Degradation of Mechanical Properties of

Cr-Mo-V and 2.25Cr-1Mo Steel Components After

Long-Term Service at Elevated Temperatures
N.S. C H E R U V U

The influence of operating temperature on in-service degradation of mechanical properties of

high temperature stream turbine components has been investigated. Material samples for this
study were taken from a Cr-Mo-V rotor and several 2 . 2 5 C r - l M o cast steel components which
had operated over 200,000 hours. The test results revealed that the degree of in-service deg-
radation of strength, toughness, and the fracture appearance transition temperature of both steels
were very sensitive to the service temperature. Both steels softened only when they were ex-
posed at a temperature greater than 454 ~ (850 ~ and the degree of softening increased with
further increase in service temperature. In Cr-Mo-V steel, the loss in strength was accompanied
by an improvement in ductility and toughness. Despite softening of 2.25Cr-1Mo steel in service,
elevated temperature exposure resulted in a marked decrease in ductility and toughness. The
loss of toughness in this steel was in part irreversible. In contrast, a severe increase in fracture
appearance transition temperature, due to reversible temper embrittlement, occurred in both
steels at a service temperature of around 427 ~ (800 ~ but not at the highest service tem-
perature. In fact, the Cr-Mo-V steel did not temper embrittle as a result of service exposure at
the highest operating temperature investigated. These results are rationalized in terms of changes
in microstructure and grain boundary chemistry that occur in service as a function of operating
temperature.

I. I N T R O D U C T I O N of impurities such as P, Sn, As, and Sb to the grain

boundaries. It is now well established that segregation
Low alloy ferritic steels of the Cr-Mo-V and C r - M o of impurities at the grain boundaries temper embrittles
type are extensively used for high temperature applica- these steels. The segregation-induced temper embrittle-
tions in the power, chemical, and oil industries, for their ment is manifested as an increase in the transition
excellent elevated temperature strength and good resis- temperature and is reversible. The changes in micro-
tance to oxidation and hydrogen embrittlement. These
structure, such as carbide coarsening and precipitation
steels are used in the power industry for components such
of more stable carbides during service, can cause soft-
as high pressure (HP) and intermediate pressure (IP) fos- ening and irreversible embrittlement. The carbide-
sil steam turbine rotors, cylinders, blade rings, nozzle
induced embrittlement is manifested as a decrease in the
chambers, steam chests, steam pipes, boilers, and head- impact toughness without affecting the ductile-to-brittle
ers. The inlet and exhaust temperatures of a typical fossil transition temperature t4'5'6] and is irreversible, unlike
HP or IP turbine element are 538 ~ (1000 ~ and 288 ~
segregation-induced embrittlement.
(550 ~ respectively. Thus, the steam turbine compo-
Recent investigations by several investigators reveal-
nents operate in the temperature range where creep and ed that Cr-Mo, tS~ 2 . 2 5 C r - l M o , t7'81 and Cr-Mo-V rotor
temper embrittlement is a concern. Though Cr-Mo-V and steels t9'~1~ had softened significantly as a result of ele-
Cr-Mo steels were believed to be immune to temper em- vated temperature exposure. Zhe e t a l . t1~ studied the in-
brittlement, the results generated over the last decade in- fluence of operating temperature on the hardness of a
dicate that they are highly susceptible to temper
rotor, which was referred to as Buck rotor, after sev-
embrittlement, tl,2,31 Thus, prolonged service exposure of
enteen years of service and reported that the steel had
these turbine components leads to degradation of their
softened when the service temperature was greater than
mechanical properties. Hence, the remaining service life
308 ~ (586 ~ However, this temperature is too low
of these components is limited by the extent of in-service to cause any significant changes in microstructure during
material degradation that has occurred. Evaluation of in-
service since the rotor forgings are normally tempered,
service material degradation after long-term service is
prior to placing them in service, at a considerably higher
therefore necessary for meaningful remaining service life temperature than the temperature at which onset of soft-
assessment. ening was reported. The observed m a x i m u m difference
When low alloy steel components are exposed to el- in hardness between the hot and cold locations of the
evated temperature during service, their properties de- Buck rotor was only 5 Rc points. Generally, a variation
teriorate due to changes in microstructure and diffusion of 70 MPa (10 ksi) in tensile strength, or 5 Rc points in
hardness, from the one end to the other is not atypical
in 1950 vintage rotor forgings, t~2] Hence, it was not clear
N.S. CHERUVU is Senior Engineer, Power Systems Division,
whether the observed variation in hardness between the
Westinghouse Electric Corporation, The Quadrangle, University Blvd.,
Orlando, FL 32826. cold and hot locations was due to softening or due to
Manuscript submitted April 22, 1988. variation in the original properties of the Buck rotor.

METALLURGICAL TRANSACTIONS A VOLUME 20A, JANUARY 1989--87

 Furthermore, the results of Nishizaka et al.~SJ and Qu Table I. Operating Metal Temperatures of Test
and Kuo [6] on service-exposed Cr-Mo and 2.25Cr-lMo- Material Locations of the HP-IP Rotor
0.25V steels have shown that these steels embrittle due
Service
to precipitation or change in morphology of carbides Temperature
during service. It was not known from the studies of Zhe
et al. ~1~ and Cheruvu et al. Bq whether or not softening Location ~ ~
of a 1Cr-lMo-0.25V rotor steel during service was as- HP dummy (hot) 985 529
sociated with precipitation-induced embrittlement. Ring 1 850 454
The results t9-1~'~3~ of service-exposed Cr-Mo-V rotors Ring 2 975 524
further revealed that these steels displayed the greatest Ring 3 950 510
susceptibility to reversible temper embrittlement in re- Ring 6 870 466
Cold 550 288
gions where the service temperature was around 427 ~
(800 ~ Though Cr-Mo steel turbine cylinders operate
over a range of temperatures, the influence of service operated at temperatures of 454 ~ (850 ~ 427 ~
temperature on temper embrittlement behavior of these (800 ~ and 538 ~ (1000 ~ respectively.
steels is not known. Therefore, the present study was The rotor steel was austenitized at 1010 ~ (1850 ~
undertaken to evaluate systematically the influence of air cooled, tempered at 666 ~ (1230 ~ and then fur-
service temperature on the softening and embrittlement nace cooled from the tempering temperature. A review
behavior of Cr-Mo-V and 2 . 2 5 C r - l M o steels: The pri- of original test results revealed that the rotor had uniform
mary objectives of this study were: tensile properties from the one end to the other (Table
(I) To determine the m a x i m u m service temperature up II). The original heat treatment details of 2.25Cr-lMo
to which no softening occurs during service for both steels. castings are not available. However, all three castings
(II) To establish susceptibility of these steels to precip- were made to the same specification in 1953. The spec-
itation or carbide-induced embrittlement. ification applicable at that time indicated that these cast-
( l i d To identify the service temperature at which severe ings were normalized and tempered at a temperature
temper embrittlement occurs in 2 . 2 5 C r - l M o steel. greater than 665 ~ to produce minimum tensile and yield
To identify reversible and irreversible embrittlements, strengths of 552 and 276 MPa (80 and 40 ksi), respec-
impact tests in the as-service-exposed and de-embrittled tively. The chemical composition of the test materials is
conditions, fractographic characterization of Charpy shown in Tables III and IV. The chemistry of the casting
specimen fracture surfaces and Auger analysis of frac- used to generate the A S T M data base t141 is also given in
ture surfaces for grain boundary chemistry were carried Table IV for comparison since original data for the sub-
out in this study. ject 2 . 2 5 C r - l M o components are not available and this
casting is referred to as the ASTM casting. The chem-
istry and tensile properties of the A S T M casting com-
II. EXPERIMENTAL PROCEDURE plied with the specification requirements. The chemistry
A . Material and Microstructure
of all components reasonably agrees with that for the
A S T M casting.
Material for this study was taken from a combined HP The microstructure of the rotor steel was essentially
and IP Cr-Mo-V rotor, and 2.25Cr-lMo steel flow guide, upper bainite with an average prior austenite grain size
blade ring, and nozzle chamber castings, which had op- of A S T M # 3 . The microstructure of the three castings
erated over 200,000 hours. From the rotor, the IP blade exhibited a duplex structure containing ferrite and bain-
groove walls 1 through 6 were machined as rings and ite, which was a typical structure for normalized and
were identified as rings 1 through 6. Rings 1, 2, 3, and tempered 2.25Cr- 1Mo steel. Typical microstructures for
6 were used for this study. In addition to matching the Cr-Mo-V and 2.25Cr-1Mo steels after service are shown
rings, trepan samples were also drilled from the HP in Figure 2.
d u m m y (hot) and cold locations of the rotor. The test
material locations are shown in Figure 1, and service B. Test Conditions
temperature at these locations is given in Table I. The
flow guide, blade ring, and nozzle chamber castings were Tensile specimens with a gage length of 35.6 m m
(1.4 in.), a diameter of 9.1 m m (0.36 in.), and standard
size V-notch Charpy specimens were machined from each
test location of the rotor, flow guide, blade ring, and
nozzle chamber. A portion of the Charpy samples ma-
chined from Rings 2, 3, and 6, the flow guide, and the
nozzle chamber were given a de-embrittling treatment of
638 ~ (1180 ~ for 1 hour and then cooled. Combi-

co~~! v~o~ nation smooth bar-notch bar type creep rupture speci-
mens, with a theoretical stress concentration factor of
approximately 4 at the notch, were machined from the
hot and cold locations of the rotor and the three cast-
ings. The dimensions of a combination creep rupture
specimen were given elsewhere. 1151 These specimens
Fig. 1 - - T e s t sample locations in the HP-IP rotor. were utilized for creep rupture tests to evaluate potential

88--VOLUME 20A, JANUARY 1989 METALLURGICAL TRANSACTIONS A

 Table II. The Original Tensile Properties at Various Locations of the HP-IP Rotor
Tensile Strength Yield Strength
Location MPa Ksi MPa Ksi Pet Elong. Pet R A
HP d u m m y (hot) 938 136 786 114 12 25
Near IP blade groove walls 931 135 779 113 12 26
Cold location 945 137 800 116 13 25

Table III. Chemistry of the HP-IP Rotor (Wt Pet)

C Mn Si Ni Cr Mo V S P Sn As Sb
After service 0.33 0.73 0.22 0.11 1.05 1.15 0.25 0.026 0.031 0.011 0.007 0.001
Original test report 0.31 0.70 0.19 0.18 1.07 1.35 0.24 0.026 0.030 -- -- --
Rotor forging heat treatment for properties: austenitized at 1010 ~ (1850 ~ cooled; tempered at 666 ~ (1230 ~ cooled.

Table IV. Chemistry of the Stationary Components ( W t Pet)

Component C Mn Si Ni Cr Mo S P Sn As
F l o w guide 0.16 0.57 0.40 0.07 2.60 0.74 0.015 0.015 0.014 0.017
Blade ring 0.16 0.60 0.44 0.07 2.63 0.76 0.020 0.016 0.013 0.016
Nozzle chamber 0.11 0.58 0.48 0.07 2.22 0.68 0.035 0.021 0.021 0.015
A S T M castings tl41 0.17 0.56 0.40 -- 2.32 0.93 0.036 0.030 -- --
Heat treatment: All castings were normalized and tempered.

Fig. 2--Microstmctures of (a) cold location of Cr-Mo-V rotor-exposed at 288 ~ (b) 2.25Cr-lMo flow guide-exposed at 454 ~ (c) 2.25Cr-
1Mo blade ring-exposed at 427 ~ (d) 2.25Cr-lMo nozzle chamber-exposed at 538 ~

METALLURGICAL TRANSACTIONS A VOLUME 20A, JANUARY 1989--89

notch-sensitive behavior of the subject steels. All tensile 250 350
~
450
--140
specimens were tested at room temperature. The Charpy
specimens were subjected to impact testing to determine
TENSILE STRENGTH
the 50 pct fracture appearance transition temperature
90C- 130
(FATT). Approximately 6 to 10 specimens were utilized
g
to generate a full Charpy curve. The creep rupture spec- v
imens were tested at different temperatures and stresses. z
~_ 825. 120 5o
A few specimens from the castings were also tested at
f=
a stress of 110 MPa (16 ksi) and different temperatures
YIELD STRENGTH 0 0 9
i
since these results were successfully used for remaining m 110 ~o ~
F- 750
service life assessment of components. [16,~7~ g
I
C. Fractography and Prior Austenite 100
~. 675
Grain Boundary Chemistry
[ ~
~ ~
REDUCTION
OAREA F
9 DII
!
The fracture surfaces of selected broken Charpy spec- 90 20
imens tested at room temperature were examined using 600
a Cambridge Scanning Electron Microscope (SEM) Model ELONGATION

5150 B to determine the fracture mode. In addition, a I

400 6O0 800 1000
~
circumferentially notched cylindrical specimen was ma- OPERATING TEMPERATURE, ~
chined from each of these Charpy specimens, and frac-
Fig. 3--Variation of tensile properties of Cr-Mo-V rotor steel with
tured in the ultra-high vacuum chamber of the Auger service temperature. The solid symbols denote the original properties"
Spectrometer at room temperature. Each fracture surface of the rotor steel at the corresponding locations.
was analyzed in the chamber for grain boundary chem-
istry at ten spots with a primary beam size of 8 /zm.
able, the present results are compared with those re-
III. RESULTS ported for an unexposed ASTM casting. I14] The tensile
properties of the flow guide and blade ring castings after
A. Tensile Properties service were higher than those of the unexposed steel,
1. Cr-Mo-V steel suggesting that the original properties of the subject cast-
ings were above the minimum specified values. These
The influence o f operating temperature on tensile
results further suggest that 2.25Cr-lMo steel does not
properties of the rotor steel is shown in Figure 3. The
soften when it is exposed at or below 454 ~ (850 ~
tensile and yield strengths were almost unaffected as a
The yield and the tensile strengths of the nozzle chamber
result of service exposure at temperatures up to approx-
steel casting were, however, significantly lower than that
imately 454 ~ (850 ~ and then dropped with further
increasing service temperature. However, the ductility, of flow guide and blade ring castings, indicating that this
steel had softened due to service exposure at 538 ~
as measured by pet elongation and reduction of area, de-
(1000 ~ In contrast to the results of Cr-Mo-V steel,
creased slightly with increasing temperature up to about
the pet elongation and reduction of area of the nozzle
454 ~ (850 ~ and then sharply increased with increas-
ing temperature. chamber casting were not improved as a result of soft-
ening. Both pct elongation and reduction of area dropped
A comparison of results of the specimens obtained from
somewhat as compared to the results of the other two
the hot location before (Table II) and after service (Fig-
service-exposed castings.
ure 3) revealed that the tensile and yield strengths de-
creased from 939 MPa (136 ksi) to 855 MPa (124 ksi),
and 786 MPa (114 ksi) to 686 MPa (99.5 ksi), respec- B. Creep Rupture
tively, and the reduction of area increased from 24.8 pet 1. Cr-Mo-V steel
to 39.1 pet (by 56 pet) due to service exposure. Creep rupture results of Cr-Mo-V steel are summa-
2. 2.25Cr-lMo steel rized in Table VI. All creep rupture specimens obtained
The results of room temperature tensile properties of from both cold and hot locations failed first in the notch,
2 . 2 5 C r - l M o steel castings after service are summarized indicating that the rotor steel was notch sensitive. These
in Table V. Since the original properties are not avail- results suggest that the notch sensitivity of the steel was

Table V. Comparison of Room Temperature Tensile Properties of 2.25Cr-lMo Steel Components

After Service with Unexposed ASTM Casting
Temperature
Reduction
Exposure Tensile Strength Yield Strength Elongation of Area
Component ~ ~ MPa Ksi MPa Ksi Pet Pet
Flow guide 454 850 578.8 84.0 405.1 58.8 31.2 69.4
Blade ring 427 800 579.5 84.1 397.6 57.7 30.2 68.1
Nozzle chamber 538 1000 516.8 75.0 250.1 36.3 29.5 55.6
Unexposed ASTM castings I~< -- -- 532.0 80.0 375.0 53.0 27.0 51.0

90--VOLUME 20A, JANUARY 1989 METALLURGICAL TRANSACTIONS A

 Table VI. Creep Rupture Results of Cr-Mo-V Rotor Steel
Rupture Time
Hours
Test Temperature StresS Notch Smooth
Location ~ ~ MPa Ksi Bar Bar Pct Elong. Pct RA
Cold 510 950 345 50 1197 3375 0.6 0*
524 975 324 47 817 3896 0.1 1.0
538 1000 310 45 868 3856** -- --
566 1050 234 34 1192 5294 0* 0*
Hot 510 950 345 50 1811 2532 5.9 9.0
510 950 345 5O 1384 2095 7.5 15.4
524 975 345 47 762 1767 3.8 7.8
538 1000 310 45 705 1585 4.0 6.8
566 1050 234 34 1010 2393 4.3 7.0
*No measureable change in dimensions was noted.
**Not failed.

not due to elevated temperature exposure during service, tent with the tensile ductility results, the creep rupture
but can be attributed to the temperature at which the ro- ductility of the rotor steel also i m p r o v e d c o n s i d e r a b l y as
tor was austenitized.t~8,19~ F o l l o w i n g the initial notch fail- a result of service exposure at 529 ~ (985 ~
ure, all these specimens were tested for a smooth-bar
failure. 2. 2 . 2 5 C r - l M o steel
Several interesting observations can be m a d e from the The creep rupture results o f 2 . 2 5 C r - l M o cast steel
creep rupture results. First, service exposure at 530 ~ c o m p o n e n t s are s u m m a r i z e d in T a b l e VII. In this case,
(985 ~ reduced s m o o t h - b a r rupture life by approxi- all specimens failed first in the gage section, suggesting
mately a factor of two, without significantly affecting that these steels were not notch sensitive. F o l l o w i n g the
notch bar rupture life, as c o m p a r e d to that of specimens initial failure in the s m o o t h - b a r section, no tests were
from the cold location. A l t h o u g h service exposure led to conducted on the notch section. In a sharp contrast to
softening at the hot location, the notch rupture life was C r - M o - V results, these steels exhibited good creep rup-
not increased as a result of softening. H o w e v e r , consis- ture ductility; reduction of area of all specimens was in

Table VII. Creep Rupture Results of 2.25Cr-lMo Steel Castings

Test Rupture
Temperature Stress Time,
Component ~ ~ MPa Ksi Hours Pct Elong. Pct RA
Flow guide 482 900 241 35 185 34.8 79.3
510 950 193 28 409 32.1 81.8
538 1000 145 21 1113 34.4 86.6
566 1050 110 16 2074 31.5 88.9
593 1100 82 12 7043 -- 82.6
593 1100 110 16 517 45.0 84.6
579 1075 110 16 1203 -- 87.2
566 1050 110 16 3144 43.2 86.1
552 1025 110 16 7019 21.5 85.2
538 1000 110 16 9000 not failed --
Blade ring 482 900 241 35 415 27.1 78.9
510 950 193 28 453 37.1 83.1
538 1000 145 21 1826 47.9 86.5
566 1050 110 16 3623 44.7 90.3
593 1100 82 12 8155 -- 87.6
Nozzle chamber 552 1025 110 16 649 49.2 84.9
566 1050 110 16 366 54.1 88.9
566 1050 110 16 320 55.6 90.3
579 1075 110 16 127 54.5 87.3
593 1100 110 16 63 43.4 88.0
538 1000 110 16 2311 53.7 83.5
579 1075 90 13 1069 47.1 89.9
566 1050 90 13 2197 44.6 88.0
593 1100 69 10 4531 50.4 92.4
621 1150 48 7 5986 35.5 89.9

METALLURGICAL TRANSACTIONS A V O L U M E 20A, J A N U A R Y 1 9 8 9 - - 9 1

the range of 79 to 92 pct. No apparent differences in
creep ductility were noted among the three castings.
Figure 4 is a plot of the logarithm of stress vs a time- FLOW GUIDE
EXPOSED AT 4540C (850~
temperature (Larson-Miller) parameter for three cast- 625
1150
ings. These results showed that the creep rupture strength NOZZLE CHAMBER
[ ] EXPOSED AT 538~ (1000~
of the flow guide and blade ring castings were almost
identical. The nozzle chamber steel casting had signifi- 600
O
o
1100
cantly lower creep rupture strength as compared to the
other two castings. This drop in rupture strength was not P-
575
accompanied by an improvement in creep rupture duc- ==
1050
tility. The lower creep rupture strength of the nozzle
chamber casting can be explained in terms of softening 550
that occurred in service.
1000
Variations of rupture time with temperature at an ap-
plied stress of 110 MPa (16 ksi) for the flow guide and 525

nozzle chamber castings are illustrated in Figure 5. At

950
all test temperatures investigated, the specimens from the
nozzle chamber exhibited an order of magnitude and I I I
100 1000 10,000
shorter rupture life than those from the flow guide. The
T I M E TO F A I L U R E , H O U R S
implication of these results is that the nozzle chamber
has a shorter remaining service life at any given tem- Fig. 5--After service rupture life of 2.25Cr-lMo cast steels as a func-
perature than the other two cast components. tion of test temperature at 16 ksi (110 MPa).

C. Impact Properties
ture toughness results reported by CheruvuJ 9] Further-
1. Cr-Mo-V steel more, it is interesting to note that FATT of the samples
Variation of FATT and upper shelf energy of the machined from the hot location exposed at the highest
Cr-Mo-V rotor steel as a function of service temperature temperature was lower than that of cold location, and
is shown in Figure 6. The F A T T increased with increas- this drop in F A T T was, in part, due to softening. The
ing operating temperature up to approximately 427 ~ improvement in upper shelf energy of the samples from
(800 ~ reached a m a x i m u m value, then sharply de- the hot location can also be attributed to the drop in ten-
creased with further increase in temperature. The upper sile strength.
shelf energy exhibited a reverse trend of FATT; it de- The FATT and upper shelf energy of the samples in
creased with increasing temperature to around 427 ~ the de-embrittled condition are also shown in Figure 6.
(800 ~ and then increased with further increase in tem- The de-embrittling treatment lowered the FATT of the
perature. The improvement in upper shelf energy above samples from Ring 6 from 249 ~ (480 ~ to 168 ~
this temperature is consistent with the plane strain frac- (335 ~ suggesting that the embrittlement that occurred
in service was reversible. In addition, the FATT and up-
T (K) (20 + log tt) x 10 .3 per shelf energies of the specimens from the hot location
18 20 22
in the service-exposed condition were comparable to those
of de-embrittled specimens.
60 2.25 Cr - 1 M o STEEL CASTINGS
A FLOW GUIDE
Figure 7 shows the typical fracture mode of broken
9 B L A D E RING Charpy specimens from the cold location, Ring 1, and
[] NOZZLE CHAMBER 320
the hot location. The fracture mode of the specimens from
Ring 1, which was exposed at 454 ~ (850 ~ was

~
204 300 400 50o 590
160
500 TRANSITION UPPER SHELF
20 TEMPERATURE ENERGY
250 OSERVICE EXPOSED 9 V SERVICE EXPOSED
9 DE-EMBRI'i'FLE D 9 DE-EMBRITrLED
# _~ 45O

80
i 400 9
10 2oo

i? ~ E4

O t=
m \\ = 40 w , ~
40 -= aso 99 -Z~ "4, t ~tu
150 300 ,17 o~ u ~z

3'0 ~'~, 3'~ 3'5 ~'~ 27 t=

T (R) (20 + l o g tf) x 10 -3 25~ 600 800 1000
LARSON - M I L L E R PARAMETER OPERATING TEMPERATURE, ~

Fig. 4 - - C r e e p rupture strength of 2.25Cr-lMo cast steel after Fig. 6--Variation of FATT and upper shelf energy of Cr-Mo-V steel
service. with service temperature.

92--VOLUME 20A, JANUARY 1989 METALLURGICAL TRANSACTIONS A

Fig. 7--Scanning electron micrographs showing variation of fracture mode in Cr-Mo-V steel samples from (a) cold location-exposed at 288 ~
(b) Ring l--exposed at 454 ~ (c) hot location--exposed at 529 ~

predominantly intergranular (Figure 7(b)). The speci-

mens from both the cold and hot locations exhibited a
primarily transgranular cleavage mode. Typical Auger
spectra of the specimens from the cold location and Ring
1 are compared in Figure 8. The spectrum of the former
specimen represented matrix composition, since the
fracture mode was transgranular. In the latter specimen, a) C O L D L O C A T I O N
- TRANSGRANULAF
a considerable amount of phosphorus, molybdenum, and
carbon was noted on the grain boundaries. The average
phosphorous concentration on the grain boundary was
calculated to be 3.8 at. pct from the peak height ratio of
phosphorus to iron. The higher molybdenum and carbon Fe

peaks in the lower spectra (Figure 8) presumably are due

to presence of carbides on the grain boundaries. The tin
l
concentration on the grain boundaries was below the de-
tectable limit of the instrument. These results suggest
that embrittlement that occurred in service at Ring 6 lo-
cation was due to segregation of phosphorus alone to the b) R I N G 1
- INTERGRANULAR
grain boundaries.
2. 2.25Cr-1Mo steel
The results of impact properties of 2.25Cr-lMo steel
Fe ~-e
in the service-exposed and de-embrittled conditions are
summarized in Table VIII. The FATT and room tem-
perature, and upper shelf impact energies of the nozzle E (eV) ~.

chamber casting in the service-exposed condition were Fig. 8 - - A u g e r spectra of Cr-Mo-V samples from (a) cold location;
lower than those of both the flow guide and blade ring. (b) Ring 1.

METALLURGICAL TRANSACTIONS A VOLUME 20A, JANUARY 1989--93

 Table VIII. Impact Properties of 2.25Cr-lMo Cast Steels

Room
Exposure- Temperature Upper Shelf Upper Shelf
Temperature Impact Energy Temperature Energy 50 Pct FATT
Component ~ ~ Joules Ft/Lbs ~ ~ Joules Ft/Lbs ~ ~
Flow guide 454 850 22 16 149 300 152 112 93 200
Flow guide
de-embrittled -- -- 132 97 93 200 174 129 4 40
Blade ring 427 800 29 21 204 400 156 115 132 270
Nozzle chamber 538 1000 8 6 204 400 111 82 82 180
Nozzle chamber
de-embrittlement -- -- 27 20 204 400 81 60 57 t35

Consistent with tensile ductility, these results suggest that the Cr-Mo-V rotor steel softens during service is signif-
the toughness of the nozzle chamber casting decreased icantly higher than what they have reported. The exis-
as a result of service exposure. The blade ring which was tence of minimum exposure temperature for softening to
exposed at 427 ~ (800 ~ exhibited the highest tran- occur during service is related to the higher tempering
sition temperature; however, the de-embrittling treat- temperatures normally used for these steels.
ment at 638 ~ (1180 ~ significantly lowered the FATI" The main factors affecting the tensile and creep strengths
and increased the impact energy of the flow guide steel, of a low alloy ferritic steel are the morphology and type
suggesting that the embrittlement that took place in of carbides, and the amount of alloying elements present
service was reversible. On the other hand, the de- in the ferrite. The carbide morphology and alloying ele-
embrittling treatment slightly lowered the FATT of the ments in solid solution control movement of dislocations
nozzle chamber steel from 82 ~ (180 ~ to 57 ~ in the ferrite grains. I2~ It has been suggested that the
(135 ~ Therefore, the embrittlement which occurred creep strength of these steels decreases as the carbides
in this steel during service was predominantly irrevers- grow and lose coherency with the matrix, and is also
ible. Though the nozzle chamber was exposed at 538 ~ affected by the level of molybdenum content in the ma-
(1000 ~ and had softened significantly, the softening trix; t21,221 the strength is directly related to the molyb-
did not improve impact toughness of this steel, as it did denum content. Williams and Wilshire I23~ have further
in Cr-Mo-V steel. shown that the hardness of a 0.5Cr-0.5Mo-0.25V steel
The broken Charpy specimens machined from the flow before and after service at elevated temperatures de-
guide and blade ring in the service-exposed condition creases with increasing inter-carbide particle spacing and
exhibited intergranular fracture mode, while the one from with decreasing alloying elements in the solid solution.
the nozzle chamber failed in cleavage and dimple modes Though the influence of service exposure on the micro-
of fracture. As expected, the de-embrittling treatment structure is not evaluated in this study, there is evidence
changed the fracture mode from intergranular to trans- in the literature to show that service exposure at elevated
granular in the flow guide and blade ring steels. Typical temperatures of 454 ~ (850 ~ and above results in
fractographs are shown in Figure 9. Auger spectrum ob- coarsening of carbides, 123,24] spheroidization of car-
tained from an intergranular fracture surface is compared bides, E7,81an increase in inter-particle spacing, t231and pre-
with the transgranular fracture of the specimen from the cipitation of more stable carbides E3-24j in the Cr-Mo-V
nozzle chamber in Figure 10. The spectrum of the for- and Cr-Mo type steels. Recent studies by several inves-
mer fracture surface showed considerable amounts of tigators [6,8,241 have shown service exposure causes pre-
phosphorus and tin on the prior austenite grain bound- cipitation of acicular carbide needles (M2C) and spherical
aries. The calculated phosphorus and tin on the bound- carbide particles (M6C) and H-type carbides. Precipita-
aries were 5.6 and 0.7 at. pct, respectively. These results tion of these carbides during service depletes the alloy-
suggest that segregation of phosphorus and tin to the grain ing element content in the ferrite matrix. In fact, the
boundaries in service was responsible for the observed chemical analyses of the ferrite phase of service-exposed
embrittlement of the flow guide and blade ring castings. Cr-Mo TMand Cr-Mo-V 161 steels have revealed a signifi-
cant decrease in molybdenum and carbon contents in sol-
id solution after service exposure. The observed softening
of Cr-Mo-V rotor steel and the 2.25Cr-lMo nozzle
IV. D I S C U S S I O N chamber casting in this study can be attributed to changes
in microstructure and depletion of alloying elements in
A. Softening
solid solution that occurred in service. Since these mi-
The results presented in the previous section demon- crostructural changes are thermally activated processes,
strate that both Cr-Mo-V and 2 . 2 5 C r - l M o steel soften the degree of softening is directly related to the service
when they are exposed at a temperature greater than 454 ~ temperature. From the present mechanical property re-
(850 ~ Contrary to the observations of Zhe et al.,tl~ sults, it can be inferred that carbides produced during
the results of this study show no appreciable softening the original high temperature tempering are extremely
in the samples taken from Ring 1 after service exposure, stable during long-term service at temperatures up to
suggesting the minimum exposure temperature at which 454 ~ (850 ~

9 4 - - V O L U M E 20A, J A N U A R Y 1989 M E T A L L U R G I C A L TRANSACTIONS A

Fig. 9--Fracture mode of 2.25Cr-lMo steel specimens from (a) the flow g u i d e - - a s service-exposed at 454 ~ (b) the flow guide service-exposed
and de-embrittled; (c) blade r i n g - - a s service-exposed at 427 ~ (d) nozzle chamber as service-exposed at 538 ~

B. Carbide-Induced Embrittlement A simple mass calculation indicates that the carbon

content of Cr-Mo-V steel is twice the amount that is
In contrast to the results for Cr-Mo-V steel, softening needed to precipitate all vanadium, molybdenum, and chro-
of 2.25Cr-lMo steel is accompanied by a loss of duc- mium as their respective carbides, unlike in 2.25Cr-lMo
tility and impact toughness. Since this toughness loss is steel. This excess carbon is present in the former steel
not fully recovered by the de-embrittling treatment, it is as cementite, and, hence, it is reasonable to assume that
reasonable to conclude that the loss is due to changes in the Cr-Mo-V steel has a larger amount of grain boundary
carbide type or morphology during service. It is not well cementite platelets than in the other steel. If coarsening
established what changes in carbide morphology cause of grain boundary cementite platelets in service was an
irreversible embrittlement. Brindley and Lindley ~4~have important factor for the toughness loss, the Cr-Mo-V steel
shown that the tensile ductility of an iron-carbon alloy should have also embrittled irreversibly as 2.25Cr-lMo
decreases with increasing grain boundary plate cementite steel did in service, contrary to the results observed in
particle size. They have further shown that cracking of the present study.
these platelet cementite particles is responsible for form- Recent carbide extraction work by Bigs and Wada ~251
ing cleavage microcracks in the ferrite. Such cleavage has revealed that service exposure causes precipitation
cracks are not observed when the cementite is precipi- of M2C carbide needles in the ferrite grain, and spherical
tated inside the ferrite grain. M2C carbide particles on the prior austenite grain

METALLURGICAL TRANSACTIONS A VOLUME 20A, JANUARY 1989--95

 at 288 ~ (550 ~ and 529 ~ (985 ~ even after long-
term service exposure. The absence of intergranular
fracture in the Charpy specimens taken from the cold
a)-NTO
RZ
AZ
NLSEC
GH
RA
AMBL
NUEARR and hot locations indicates that the grain boundaries are
not weakened by the segregation of impurities (partic-
ularly phosphorus for Cr-Mo-V steel) during service. At
the cold location, the service temperature is too low for
~ e significant levels of impurities to diffuse to the bound-
T aries, while at the hot location the temperature is too
high to build up a critical concentration of impurities
needed at the grain boundaries to weaken them. This type
of temperature dependency of embrittlement is normally
referred to as C-type behavior for reversible temper em-
b)B LA
- IN DR
TEER
GIR
NA
GNULAR brittlement, t26] Consistent with the present results, evi-
dence for the occurrence of severe embrittlement of
Ni-Cr-Mo-V steel at a temperature around 454 ~ (850 ~
can be found in the literature. [26,271
Results of past studies suggest that the original F A T T
and temper embrittlement behavior of a Cr-Mo-V rotor
steel are also a function of tensile strength of the steel. E1'151
It is clear from the data of the de-embrittled specimens
E (eV)
(Figure 6) that softening, which resulted from service
Fig. 10--Auger spectra of service-exposed samples from (a) nozzle exposure, has slightly lowered the FATT of the steel.
chamber; (b) blade ring. The temper embrittlement susceptibility, as measured by
the difference in F A T T (A FATT) in the service-exposed
boundaries in 1Cr-0.5Mo steel. They have postulated that and de-embrittled conditions, tends to decrease with in-
the precipitation of these carbides lowers the impact creasing service temperature above 466 ~ (870 ~ The
toughness of the steels. The precipitation of molybde- difference is estimated to approach zero (Figure 6) at a
num carbides (M2C) can occur during service in both Cr- temperature of approximately 552 ~ (1025 ~
Mo-V and 2 . 2 5 C r - l M o steels, but the temperature at This A FATT actually represents a shift in FATT due
which these carbides form and the kinetics of formation to service exposure and any change in FATT as a result
are expected to be different for these steels. Qu and Kuo [6] of slow cooling from the original tempering temperature.
have suggested that the presence of vanadium decreases As stated earlier, the rotor forgings are typically air cooled
the rate of formation of molybdenum-rich carbides and from the austenitizing temperature and furnace cooled
favors the precipitation of M6C o v e r M2C during service from the tempering temperature during manufacturing,
at 540 ~ (1004 ~ Their results have shown that pre- and, thus, are slowly cooled through the critical tem-
cipitation and coarsening of molybdenum-rich carbides perature range (300 to 550 ~ where reversible temper
(M6C) have a detrimental effect on impact toughness of embrittlement normally occurs. Due to the large diam-
2.25Cr-lMo-0.4V steel. The present results, coupled with eter of these forgings, the occurrence of some temper
the results of Qu and K u o , [61 indicate that Cr-Mo-V steel embrittlement during cooling is unavoidable. The results
is not susceptible to carbide-induced embrittlement dur- obtained on the Buck rotor rml have shown that the shift
ing service at a temperature below 540 ~ in FATT due to slow cooling of the forging from the
From the results of Bigs and Wada f251and Qu and Kuo, [61 original tempering treatment can be as high as 50 ~
it can be inferred that precipitation of molybdenum car- (90 ~ Therefore, the observed difference in FATT in
bides (M2C or M6C ) during service is detrimental for the the as-received and de-embrittled condition at the hot lo-
toughness of Cr-Mo(V) type ferritic steels. However, the cation (Figure 6) is significantly ( ~ 2 0 ~ lower than the
mechanism by which carbides affect toughness is not F A T T shift expected as a result of slow cooling during
understood at this time. Hence, a systematic investiga- manufacturing. Hence, these results suggest that service
tion of all the variables such as carbide type morphology exposure at 530 ~ (985 ~ or above negates any det-
and distribution and service temperature is needed to un- rimental effects of slow cooling through the critical tem-
derstand the role of carbides on the toughness. perature range during the original heat treatment by
desegregating the phosphorus from the grain boundaries
in service. In fact, the occurrence of de-embrittlement
C. Temper Embrittlement
during service was confirmed by performing additional
In both Cr-Mo-V and 2 . 2 5 C r - l M o steels, the most se- Charpy tests on the specimens from the coupling loca-
vere temper embrittlement occurs in service when they tion of the rotor where service temperature was 80 ~
are exposed to a temperature near 427 ~ (800 ~ The (175 ~ The FATT of the hot location is 22 ~ (40 ~
observation of a maxima in F A T T vs service temperature lower than that of the coupling location of the rotor.
for the Cr-Mo-V rotor steel (Figure 6) is consistent with Additions of molybdenum up to 0.7 pct is known to
the findings of Hirota e t al. L13~and Zhe et al. tm~ The FATT reduce temper embrittlement susceptibility of 2.25Cr-Mo
data coupled with fractography results (Figure 7) suggest type steel321 Molybdenum suppresses embrittlement by
that the rotor steel does not appreciably temper embrittle scavenging phosphorus through the formation of a

96--VOLUME 20A, JANUARY 1989 METALLURGICAL TRANSACTIONS A

Mo-P compound and thus decreases phosphorus segrega- ACKNOWLEDGMENTS
tion to the grain boundaries, t2,281 McMahon and his co-
The author wishes to express sincere thanks to Drs.
workers t3,1~ and Qu and Kuo ~6~ have s h o w n that
R.L. Novak, B.B. Seth, B. Shaw, and T. Wada for crit-
any decrease in molybdenum content in the matrix
ically reviewing the manuscript and giving constructive
due to formation of molybdenum carbides (MzC or M6C)
comments.
during tempering or aging significantly enhances em-
brittlement susceptibility of 2 . 2 5 C r - l M o and Cr-Mo-V
steels. Though service exposure has resulted in a de- REFERENCES
crease of molybdenum content in the matrix (as evi- 1. R. Viswanathan and A. Joshi: Metall. Trans. A, 1975, vol. 6A,
denced by softening), the observation of no enhancement pp. 2289-97.
in temper embritflement susceptibility at the hot location 2. T.' Wada and W.C. Hagel: Metall. Trans. A, 1976, vol. 7A,
also indicates that the rotor steel does de-embrittle pp. 1419-26.
3. J. Yu and C.J. McMahon, Jr.: MetaU. Trans. A, 1980, vol. l l A ,
during service at this temperature to negate the
pp. 277-89.
phosphorus scavenging effect. 4. B.J. Brindley and T.C. Lindley: J. Iron and Steel lnst., 1972,
vol. 23, pp. 124-25.
5. Y. Nishizaka, Y. Hara, A. Hori, T. Tsukahara, K. Miyano,
T. Wada, and T.B. Cox: ASME Proc. Pressure Vessel Piping
V. C O N C L U S I O N S Conf., 1985, PVP-vol. 98-1, pp. 125-35.
6. Z. Qu and K.H. Kuo: Metall. Trans. A, 1981, vol. 12A,
The influence of service temperature on softening and pp. 1333-37.
embrittlement behavior of Cr-Mo-V and 2.25Cr-lMo steel 7. T. Wada and V.A. Bigs: Metall. Trans. A, 1983, vol. 14A,
has been studied. From the results of this study the fol- pp. 845-55.
lowing conclusions can be drawn: 8. L.H. Toft and R.A. Marsden: J1SI & JIM Proc. Intl. Conf. on
Structural Process in Creep, 1963, pp. 276-93.
1. Both Cr-Mo-V and 2 . 2 5 C r - l M o steels soften only 9. N.S. Cheruvu: ASM Proc. Intl. Conf. on Nuclear Power Plant
when they are exposed at a temperature greater than Aging, Availability Factor and Reliability Analysis, V.S. Goel,
454 ~ (850 ~ The degree of softening of Cr-Mo-V ed., 1985, pp. 575-80.
10. O. Zhe, S.C. Fu, and C.J. McMahon, Jr.: EPRIReport CS2242,
increases with increase in service temperature above Project 559, February 1982.
454 ~ (850 ~ In this case, the softening is accom- 11. N.S. Cheruvu, W.E. Howard, and O.M. Beasley: ASME Joint
panied with an improvement in tensile and creep duc- Power Generation Conf., paper #85-JPGC-Pwr-29, 1985.
tility and upper shelf energy. 12. H.C. Argo and B.B. Seth: Case Studies in Fracture Mechanics,
Army Materials Research Center, MS77-5, T.P. Rich and D.J.
2. The 2 . 2 5 C r - l M o steel has softened after service Cartwright, eds., 1977, pp. 3.2.1-11.
exposure to 538 ~ (1000 ~ In addition to soften- 13. Y. Hirota, Y. Kadoya, T. Goto, M. Wake, and H. Fuji:
ing, service exposure at 538 ~ (1000 ~ results in Mitsubishi Heavy Industries Ltd. Technical Review, October 1982.
a decrease in tensile ductility and toughness of this 14. Supplemental Report on The Elevated Temperature Properties of
steel. This loss in toughness is in part irreversible and Cr-Mo Steels, ASTM Data Services Publications No. DS6S1,
American Society for Testing and Materials, J.A. Van Etho and
is due to changes in carbide type or morphology dur- W.F. Simmons, eds., 1966, p. 52.
ing service. 15. L. Tu and B.B. Seth: Metall. Tech., 1978, vol. 5, pp. 79-91.
3. The Cr-Mo-V steel does not show any evidence of 16. R.V. Hart: Metall. Tech., 1976, vol. 3, pp. 1-7.
carbide-induced irreversible embrittlement after long- 17. N.S. Cheruvu and L.R. Malmfeldt: ASME Joint Power Gener-
ation Conf., paper #87-JPGC-Pwr-53, 1987.
term exposure at 529 ~ (985 ~ The present results, 18. R. Viswanathan and J.I. Jaffee: Turbine Rotor Life Assessment
coupled with the results reported in the literature, in- and Improvement, EEl Prime Movers Committee, paper R9,
dicate that there is a critical service temperature Miami, FL, 1984.
(540 ~ for Cr-Mo-V steel to embrittle due to changes 19. R. Viswanathan: Metall. Tech., 1974, vol. 1, pp. 283-88.
20. J. Nutting: A1ME Proc. Topical Conference of Ferritic Alloys for
in carbide morphology or type in service.
Use in the Nuclear Energy Technologies, J.W. Davis and D.J.
4. The notch rupture life o f Cr-Mo-V steel is not im- Michel, eds., 1983, pp. 3-16.
proved by softening as a result of service exposure. 21. J.D. Murray: in Precipitation Process in Steels, ISI Special Re-
5. Service exposure at 538 ~ (1000 ~ has lowered the port no. 64, 1959, p. 285.
creep rupture life o f 2 . 2 5 C r - l M o steel by an order of 22: J.D. Baird, A. Jamieson, R.R. Peterson, and R.C. Cochrane:
Proc. Conf. on The Creep Strength in Steel and High Temper-
magnitude as compared to that of the steel exposed ature Alloys, Sheffield, Metals Society, London, 1973,
at 454 ~ (850 ~ pp. 207-15.
6 . Both steels severely temper embrittle during service 23. K.R. Williams and B. Wilshire: Materials Sci. Eng., 1981,
at a temperature around 427 ~ (800 ~ This em- vol. 47, pp. 151-60.
brittlement is due to segregation o f residual elements 24. K.R. Williams and B. Wilshire: Materials Sci. and Eng., 1977,
vol. 28, pp. 289-96.
to the grain boundaries in service. 25. V.A. Biss and T. Wada: Metall. Trans. A, 1985, vol. 16A,
7. The Cr-Mo-V steel does not temper embrittle during pp. 109-14.
service at 529 ~ (985 ~ It appears that service ex- 26. G.C. Gould: American Society for Testing and Materials, Spec.
posure at this temperature can negate any detrimental Tech. Publication 407, 1968, pp. 90-105.
27. R. Viswanathan and T.P. Sherlock: Metall. Trans., 1972,
effects of slow cooling through the critical tempera- vol. 3, pp. 459-68.
ture range for temper embrittlement during original 28. C.J. McMahon, A.K. Cianelli, and H.C. Feng: Metall. Trans.
heat treatment of the forging. A, 1977, vol. 8A, pp. 1055-61.

METALLURGICAL TRANSACTIONS A VOLUME 20A, JANUARY 1989--97