J Fail. Anal. and Preven.

(2019) 19:792–801
https://doi.org/10.1007/s11668-019-00661-7

TECHNICAL ARTICLE—PEER-REVIEWED

Investigation of Premature Failure of a Coal-Fired Boiler

Superheater Tube
Biraj Kumar Sahoo . Snehashish Tripathy . M. Ghosh .
Swapan Kumar Das . Goutam Das

Submitted: 21 November 2018 / in revised form: 22 March 2019 / Published online: 12 June 2019
Ó ASM International 2019

Abstract A metallurgical investigation has been carried Introduction and Background Information
out to ascertain the cause of premature failure of boiler
superheater tube. The tube material was SA213-T22-grade Thermal power plants are the major source of power gen-
steel. Visual observation of fractured region of failed eration worldwide. The most common fuel to produce
component revealed thick-lip fracture, parallel cracks steam in thermal power plant is coal. In USA, 70% of the
along the longitudinal direction, bulging, and non-uniform electricity is produced from fossil fuel, and in India, it is
wall thickness. Microstructural investigation showed * 80% [1, 2]. The coal-fired boilers are exposed to severe
complete spheroidization of second phase, formation of service environment like high temperature, stress, corro-
thick scale to the tune of * 1 mm, the presence of creep sive atmosphere, and erosion, which lead to degradation of
voids, and significant decrease in hardness of the tube alloys. Boiler tube failure due to creep burst is the most
material. These signatures are indicative of overheating common problem and shares nearly * 30% of the whole
and creep of the alloy during service exploitation. Fracture spectrum [3, 4]. The burst rupture resembles as ‘‘fish-
surface predominantly contained the signature of inter- mouth’’-type feature. The ‘‘thin lip’’ of ‘‘fish mouth’’
granular failure. The microstructure near-fracture region indicates substantial deformation during creep rupture,
exhibited a large number of creep voids and their coales- whereas the ‘‘thick lip’’ designates minimal elongation
cence along grain boundaries. Under the hoop’s stress, the under low stress [5, 6]. Exposure of the tube material at
coalescence of these creep voids triggered intergranular elevated temperature causes microstructural changes [7].
fracture leading to rupture of the tube. Operating temper- Change in structure is progressive in nature, either over a
ature was estimated from steam side oxide-scale thickness period of time under given design conditions or within a
using kinetic model equations. It indicated that for the short span of time due to operation above design
formation of * 1-mm scale over a service span of recommendation.
9 months required operating temperature of * 677 °C. The present study focuses on exploring the reason of
However, the material is designed to withstand a maximum early failure in a superheater tube (final superheater zone)
temperature of 540 °C. It was thereby concluded that of a pulverized coal-fired boiler in a 150-MW thermal
overheating during the operation led to accelerated creep power plant. The designed life of the superheater tube is
and subsequent failure within a short period. around 10 years; however, the tube failed within 9 months
of service. The alloy designation for tube material was SA-
Keywords Superheater tube  Creep deformation  213 T22. The operating temperature and pressure of the
Intergranular fracture  T22 steel  Overheating boiler unit was: flue gas temperature 611 °C, steam tem-
perature 536 °C, and steam pressure 114 Kgf/cm2.

B. K. Sahoo (&)  S. Tripathy  M. Ghosh  S. K. Das  G. Das

Materials Engineering Division, CSIR - National Metallurgical
Laboratory, Jamshedpur 831007, India
e-mail: biraj@nmlindia.org

123
J Fail. Anal. and Preven. (2019) 19:792–801 793

Experimental Methods carried out in scanning electron microscope (FEI Nova

NanoSEM 430) equipped with energy-dispersive spec-
The failed superheater tube was collected from a coal-fired troscopy (EDS) for elemental analysis.
thermal power plant for present investigation. Visual
examination was done in naked eyes, and photographs
were taken by digital camera. The burst tube in the final Results
superheater zone is shown in Fig. 1a. For our investigation,
the failed tube has been divided into two regions. The Visual Inspection
undamaged part (but exposed to service) is denoted as
Region-1, and the fractured part is denoted as Region-2 as Figure 2 shows the radial section of the tube for Region-1.
shown in Fig. 1. Subsequently, specimens were cut from The thickness (including oxide scale) of the tube in this
the failure region and away from failure for study. The region was more or less uniform (* 9.6–9.8 mm).
dotted rectangle is Figs. 2 and 3 denotes the section from Figure 3 shows the details of the fractured Region-2.
where samples were taken for investigation. The chemical The fracture surface revealed thick-lip fracture (Fig. 3a,
composition of the tube was analyzed by using direct upper region). The thickness of the fracture lip was found
reading spectrometer (DRS). Hardness measurement was to be around 4–6 mm. Fracture was accompanied by bul-
done using Vickers hardness testing machine, with a load ging (Fig. 3a).The diameter of the bulged region was
of 10 kg. Subsequently, for microscopy, sample prepara- * 65 mm. Parallel cracks on the outer surface of the tube
tion was done using the standard metallographic technique along longitudinal direction were detected (Fig. 3a). On the
involving grinding up to 2000-grit SiC paper and polishing opposite side of the fractured region, bending was observed
with alumina and colloidal silica. Samples were etched with exfoliation of scales (Fig. 3b).
with 2% Nital (2% HNO3–98% Ethanol), followed by The tube was cut at a close vicinity of fracture region as
cleaning with ethanol and drying in hot-air stream. shown in Fig. 3c. The thickness of the pipe was measured.
Microscopic examination of the failed tube was done with At c.s.2(a) and c.s.2(b), it was found that the thickness of
optical microscope (LEICA DM2500). Following it, the the tube was not uniform. Thinning was observed at
detailed microstructural studies and fractography were c.s.2(a) and c.s.2(b). Thickness was reduced to the extent of
* 7.3–7.5 mm, against the same of other regions as
* 9.8 mm. On further observation, it was explored that
overheating was from outer diameter of the tube as indi-
cated by thinning that has taken place along the outer
surface of the tube.

Fig. 1 The burst of the superheater tube (a) the burst tube in the final
superheater zone (b) As-received failed superheater tube showing
undamaged region (Region-1) and fractured region (Region-2) Fig. 2 Cross section of the tube at Region-1

123
794 J Fail. Anal. and Preven. (2019) 19:792–801

Fig. 3 Macro-image of failed Region-2 (a) Fracture region with signatures of failure, (b) bending and exfoliated scales over the opposite side of
fractured region and (c) thickness of the tube at c.s.2(a) and 2(b) showing thinning and non-uniform wall thickness

Chemical Analysis Table 1 Compositional analysis of tube material

Composition
From the supplier, it was obtained that the tube material (wt.%) C Mn P S Si Cr Mo
belonged to the specification SA213-T22. The composition
obtained from DRS is given in Table 1. The composition SA213- 0.05– 0.3–0.6 0.025 0.025 0.5 1.9–2.6 0.87–
T22 (max.) 0.15 1.13
complied with the standard, and no discrepancy was
Failed tube 0.13 0.59 0.011 0.004 0.17 2.57 0.82
detected. Though Mo content exhibited minor deviation in material
second decimal places, it might be considered as an
inherent experimental variation.

123
J Fail. Anal. and Preven. (2019) 19:792–801 795

Hardness Measurement The thickness of the oxide scale on the outer and inner
surfaces was measured. The scale thickness was close to
The hardness of the tube was measured at Region-1 and each other for both surfaces (* 1 mm, Fig. 7a, b). This
Region-2 of both flue gas and steam side. The values thickness of scale was very high taking into account the
(average of five readings) are collated and compared service life of 9 months under normal operating conditions.
against standard in Table 2. SEM micrographs near the interface between scale and
tube material on the flue gas side also illustrated the
Microstructural Investigation presence of creep voids (Fig. 7c).
The EDS mapping of the scales on flue gas side and
A systematic and intensive microstructural investigation steam side is shown in Fig. 8a and b, respectively. EDS
was carried out in regions 1 and 2 to ascertain the cause of analysis shows that the scale on flue gas side and steam
failure. side is primarily iron oxide. Other elements like Cr, Mo, Si,
and Al were found to be present in traces.
Microstructural Investigation in Region-1 From the overall experimentation on Region-1, it can be
said that the microstructure of the material has undergone
The optical micrographs of the radial and long transverse excessive degradation in short span of service. Such high
section of Region-1 are given in Fig. 4. It showed ferrite degree of spheroidization, the presence of creep voids, and
grains with second phases both inside grains and along the formation of excessive scale undoubtedly were the
grain boundaries. signatures of overheating that might have led to accelerated
The same specimen was also examined in scanning creep in the material [7].
electron microscope to reveal finer structural details
(Fig. 5). The matrix consisted of polygonal ferrite. Large Microstructure Investigation in Near-Fracture Region:
carbides were present along grain boundary (Fig. 5a). Region-2, Cross Section 2(a)
Pearlite/bainite, which is common microstructural feature
of this grade, was completely absent. The cementite To explore the mechanism of creep in the failed region,
lamellae of the pearlite colonies were spheroidized and SEM investigation was done near Region-2, c.s.2(a) of the
dispersed in the matrix. Along the flue gas side of the tube, tube.
a significant number of creep voids were detected (Fig. 5b, Figure 9 shows the microstructure at Region-2, c.s.2(a),
c). On the steam side of the tube, these kinds of voids were where thinning of tube wall was severe. Through thickness
absent (Fig. 5d). micrographs were taken from steam side, center, and flue
Compositional analysis was performed near microvoids gas side. Images revealed that at steam side region and at
V1 and V2 (Fig. 6a). The precipitates near voids were the center of the tube creep voids were absent. Along the
essentially Cr- and Mo-rich carbides (Fig. 6b, c). At the flue gas side of the tube, drastic change in the
triple point, the carbides were coarse and decohesion microstructure has been noticed. Oxidation was traced
occurred between matrix and second phase, resulting in the along intergranular position as shown in Fig. 10a. Massive
formation of void. Other carbides along grain boundaries isolated creep voids and coalescence of voids along the
(C1 and C2) were also enriched in Cr and Mo (Fig. 6d, e). boundaries have been found (Fig. 10b).
Compositional mapping (Fig. 6f) was done through EDS
which clearly showed that the precipitates along grain Microstructural Investigation in Region-2
boundaries are enriched with Cr, Mo, and C indicating Cr
and Mo carbides. Fractography of upper and lower lips of Region-2 was
performed to assess the nature of failure (Fig. 11). The
failure was intergranular in nature (Fig. 11a and c). The
Table 2 Hardness of failed tube at different locations fracture was along the grain boundaries such that the grains
have been pulled off leaving a trace on the surface
Region-2 Region-2
SA Region-1 Region-1 (fracture (fracture
(Fig. 11b). To further ascertain the cause of fracture, SEM
213- (steam (flue gas surface—upper surface—lower image was taken from the subsurface of fracture location
T22 side) side) lip) lip) (Fig. 11d). A significant number of creep voids were
(HV) (HV) (HV) (HV) (HV) found. Thus, it corroborates pervious inference of grain
170 122 122 108 109 boundary weakening due to void formation.

123
796 J Fail. Anal. and Preven. (2019) 19:792–801

Fig. 4 Optical micrographs of

Region-1 (a) and (b) radial
section, (c) and (d) long
transverse section;
microstructure showed ferrite
grains containing carbides
within grain body and along
grain boundary

Fig. 5 SEM images in Region-

1 of failed tube: (a) extensive
spheroidization of the ferritic–
pearlitic microstructure, (b)
arrows showing distribution of
creep voids along the flue gas
side of tube, (c) magnified
image of the voids present at
grain triple junction on the flue
gas side, and (d) microstructure
along the steam side of the tube
where no creep voids are
detected

123
J Fail. Anal. and Preven. (2019) 19:792–801 797

Fig. 6 X-ray characteristic spectrum near voids and carbides in carbides at C1 and C2 revealing high concentration of Cr and Mo, (f)
Region-1 along flue gas side of the tube: (a) location of voids and Compositional mapping showing precipitates along grain boundaries
carbides, (b) and (c) EDS spectrum near voids V1 and V2 showing enriched in Cr, Mo, and C
high concentration of Cr and Mo, (d) and (e) EDS spectrum of the

Estimation of Operating Temperature Based on Steam Discussion

Side Oxide-Scale Thickness
From chemical analysis (Table 1), it is observed that no
The steam side oxide scale grew over low-alloy ferritic deviation of composition from the standard was detected.
steel (1–3 wt.% Cr) as function of temperature and time; Hence, failure due to discrepancy in composition of the
this can be expressed by the equation below [2], material was ruled out. Considerable decrease in hardness
Log X ¼ 6:8398 þ 2:83  104  T of the material from as specified in the standard (* 170
 ð13:62 þ logtÞ ðEq 1Þ HV) was observed in the Region-1 of steam side and
Region-1 of flue gas side (* 122 HV) (Table 2). The
where X = steam side oxide-scale thickness in mils hardness value in the Region-2 was * 108–109 HV,
(1 mm = 40 mils); T = mean operating temperature in °R which is substantially low as compared with the standard
(R = °F ? 460); t = service time in hours value (= 170 HV). This indicates that significant degrada-
The above correlation has been developed based on a tion has taken place in the tube material.
database with exposure times representing a small fraction The visual observation (Fig. 3) in the fractured region
of service life (\ 15%). However, it gives a qualitative showed the occurrence of thick-lip-type fracture, bulging,
estimation of the temperature. Considering the failure in and parallel cracks, which are indication of overheating of
this case: X = 1 mm = 40 mils; T = 9 months = 6480 h, the material [7]. The microstructure in Region-1 (Fig. 5)
the operating temperature became * 677 °C. shows that the material has undergone complete

123
798 J Fail. Anal. and Preven. (2019) 19:792–801

Fig. 7 (a) Scale thickness on

the outer or flue gas side of the
tube, (b) scale thickness on the
steam side of the tube, (c)
microstructure at the interface
between the scale and flue gas
side of the tube

spheroidization with marked increase in the size of carbide accelerated creep in the material. This was confirmed by
particles. The complete spheroidization of pearlite and the presence of large number of creep voids along the flue
coarsening of carbides occur due to rise in temperature gas side of the tube in near-fracture region (Figs. 9, 10)
above the normal operating temperature or long-term ser- [10]. After coalescence, the fully grown voids weaken the
vice exposure [8]. The presence of creep voids were found boundary of the grains such that the grain boundary sliding
along the flue gas side of the tube in Region-1. The pres- becomes dominant leading to creep failure under hoops
ence of creep voids along flue gas side is an indication of stress and consequent thinning of the tube. Thus, the creep
initiation of creep from the flue gas side of the tube [3]. mechanism operating in the fractured region can be
Excessive scale (* 1 mm) formation was also identified in sequentially illustrated as:
such short period of service (Fig. 7a, b). This can happen if Overheating of the tube ? creep void formation along
operating temperatures are higher than normal operating flue gas side of the tube ? void coalescence at grain
temperature. These evidences are indicative of high-tem- boundaries ? weakening of grain boundary ? intergran-
perature exposure of material during service that may have ular crack ? intergranular fracture under hoop’s
led to premature failure of the tube within a short period of stress ? rupture of the tube.
9 months. The operating temperature estimated based on steam
The nature of fracture in Region-2 (Fig. 11) was found side oxide-scale thickness comes to * 677 °C. But the
to be of intergranular type. This kind of fracture phe- tube material, i.e., SA213-T22 grade, is designed to with-
nomenon occurs when the grain boundaries become stand maximum temperature of * 540 °C [11]. Therefore,
relatively weak and are unable to bear the load/stress, it can be inferred that this superheater tube has been sub-
thereby breaking away [9]. It must have been caused from jected to localized overheating (Temperature  540 °C)
the weakening of grain boundary due to void formation and during its operation for 9 months, resulting in degradation
coalescence at elevated temperatures, leading to of tube material and, subsequently, it failure.

123
J Fail. Anal. and Preven. (2019) 19:792–801 799

Fig. 8 EDS mapping of the scales on (a) flue gas side of the tube and (b) steam side of the tube

Fig. 9 Microstructure at c.s. 2(a) in Region-2: (a) no voids present on inner or steam side of the tube, (b) no voids present at center, (c) presence
of voids detected on the outer or flue gas side of the tube

Summary and Conclusion inner and outer walls of the tube was observed. The pres-
ence of creep voids and their coalescence along grain
The composition of the tube material complied with the boundaries were noted. SEM fractography revealed that the
standard. Microstructural degradation was prominent in the mode of fracture was intergranular. Based on all the
failed superheater tube along with the reduction in bulk experimental evidence, it was postulated that the super-
hardness of the material. Thick oxide scale (* 1 mm) on heater tube failed within a short span of 9 months due to

123
800 J Fail. Anal. and Preven. (2019) 19:792–801

Fig. 10 Microstructure at c.s.

2(a) in Region-2, flue gas side
of the tube: (a) and (b) voids
coalescence and growth,
showing creep taking place

Fig. 11 Fractographs of
Region-2 showing intergranular
fracture: (a) and (b) upper lip of
the fracture, (c) lower lip of the
fracture, and (d) SEM image
just below the fractue surface

accelerated creep because of rise in metal temperature Acknowledgments The authors are thankful to Director, CSIR-
much above the operating temperature. Estimation of rise National Metallurgial Lab, for giving permission to publish this work.
in temperature based on steam side oxide-scale thickness
was made, and it has been found that the material was
References
exposed to temperature as high as 677 °C, which is much
above than the designed maximum operating temperature 1. D.N. French, Metallurgical Failure in Fossil Fired Boilers, 2nd
(540 °C) of the tube material grade. Based on above evi- edn. (Wiley, London, 1993)
dence, it can be concluded that due to localized 2. R. Viswanathan, Damage Mechanism and Life Assessment of
overheating, there was an accelerated creep in the material High Temperature Components (ASM International, Metals Park,
1989), p. 228
followed by rupture under hoop stress.

123
J Fail. Anal. and Preven. (2019) 19:792–801 801

3. D.R.H. Jones, Creep failures of overheated boiler, superheater remaining life assessment of aged components in thermal power
and reformer tubes. Eng. Fail. Anal. 11, 873 (2004) plants and petrochemical industries, vol. 2 (Jamshedpur, NML,
4. M.M. Rahman, J. Purbolaksono, J. Ahmad, Root cause failure 2008), p. 43
analysis of a division wall superheater tube of a coal-fired power 9. W.D. Pridemore, Stress-rupture characterization in nickel-based
station. Eng. Fail. Anal. 17(6), 1490–1494 (2010) superalloy gas turbine engine components. J. Fail. Anal. Prev. 8,
5. B.J. Benac, V.P. Swaminathan, Elevated-temperature life 281 (2008)
assessment for turbine components, piping, and tubing, in Failure 10. A.S.H. Makhlouf, M. Aliofkhazraei, Progressive failures of
Analysis and Prevention, ASM Handbook, vol. 11, ed. by W.T. components in chemical process industry: case history investi-
Becker, R.J. Shipley (ASM International, Metals Park, 2003), p. gation and root-cause analysis, in Handbook of Materials Failure
290 Analysis With Case Studies from the Chemicals, Concrete, and
6. S.K. Dhua, Metallurgical investigation of failed boiler water-wall Power Industries, p. 1
tubes received from a thermal power station. Eng. Fail. Anal. 11. K. Singh, Advances in materials for advanced steam cycle power
17(7), 1572–1579 (2010) plants. Bhel J. 27(2), 1–19 (2006)
7. S. Chaudhuri, R. Singh, High temperature boiler tube failures-
case studies, in Proceedings, COFA- NML Jamshedpur, 1997, p. Publisher’s Note Springer Nature remains neutral with regard to
107 jurisdictional claims in published maps and institutional affiliations.
8. B.B. Jha, S.N. Ojha, B.K. Misra, Residual life estimation of high
temperature tubings based on oxide scale thickness measurement,

123