jan amp features_am&p master template new QX6.

qxt 12/18/2012 5:38 PM Page 23

Managing Steam-Side Oxidation

and Exfoliation in USC Boiler Tubes
J.P.

S
uperheaters (SH) and reheaters (RH) in
modern supercritical (SC) and ultrasu- Epoxy
Shingledecker*
Electric Power percritical (USC) coal-fired steam
Research Institute, power plants operate with maximum steam
Charlotte, N.C. temperatures between 580 and 620°C and pres- Scale
sures up to 28 MPa. A single unit can contain
B.A. Pint* miles of tubing, therefore the formation
and A.S. Sabau (growth) of steam-grown oxide scales on the in-
Oak Ridge National side of this tubing and the potential for these
Alloy: T91 50 mm
Laboratory, scales to exfoliate during operation can have a
As polished
Oak Ridge, Tenn. significant impact on plant operations. In 2012,
National Physical Laboratory (NPL) and Elec-
A.T. Fry tric Power Research Institute (EPRI) held a joint
National Physical workshop to review recent experience in the
Laboratory, practical management of exfoliation, progress
Teddington, UK in understanding the phenomenon, and devel-
opment of models intended to provide the basis
I.G. Wright for improved approaches. This article briefly
TP347H 50 mm
Wright HT, describes some of the characteristics of oxide
Denver, Colo. growth and exfoliation, the impact of these ox- Fig. 1 — Structures of oxide scales (polished cross
ides in modern power plants, and highlights sections) observed on T91 alloy steel after about
findings from the workshop and requirements eight years of service at a steam temperature of
Materials are for additional research. approximately 541°C, and on TP347H stainless
steel after about six months service in steam at a
the key temperature of approximately 580°C. Alloy
enabling Oxide growth composition plays an important role in the
The growth and exfoliation of thermally structure/morphology of the oxide scale.
technology
grown oxides, or scales, in steam is a complex
driving Ni coating
phenomenon that depends on alloy composi-
development of tion, microstructure (including surface condi-
high-efficiency tion), temperature, pressure, and plant Fe2O3
power plants, operation. Steels and stainless steels used in SH Voids gap
and RH tubing have varying chemical composi- Fe3O4
which requires
tions with 2 to 25 wt% chromium and 0 to 25%
better nickel, and can be austenitic, bainitic, or
fundamental martensitic.
(Fe,Cr)3O4 + Fe3O4

and practical Figure 1 shows the difference in oxide scale

understanding morphology between martensitic T91 alloy
of the effects steel (9Cr-1Mo-V) and austenitic TP347H
stainless steel (18Cr-8Ni-Nb) after service in
of operating the superheaters of utility steam boilers. Scale
environments. on T91 forms as layers of uniform thickness,
and separation occurs between the main inner
and outer layers (with the outer layer lifting
off ). The inner layer of the essentially double-
layer scale formed on TP347H has a very irreg-
ular thickness, and while separation also occurs
between the two main layers, the outer layer ex-
foliates completely. Figure 2 shows the com-
plexity in the types of oxides and defects that
can form in a 9% Cr martensitic steel.
*Member of ASM Results of exfoliation
International When oxide scales fail (crack), separate, and
FeO + Cr2O3
Alloy 100 mm
Fig. 2 — Cross section of 9% Cr steel after
laboratory exposure for 10,000 hours in dry steam
at a temperature of 625°C showing the complexity
of the oxide scale.
are lost (exfoliate) from tube surfaces, there is
potential for tube overheating (thermal insula-
tion from a lifted scale), tube blockage from ac-
cumulation of oxides, and erosion damage to
downstream components. Figure 3a shows a
tube blocked from oxides that accumulated in
the lower loop of a superheater after a plant
shutdown. The blockage severely restricted
steam flow and caused a short-term overheat
tube failure during the unit restart. The figure

ADVANCED MATERIALS & PROCESSES • JANUARY 2013 23

jan amp features_am&p master template new QX6.qxt 12/18/2012 5:38 PM Page 24
(a) (b)
Fig. 3 — Tube blockage (a) and erosion damage to a valve stem (b), both
due to the presence of exfoliated steam-grown oxides in utility boilers.
also shows erosion damage to a valve stem due to exfoli-
ated oxides. A 2011 EPRI survey of U.S. utilities showed
that over 50% of respondents experienced exfoliation-re-
lated damage in their power plants including tube failures,
erosion of drain lines, reduced steam turbine performance
due to erosion of blading, and increased maintenance of
valve components.
Modeling the process
To minimize the likelihood of tube failures, manage
exfoliation, and select suitable materials for different
operating conditions, EPRI conducted modeling work
on the process of oxide growth and exfoliation taking
into consideration the accumulation of strain in the
oxide or oxide layers[1]. Strain accumulates during oper-
ational transients, and is a function of many factors in-
7th International Conference on
Advances in Materials Technology
for Fossil Power Plants
October 22-25, 2013
Hilton Waikoloa Village, Hawaii
Materials are the key enabling technology that drives the development
of high-efficiency power plants. To effectively achieve desired life, man-
age high-temperature degradation, and select proper materials for de-
manding high-temperature, high-pressure, and corrosive conditions,
improved fundamental and practical understanding is needed. EPRI sup-
ports and sponsors expert workshops, projects, and large group projects
to address key industry needs in many of these areas. This conference
brings many of these individuals, organizations, and EPRI collaborators
together to review the most recent advancements and help shape fu-
ture research.
This conference is the 7th in a series of conferences on the subject of
materials for advanced plants held every three years by EPRI. It is in-
tended to continue the promotion of information exchange between sci-
entists and engineers on an international scale. Invited keynote speeches
by prominent scientists, as well as submitted papers, will be presented.
Conference participants are likely to represent domestic and interna-
tional utilities, equipment manufacturers, alloy materials vendors, forge
shops, casting houses, universities, national laboratories, and consult-
ing/research organizations. For more details on the conference, visit the
calendar of events on www.epri.com.
EPRI teamed with ASM International to publishing proceedings of the
4th, 5th, and 6th conferences, and will collaborate again to publish pro-
ceedings of this 7th conference.
24 ADVANCED MATERIALS & PROCESSES • JANUARY 2013
Mean coefficient of thermal expansion (a), x10-6 C-1
24 Fe3O4
TP347
20
FeCr
spinel
16
Fe2O3
12
8
0 200 400 600 800 000 1200 1400
Temperature, °C
Fig. 4 — Coefficient of thermal expansion of various iron oxides
compared with that of TP347H stainless steel[4].
cluding scale chemistry and structure, physical proper-
ties of the scale, oxide growth stresses, creep, and phys-
ical tube dimensions. One major driver for strain
accumulation and eventual scale failure is the mismatch
in coefficient of thermal expansion (CTE) between
oxide layers and substrate alloys. Figure 4 shows the
CTE as a function of temperature for various iron ox-
ides compared with TP347 stainless steel.
Modeling oxide growth and strain accumulation
(which is particularly high during unit start-up and shut-
down events) enables identification of the point when
oxide exfoliation is expected to occur. Applying this
methodology to the operation of a superheater (includ-
ing temperature and heat-flux gradients) can provide
guidance on the amount of scale lost by exfoliation and
the potential for tube blockage. Figure 5 shows model-
ing results for a stainless steel superheater. Tube block-
age is predicted during the first two to three years of
operation (8000 to 20,000 hours), and the likelihood of
further blockages decreases with time assuming that ex-
foliated scale is removed after each shutdown event (i.e.,
the scale is blown through the system), which is consis-
tent with field experience.
EPRI-NPL workshop findings
Challenges with steam-side oxide growth and exfolia-
tion management are being experienced worldwide, and
an “experts workshop” was held at NPL in January 2012[2]
to bring together practical experience and current research
to identify key knowledge gaps and research needs. A
number of issues were identified including:
• Boilers with nominally the same materials and
operating conditions had vastly different experiences
with exfoliation, and the role of boiler design has not
been clearly identified.
jan amp features_am&p master template new QX6.qxt 12/18/2012 5:38 PM Page 25

• In the case of modeling, the failure criteria for

1
exfoliation need further refinement, and scale defect Outlet steam temperature, °C

Predicted blocked area fraction

sizes that are key to developing failure criteria are
currently unknown. 0.8
• Many discrepancies exist between laboratory test
results and those obtained from field experience. This
is partly due to difficulties in simulating some key 0.6
aspects of plant operation in laboratory testing, and
to the fact that mass gain and not oxide thickness are 0.4
often reported from laboratory examinations. This is Porosity of deposit
critical because laboratory testing is currently being = 25%
used to validate materials for future higher- 0.2
temperature advanced USC steam designs.
• There may be a need to define temperature use 0
limits for alloys based not only on high-temperature 0 10,000 20,000 30,000 40,000
Time, h
strength, but also on exfoliation considerations.
Based on this feedback, a major outcome of the work- Fig. 5 — EPRI modeling results for exfoliation blockage in a stainless steel
shop was the need for development of an atlas of steam- superheater, assuming six-month (~4000 h) operational cycles and
grown oxide structures containing the necessary details to assuming removal of all exfoliated scale after each shut-down event (shown
by data points). The propensity for tube blockage decreases after two years
provide practical diagnostic guidance to power-plant oper-
of operation, which is consistent with industry experience.
ators dealing with oxide exfoliation, a consistent set of data
to use for modeling activities, and accepted field mi- Fig. 6 —
crostructures for laboratory researchers to validate their Epoxy Secondary
electron
experimental approaches. EPRI and Oak Ridge National image (top)
Laboratory have started this effort[3]. Inner oxide and
characteristic
New research x-ray images
(bottom) for
Work also continues through the use of advanced char- Fe, Cr, and O
acterization techniques to further clarify mechanisms and in an oxide
long-term stability of steam-grown oxides scales. Figure 6 347H formed on a
29,672 h at 582oC stainless steel
shows an electron probe secondary electron image and a superheater
color compilation of Fe, Cr, and O x-ray maps from the tube. Regions
inner scale formed on a stainless steel superheater after of Cr
~3.5 years of service in a SC steam boiler. The formation of enrichment
10 mm (light blue)
a chromium rich scale is expected to slow the oxide growth and depletion
rate significantly, but the x-ray map clearly indicates re- (red) are
gions of oxide where no Cr-rich oxide layer has formed ad- observed.
jacent to the metal and regions in the metal with
chromium depletion. Typically, a uniform Cr-rich layer is
present at the reaction front, and this result suggests there
is nonuniform growth, perhaps initiated by Cr depletion.
Measurements of these various layers with time will pro-
vide a more detailed picture of the reaction mechanism
during long-term service in supercritical steam.
Cr-depleted alloy
References
1. A.S. Sabau, I.G. Wright, and J.P. Shingledecker, Oxide scale
exfoliation and regrowth in TP347H superheater tubes, Ma- Cr-rich oxide
terials and Corrosion, 63 (10), p 896–908, 2012. 10 mm
2. Second EPRI-NPL Workshop on Scale Exfoliation From
Steam-Touched Surfaces, Proc. Workshop held at Bushy
House, National Physical Laboratory, Teddington, London,
January 17-18, 2012; EPRI, Palo Alto, Calif.: 1026663. spalling of oxide scale, Materials for Advanced Power Engi-
3. B.A. Pint, J.P. Shingledecker, and I.G. Wright, Characteriza- neering, 21, p 1169-1179, 2002.
tion of Steam Oxidation Products from Field-Exposed Tubes,
in Advances in Condition and Remaining Life Assessment for For more information: John Shingledecker, Ph.D., is senior
Fossil Power Plants, EPRI, Charlotte, N.C., presented Hilton project manager, Fossil Materials & Repair Program (P87),
Head, S.C., Oct. 2012. Electric Power Research Institute (EPRI), Charlotte, N.C.,
4. S.J. Osgerby and L.N. McCartney, Steam oxidation of 9- tel: 704/595-2619; email: jshingledecker@epri.com; www.epri.
12Cr martensitic steels: characterisation and modeling the com.

ADVANCED MATERIALS & PROCESSES • JANUARY 2013 25