WELDING RESEARCH

SUPPLEMENT TO THE WELDING JOURNAL, OCTOBER 2006

Sponsored by the American Welding Society and the Welding Research Council

Structural Stability of Dissimilar Weld

between Two Cr-Mo-V Steels
Results are presented of an experimental study and modeling of the structural
stability of 6Cr-Mo-V 8-3-2 and X12Cr-Mo-Nb 10-1 steels
in a temperature interval from 600° to 900°C
BY R. FORET, B. ZLAMAL, AND J. SOPOUSEK

ABSTRACT. Heterogeneous weld joints creep-resistant steels with ferritic matrix is different kinds of creep-resistant steel. A
of creep-resistant steels are used with in- the study of the degradation processes oc- redistribution of interstitials (C, N, and H)
creasing frequency in the construction of curring during their operation in power- in dissimilar weld joints of austenite/fer-
new or repair and reconstruction of exist- generating and chemical plants, inclusive rite type and ferrite/ferrite type can be ob-
ing power-generating and chemical indus- of the structural instability of both the served within welding and subsequent
try facilities. These weld joints show struc- steels themselves and their weld joints. PWHT or in the course of the welds ex-
tural instability during postweld heat The term structural instability usually ploitation at elevated temperatures. The
treatment (PWHT) and in the course of refers to the successive microstructural interstitials redistribution can cause the
subsequent operations. Carbon redistrib- changes toward equilibrium state, when formation of carbon-depleted zones
ution and subsequent sequence of changes the Gibbs energy is minimal, and with (CDZ) in low-alloy steels, and on the con-
in the structure lead to the formation of a creep-resistant steels, the metastable mi- trary, the formation of carbon-enriched
carbon-depleted zone (CDZ) in low-alloy nority phases dissolve and the stable ones zones (CEZ) in high-alloy steels, both of
steel. On the contrary, a carbon-enriched precipitate, and the density of dislocations them in a close vicinity to the fusion zone.
zone (CEZ) appears in high-alloy steel, and the size of grain boundary area (phase The appearance of the above zones is de-
with the differences in chromium content interface) decrease. An example of the ef- termined by the type of steels being
having the greatest significance. fect of structural instability on the creep welded, by the conditions of PWHT, and
This paper presents the results of a properties of some modified 9–12% Cr by the exploitation itself. The structure of
study of structural changes in laboratory steels is their sigmoidal behavior (Ref. 1). the carbon-depleted zone is usually
welds of 6Cr-Mo-V 8-3-2 (T25) and In view of the exhausted service life of formed by ferritic grains without any ap-
X12Cr-Mo-V-Nb 10-1 (P91) steels an- a number of power-generating blocks (2 × parent carbide precipitate (Refs. 4–6), in
nealed at temperatures from 600° to 900°C 105 and more hours of operation) and with which the localization of plastic deforma-
(1112° to 1652°F). Carbon redistribution regard to the growing power demand, in- tion may occur and the state of triaxial
measurements by the EPMA method creased capital investment is expected in stress appears.
were complemented with detailed struc- the area of constructing conventional In Ref. 7 it is stated that three principal
tural analyses aimed at the phase and power plants. It can be expected that in the cracking mechanisms affecting Cr-Mo-V
chemical compositions of coexisting car- construction of new power-generating welds are reheat cracking, Type IV crack-
bides and carbonitrides. blocks and in the reconstruction of exist- ing, and Type IIIa cracking. In Ref. 4, a de-
The results of experimental work were ing blocks attention will focus on increas- tailed description is given of the differ-
compared with thermodynamic and ki- ing their efficiency, which will entail con- ences between reheat cracking and Type
netic calculations using the Thermo-Calc tinuous development of creep-resistant IIIa cracking, adding that Type IIIa crack-
and DICTRA software. A very good agree- steels with a ferritic matrix (Refs. 2, 3). ing is due to the above carbon migration.
ment between the calculations and the ex- Both nuclear and conventional power Helander, Andersson, and Oskarsson (Ref.
periments was obtained, in particular for plants are complex systems in which it is 8) observed that the influence of the CDZ
the phase composition of individual areas practically impossible to rule out welding on the creep strength of the weld has been
of the weld joints. the subject of some controversy over the
years. Some authors suggest that it has a sig-
Introduction nificant influence on the mechanical prop-
erties (Ref. 9), while others considered the
An integral part of the development of KEYWORDS influence negligible (Ref. 10). Also ambigu-
ous are the results of the effect of the time
R. FORET ([email protected]) and B. ZLA- Creep-Resistant Steel of operation (creep tests) on the occurrence
MAL are with Brno University of Technology, Weld Joints of Type IIIa failure. The behavior of a weld
Faculty of Mechanical Engineering, Department Structural Analyses of the type of 1%Cr-Mo-V/12%Cr-Mo-V
of Materials Science and Engineering, Brno, Thermodynamic Modeling studied in (Ref. 9) is determined in short-
Czech Republic. J. SOPOUSEK is with Masaryk Kinetic Modeling time creep testing by heterogeneous mi-
University Brno, Faculty of Science, Department
of Theoretical and Physical Chemistry, Brno,
crostructural features. In the case of welds
Czech Republic. of 2.25%Cr-Mo/12%Cr-Mo-W-V steels,

WELDING JOURNAL 211 -s

WELDING RESEARCH
A B

C is also necessary to ex- process of the phase transformation rate

amine the occurrence (Refs. 15, 16), the theory of multicompo-
of carbides or nitrocar- nent bulk diffusion (Ref. 17), and of
bides not only in the course, the gradients of chemical poten-
CDZ but also in the tials/activities of the elements are consid-
carbon-enriched ered for mass flux evaluations.
(hard) zone of high- This paper includes the results of a
alloy steel. Then the study of structural stability in laboratory
mechanical properties weld joints of T25 (steel under develop-
of the weld joint as a ment, described in Ref. 18) and P91 steels.
whole should be evalu- A comparison of the results of experimen-
ated, i.e., with the cor- tal work and the results of theoretical
responding stress con- calculation is presented.
centration and level of
internal stress. The Experimental Procedures
need for an appropri-
ate prediction of the The structural stability of dissimilar weld
redistribution of inter- joints of two types of creep-resistant steel
stitials and subsequent was studied. The chemical composition of
structural changes is the steels examined is given in Table 1.
thus obvious. Cylindrical samples (diameter 12 mm,
The CALPHAD ap- height 4 mm) were machined from both
proach (Ref. 12) can materials. Each of these samples was
Fig. 1 — Comparison of simulated and measured redistribution of carbon
be used for the solu- ground and polished, and pairs were
in the weld joint T25/P91 annealed at different conditions (simulation mea-
tion of both local and welded to form “sandwich-type” couples.
surement by EPMA method). A — 600°C/240 h; B — 700°C/56 h; C —
global phase equilib- The laboratory weld joints were prepared
900°C/18 h — (simulation, ...measument by EPMA method).
rium problems con- via AC heating to a temperature of 800°C
cerning the base ma- (1472°F) within ~20 s, with subsequent
Type IIIa failure was observed only after terial and weld joints. The approach heating to a temperature 1050°C (1922°F)
longer creep stress times (Ref. 8). Likewise, enables the calculation of chemical compo- within ~5 s. Welding was conducted in an
Albert et al. (Ref. 11) reported that the oc- sitions of equilibrated phases at a given atmosphere of argon under a thrust force
currence of CDZ and CEZ in heterogenous temperature as well as any of the phase di- of 200 N. These samples were sealed into
weld joints is of pronounced negative effect agram cross sections. Also important is the evacuated quartz ampoules and annealed
on their toughness at room temperature. As possibility to calculate the chemical poten- as shown in Table 2.
regards the creep characteristics, this effect tials/activities of elements. After annealing, the joints were cut
is ambiguous in the case of type ferrite/fer- The DICTRA program (Ref. 13), which perpendicularly to the weld interface.
rite welds, while in the case of type fer- contains subroutines for the CALPHAD Analyses of microstructure were then car-
rite/austenite welds the degradation of method, is used as a conductive tool to ried out on these surfaces, with Vilella’s
these characteristics predominates. simulate processes in dissimilar welds. reagent and 2% nital reagent being used
The effect of CDZ on mechanical prop- The benefits of using the DICTRA pro- to produce the structure.
erties, inclusive of the creep characteristics, gram are described in Ref. 14. This pro- The carbon motion across the welds
must be considered comprehensively. Be- gram embodies the assumption of local was measured by means of wavelength dis-
sides the carbon depletion itself, assessed by condition of phase equilibrium, the as- persive X-ray analysis (WDX) in a Jeol
the zone width and carbon concentration, it sumption that diffusion is the controlling JXA-8600 electron probe microanalyzer

212 -s OCTOBER 2006

WELDING RESEARCH
A

Fig. 2 — Microstructure of the weld zone, steel T25 is located on upper part
(SEM micrograph). B

(EPMA). A transmission electron micro- occurrence of carbide

scope Philips CM12 working in the scan- phases, it was not possible
ning mode (STEM) equipped with energy to determine reliably the
dispersive X-ray (EDX) Phoenix analyzer, carbon content in the
and selected area electron diffraction CDZ. The measurement
(SAED) were used to determine the char- of CDZ widths is also only
acteristics of microstructure from carbon approximate. The maxi-
extraction replicas. A combination of dif- mum carbon concentra-
fraction data and chemical composition tions in the CEZ (Cmax)
was used for phase identification, espe- are given in Table 2, to-
cially for the carbides. gether with estimates of
Model calculations were conducted CDZ and CEZ widths. If
using the Thermo-Calc and the DICTRA during annealing the ma-
software and utilizing the STEEL 16.tdb trices of the two steels are
database (Ref. 19) and the Dif.tdb kinet- austenitic (a temperature
ics database (Ref. 20). of 900°C), the redistribu- Fig. 3 — A — Microstructure of the CDZ and CEZ (annealed 700°C/56h)
tion of carbon is little pro- with contamination path from EPMA measurement (light microscopy);
Experimental Results nounced. If, however, the B — detail of the weld joint interface from Fig. 3A (SEM micrograph).
above matrices are ferritic,
The measured carbon concentration the redistribution of car-
profiles across the weld interface are given bon is considerable, with the Cmax concen- able increase in the width of carbon-en-
in Fig. 1. As expected, carbon diffused from tration values in the interval 0.6–0.7 wt-% riched zones, which means that the
the T25 low-alloy steel into the P91 high- and in agreement with the values measured amount of redistributed carbon increases.
alloy steel. In view of the low carbon con- by Buchmayr et al. (Ref. 9). With increas- The microstructure of the weld inter-
centration of the two steels (in the CDZ in ing annealing temperature in the ferrite face depends on the annealing tempera-
particular) and considering the scatter of region, there is a tendency for the Cmax ture and on the amount of redistributed
the values measured, which is due to the values to grow while there is a consider- carbon. At annealing temperatures of

Table 1 — Chemical Composition of Steels (wt-%)

Steel C Mn Si P S Cr Mo V B Al N Nb Fe
T25 0.06 0.42 0.34 0.012 0.011 1.91 0.31 0.22 0.003 0.009 0.024 0.058 bal.
P91 0.12 0.38 0.44 0.010 0.003 9.96 0.89 0.22 — 0.010 0.069 0.070 bal.

Table 2 — Heat Treatment and Basic Characteristics of Carbon Redistribution

Sample No. Annealing Experimental Results Results of Modeling

T[°C (°F)] t(h) Cmax (wt-%) width of CDZ Cmax (wt-%) Cmin (wt-%) width of CDZ
and CEZ (mm) and CEZ (mm)
3 900 (1652) 18 0.19 — 0.55 0.125 0.055 1.2 1.0
6 700 (1292) 56 0.65 1.0 0.30 0.33 0.01 1.3 0.5
7 600 (1112) 240 0.60 0.4 0.20 0.31 0.01 0.7 0.3

WELDING JOURNAL 213 -s

WELDING RESEARCH

Fig. 4 — Phase diagram for base T25 steel. Red line represents carbon con- Fig. 5 — Phase diagram for base P91 steel. Red line represents carbon con-
tent. Phase fields: 1…BCC + MX, 2…BCC + MX + M23 + M7, 3…BCC tent. Phase fields: 1…BCC + FCC + MX, 2…BCC + MX + M6 +
+ MX + M7+M6, 4…BCC + MX + M6, 5…BCC + MX + M23 + M6. LAVES, 3…BCC + MX + M6 + M23.

CDZ. In comparison two groups. The phase denoted MX(2) is a

with the nonaffected nitride or a carbonitride with vanadium
region, a markedly in- predominating (35–41 at.-%), and it also
creased density of car- contains small amounts of Nb (4–7 at.-%)
bides can be observed and Cr (1—10 at.-%). The composition
in the CEZ of the P91 obtained by calculation is almost identical.
steel — Fig. 3B. The composition of the metallic part of
Results of the analy- phase MX(2) does not depend on steel
ses of minority phases composition and annealing temperature.
in individual regions of The phase denoted MX(1) occurs in the
the weld joints under T25 steel and in comparison with MX(2), it
study are given in Table has a lower content of V (27–30 at.-%) and
3. At a temperature of a higher content of Nb (15–21 at.-%).
900°C (1652°F) only Judging from the chemical composition, it
phase MX (X = C,N) is a transition type between NbX and
occurs in the T25 steel MX(2). In this case simultaneous analyses
while M23C6 and MX of both particles or heterogeneous nucle-
could be observed in ation of MX particles on nondissolved pri-
the P91 steel. At a tem- mary NbX particles cannot be ruled out.
perature of 600°C
(1112°F), M7C3 and Equilibrium Predictions
MX are found in non-
Fig. 6 — Temperature dependence of carbon activity in base T25 and P91 affected T25 material;
The CALPHAD approach (Ref. 12)
steels. carbon depletion led to and the STEEL thermodynamic database
the dissolution of car- (Ref. 19) were used for the solution of
900°C (1652°F) and higher, when the matrix bides M7C3. Thus only the MX phase occurs both local and global phase equilibrium
of the two steels is austenitic, subsequent in the CDZ, which was also established for problems concerning the base material
cooling in air results in martensitic transfor- annealing temperatures of 500° and 700°C (T25, P91), and weld joints of these steels.
mation of the P91 steel and predominantly (932° and 1292°F). In the P91 steel at 600°C The phases that can be found in the system
bainitic transformation of the T25 steel. In (1112°F), the minority phases are formed by are different and the thermodynamic de-
these structures, carbon depletion and car- M23C6 + MX + M2X, and carbon enrich- scription is given in Ref. 20. A list of
bon enrichment are difficult to observe, as ment does not alter the phase composition. phases is given in Table 4, which also gives
can be seen in Fig. 2. In the two steels an NbX primary phase the abbreviations used for phases in this
A typical structure of the weld interface occurs, not dissolved in the course of paper. The calculated stable phase dia-
of a sample annealed at a temperature of austenitization, in which the content of Nb gram cross sections for the base T25 and
700°C (1292°F) can be seen in Fig. 3. The ranges between 27 and 39 at.-%. This P91 steels, whose composition is given in
CDZ of the T25 steel is formed by ferritic phase also contains 7–22 at.-% of V. The Table 1, are given in Figs. 4, 5.
grains of irregular shape (Fig. 3A), in which chemical composition of secondary MX It is evident from the values given in
relatively coarse precipitates can be seen to phases varies considerably, which is prob- Table 3 that the experimental results tally
occur — Fig. 3B. At an annealing tempera- ably related to the difficulty with which with the thermodynamic calculations ex-
ture of 600°C (1112°F), the initial bainite equilibrium states are achieved. The sec- cept for the P91 base material, where M23C6
morphology remains partly preserved in the ondary MX phases could be divided into + MX should have occurred by calculation.

214 -s OCTOBER 2006

WELDING RESEARCH
Table 3 — Coexisting Phases and Their Chemical Composition (at.-%)

Experimental(a) Calculation
Phase V Cr Fe Nb Mo Phase V Cr Fe Nb Mo C N

MX(2) 38.5 2.4 0.9 5.4 3.0 MX(2) 40.2 1.0 0.0 11.4 0.2 33.0 14.4
BMT MX(1) 27.1 1.2 1.1 17.7 3.0
NbX 16.1 0.4 0.4 32.1 1.1

MX(2) 40.8 2.2 0.5 4.8 1.8 MX(2) 39.6 0.5 0.0 12.5 0.2 0.5 46.9
CDZ MX(1) 27.2 0.7 0.3 21.1 0.7
900°C (1650°F)

NbX 17.3 0.6 0.5 30.6 1.1

M23C6 1.7 47.5 25.5 0.6 4.0 M23C6 1.1 47.5 27.7 0.0 2.8 20.7 0.0
CEZ MX(2) 35.1 7.5 0.7 6.2 0.7 MX(2) 35.3 8.9 0.0 7.6 0.1 7.3 40.1
NbX 7.0 1.9 1.0 38.8 1.4

M23C6 1.3 47.3 26.3 0.5 4.0

MBP MX(2) 35.9 6.1 0.3 7.3 0.6 MX(2) 36.1 8.2 0.0 7.6 0.1 6.5 41.6
NbX 7.5 2.6 0.9 37.9 1.3

M7C3 4.1 35.3 28.9 0.6 1.1 M7C3 1.9 54.7 10.7 0.0 0.8 30.0 0.0
MX(2) 33.6 6.0 1.5 4.1 5.0 MX(2) 44.9 1.9 2.9 4.9 0.0 26.6 21.6
BMT MX(1) 29.7 2.4 0.6 15.0 2.4 MX(1) 26.8 0.1 0.0 20.4 3.7 46.4 1.8
NbX 16.6 0.7 0.4 31.2 1.2

MX(2) 39.9 3.2 0.6 3.7 2.7 MX(2) 39.3 0.5 0.0 13.7 0.0 5.9 40.6
CDZ NbX 21.6 0.7 0.3 26.6 1.0
600°C (1100°F)

M23C6 1.0 52.8 20.8 0.6 4.1 M23C6 0.3 56.5 18.8 0.0 3.3 20.7 0.0
CEZ MX(2) 34.9 10.2 0.65 3.6 0.6 MX(2) 35.9 7.7 0.0 7.2 0.0 2.1 47.2
M2X 9.6 54.3 1.1 0.4 1.3

M23C6 1.1 51.8 21.7 0.6 4.1 M23C6 0.4 62.8 12.7 0.0 3.1 20.7 0.0
MX(2) 35.9 9.8 0.6 3.2 0.6 MX(2) 38.1 5.8 0.0 7.3 0.0 1.1 47.7
BMP NbX 7.4 2.4 1.0 38.0 1.3
M2X 9.3 53.1 1.1 0.7 2.5

(a) Chemical composition of metallic part, BMT — base material, steel T25; BMP — base material, steel P91.

Table 4 — Phase Abbreviations Used in the Figures

Phase Austenite Ferrite M23C6 M7C3 M3C M6C Laves Phase M2X Carbonitride MX Carbonitride
Abr. FCC BCC M23 M7 M3 M6 LAVES M2 MX

In the case of experiments, an analysis of the two compositions are either the same or and this diffusion flux is roughly propor-
nonequilibrium state is probably involved the calculated composition exhibits a higher tional to the absolute value of the activity
here since M2X is usually substituted by the content of elements of higher affinity to car- difference. In the case of the microstruc-
MX phase only after longer periods of an- bon or nitrogen (higher content of C at the ture evolution of dissimilar welds of steels,
nealing. It follows from Table 3 that, ac- expense of Fe in the case of carbides M23C6 interstitial elements (such as carbon, ni-
cording to the calculation, the MX phase and M7C3, and a higher proportion of V at trogen, and hydrogen) are of utmost im-
in the two steels is generally a carboni- the expense of Cr and Mo in the MX car- portance because they diffuse more
tride, with the ratio of C content to N con- bonitrides). The above differences are quickly than substitution elements. Thus
tent depending on the type of steel and on again related to the insufficient annealing the diffusion of the metal elements can be
the preceding thermal history. In the T25 times for reaching equilibrium states at neglected. Along with the carbon redistri-
steel we are concerned with carbonitrides lower annealing temperatures. bution, the redistribution of Fe, Cr, and
with predominant proportion of C, the An important result of phase equilib- Mo was also measured. These analyses
content of which varies from 27 to 46 at.- rium calculations is the evaluation of the confirmed that width of the fusion zone
% while in the P91 steel almost pure ni- activities of individual elements in the and the diffusion zone of substitution ele-
trides are involved, which is confirmed by steels with respect to standard element ments does not exceed the value 5 µm and
literature searches conducted in Ref. 21. reference (SGTE) (Ref. 22). The activity can thus be neglected.
For annealing at a temperature of 900°C difference of the given element in two dif- The temperature dependence of the car-
(1652°F) the calculated composition and ferent materials can be used as a first ap- bon activity for the examined steels is given
the measured composition of the occur- proximation for weld joint stability assess- in Fig. 6. In spite of the higher carbon con-
ring minority phases are in principle iden- ment (Ref. 23) because each element tent (0.12 wt-%) in the P91 steel and the
tical. For a temperature of 600°C (1112°F), diffuses to a place with its lower activity lower carbon content (0.06 wt-%) in the

WELDING JOURNAL 215 -s

WELDING RESEARCH
A B

trary, the experimen- fusion of carbon and nitrogen.

C
tal values of the CDZ The phase profile calculated for anneal-
and CEZ widths are as ing temperatures of 600° and 900°C (1112°
much as twice lower. and 1652°F) using the DICTRA software
Similar differences, (Fig. 7) is almost identical to the composi-
although not so tion calculated for individual regions of the
marked, were ob- two steels (Table 3) directly by means of the
served by Helander et ThermoCalc program, and for carbon con-
al. (Ref. 8). tents established experimentally. The DIC-
It would therefore be TRA program is an extension of the Ther-
opportune to modify moCalc program so that the agreement
the carbon mobility val- observed can be expected. A difference was
ues in the kinetics data- observed only in the case of the CEZ of the
base used and thus to P91 steel for an annealing temperature of
have a better agree- 900°C (1652°F). This difference can be ex-
ment between calcula- plained by the different carbon contents
tions and experiments. (calculation– 0.125, experiment– 0.19 wt-
In view of the limited %) since, as is obvious from Fig. 5, a carbon
Fig. 7 — Phase profiles in the weld joint T25/P91 annealed at different con- extent of experimental concentration of 0.125 wt-% (calculated)
ditions. A — 600°C/240 h; B — 700°C/56 h; C — 900°C/18 h. work and in view of the relates to austenite and MX, while at a car-
difficulties involved bon concentration of 0.19 wt–%, it is
when measuring by the austenite + M23C6 + MX that occur.
T25 steel, Fig. 6 predicts that the carbon will EPMA method the low carbon concentra- The question arises whether it is pur-
diffuse from T25 to P91 at low tempera- tions in multiphase systems, this change poseful to continue building the two data-
tures, ranging from 500° to 870°C (i.e., car- would at present be only formal. Another bases, and possibly, make the model used
bon up-hill diffusion). At temperatures possible cause of the disagreement ob- more specific from the viewpoint of prac-
higher than 870°C (1598°F), when the ma- served may be the assumption of phase tical applications. From this viewpoint the
trices of the two steels are completely equilibrium in every computation cell, calculation of the CDZ width is of primary
austenitic, the differences in activities are which is not fulfilled for low temperatures importance by reason of
substantially smaller; consequently, the car- and short annealing times. • A direct measurement of the CDZ
bon redistribution is also markedly lower, as The accuracy of the thermodynamic and width by the EPMA method is costly,
was experimentally confirmed — Fig. 1. kinetic calculations used is currently given mostly impossible, and inaccurate (prob-
primarily by the accuracy of the databases lems of the detectability of low C contents
Discussion used. The above good agreement between and in multiphase systems);
the calculated and the experimentally veri- • the CDZ width might correlate with
The weld joint of the two advanced fied phase composition (inclusive of the the critical crack length from the view-
creep-resistant steels is structurally unsta- chemical composition of individual phases) point of brittle, fatigue, and creep
ble due to the carbon redistribution, and the semiquantitative agreement in car- fractures.
which continues to proceed even at a tem- bon distribution across the weld interface In real heterogeneous weld joints of
perature of 800°C (1472°F). The depen- prove the consistency of the databases that creep-resistant steels there will be at least
dence relations given in Fig. 1 confirm the are used and the building of which should two fusion zones, namely between the
conclusions resulting from the tempera- be continued. Advanced creep-resistant base metals and the filler metal whose
ture dependence of carbon activity in the steels contain, among other things, nitro- composition will be between the composi-
matrices of the two steels and are in gen, which due to its high diffusion mobility tions of both base metals so that the car-
agreement with experimental results. The and low solubility in ferrite behaves analo- bon activity gradients in individual matri-
Cmax values measured are approximately gously to carbon, and in the case of steels, ces will be lower than in the joint types
twice the calculated value; on the con- attention must be given to simultaneous dif- under evaluation. It also needs to be con-

216 -s OCTOBER 2006

WELDING RESEARCH
sidered that both dissolution of minority served in the CDZ at temperatures of 700°C karsson, M. 2000. Structural changes in 12-2.25
phases in low-alloy steel and, on the con- (1292°F) and higher. In the CEZ of steel weldments — an experimental and theoretical
trary, precipitation of carbides and ni- P91 further precipitation of carbides M23C6 approach. Materials at High Temperatures 17(3):
trides in high-alloy steel, result in solid so- could be observed, provided the matrix of 389-s to 396-s.
lutions being enriched with and depleted this steel was at least partially ferritic. 9. Buchmayr, B., et al. 1990. Experimental
of carbide- and nitride-forming elements, The experimentally established chemi- and numerical investigations of the creep behav-
which leads to the difference in carbon ac- cal and phase compositions of annealed ior of the dissimilar weldment GS-17 CrMoV 5
tivities being reduced on either side of the weld joints were confirmed by thermody- 11 and X 20 CrMoV 12 1. Steel Research 61(6):
weld joint. With increasing annealing time namic and kinetic calculations using the 268-s to 273-s.
(exploitation) the redistribution of carbon Thermo-Calc and the DICTRA software, 10. Roberts, D. I., Ryder, R., and
and nitrogen will thus be less intensive. with very good agreement being obtained Viswanathan, R. 1985. Performance of dissimilar
The appearance of the CDZ and CEZ in for the phase composition. The calculated welds in service. Journal of Pressure Vessel Tech-
heterogeneous weld joints of creep- and the experimentally established carbon nology 107(3): 247-s to 254-s.
resistant steels operated at elevated tem- concentration profiles exhibit sufficient 11. Albert, S. K., et al. 1997. Soft zone forma-
peratures is a regular process that can even agreement. The observed lack of agree- tion in dissimilar welds between two Cr-Mo
lead to the appearance of Type IIIa crack- ment was connected with the analyses of steels. Welding Journal 76 (3): 135-s to 142-s.
ing. The possible methods of suppressing nonequilibrium structural states and with 12. Saunders, N., and Miodovnik, A. P. 1998.
the occurrence and development of this neglecting a probably simultaneous diffu- CALPHAD (Calculation of Phase Diagram — A
structural instability are as follows: sion of carbon and nitrogen. Comprehensive Guide). Amsterdam, Elsevier
• use a diffusion barrier, e.g. a nickel Science Publishing.
alloy; Acknowledgments 13. Anderson, J. O., et al. 2002. Thermo-Calc
• minimize the PWHT temperature and DICTRA, computational tools for materials
and time; The present work was supported by the science. CALPHAD. 26: 273-s to 312-s.
• choose the chemical composition of Grant Agency of the Czech Republic under 14. Borgenstam, A., et al. 2000. DICTRA, a
filler metals such that the differences in project No. 16/03/0636, and by the Ministry tool for simulation of diffusional transformations
carbon activities are distributed uniformly of Education, Youth, and Sports, Project in alloys. Journal of Phase Equilibria 21(3): 269-s
among individual joint welds. No. MSM0021630508. to 280-s.
Thermodynamic calculations have 15. Kirkaldy, J. S., and Young, D. J. 1985. Dif-
shown that in the steels under study the car- References fusion in the Condensed State, London, The Insti-
bon activity is most affected by the tute of Metals.
chromium content. The alloying base of 1. Strang, A., and Vodarek, V. 1998. Micro 16. Engstram, A., Haglund, L., and Ägren, J.
both the base metals and the filler metals structural degradation of martensitic 12%Cr 1994. Computer simulation of diffusion in multi-
should be chosen also with respect to the power plant steel during prolonged high temper- phase system. Metallurgical and Materials Trans-
thermodynamic stability of individual ature creep exposure. Proc. of the 6th Liege Con- actions A. 25A(2): 1127-s to 1134-s.
phases such that either the nondissolved ference Materials for Advanced Power Engineering, 17. Andersson, J. O., et al. 1990. Computer
phases remain in the CDZ or they dissolve eds. J. Lecomte-Beckers et al., p. 601, Jülich: simulation of multicomponent diffusional trans-
for as long as possible. The knowledge ob- Forschungszentrum. formations in steel. Proc. Fundamentals and Ap-
tained to date shows that nitrides or car- 2. Masuyama, F. 2002. Trends in power engi- plications of Ternary Diffusion, ed. G. R. Purdy,
bonitrides of V and Nb are considerably neering in Japan and requirements for improved pp. 153–163. Pergamon Press, New York.
more stable than carbide phases in low- materials and components. Proc. of the 7th Liège 18. Foldyna, V., et al. 2002. Microstructure
alloy Cr-Mo steels of the type 2.25Cr-1Mo. Conference Materials for Advanced Power Engi- and properties of modified 3%Cr steels. Proc. of
neering, eds. J. Lecomte-Beckers et al. pp. the 7th Liège Conference Materials for Advanced
Conclusions 1767–1782. Jülich Forschungszentrum. Power Engineering. Eds. J. Lecomte-Beckers et
3. Mayer, K. H., et al. 2002. Development step al., pp. 1477–1486. Jülich Forschungszentrum.
The structural analyses conducted to- soft new steels for advanced steam power plants. 19. Kroupa, A., et al. 2001. Phase diagram in
gether with thermodynamic and kinetic Proc. of the 7th Liège Conference Materials for Ad- the iron-rich corner of the Fe-Cr-Mo-V-C system
model calculations have confirmed that het- vanced Power Engineering, eds. J. Lecomte-Beck- below 1000 K. Journal of Phase Equilibria 22(3)
erogeneous weld joints of creep- ers et al. pp. 1385–1397. Jülich Forschungszen- :312-s to 323-s.
resistant steels are usually structurally un- trum. 20. Sopousek, J., Jan, V., and Foret, R. 2004.
stable. This instability is in essence due to 4. Brett, S. J. 2004. Type III cracking in Simulation of dissimilar weld joints of steel P91.
the different carbon activities in individual 1/2CrMoV steam pipe work systems. Science and Science and Technology of Welding and Joining
base materials of the weld joint. In the given Technology of Welding and Joining 9(1): 9(1): 59-s to 63-s.
case, carbon diffused by the so-called uphill 41-s to 45-s. 21. Kubon, Z., Foldyna, V., and Vodarek, V.
diffusion from the T25 low-alloy steel (6Cr- 5. Foret, R., et al. 2001. Structural stability of 1998. Optimized chemical composition of 9-
Mo-V 8-3-2) into the P91 high-alloy steel dissimilar weld joints of steel P91. Science and 12%Cr steels with respect to maximum creep
(X12Cr-Mo-V-Nb10-1). The redistribution Technology of Welding and Joining, 6(6): 405-s to resistance. Proc.of the 6th Liege Conference Mate-
of carbon, which is substantially larger in the 411-s. rials for Advanced Power Engineering, eds. J.
ferritic matrix than in the austenitic matrix, 6. Sudha, C., et al. 2002. Systematic study of Lecomte-Beckers et al., pp. 311–325, Jülich:
led to the appearance of CDZ in the T25 formation of soft and hard zones in dissimilar Forschungszentrum.
steel and CEZ in the P91 steel. weldments of Cr-Mo steels. Journal of Nuclear 22. Dinsdale, A. T. 1991. SGTE data for pure
In the CDZ of steel T25, there was a Materials 302(4): 193-s to 205-s. elements. CALPHAD 15(4): 317-s to 425-s.
complete dissolution of carbides M7C3 and 7. Bret, S. J. 1998. In-service cracking mecha- 23. Jan, V., Sopousek, J., and Foret, R. 2004.
partial dissolution of phase MX. No com- nism affecting 2CrMo welds in1/2CrMoV steam Weld joint simulation of heat-resistant steels.
plete dissolution of the MX phase was ob- pipe work system. Proc. of Int. Conf. Integrity of Archives of Metallurgy and Materials 49(3):
served in the T25 steel, not even at a tem- High-Temperature Welds. pp. 3–14. Professional 469-s to 480-s.
perature of 900°C (1652°F), but in spite of Engineering Publishing Ltd., London.
this, the recrystallization of ferrite was ob- 8. Helander, T., Andersson, C. M., and Os-

WELDING JOURNAL 217 -s