Escolar Documentos
Profissional Documentos
Cultura Documentos
CITATIONS READS
7 466
3 authors, including:
Iman El Mahallawi
Cairo University
86 PUBLICATIONS 429 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Friction Stir Welding of Similar and Dissimilar Metal Alloys: Adopting the New
Joining Technology for Industrial Applications in Egypt View project
All content following this page was uploaded by Iman El Mahallawi on 16 March 2015.
Element C Si Mn P S Ni Cr Mo Al V Nb N
Standard 0.07–0.14 0.2–0.5 0.3–0.6 0.02 max. 0.01 max. 0.4 max. 8.0–9.5 0.85–1.05 0.04 max. 0.18–0.25 0.06–0.1 0.03–0.07
Actual 0.11 0.26 0.44 0.012 0.002 0.17 8.5 0.88 0.002 0.215 0.077 0.04
Table 3 Chemical composition of GTAW rod ER90S-B9 and SMAW electrode E9015-B9 used during welding/wt-%
Electrode/
Rod C Si Mn P S Ni Cr Mo Cu Al V Nb N
ER90S-B9 0.09 0.2 0.49 0.004 0.003 0.66 8.7 0.9 0.03 0.006 0.19 0.08 0.05
E9015-B9 0.1 0.19– 0.66– 0.01 0.005– 0.27– 8.9– 0.9– ,0.1 0.001– 0.21– 0.04– 0.03–
0.38 0.69 0.009 0.75 9.3 0.98 0.002 0.26 0.06 0.04
ferrite content was measured in the weld zone, HAZ and Results and discussion
base metal for each condition using Fisher ferrite scope
FMP-30. The ferrite scope measurement is based on the Base metal
difference in magnetic permeability between the mar- In the as received (normalised and tempered condition),
tensite and ferrite phases, and was used to evaluate the the microstructure of the P91 base metal consists of fully
changes that occur in the microstructure as a function of tempered martensite (Fig. 2a). Small precipitate parti-
the heat input and heat treatment. cles could be identified on the prior austenite grain
For the sake of traceability of the various samples of boundaries and within the grains (Fig. 2a). SEM
different heat input/heat treatment conditions a coding investigation of the base metal (Fig. 2b) shows the lath
system described in Table 4 will be used throughout the martensitic structure, as well as the precipitate particles
following text. embedded along the lath boundaries, which are possibly
of the M23C6 type as identified in earlier studies.8,9,10
EDX analysis (Fig. 3 and Table 5) confirm the presence
of M23C6 (M5Cr, Fe, Mo) precipitates across different
areas within the tempered martensite. However, much
Table 4 Coding system for different sample conditions finer precipitate particles of niobium rich or vanadium
(heat input/heat treatment) rich compounds were also observed both along the
boundaries and within the laths.11 Carbides of type MC,
Sample Code Condition M2C have been identified in Grade 91 steels.11
L 350 Low heat input condition followed by
intermediate PWHT at 350uC for 2 h Weld metal and HAZ
(as welded condition) Figure 4, illustrates the expected microstructure of the
L 760 Low heat input condition followed weld metal and HAZ as a function of the peak
by PWHT at 760uC for 3.5 h temperature gradient along the P91 weldment.12 The
L1050 Low heat input condition followed by N&T
M 350 Medium heat input condition followed
optical micrographs of the three welding heat inputs of
by intermediate PWHT at 350uC for 2 h as welded conditions are shown in Fig. 5 which show
(as welded condition) a typical untempered martensite structure with a
M 760 Medium heat input condition followed by
PWHT at 760uC for 3.5 h Table 5 Results of EDX analysis related to precipitates
M1050 Medium heat input condition followed indicated through Fig. 3/wt-%
by N&T
H 350 High heat input condition followed by Element C Si V Cr Mn Fe Ni Nb Mo
intermediate PWHT at 350uC for 2 h
(as welded condition) Spec.1a 42.98 … 0.2 6.13 0.39 49.23 0.29 … 0.79
H 760 High heat input condition followed by Spec.2a 39.02 0.73 0.26 5.43 0.77 52.37 0.25 0.23 0.95
PWHT at 760uC for 3.5 h Spec.1b 40.30 0.17 0.15 7.72 … 50.56 … 0.05 1.05
H1050 High heat input condition followed by N&T Spec.3b 42.10 0.21 0.16 5.17 0.23 51.83 … … 0.28
4 Schematic representations of microstructures developed in weld metal and HAZ as function of peak temperature dur-
ing welding12
considerable amount of acicular ferrite. A reheating volume fraction came as a result of the decomposition of
effect results from multilayer welding leading to the the untempered martensite during the subcritical PWHT
formation of a microstructure similar to that expected (Fig. 7). The obtained results mean a significant
for the HAZ (Fig. 6). An increase in the ferrite volume enhancement in the homogeneity of the resulted micro-
fraction was found to occur with the increase in heat structure of the weld metals as the volume fraction of the
input, where the ferrite content measured by the ferrite measured ferrite and martensite within the microstruc-
scope was from 75 to 77?5 and up to 82% respectively in ture are nearly the same irrelevant of the heat input.
the weld zone. On the other hand, ferrite content after Similarly, a higher degree of microstructure homo-
subcritical PWHT shows a significant increase in the geneity (homogenous microstructure of tempered mar-
measured ferrite volume fraction for all heat input tensite) was obtained for all heat input conditions after
conditions, where the ferrite content became 95, 92 and normalising and tempering treatment as a result of
93% respectively. This significant increase in the ferrite recrystallisation effect of the normalisation treatment
(Fig. 8). Figure 8c reflects also the slightly finer grain A particularly important feature of interest in the P91
size resulting from the high heat input condition weld metal is the occurrence of patches of d-ferrite.
compared to that obtained for both low and medium Figure 9 show different shapes of formed d-ferrite
heat input conditions (Fig. 8a and b). within the weld metal for different heat inputs of the
as welded condition. It was observed that the volume microstructure after subcritical PWHT as it formed
fraction of d-ferrite decreased with the increase in heat during welding as a result of the high peak temperature
input. The volume fraction of the d-ferrite in the weld at the welding pool and was retained after cooling
metal of the low, medium and high heat input of the as (Fig. 10). Decarburisation like effect was also observed
welded conditions were found to be about 0?013, 0?0088 at the boundary area between d-ferrite and martensite
and 0?0033% respectively. The resulted volume fractions (lighter martensite around ferrite as in Fig. 10), and this
presented the ratio of the total area of d-ferrite indicated may be due to the retarded diffusion of carbon atoms
along the weld cross section to the total area of the weld and other alloying elements during rapid solidification
cross section. Arivazhagan observed that the volume after welding solidification. This may also be the reason
fraction of d-ferrite increased with the increase of heat of considerable increase in the precipitation of carbides
input in a low heat input ranging from 0?52 to observed within d-ferrite grains after subcritical PWHT.
0?88 rev min21.13 Combination of this published results d-ferrite is also formed after solidification at the fusion
with that from our research leads to the conclusion that line area (solid–liquid transition area) and may be
volume fraction of d-ferrite increases with the increase in formed at the course grain region in the heat affected
heat input till a critical value of heat input and then zone CGHAZ located just beside the fusion line
decreases with the increase of heat input. This critical (Fig. 11). On the other hand, smaller traces of d-ferrite
heat input value is still within the low heat input range. grains within the weld metal are present only after
On the other hand, d-ferrite is still present in the normalising and tempering treatment with a significant
precipitation within the d-ferrites grains increasing with and by the incomplete dissolution of carbides.
the increase of heat input (Fig. 12a and b). However, Fine grained austenite is produced, which
low heat input shows the highest volume fraction of d- subsequently transforms into low carbon slit
ferrites grains. softer martensite (Fig. 13b)
The following different microstructures were identi- (iii) inter critical region (ICHAZ): here the peak
fied along the HAZ toward the base metal in both as temperature Tp is lower than Ac3 but higher
welded and subcritical PWHT conditions: than Ac1; where Ac1 is the temperature at which
(i) coarse grain region (CGHAZ): area near the transformation of ferrite into austenite is started
fusion boundary that reaches a temperature well on heating, and this results in partial reversion
above Ac3 during welding; where, Ac3 is the to austenite on heating. The new austenite
temperature at which transformation of ferrite nucleates at the prior austenite grain boundaries
into austenite is completed upon heating. and the martensite lath boundaries; whereas the
Carbides which constitute the main obstacle to remainder of the microstructure is simply
growth of the austenite grains dissolve resulting tempered. The austenite transforms into untem-
in coarse grains of austenite. This austenite pered martensite on cooling (Fig. 13c)
transform into high carbon and hard martensite (iv) over tempered region: with Tp below Ac1 the
on cooling (Fig. 13a) original microstructure of the plate material
(ii) fine grain region (FGHAZ): away from the undergoes further tempering (Fig. 13d–e).
fusion boundary where the peak temperature In fact, the obtained microstructure within the weld and
(Tp) is lower, but still above Ac3. Austenite HAZ depends strongly on the obtained peak tempera-
grain growth is limited by lower temperature ture and the cooling rate. However, the maximum
a CGHAZ; b FGHAZ e; c ICHAZ; d over tempering zone (OT); e boundary between over tempered (OT) and unaffected
base metal (UA); f unaffected base metal
13 Different areas at HAZ of low heat input (L350)
a HAZ area near base metal of high heat input condition H760; b base metal affected by double tempering at 760uC
15 Image (SEM) showing bainite like grains in subcritical PWHT condition
obtained peak temperature at welding pool, its gradient complete dissolution of carbides within the
effect, cooling rate and the width of formed HAZ are all austenite grains
directly dependent on the heat input used during the (iii) the wider the CGHAZ, FGHAZ, ICHAZ and
welding cycle, the higher the heat input, the higher the over tempered HAZ.
peak temperature, the lower the cooling rate, and the However, normalising and tempering treatment mini-
larger the temperature gradient effect and the wider the mize the difference between the HAZ, weld and base
HAZ.14 metal structures for the same heat input as a result of
To sum up, the higher the heat input: recrystallisation effect during normalising treatment, but
(i) the lower the cooling rate of weld pool leading to the differences between the HAZ microstructures of
the formation of higher volume fraction of ferrite different heat inputs exist. Figure 14 shows the optical
and lower volume fraction of martensite micrograph obtained from the HAZ of the three heat
(ii) the coarser the CGHAZ grains and the harder inputs after normalising and tempering treatment.
the martensite formed after cooling as a result of However, average ferrite content measurement conducted
16 X-ray diffraction pattern of base metal and weld metals of all heat/treatment conditions
by ferrite scope on the HAZ area of the three heat inputs 8. Bainite may appear in the microstructure of P91
after normalising and tempering treatment shows slight steel in weld, HAZ and base metal as a result of double
differences in measured ferrite content between low and reheating effects during multiple layers welding and
medium heat inputs compared to that of the high heat tempering treatments.
input condition, 90, 90?5 and 92% respectively. 9. Retained Austenite was not detected by X-ray
Another particularly important feature of interest is the diffraction analysis.
possibility of bainite formation as a result of martensite 10. The formation of a soft zone is expected to be
decomposition during multiple tempering at 760uC. A formed in the HAZ portion subjected to over tempering
percentage of bainitic ferrite was detected at base effect and may act as the source of earlier creep failure
materials and HAZ/base boundaries in the subcritical during service.19
PWHT conditions (Fig. 15). The presence of bainitic
ferrite is more evident in the high heat input condition
which may indicate an increase in the probability of
Acknowledgement
bainite formation with the increase of heat input. In fact, The authors wish to deeply thank the Project Services
the continuous cooling transformation (CCT) diagram Company (PSC) for offering the materials and facilities
published for P91 material did not indicate the prob- available at their plant, without which this work would
ability of bainitic transformation.15,16 However, Mile have not been possible.
demonstrated that in P91 steel a microstructure which
appeared to be of bainitic character.17 This could be the References
case if multiple heat treatments during a manufacturing
1. B. Arivazhagan, R. Prabhu, S. K. Albert, M. Kamaraj and S.
process (e.g. bending) took place.17 Milović reported the Sundaresan: ‘Microstructure and mechanical properties of 9Cr-
formation of small amounts of bainite in P91 steels during 1Mo steel weld fusion zones as a function of weld metal
cooling from austenisation temperature as the martensite composition’, ASM International, July 2008.
formation was preceded by separation of a smaller 2. A. Czyrska-Filemonowicz, Aielinska-Lipiec and P. J. Ennis: ‘Modified
quantity of bainite.18 He reported also the formation of 9% Cr steels for advanced power generation: microstructure and
properties’, J. Achiev. Mater. Manuf. Eng., 2006, 19, (2), 43–48.
bainite in addition to martensite in simulated HAZs using 3. H.G. Kang and S. Y. Bae: ‘Fatigue crack growth rate in low DK
simulation temperature up to 1150uC and with cooling range according to the microstructure and temperature in P122 alloy
time t8/5 of 40 s followed by PWHT at 730uC.18 steel’, Graduate School of Mechanical Engineering, Sungkyunkwan
However, X-ray diffraction patterns of the weld University, Korea, 2006.
4. H. Kenji, K. Tomoya, O. Takao, I. Rinsei and H. Kotaro: ‘Steel
metals indicated similar patterns to that obtained from products for construction, industrial machinery and plant’, JFE
the base material (martensite/ferrite phase) and no Technical Report no. 2, 2004.
retained austenite was indicated (Fig. 16). 5. R. Blume, K. E. Leich, H. Heuser and F. W. Meyer: Stainless Steel
Europe, Apr. 1995, 49–53.
6. S. Dittrich, V. Gross and H. Heuser: Proc. Nat. Welding Seminar,
Conclusions Jamshedpur, 1994.
7. P. Alto: ‘Optimal hardness of P91 weldments’, CA:1004702,
1. Heat input and heat treatment were found to have Electric Power Research Institute, EPRI, 2003.
a great influence on the obtained microstructure of P91 8. V. K. Sikka, C. T. Ward and K. C. Thomas: Proc. Int. Conf. on
‘Ferritic steels for high temperature applications’, (ed. A. K. Khare),
weldments which may have direct effect on the obtained 65–84; 1983, Metals Park, OH, ASM.
mechanical properties and weldments serviceability. 9. W. B. Jones: Proc. Int. Conf. on ‘Ferritic steels for high temperature
2. Martensite and ferrite were the main observed applications’, A. (ed. K. Khare), 116–130; 1983, Metals Park, OH, ASM.
10. W. B. Jones, C. R. Hills and D. H. Polonis: Metall. Trans. A, 1991,
phases through different areas of P91 weldments, and 22A, 1049–1058.
bainite and d-ferrite can be formed under certain 11. C. Panait, W. Bendick, A. Fuchsmann, F. Gourgues and J. Besson:
conditions. ‘Study of the microstructure of the Grade 91 steel after more than
3. In the as welded conditions; ferrite volume fraction 100,000h of creep exposure at 600uC’, Int. J. Pressure Vessels Pip.,
of the weld metal increases with the increase in heat 2010, 87, 326–335.
12. H. K. D. H. Bhadeshia: ‘Design of ferritic creep resistant steel’,
input. ISIJ Int., 2001, 41, (6), 626–640.
4. Subcritical PWHT and normalising/tempering 13. B. Arivazhagan, G. Srinivasan, S.K. Albert and A. K. Bhaduri: ‘A
treatments leads to a homogenous microstructure within study on influence of heat input variation on microstructure of
the weld metal with a near similar volume fraction of reduced activation ferritic martensitic steel weld metal produced by
GTAW process’, Fusion Eng. Des., 2011, 86, 192–197.
ferrite in all heat input conditions. 14. KlasWeman: ‘Welding processes handbook’; 2003, New York,
5. d-ferrite was observed in the weld metal at all heat CRC Press LLC.
input/treatment conditions, and its volume fraction 15. A. Di Gianfrancesco, L. Cipolla, F. Cirilli, C. Sviluppo, G. Cumino,
increased with increase in heat input till a critical value S. Caminada and T. Dalmine: ‘Microstructural stability and creep
data assessment of Tenaris grades 91 and 911’, Italy, 2005.
lower than 1?15 kJ mm21, then decreased with the 16. G. Guntz, M. Julien, G. Kottmann, F. Pellicani, A. Pouilly and J.
increase in heat input. C. Vaillant: ‘The T91 book’, Vallourec Industries, France, 1990.
6. d-ferrite was observed at the fusion boundary and 17. K. Maile: ‘Evaluation of microstructural parameters in 9–12% Cr
CGHAZ of all as welded and subcritical PWHT steels’, Int. J. Pressure Vessels Pip., 2007, 84, 62–68.
conditions, but was not reported for normalised and 18. L. Milovic, T. Vuherer, M. Zrilic, D. Momcilovic and S. Putic:
‘Microstructural analysis of simulated heat affected zone in creep
tempered conditions. resisting steel’, Proc. 3rd Int. Conf. on ‘Deformation processing
7. Normalising and tempering treatment reduces the and structure of materials’, Belgrade, Serbia, September 2007,
amount of retained d-ferrite within the weld metal, and Association of Metallurgical Engineers of Serbia. Internet: http://
results the optimum degree of microstructure homo- elektron.tmf.bg.ac.rs/dpsm2007/articles.htm
19. M. Abd El-Rahman, M. R. El-koussy and I. El-Mahallawi:
geneity along the weldments of all heat input conditions ‘Influence of heat input and post-weld heat treatment on boiler
as a result of recrystallisation effect during normalising steel P91 (9Cr-1Mo-V-Nb) weld joints, II. Mechanical properties’,
treatment. Int. Heat Treat. Surf. Eng., 2013, 7, 32–37.