective pickling of parts, or
temperature should be discretionary, weight loss acceptance criteria too. nitrogen loss from the root runs of welds
applied only to the “arduous” ASTM 923 allows no more than 10mg/ say, the custodians of the standard fall
applications rather than the “commodity” decimeter2/ day (10mdd), which back on to the scope of the standard,
application that would be covered by the equates to 1g/m2 in a 24-hour test while arguing that it was dedicated only to the

2 Stainless Steel World January/February 2022 www.stainless-steel-world.net

 [ Duplex ]
detection of intermetallic phases that
Primary austenite
cause significant loss in toughness or
corrosion resistance, recognising that
the test methods will not necessarily Reformed austenite
detect loss of toughness or corrosion
resistance attributable to other causes.
The standard has been modified over
the years, but changes were slow to
come about and did not always meet
oil and gas companies’ requirements.
Hence, the oil industry users sat down
to write a more robust standard based Intermetallic
on their own requirements, under particles in dendrite
the auspices of ISO. This became ISO arm spacings
177819, first issued in 2017.

What’s in the ISO standard

The ISO standard addresses the quality
of all grades of DSS, lean duplex,
Figure 2. Showing typical location of precipitates in the root run of a duplex stainless
standard duplex, superduplex and
steel weld.
hyper duplex, as well as welds of these
alloys. It covers all major production
routes, including wrought, cast and The occasional third phase particle and superduplex the test temperature
HIP. The document describes in detail is not necessarily a fail, provided the is -46°C. The pass/fail criteria are
how test samples should be taken, material passes the other two tests. In specified for different product forms
particularly for thicker section products, these circumstances it is suggested of the different grades, and some of
so that the tests represent the thickest that the microstructure of the corrosion these are split into two, with higher
material. The standard requires three and impact test samples be checked energies required for more demanding
different tests. to ensure similitude between these applications. As-welded joints have their
The first is a microsection and the and the original microstructure check own special requirements.
standard says where and how it should be sample. The microsection is also used to The third test is an ASTM G48 type
sectioned, and how it should be polished determine the phase balance, and the corrosion test for higher alloys and an
and etched. The most common etch is ferrite content must be in the range 35 to ASTM 1084 type for lean duplex alloys.
a two stage, in 10% oxalic and 20% to 60% for parent metal and 30 to 70% for The test temperature is specified for each
40% NaOH or KOH. These etches will welds in the as-welded condition. grade and the pass/fail criteria include
show nitrides, sigma and chi phases. This The second test is a Charpy impact a maximum weight loss, in addition
double etch must be specified, as a single toughness test, with the test temperature to having no visible pitting at x 20
NaOH etch is also an option. We find the specified for different duplex grades. magnification. This is because nitrides or
double etch good for 22% and 25% Cr The standard is also particular about ineffective acid pickling of the material
duplex grades. Outokumpu argue that the where the samples are taken and can cause a high weight loss without
oxalic acid etch encourages transpassive their orientation. This is because showing any pitting. The standard also
attack, which exaggerates the apparent the toughness varies significantly includes tests for as-welded welds with
size of precipitates and causes “ditching” depending on whether it is orientated specific requirements on sample location
of grain boundaries, so they do not in a longitudinal or transverse direction and preparation, as well as different
recommend it for etching of 2507. They with respect to the grain structure of temperatures and pass/fail criteria for
recommend etch V2A, (50ml hydrochloric the steel. For parent standard duplex each grade. Table 1 summarises the test
acid, 5ml nitric acid and 50ml water)10.
The microsection is examined firstly at Table 1. Test requirements for some duplex stainless steel welds (taken from ISO 17781).
low magnification, scanning the whole
TEST
area of the sample. Any areas that are ALLOY
then thought to be possibly affected are TOUGHNESS CORROSION MICROSECTION
then examined at high magnification Test Charpy impact test at 20°C ASTM A1084 for 24h +
Polished and double etched*
to confirm that they are not etching Lean Duplex
45J Min average; No pitting at ×20; Ferrite 30 to 70%;
artefacts but are indeed third phases. Pass
35J single minimum Wt. loss < 4g/m2 No significant third phases
This is usually done by considering the
Test Charpy impact test at –46°C ASTM G48 at 22°C for 24h Polished and double etched*
location of the particles. Intermetallics
2205
tend to precipitate at the austenite/ferrite 50J Min average; No pitting at ×20; Ferrite 30 to 70%;
Pass
40J single minimum Wt. loss < 4g/m2 No significant third phases
grain boundaries and grow in to, and
consume the ferrite phase. Locations like Test Charpy impact test at –46°C ASTM G48 at 35°C for 24h Polished and double etched*
grain triple points and interdendrictic Superduplex 50J Min average; No pitting at ×20; Ferrite 30 to 70%;
Pass
spacings in castings or welds are prime 40J single minimum Wt. loss < 4g/m2 No significant third phases
locations for precipitation. Figures 1 and * Double etching = electrolytic etching first in oxalic acid and then in NaOH.
2 illustrate this. + No temperature specified

www.stainless-steel-world.net Stainless Steel World January/February 2022 3

[ Duplex ]
requirements for some duplex welds.
References
The tests are the same for parent metal,
1) ASTM A923, “Standard Test Methods for Determining Deleterious Intermetallic Phase in Duplex
but the test conditions and pass/fail Austenitic/Ferritic Stainless Steels” (West Conshohocken, PA, USA: ASTM International).
criteria are higher, and they vary with 2) G Byrne et al “Meaningful Testing for the Quality of Super Duplex Stainless Steels” Paper 10876,
the product form. Proc. Conf. Corrosion 2018, Phoenix, AZ, USA. May 15th to 19th 2018, (NACE International Houston,
TX, USA).
3) R M Davidson and J Redmond, “Development of Qualification Tests for Duplex Stainless Steel Mill
The importance of the tests
Products”. Corrosion ‘91. Paper 302, Cincinnati, OH, USA, (NACE International, Houston, TX, USA,
The tests in ISO 17781 were designed 1991)
to prevent sub-standard material being 4) RN Gunn, “Intermetallic Formation in Super Duplex Stainless Steel Heat Affected Zones” Proc. Conf
supplied to the oil and gas industry. Stainless Steel World ’97. Paper D97-029, Page 335. Maastricht, the Netherlands. October 1997.
However, there have been failures of 5) NORSOK M-630, “Material Data Sheets for Piping,” Edition 5 (Oslo, Norway: Standards Norway,
DSS due to poor quality in many other 2010).
6) P Woollin, “Ferric Chloride Testing for Weld Procedure Qualification of Duplex Stainless Steel
industries. The authors have seen failures
Weldments” Proc. Conf. UK Corrosion and Eurocorr ’94. Bournemouth, UK, 31st October to
in the desalination, mineral processing 3rd November 1994. (ICorr, Northampton, UK).
and chemical industries. The document 7) T Mathiesen and A Andersen, “Challenges in Prequalification Corrosion Testing of CRA’s Based on
is quite demanding in its testing G48A” Corrosion 2014. Paper 4272 San Antonio, TX, USA ( NACE International, Houston, TX, USA)
requirements and pass/fail criteria, but 2014.
8) R Colwell and J Grocki “The Validity of Using ASTM A923 Test Method C Corrosion Test for Weld
this is based on the experiences of the
Procedure Qualification of 25% Chrome Duplex Stainless Steel” Corrosion 2017. Paper 8838,
oil and gas industry and what it takes New Orleans, Louisiana, USA March 2017.
to be sure that the alloy will not fail 9) ISO 17781, “Petroleum, Petrochemical and Natural Gas Industries—Test Methods for Quality Control
prematurely. Failures due to poor quality of Microstructure of Ferritic/Austenitic (Duplex) Stainless Steels,” June, 2017, (Geneva, Switzerland:
microstructures have cost tens of millions International Standards Organization).
of dollars to rectify in some instances11-13, 10) J Y Jonsson, Outokumpu, Private communication.
11) EEMUA Publication 218, “Quality Requirements for the Manufacture and Supply of Duplex Stainless
so a little extra money spent up front on
Steels” (London, United Kingdom: Engineering Equipment and Materials Users Association, 2010).
quality is felt to be justified. 12) E. Ryengen and C. Wintermark, “Lessons Learned from Heat Treatment of Components in 22Cr and
Because of the experiences that drove 25Cr Duplex Stainless Steel (and other materials),” Duplex Stainless Steels. Beaune, France. 13th to
this standard to be created, there is 15th October 2010. Paper III.C.2. page 961 to 970. Zutphen, Netherlands: KCI Publishing.
no reason for it not to be adopted by 13) R. Howard, J. Marlow and S. Patterson “Improving the Quality of Duplex Stainless Steel Components”.
Proc. Conf. Duplex Stainless Steel. Beaune, France. 13th to 15th October 2010. Paper III.C.1. page 953
other industries when purchasing, or
to 960. Zutphen, Netherlands: KCI Publishing.
fabricating, duplex stainless steels.

4 Stainless Steel World January/February 2022 www.stainless-steel-world.net