Deteeting Transformation

of Delta-Ferrite to Sigma-Phase
in Stainless Steels
by Advaneed Metallographie Teehniques
R. J. Gray, V. K. Sikka, and R. T. King
Oak Ridge National Laboratory'
Oak Ridge, Tenn.

Magnetie etehing is a useful metallographie tool for

distinguishing between delta.ferrite, sigma"phase, and
austenite in stainless steels. The teehnique is limited
only by the resolution of the optieal mieroseope.

C containing
ertain mechanical properties of ferrite-
austenitic stainless steel base
sigma-phase.4 It has been reported5 that the
transformation of delta-ferrite to sigma in some
metal and weldments are usually adversely af- austenitic steels is related to the ferrite remov-
fected by prolonged exposure to temperatures ing chromium from the austenitic matrix dur-
in the 480-900C (900-1650F) range. One cause ing aging in the 450-900C (900-1650F) range,
of the property alteration is related to the trans- thus rendering the chromium-enriched ferrite
formation of relatively ductile delta-ferrite to susceptible to sigma formation. Frequently, the
less-ductile sigma-phase. composition of stainless steel castings and
This investigation is concerned with the mi- welds is adjusted intentionally in cast austeni-
crostructural responses of the ferromagnetic tic heat-resistant steels and in stainless steel
delta-ferrite and the paramagnetic sigma-phase welds to give rise to the formation of delta-
to a ferromagnetic iron colloid in a magnetic ferrite to avoid hot-tearing difficulties. The sub-
field. Such positive or negative responses ofthe sequent transformation of delta-ferrite to
two phases to the colloid offer a more definitive sigma-phase is of importance in commercial
identification. With this technique, the iden- applications because it usually adversely af-
tification of small amounts of these phases in fects the mechanical properties. 6 ,7
the microstructure is limited only by the high- A present concern is that the prominent con-
es~ magnification and resolution of the optical
struction materials considered in the design of
mlCroscope. liquid-metal fast-breeder reactor systems are
Sigma-phase was first detected in Fe-Cr-Ni types 304 and 316 stainless steel,'~-IO and pro-
alloys and reported in 1927.1 This component jected operating temperatures for much of the
has been identified as a hard, brittle, non- primary and secondary heat-transport systems
magnetic, intermediate phase with a tetragonal are 480C (900F) to -560C (1040F). While type
crystal structure, containing 30 atoms per unit 316 stainless steel and its associated type 316
cell, space group P42/mnm, occurring in many and 16-8-2 (16% Cr, 8% Ni, 2% Mo) weId metals
binary and ternary alloys of the transition ele- are generally preferred for use above 425C
ments.2 Although Fe-Cr sigma-phase is the (800F), some type 304 stainless-steel compo-
common form in steels, some 53 examples of nents welded with type 308 filler metal will
binary sigma-phase in various metallic systems probably operate above 480C (900F). Further,
have been reported. 3 the formation of sigma-phase is a generic prob-
Many ferritic and austenitic stainless steels lem, and many observations made on types 304
are capable of forming sigma-phase after pro- and 308 are relevant to type 316 and its weId
longed heat treatment, in the 450-900C (900- metal. It is therefore appropriate to examine
1650F) range. The presence of small amounts of types 304 and 308 stainless steels which have
delta-ferrite in an austenitic stainless steel has ?een thermally exposed to form sigma-phase,
been shown to accelerate the formation of In order to attempt to understand more about
the detailed mechanisms of the sigma-phase
* Operated by Union Carbide Corporation under contract W-7405-ENG-26
transformation and how it might affect impor-
with the U.S. Dept. of Energy. tant properties.
18-J. METALS, NOVEMBER 1978
 Optical metallographie techniques for the with a chemical staining technique so that di-
identification of delta-ferrite and sigma-phase reet labeling ofthe two phases can be made. In
have been confined primarily to chemical stain- addition, the heretofore unreported identifica-
ing etchants in the past. The delta-ferriteand tion of a third phase, austenite, in the delta-
sigma-phases are stained different colors but ferrite to sigma transformation is demon-
definition of the phases is not positive. Com- strated.
plementary application of another identifica- Two separate groups of specimens were sur-
tion technique would be very helpful to relate veyed in this study. These groups ofspecimens
specific colors to the delta-ferrite and sigma. consisted of one group of type 308 stainless
This report describes the use of another tech- steel weId metal and another group oftype 304
nique which offers a different approach to iden- stainless steel plate specimens. The results are
tification. The technique is used in conjunction reported separately.

METALLOGRAPHIC TECHNIQUES
The applieation of three analytieal metallographie tech-
niques on two separate groups of types 308 and 304
stainless-steel specimens is presented and correlated where
possible. These techniques are color metallography,
magnetic etching, and microprobe analysis.

Color Metallography
The identification of microstructural components
through the use of color staining 'has been reported previ-
ously.1112 Such techniques are often applicable to indicate
qualitative chemical changes in the microstructure.
In this investigation, compositional changes in the mi-
crostructure due to the sigma transformation were detected
in color through the use of an alkali ne potassium ~COVER SLIP
ferricyanide etchant at 90-95C (195-205F) (modified
Murakami). The composition of the etchant is as follows:
Figure 1. Application of Ferrofluid. Less than V4 drop (-5
30 g KOH, 30 g KaFe(CN)6, 150 ml H20 lambda [Al) of the iron colioid is applied to the specimen. A
No. 1 glass cover slip is placed on the surface and care-
The specimen was immersed in the etchant for -15 sec. The fully pressed down to obtain minimum fluid thickness.
matrix austenite and the ferrite phase are stained light and
medium tan colors, respectively, and the sigma-phase is
stained a reddish-brown color.

Magnetic Etching
Microscopic magnetic patterns on ferromagnetic phases
by means of a suspension of collodial magnetite have been
observed for many years. The technique was first applied in
a comparatively coarse manner in 1932 by Bittef 3 and inde-
pendently by Hamos and Thiessen.14 In the interim period,
reports1>-22 on various specific applications of magnetic
etching have demonstrated the procedure to be a valuable
laboratory tool.
Magnetic etching consists of placing a thin film of iron
colloid (sub-mieron iron-oxide particles suspended in a
fluid) on a suitably prepared specimen surface and observ-
ing attraction of the partieIes to phases in the presence or
absence of an externally applied magnetic field. Since
delta-ferrite is ferromagnetic and sigma is paramagnetie, Figure 2. Large coil magnet 'or magnetic etching. The
attraction ofthe colloid particles to delta-ferrite is very pro- aluminum spool (Va in. wall, 4Y2 in. ID, 2'l'a in. high) has 800
nounced and easily observed by optical metallography by turns of #18 AWG glass-covered copper wire. Power is
switching the external field on and off. Some residual supplied by a rectifier controlled by an AC Variac. Normal
magnetic response of the ferromagnetic delta-ferrite to the operation is at 4 A DC, 28 V. The contro! box contains the
ferromagnetic colloid particles in the absence ofthe magne- rectifier, choke, on-off switch, and a polarity switch. Flux lines
tic field is readily apparent. Previous work on etching of are normal to specimen surface, as shown by iron filings
ferromagnetic delta-ferrite to show domain patterns has display.
been reported.23 Extended applications of this technique
have included the revelation of strain-induced martensite SPECIMEN
UNT
in austeni tic stainless steels. 24 .25 In this report, the
maximum resolution of the optical metallograph is re-
quired to observe the transformation of delta-ferrite to
sigma-phase because the microstructural changes are rela-
tively minute.
The method of applying the colloid to an electropolished
specimen is shown in Figure 1. A magnetic coil is po-
sitioned on the metallograph stage as shown in Figure 2 OlL
with the specimen inverted on the stage in the center of the LENS
coil. A sketch showing the specimen ready for examination
is shown in Figure 3. The magnetic field is activated by a
double-pole/double-throw switch which controls the polar-
ity of the direct current to the coil. Figure 3. Schematic drawing of magnetic etching.

NOVEMBER 1978, J. METALS--19

 V
Ja
:
28 i /
26
I
I ./ // / V
"'-.. ! V
~/
1
<:
, ~.'
~ 24 Ausltnilt , '1' ./

.~! ! ! I ..~/v
L
.
It)
022

o
..........
1
'
I y I yf /~ fit
! Vy /~ f I,\~/
20
N

g
... /8 ~ ""'- i'., .--~______
, V ef

i'-..
,/
r-~~
..
;;;
/6 A+M
~~-~~'~v ~./ ..--
N
"
14 1-
r-
i,
I
""" ~
1

! ~
:
i~ A'''A A +F
,/Kl)4' A YI ~ ----
! / V .1 f"'~ I

"-/ "<,,(/
--
/ / y > < / ..-- ~ ~."il. f.--":"
1\
Morltns,tt
+ ~:; ~ p;;;-
\ /
/
~_. __ !~f-- Figure 4. Location of the
composition of the type
F\ V M+F ,
__..--- I Fturift

--+---r
,
+ .. 308 stalnless steel weid
Q
024
M
1\ /
10 /2 /4 /6 /8 20 22
I
24 26 28 Ja J2 J4 J6 J8 40
metaion the Schaeffler
diagram. Note the ferrite
Chromium E:quivolent. %er + %Mo + /.5. %Si + Q5 % Cb range of -3-18%.

Mleroprobe Analysis Note the sharp, angular outline ofthe phase. Magnetie eteh-
The electron beam mieroprobe as a praetieal analytical ing of the as-welded speeimen is shown in Figure 6. The
tool was developed by Castaing and Guinier in 1949.28 It is magnetic attraetion of the eolloid to the ferromagnetie
based on physieal prineiples diseovered by Moseley in 1913 delta-ferrite is very evident.
and developed over the yearS by Coster and Von Hevesey in
1923, Hamos in 1938, and Hillier in 1943.27. 30
The electron beam ean be seanned like a television raster
or held stationary on a given point, while the sam pIe ean be
moved independently. The seanning beam is used to pro-
duee eomposition maps on a eathode-ray tube. The station-
ary beam is used to aRalyze specifie areas such as a seeond
phase or interface; e.g., diffusion eouple.

SIGMA TRANSFORMATION FROM DELTA-FERRITE IN A TYPE

308 STAINLESS STEEL WELD
Specimen Information
The eomposition of the type 308 stainless-steel weId
metal (wt%) was 0.057 C, 1.40 Mn, 0.027S, 0.83 Si, 20.6 Cr, 10.1
Ni, 0.29 Mo, 0.22 Cu, 0.058 N. This eomposition when plotted
on the Sehaeffler diagram,31 as shown in Figure 4, gives an
expeeted ferrite eontent for the weId as -3-18%.
Four type 308 stainless steel welds with varying histories
were examined. The post-weId treatments were as folIows:
Flgure 5. Microstructure of type 308 stainless steel weid, as
SpeciJmen deposited. Etchant: modified alkaline potassium ferricyanide.
No. History
1 As-welded speeimen. Shielded-metal-are (SMA)
weId made in 1.27 em (0.5 in.) thiek type 304 stain-
less steel base metal. A 75 included angle single-V
joint configuration was used. The eleetrodes were
3.97 mm (5/32 in.) diameter. The average welding
eonditions were 28 V and 130 A.
2 Creep-test speeimen of the same material listed
under (1), 523.9 hr at 6.4% strain vllith a load of 110
MPa (16,000 psi) at 6500C (1200F). The reduetion in
area was 9.9%. The mierostruetural study was
made near the major fraeture.
3 From an SMA test weId which joined 15 x 20 em
seetions of type 304 stainless steel plates 6 em
thiek (6 x 8 in., 2% in thick). The test welds were
made in a double-U joint design with 0.635 em
(0.25 in.) diameter eontrolled-residual-elements
type 308 stainless steel weId metal eleetrodes. The
aging treatment was for 20,000 hr at 480C (900F).
4 Speeimen from the same test material as No. 3; the
agingtreatment was for 20,000 hr at 595C (1100F).
Metallographie Analysis
The microstrueture of the as-reeeived material (Speei- Figure 6. Magnetic etching of as-welded type 308 stainless
men No. 1) is shown in Figure 5. At 1,500X, the eellular steel. a) Note the colloid streaming from the ferromagnetic
dendritie eharaeteristies of the delta-ferrite ean be seen. delta-ferrite. b) Colloid attraction is close to the delta-ferrite.

20--1. METALS, NOVEMBER 1978

 The microstructure of the creep specimen (No. 2) in the seen. Associated with this transformation, some dispropor-
fracture area is shown in Figure 7. A minor fracture area tionation has occurred in the parent delta-ferrite to form
selected for detailed examination is shown in Figure 8 after austenite. Magnetic etching verifies this analysis, as shown
etching with alkali ne potassium ferricyanide. This minor in Figure 11.
fracture, which is oriented 90 to the stress axis, follows a Microprobe analyses were not attempted on these type
specific line of transformed sigma. 308 stainless steel weldments. The delta-ferrite and sub-
Some untransformed, cellular, dendritic delta-ferrite can sequent transformation products were smaller than the ex-
be seen in Figure 8 bot~ to the left and right of the minor citation area from the electron beam, so accurate analyses
fracture. The delta-ferrite in these areas still shows the would be impossible.
characteristic sharp angular outline of the phase pointed
out previously (see Figure 5). The presence of both the
untransformed delta-ferrite areas and the transformed
sigma area in Figure 8 shows that transformation has oc-
curred in very small patches. Furthermore, incomplete
transformation of individual delta-ferrite islands to sigma-
phase can be seen. Evidence ofthis incomplete transforma-
tion is based on the presence of both the medium-tan color
of delta-ferrite as weH as the reddish-brown color within the
same island boundary. The ferrite-to-sigma transformation
process induces a morphological change from sharp angu-
lar dendritic ferrite to a more-rounded sigma. Some of the
sigma-phase particles contain microfissures that are
oriented 90 to the stress axis ofthe specimen. Some ofthe
sigma-phase is split by the microcrack and can be seen on
the two edges of the fracture. Additional information on
this type 308 stainless steel weId has been published.32
Another type 308 stainless steel weId specimen (No. 3)
was examined after thermal aging at 480C (900F) for 20,000
hr to determine the extent and characteristics of delta- Figure 7. Fracture area of type 308 stainless steel weid creep
ferrite to sigma transformation. Only the magnetic etching specimen. Arrow points out location of examination at higher
technique was used to determine the extent of transforma- magnification.
tion. Photomicrographs exhibiting the colloid response to
ferromagnetic areas are shown in Figure 9. The mi-
crostructures are shown after the specimen was elec-
tropolished, so both the ferrite and sigma-phases displaya
clear-white color. The microstructures are shown with the
magnetic field on and off. If the magnetic field is off, re- , '. {2g
sidual magnetic locations in the microstructure cause
"streaming" of the colloid particles from these minute '"
~

areas. By closing the electric current to the coil, the magne-

tic field causes the coHoid to collect in these ferromagnetic ~x[l

,.. ~t'!!~
1, '-

areas. Note the lack ofresponse ofmost ofthe second phase '... '. ji!8 r
in the microstructure to the colloid, indicating nearly com-
plete transformation to sigma-phase. As described earlier I ','\ "~i '.
. .."..'"
for Specimen 2, the sigma-phase has assumed a rounded

~.~y~..7.t'! j'.l~.I'i'.; ~~~j','!! ~

shape. U sing an electrolytic extraction and x-ray diffraction
technique, the extracted particles showed only Fe304.33 The
ferromagnetic areas seen in Figure 9 could be untrans- t -
formed delta-fenite which is undetected by the extraction .}/ "'j., .. . ' -
technique, or minute amounts ofFe304, since both compo- -mc... .,,'~.
. ~
nentsare magnetic. ._ J -,-

The microstructure of a weldment specimen which was

aged 20,000 hr at 595C (1100F) (Specimen No. 4) is shown in Flgure 8. Relationship of delta-ferrite to sigma transforma-
tion and crack propagation during creep test of type 308
Figure 10 after etching with alkali ne potassium stainless steel weid at 650C (1200F). Microstructure Iden-
ferricyanide. It is evident that considerable dissolution of tification: (T = sigma, f = ferrite, a = austenite, mc = mi-
the original delta-ferrite in the austenitic matrix has oc- crocrack, f, (T = incompletely transformed sigma, (T* = sigma
curred. Within the remaining undissolved delta-ferrite is-
lands, more than 50% transformation to sigma-phase can be on both sides of crack, - stress axis.

15 M'CRONS 45
---'---+,'

./.~. ~
---'--1500 X--L,----L,---'-0-.0..,0,2--1
01

.".---
0.0005 INCHES

~ .--'
.{
"~~:i,"i: . . ".tr',, ~

6" ~~
. / 'Y
AUSTENITE
I' ~~~
, ~. ,<u
,/ er

.. I .~
-. "":
~~~--. (
Figure 10. Type 308 stainless steel weid, aged 20,000 hr at Figure 9. Magnetic etching of type 308 stainless steel weid
595C (1100F). Transformed sigma ((T), austenite (y), and re- after aging 20,000 hr at 480C (900F). Almost complete trans-
tained delta-ferrite () are identified. formation of the delta-ferrite to sigma-phase has occurred.

NOVEMBER 1978, J. METALS-21

 Figure 11. Magnetie etehing of type 308 stainless steel weid of cellular dendritic delta-ferrite () to sigma (u) arid austenite
after aging 20,000 hr at 595C (11 OOF). Partial transformation (y) can be seen.

TRANSFORMATION OF DELTAFERRITE STRINGERS TO retained delta-ferrite. Verification of the identification of

SIGMAPHASE IN TYPE 304 STAINLESS STEEL BASE PLATE these phases is continued in Figures 16b and c with magne-
Speelmen Information tic etching. The paramagnetic sigma and austenite do not
respond to the colloid. The retained delta-ferrite is iden-
This study is concentrated on one heat of type 304 stain- tified by the magnetic response ofthe colloid. Domain pat-
less steel plate. The chemical composition (wt %) of this terns formed by the colloid can be seen in the delta-ferrite.
heaP4 (No. 807) was 0.029 C, 1.26 Mn, 0.024P, 0.023 S, 0.50 Si, The sigma-phase has a platelet morphology that is crys-
18.8 Cr, 9.67 Ni, 0.20 Mo, 0.11 Cu, 0.03 Co, 0.02 W, 0.0005 Ta, tallographically oriented according to the host single-
0.01 Pb, 0.01 Sn, 0.012 V, 0.0015 Nb, 0.002 Ti, 0.0021 N, 0.005B, crystal ferrite island. Consequently, the sigma-phase can
0.010 0, 0.0012 H. appear to be a needle or a platelet depending on the plane of
This heat was selected for metallographic study because examination of the specimen. A ferrite island oriented so
it contained ~5% of relatively large delta-ferrite stringers; that the width ofthe platelets can be seen is shown in Figure
the amount of stringers was measured with a quantitative 15. Another ferrite island with the platelets on edge is
measuring microscope (Quantimet B). Group A specimens shown in Figure 17.
were aged for 10,000 hr at 480C (900F), 595C (1100F), and The fractured edge of a tensile SpecimE!ll was examined to
650C (1200F). Group B specimens were aged at the same determine the infiuence of the transformed sigma or frac-
temperatures for 31,000 hr. ture characteristics. A polished plane of a tensile specimen
After the aging treatments, tensile specimens from the aged at 650C (1200F) is shown in Figure 18. The fracture
Group A specimens were tested hot at the same aging tem- appears to follow the original ferrite-austenite. matrix in-
peratures for comparative mechanical property data with terface which has become embrittled by the presence ofthe
unaged specimens. Specimens were cut for metallographic sigma. The presence ofthe sigma-phase in the delta-ferrite
examination. Group B specimens were examined onlyafter appears to playa significant role in the fracture formation
the aging treatments. and propagation within the delta-ferrite island, as shown in
Figure 18. This confinement ofthe sigma within the isolated
Metallographie Analysis stringers of delta-ferrite exercised relatively small
The microstructure of the as-received type 304 stainless infiuence on mechanical properties. The results of mechan-
steel is shown in Figure 12 after etching with alkali ne potas- ical property tests on this heat (No. 807) as weIl as various
sium ferricyanide; the microstructure using the magnetic other heats after aging have been compared and reported.'"
etch is shown in Figure 13. Although pronounced microstructural changes could be
An electron microprobe continuous scan across one of seen after 10,000 hr aging at 480C (900F), 595C (1100F),
these delta-ferrite stringers and the austenite matrix in the and 650C (1200F), extended aging time at these same tem-
as-received material to analyze for iron, nickel, and peratures would be of great value to study. For this reason,
chromium produced the results shown in Figure 14. The this heat of stainless steel was aged for 31,000 hr at these
three temperatures; the microstructures are shown in Fig-

-
delta-ferrite contains slightly less iron, less nickel, and
more chromium than the matrix. Static electron be am ures 19, 20, 21.
counts of the delta-ferrite and matrix show nominal
compositions of these three elements.
Representative microstructures of the Group A tensile
specimens aged at 480C (900F), 595C (1100F), and 650C
(1200F) for 10,000 hr are shown in Figure 15 after etching
with alkali ne potassium ferricyanide reagent. The
reddish-brown sigma-phase is dark in a black-and-white
....
photomicrograph. A light-tan colored phase which tested to
be paramagnetic with the magnetic etch is identified as
austenite. The transformation of delta-ferrite after aging at
595C (1100F) and 650C (1200F) is, in fact, weIl demon-
strated to be a disproportionation: delta-ferrite to sigma +
austenite.
Photomicrographs of some delta-ferrite islands in a
specimen aged at 650C (1200F) for 10,000 hr are shown in
Figure 16. The reddish-brown sigma-phase appears very
dark in the black-and-white photomicrograph after etching Figure 12. Delta-ferrite stringers in asreceived type 304
with alkaline potassium ferricyanide. The transformed aus- stainless steel plate. Etchant: alkaline potassium
tenite within the ferrite island perimeter is lighter than the ferricyanide.

22-J. METALS, NOVEMBER 1978

 ~.III." ~. --, bUiAfthW'. :11.

Figure 13. Magnetic etching of delta-ferrite stringers in type 304 stainless steel plate. a) Colloid applied, magnet off; b) Colloid
applied, magnet on "+") c) Colloid applied, magnet on "-". Note domain pattern changes (b, c) with change in polarity.

~
AUSTENJJE

INCREASEe>
I
AUSTENITE
AUSTENIl

Ni Cr ~ Fe

Figure 14. Electron mi-

* CONTlNUOUS UNE SCAN*t
NORMALIZED
croprobe analysis of
delta-ferrite and austen-
ite in as-received type
AUSTENITE
FERRITE
Fe
63.9
62.9
er
24.4
29.9
~j

9.5
5.5
MICROSTRUCTURE '"
304 stainless steel.

Figure 15. Transformation of delta-ferrite in type 304 stain- a. 480C (900F)

less steel after aging for 10,000 hr at indicated temperatures. b. 595C (11 DooF)
Etched with alkaline potassium ferricyanide. c. 650C (1200F)

NOVEMBER 1978, J. MET ALS--23

 0'
Figure 16. Transformation delta-ferrite to sigma and austenite in type 304 stainless steel after 10,000 hr at 650C (1200F).
a) As electropolished and chemically etched. b) Magnetic etch; electromagnetic coil off. c) Magnetic etch; electromagnetic coil
on.

Figure 17. Transformation of delta-ferrite to sigma-phase in Figure 18. Major and minor fracture characteristics in trans-
type 304 stainless steel after 10,000 hr at 650C (1200F). Note formed delta-ferrite to sigma-phase, type 304 stainless steel
that the sigma-phase is oriented in this ferrite island so the after 10,000 hr at 650C (1200F). Note that the fractu res follow
edges of the platelets are visible. sigma-phase at original delta-ferrite/austenite interface, and
that trans-delta-ferrite island fractures propagate in sigma-
phase.

DISCUSSION into the delta-ferrite from the austenitic matrix during-the

Type 308 Stainless Steel Weldments aging process. In our type 304 specimen, ihe electron mi-
croprobe scan analysis as seen in Figure 14 shows that
Maximum magnifications of 1,000X-l,500X with the op- lower nickel, higher chromium, and slightly lower iron val-
tical metallograph were essential to resolve the fine details ues occur in the delta-ferrite than in the austenitic matrix
of the delta-ferrite and subsequent transformation prod- prior to aging. The aging treatment tl,en results in dispro-
ucts. The role of sigma transformation and crack propaga- portionation of the delta-ferrite to sigma + austenite. As is
tion in a creep specimen after 593.9 hr at 650C (1200F) is apparent from the reduction ofthe amount of delta-ferrite
demonstrated in Figure 8. An "hour-glass" shaped area con- stringers after long aging periods at 59.5C (llOOF) and
taining a network of sigma-phase is shown with a crack 650C (1200F), some delta-ferrite to austenite (matrix) dif-
propagating down the center of the network in the sigma- fusion is occurring.
phase. Some sigma-phase can be seen on the sides of the
After aging at 480C (900F), the delta-ferrite stringers
crack in severallocations. The sigma network is sufficiently show only faint evidence of transformation to sigma. The
compact to allow a critical crack size to form. Although we
show only one ofnumerous minor fractures, the major frac- transformation is quite pronounced after aging at 595C
(1lQOF) and 650C (1200F) (see Figure 15), and a third phase
ture undoubtedly propagated in a similar manner.
Accompanying this transformation ofthe angular, cellular identified as austenite can be seen. Preferential diffusion on
certain crystallographic directions is possibly occurring,
dendritic ferrite to sigma is a morphological change to a
since austenite and sigma appear to form on the same habit
more-rounded shape--another clue that must be remem- plane ofthe host delta-ferrite. In certain ferrite islands, the
bered when determining the transformation of delta-ferrite transformation process has resulted in fine sigma precipi-
to sigma. tates in a matrix of austenite surrounded by untransformed
Almost complete transformation ofthe cellular dendritic
delta-ferrite, as shown in Figure 15c. Other regions of aus-
delta-ferrite to sigmais shown in Specimen 3 (Figure 9) after tenite can be seen in the delta-ferrite in this same photomi-
applying the magnetic-etch technique. Again, the crograph. This matrix austenite was present in the ferrite
morphological change to a rounded profile is very pro- island in the as-received state like a "hole in a doughnut."
nounced.
Specimen 4 shows an intermediate stage of sigma trans- Further evidence of delta-ferrite to sigma and austenite
formation in Figure 10. The retained delta-ferrite, sigma, disproportionation is shown in Figure 16. Magnetic etching
and a third phase, assumed to be austenite, can be seen after was applied after etching with alkali ne potassium
etching with alkaline potassium ferricyanide. ferricyanide. By following this procedure, the ferromagne-
tic delta-ferrite and paramagnetic sigma and austenite
Type 304 Stainless Steel Plate within the original ferrite boundary could be identified.
A reference 5 cited earlier postulated that the transforma- This retained delta-ferrite displays some domain patterns.
tion of delta-ferrite to sigma was due to chromium diffusing The need to apply only a minimal amount of colloid is

24--J. METALS, NOVEMBER 1978

 most evident in Figure 16. If the collold layer is very thin,
magnetic microstructures may be clearly defined without
energizing the magnetic coil, as shown in Figure 16b. The
magnetic response can be related to the staining charac-
teristics with the alkaline potassium ferricyanide. Once
these identifications are made, other specimens can be
examined using only the alkali ne potassium ferricyanide
and identification can be made with confident accuracy.
In Figures 15 and 16, ferrite islands were chosen that
show the sigma having relatively large width. Ferrite is-
lands could be found that contained areas of sigma-phase
that were much narrower, as shown in Figure 17. We can
assume that the sigma~phase developed in the form of
platelets that were oriented on specific planes of the delta-
ferrite crystallographic lattice. Viewing the edge or the flat
of the platelets would depend on the orientation of the
delta-ferrite and the plane of polish.
The fracture edge of the ht-tensile specimen which was
aged 10,000 hr at 650C (1200F) showed that the fracture
propagated along the ferrite-matrix interface (Figure 18).
Figure 19. Transformation of delta-ferrite to sigma in type Some sigma-phase fragments could be found along the
304 stainless steel after 31,000 hr at 480C (900F). Note the fracture edge. Although the sigma-phase played a leading
fine precipitate of sigma in the delta-ferrite stringers. role in the crack propagation within the isolated ferrite
stringers, as seen in Figure 18, the overall negative
influences ofthe sigma-phase on the mechanical properties
of this heat of type 304 stainless steel was not very pro-
nounced.
The extended aging of this type 304 stainless steel for
31,000 hr produced definite disproportionation of the
delta-ferrite to sigma plus austenite (B ~ CF + 1'). At 480C
(900F), the phases produced are extremely fine. At the
higher temperatures of 595C (1100F) and 650C (1200F), as
one might expect, the resulting sigma platelets and the
austenite are larger, with less delta-ferrite. The size of the
sigma platelets after the 31,000 hr treatment at 595C
(llOOF) appears to be very similar to the platelet size and
amount observed after 10,000 hr at 650C (1200F).
In addition to the sigma-producing transformation of the
delta-ferrite, sigma-phase also appears to be produced as a
precipitation product at matrix grain boundaries after
aging 31,000 hr at both 595C (llOOF) and 650C (1200F)-
more being evident at the higher temperature. In addition,
the longer time at the highest temperature has caused some
precipitation of sigma that follows a Widmanstatten pattern
in the austenitic matrix.

Flgure 20. Transformation of delta-ferrite to sigma and aus-

tenite in type 304 stalnless steel after 31,000 hr at 595C CONCLUSIONS
(1100F). Note the sigma, retained delta-ferrite, and austenite Three metallographie techniques-color etching, magne-
in the original delta-ferrite stringers. So me grain-boundary tic etching, and microprobe analysis-have been applied in
precipitate in the austenitic matrix can be seen. a study of the transformation of delta-ferrite to sigma and
austenite phases. Although staining techniques have been
used in the past to great advantage in showing delta-ferrite
and sigma-phases in color, unequivocal identification of
these phases has not been possible; however, with the addi-
tion of the magnetic etching technique, direction relation-
ship of colors and phases can be made.
Associated with the transformation of delta-ferrite to
sigma is evidence ofan additional disproportionation ofthe
parent phase to austenite (B ~ CF + 1').
The alteration ofthe small 2p.m) delta-ferrite islands in
the welded specimens after aging for 10,000 hr at 595C
(llOOF) and 650C (1200F) could be shown. Although this
evidence taxed the limitations ofthe optical metallograph,
color changes due to stain etching can be developed that
show some delta-ferrite islands only partly transformed to
sigma.
The more massive delta-ferrite stringers in the type 304
stainless steel plate allowed transformations to be resolved
more easily. The 10,000 hr aging treatment at 480C (900F)
produced a new precipitate within the delta-ferrite near the
matrix boundary that only suggested a transformation to
sigma. After 31,000 hr, however, entire delta-ferrite islands
displayed a very fine precipitate that could just be resolved
Figure 21. Transformation of delta-ferrite to sigma and aus- at 1,500X. We can presume that this precipitate is sigma.
tenite in type 304 stainless steel after 31,000 hr at 650C The higher-temperature aging treatments of 595C
(1200F). Note the sigma, retained delta-ferrite, and austenite (1100F) and 650C (1200F) allowed color etching technique
in the original delta-ferrite stringers. Some grain-boundary to be used to show sigma, retained delta-ferrite, and austen-
precipitate and Widmanstatten precipitate in the austenitic ite with relative ease. Transformation characteristics ofthis
matrix can be seen. type 304 stainless steel plate after 31,000 hr at 595C (llOOF)

NOVEMBER 1978, J. METALS----25

 appear to be very similar to the microstructure after 10,000 28 D. Coster, and G. von Hevesy, Naturwissenschaften, 11 (1933) p. 133
L. Hamos, J. Sei. Inst., 15 (1938) p. 87
hr at 650C (1200F). The extended aging of31,OOO hr at 650C 29

30 J.Hillier, U.S. Pa\. No. 2,418,029, 1947

(1200F) produced a greater amount of grain boundary pre- 31 Anton L. Schaeffler, "Constitution Diagram for Stainless Steel Weid Metai,"
cipitate as weH as a Widmanstatten precipitate within the Meta) Progress (Nov. 1949) pp. 680-683
matrix grains. We can only presume at this time that these 32 R. G. Berggen, N. C. Cole, G. M. Goodwin, J. O. Stiegler, G. M. Slaughter, R. J.
Gray, and R. T. King, "Structure and Elevated Temperature Properties ofType
grain-boundary and intragranular precipitates are sigma. 308 Stainless Steel," Report ORNL-TM-5145, Oak Ridge National Laboratory,
Oak Ridge, Tenn.
ACKNOWLEDGEMENTS 33 James M. Leitnaker, Ronald L. Klueh, and William R. Lang, "The Composi-
tion ofEta Carbide Phase in 211. Cr-l Mo Steel," Metallurgical Transactions, 6A
The authors are grateful for the electron microprobe (October 1975)
analysis performed by R. S. Crouse and the technical 34 V. K. Sikka, C. R. Brinkman, and H. E. McCoy, "Effect ofThermal Aging on

suggestions from J. M. Leitnaker, J. L. Scott, and G. M. Tensile and Creep Properties ofTypes 304 and 316 Stainless Steel," pp. 316-350
in Symposium on Structural Materials for Elevated Temperature Service in
Adamson. Felicia M. Foust played a valuable role through Nuclear Power Generation Service ed. by A. O. Schaefer, American Society of
her suggestions in the preparation of the manuscript. Mec;hanical Engineers, New York, NY, 1975

References
1E. C. Bain and W. E. Griffiths, AIMME, 75 (1927) p. 166
2 "Definitions Relating to Metals and Metal Working," in Metals Handbook, ABOUT THE AUTHORS
Vol. I, 8th ed., p. 34: Ameriean Society for Metals, Metals Park. Oh.,
3 E. O. Hall and S. H. Algie, Met. Review, 11, (1966) p. 61
R. J. Gray has been Supervisor, Metallog-
F. R. Beckitt, "The Formation of Sigma-Phase from Delta-Ferrite in a Stain-
less Steel," J. of Iron and Steel Institute, 207 (1969) pp. 632-638 raphy Laboratories, Oak Ridge National
"Wrought Stainless Steels," in Metals Handbook, 8th ed., Metals Park, Ohio, Laboratory, for 30 years. He is president of
pp. 419-422: American Society for Metals, Metals Park, Oh., the International Metallographie Soeiety
6 L. R. Poole, "Sigma Phase-An Unwanted Constituent in Stainless Weid
Metai," Metal Progress, 65 (6) (1954) pp. 108-112 and is a Fellow of Ameriean Soeiety for
7 J. O. Stiegler, G. M. Goodwin, and R. T. King, "Effect ofResidual Elements on Metals. He has been a strang proponent of
Fracture Characteristics and Creep Ductility ofType 308 Stainless Weid Metai," the use of color metallography and
J. Eng. Mater. Techno!., 97, (3) (1975) pp. 245-250
Donald F. Baxter, Jr., "A Look at the Chromium Situation," Metal Progress,
magnetie etehing as diagnostie tools in
109 (6) (1976) pp. 24-31 metallography. Mr. Gray has reeeived
9 John Graham, "Nuc1ear Safety Design of the Clinch River Breeder Reactor numerous metallographie awards, inelud-
Plant," in Nud. Safety 16 (5) (1975) ing the Franeis F. Lueas Award in 1953,
10 V. K. Sikka, H. E. McCoy, Jr., M. K. Booker, and C. R. Brinkman, "Heat-to-
Heat Variation in Creep Properties ofTypes 304 and 316 Stainless Steels," J. of 1958,1964, and 1971. He is the author of more than 85 teehni-
Pressure Vessel Technology (November 1975), pp. 243-251 eal papers on the development and applieation of
11 R. J. Gray, "Modern Metallographie Techniques," in Fifty Years of Progress metallographie teehniques and investigations. He reeeived
in MetaUographic Technique, STP 430, ASTM, Philadelphia, Pa
12 R. S. Crouse, B. C. Leslie, and R. J. Gray, "Applications ofColor in Metallog-
his BA degree in bioehemistry from Sterling College, Sterling,
raphy and Photography," in Interpretive Techniques in Microstructural Kansas in 1941.
Analysis, ed. by JamesL. McCall and E. J. Myers, Plenum Press, New York, NY
13 F. Bitter, "On Inhomogeneities in the Magnetization ofFerromagnetic Mate-
rials," Phys. Rev., 38 (1931) p. 1903 R. T. King has been a staff member at Oak
,. L. von Hamos, and P. A. Thiessen, "Uber die Sichtbarmachung von Bzirken Ridge National Laboratory sinee 1965. He
verschiedeman ferromagnetischen Zustandes festen Korpen," Z. Phys., 71 reeeived his BS, MS, and PhD degrees
(1931) p. 442
15 F. Bitter, "Experiments on the Nature of Ferromagnetism," Phys. Rev., 41
from Carnegie-Mellon University in Metal-
(1932) p. 507 lurgy and Materials Seience. He has pub-
16 L. W. McKeehan, and W. C. Elmore, "Surface Magnetization in Ferro Magne- lished more than 50 teehnieal reports and
tic Crystals," Phys. Rev., 46 (1934) p. 226 papers on the meehanieal and physieal
17 H. S. Avery, V. O. Homerbert, and E. Cook, "Metallographie Identification of
Ferro Magnetic Phases," Metals Alloys, Vol. 10 (1935) pp. 353-355 metallurgy of struetural materials for nu-
18 W. C. Elmore, "Ferromagnetic Colloid for Studying Magnetie Phases," Phys. elear reaetor service, and more reeently,
Rev., 32 (1938) pp. 309-310 for fossil energy systems. He manages the
19 E. A. M. Harvey, "Metallographie Identification of Ferro-Magnetic Phases,"
Metallurgia, 32 (June 1945) pp. 71-73 Fossil Energy Materials Program in the
20 P. F. Weinrich, "Microferrographic Technique," Australasian Engr. (N 0- Metals and Ceramies Division at ORNL, and is aetive in a
vember 1948) pp. 42-44 number of eommittees for ASME, the ASTM. and the Metals
21 George Fisinai, "Magnetie Oxide 'Etchant'," Metal Progress 17 (October
Properties Couneil.
1956) pp. 120-122
22 R. Carey, and E. D. Isaac, eds., Magnetic Domains and Techniquesfor Their
Observation, Academic Press, New York, 1966 v. K. Sikka (MS, PhD, Metallurgieal Engi-
23 R. J. Gray, "Revealing Ferromagnetic Microstructures with Ferrofluid," Pro- neering, 1970, 1973, University of Cinein-
ceedings of the International Microstructural Analysis Society, (September
1971) pp. 21-23 nati, Ohio) joined Oak Ridge National Lab-
2. R. J. Gray, "The Detection of Ferromagnetic Phases in Types 304 and 301 oratory in 1974 and has been working on
Stainless Steels by Epitaxial Ferromagnetic Etching," in Microstructural Sei- the elevated-temperature meehanieal
ence, Val. 1 ed. by R. J. Grayand J. L. McCall, American Elsevier Publishing
Co., Inc., New York, NY, 1973 properties of austenitie and ferritie stain-
25 R. J. Gray, "Magnetic Etching with Ferrofluid," in MetaUographic Specimen less steels. His MS and Ph.D. work are re-
Preparation, ed. by J. L. McCall and W. M. Mueller, Plenum Publishing Corp., lated to high-temperature oxidation and
New York, NY
26 R. Castaing, and A. Guinier, Proc. First Internat. Conf. EZectron Microscopy,
irradiation damage of metals and alloys.
Delft, 1949, p. 60 He has published more than 40 teehnieal
27 H. Moseley, Phi!. Mag., 26 (1913) p. 26; ibid 27 (1914) p. 704 reports and papers.

26---,-J. METALS, NOVEMBER 1978