-..

' 2; i
CAUSTIC STRESS CORROSION CRACKING OF
CARBON STEELS
A Supplement to
STRESS CORROSION CRACKING OF
.(~ . 'CONTINUOUS DIGESTERS
iI
? Project 3589
| /Final
Report
( -/A--lProgress Report
to
DIGESTER CRACKING RESEARCH COMMITTEE
't r October 17, 1986
1::
L
NOTICE & DISCLAIMER
The Institute of Paper Chemistry (IPC) has provided a high standard of professional service and has exerted its best efforts
within the time and funds available for this project. The information and conclusions are advisory and are intended only for
the internal use by any company who may receive this report. Each company must decide for itself the best approach to solv-
ing any problems it may have and how, or whether, this reported information should be considered in its approach.
IPC does not recommend particular products, procedures, materials, or services. These are included only in the interest of
completeness within a laboratory context and budgetary constraint. Actual products, procedures, materials, and services used
may differ and are peculiar to the operations of each company.
In no event shall IPC or its employees and agents have any obligation or liability for damages, including, but not limited to,
consequential damages, arising out of or in connection with any company's use of, or inability to use, the reported informa-
tion. IPC provides no warranty or guaranty of results.
MA~Y
2 01987
RESEARCH.
TECH..I.C.AL
IIILES
THE INSTITUTE OF PAPER CHEMISTRY
Appleton, Wisconsin
CAUSTIC STRESS CORROSION CRACKING OF CARBON STEELS
A Supplement to
STRESS CORROSION CRACKING OF CONTINUOUS DIGESTERS
Project 3589
by
R. Yeske
Section Leader
Corrosion and Materials Engineering
Final Report
A Progress Report
to
DIGESTER CRACKING RESEARCH COMMITTEE
October 17, 1986
TABLE OF CONTENTS
Page
SUMMARY 1
AN INTRODUCTION TO CAUSTIC CRACKING 2
ACCELERATED TEST METHODS 8
Controlled Potential Testing 8
Polarization Curves 14
Fast-Slow Polarization as an Accelerated SCC Test 16
Slow Strain Rate Testing 18
Crack Growth Rate Testing 23
Circular Patch Weld Test Coupons 28
LITERATURE CITED 31
GLOSSARY OF TERMS 33
THE INSTITUTE OF PAPER CHEMISTRY
Appleton, Wisconsin
CAUSTIC STRESS CORROSION CRACKING OF CARBON STEELS
A Supplement to
STRESS CORROSION CRACKING OF CONTINUOUS DIGESTERS
SUMMARY
This report supplements the final report to the Digester Cracking
Research Committee on stress corrosion cracking of Kamyr continuous digesters.*
It provides an introduction to stress corrosion cracking for those who are unfa-
miliar with the concepts and terminology of this topic. Included in the report
are an introduction to caustic cracking, a summary of test methods used in the
DCRC program to investigate stress corrosion cracking, and a glossary of terms.
R. Yeske, "Stress Corrosion Cracking of Continuous Digesters: A Summary of
Digester Cracking Research Committee Projects," Final Report on Project 3589
(October, 1986), available from The Institute of Paper Chemistry, Appleton, WI.
Project 3589 -2- Final Report
AN INTRODUCTION TO CAUSTIC CRACKING*
Carbon steels are known to'be susceptible to a form of stress corrosion
cracking known as caustic cracking or caustic embrittlement. In caustic
cracking, cracks initiate and grow along microstructural features called grain
boundaries in the carbon steel. Extensive crack branching is also a distinguish-
ing feature of caustic cracking (Fig. 1). The driving force for crack propaga-
tion is a high tensile stress applied externally to the steel and/or residing
within the steel as a result of fabrication. There have been numerous incidents
of caustic cracking failures reported in several different industries over the
years.
As with other forms of stress corrosion cracking, caustic cracking will
occur only when several conditions are satisfied. First, a high tensile stress
must be present in the steel while it is exposed to the corrosive medium.
Second, the material must be susceptible to stress corrosion cracking in
caustic. (Nearly all carbon steels, including base metal and weld filler metal,
are thought to be susceptible to caustic cracking.) Next, the corrosive environ-
ment must spontaneously react with the steel to form a film on the wetted sur-
face that offers only limited corrosion protection to the underlying metal.
Finally, the electrochemical potential (i.e., the voltage) spontaneously devel-
oped on the surface of the wetted steel must remain within a specific range.
This voltage is thought to condition the surface film to allow crack initiation
and growth. All of these conditions must be fulfilled at the same time in order
for caustic cracking to proceed.
*A detailed review of caustic cracking is available in an earlier report.!
** ' , ' .. .'.. . , ' ' 1,'
Final Report
Figure 1.
K . ' : " I
,j'- *|^-i:'^ fSS *. < f'2j
LL" i:. .. r:..i<. :;
-~ 'Kr, -
A photomicrograph showing caustic cracking along microstructural
features in steel as revealed by metallographic etching.
Tensile stresses can arise in carbon steel pressure vessels from
various sources. The principal source of stress in welds is the differential
contraction that occurs as welds cool under restraint during vessel fabrication.
If a stress relieving heat treatment is not performed on the steel after welding
the residual stresses remaining at the weld are on the order of the tensile
strength of the steel - as high as 60,000 pounds per square inch.
2
These
residual stresses remain in the weld and the adjacent base metal unless they are
relaxed by a stress relief heat treatment or by cracking. Operating stresses
can also contribute to the driving force for crack growth, especially at discon-
tinuities in the metal that can raise the effective local stress. In continuous
digesters, as in all ASME Code vessels, the nominal stress arising from
pressurization is on the order of one-fourth of the tensile strength of the
steel, but the local stress can be much higher.where.surface imperfections,
Project 3589 -3-
Project 3589
scratches, and other defects concentrate the stress. Small fluctuations in
stress - perhaps arising from pressure fluctuations during routine operation
-are thought to increase the propensity for cracking.
All plain carbon steels used in fabrication of pressure vessels are
thought to be susceptible to caustic cracking, including both base metals and
weld filler metals.
1
This list of susceptible steels includes A516 Gr.70, A285
Gr.C, and A212 Gr.B pressure vessel steels, and E7018 and E6010 weld filler
metals, all of which were used in the construction of Kamyr digesters. Stain-
less steels appear to be resistant to cracking in kraft cooking liquors at
impregnation zone temperatures, although cracking can occur when the tem-
peratures are higher, when the caustic concentrations are increased, or when the
stainless steel microstructure is sensitized by an improper heat treatment.
Caustic cracking of carbon steel occurs over a wide range of caustic
concentrations. The lower limit for cracking appears to be in the range of a
few percent (by weight) of caustic as NaOH. In kraft cooking liquors, the con-
centration of hydroxides at the site of cracking is typically on the order of 3
to 5 percent,
3
which is sufficient to induce cracking of carbon steels if all
other conditions for cracking are satisfied.
If caustic cracking is to proceed, the potential (i.e., the voltage)
developed spontaneously at the wetted metal surface must remain in a critical
range while the stress is being imposed. The range of potentials that will
cause cracking depends on the caustic concentration
4
, as shown in Fig. 2.
Concentrated caustic solutions will induce caustic cracking of carbon steels
over a wide potential range. As the concentration of NaOH decreases toward five
percent (by weight), the width of the potential zone for caustic cracking
-4-
Finali Report
1
l u a e u U O I T A U O 9 q j J O U 0 o 3
- P a l u a o u o Z z T s v a q J O u p o j u n l j v S p ' 2 n D ~ o u p o B u T ) I . 1 o
o p s n o 9 8 9 1 4 A 9 a g u v a T U T ~ u a o d a q p 2 u p m o q s 3 o i d O l w ~ t q z s V
- I B c = s a e N .
0 9 I t v
H O B N ( %
O L 0
~ ~ i o d ~ ~ ~ ~ ~ j T I U T ~ ~ ~ ~ ~ ~ ~ ~ 8 : O P 2 Z d
0 9
m
C , )
C l )
C , )
. m
- Z 9 j n 2 T d
L ,
l a o d B U T U T d
Project 3589
decreases dramatically. Below a few percent of dissolved caustic, the potential
range shrinks to zero and caustic cracking does not occur.
In general, the electrochemical potential measured at the surface of
carbon steel wetted by kraft cooking liquors has been found to be close to the
potential range associated with caustic cracking. However, the factors which
determine this potential are not well understood.
The critical requirements for caustic cracking can be rationalized in
terms of a generally accepted mechanism of crack growth (Fig. 3). This model
for caustic cracking, often referred to as the slip-dissolution model, envisions
crack growth as a series of repetitive dissolution events occurring at the tip
of a growing crack. The film that forms on most of the wetted steel surface
prevents active corrosion of most of the surface wetted by the caustic solution.
However, tensile stresses that deform the steel slightly at the tip of the crack
rupture the attached film at the crack tip. Once the protective film is
destroyed, rapid corrosion occurs at the crack tip, resulting in apparent
penetration of the crack. Eventually, the film reforms at the crack tip, only
to be ruptured again by the tensile stresses and followed by another dissolution
event. In this way, cracks grow by repeated film rupture and dissolution events
that take place at the crack tip.
Caustic cracking will generally follow a preferred path through the
steel microstructure as shown in Fig. 1. The reason for this preferred path of
cracking remains controversial, but is probably associated with either unique
electrochemical conditions or unique strain conditions that develop at the
microstructural boundaries in the steel. These conditions exist because of
the unique composition and structure associated with grain boundaries.
Final Report -6-
L
U
~
~
~
~
~
~
~
~
L

0
r
4
1
.
5
4

c
n
a
.
)
.
0
V
.I
.-
4
0
-
J
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
J
0
0
~
~
~
~
~
~
~
~
-
*
5
.
.
.

C
A
.
-
'~
~
~
~
~
~
~
~
~
~
~
~
0
C
a
-
0
~
~
~
~
~
~
~
~
~
b

L
U

4
.4
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
.
-
-
'
i
"
.
~
~
~
~
~
~
~
~
~
~
z

I
'
:

_
_
_
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
ir
0
~
~
~
~
~
~
~
~
~
'

*
5
.
"
'*
.
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
H
c
.
I
*
'"

'j
''
.
.
.
*
.

.
.
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~

*
*

b
5
.
1
.
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
0
0
r
Project 3589
ACCELERATED TEST METHODS
Several different accelerated test methods were used to investigate
stress corrosion cracking in the DCRC programs.
5 - 12
Accelerated tests are
necessary because the incubation time for cracking at welds exposed to cooking
liquors can be quite long. Consequently, accelerated tests, such as the slow
strain rate test (SSRT) and fracture mechanics testing for crack growth, were
employed in many phases of the DCRC program. The objective of this brief sub-
section is to describe these accelerated tests.
CONTROLLED POTENTIAL TESTING
The potential difference (or voltage) found at the wetted metal surface
is an important variable in stress corrosion cracking studies because it has
such a strong effect on cracking susceptibility. Cracking does not occur
unless the potential developed at the metal/electrolyte interface remains within
a specific zone. To identify this cracking zone, tests must be conducted to
determine the cracking susceptibility at various potentials. The zone where
cracking is observed can then be compared with the potential that spontaneously
develops at a wetted metal surface in the laboratory or in the field. This sub-
section addresses the origins of electrochemical potentials and briefly con-
siders how potentials can be measured and controlled.
Every metal immersed in an electrolyte will spontaneously develop a
voltage (or potential) at its wetted surface. In a lead-acid storage battery,
for example, the voltage difference between the positive and negative terminals
arises because plates of two different materials are immersed in the battery
acid.. A separate voltage develops at each of the plates but neither of these
separate voltages can be easily measured. However, the voltage of the battery
Final Report
-8-
Project 3589
can be easily measured as the algebraic difference between the potentials devel-
oped at each of the two plates. Pressure vessel steels immersed in kraft
cooking liquors also develop their own characteristic potentials.
The potentials that exist on wetted metal surfaces are easily measured
using a high-resistance DC voltmeter and a reference electrode, as shown in Fig.
4. One terminal from the voltmeter is connected to the metal and the other ter-
minal (the low or common terminal) is connected to a reference electrode immersed
in the same electrolyte. The voltmeter measures the algebraic difference between
two voltages - the voltage at the metal surface and the voltage measured at the
reference electrode surface. The reference electrode provides a stable, repro-
ducible, well-known voltage as a benchmark against which other potentials can be
measured. (In some respects, the reference potential for corrosion measurements
resembles earth ground potential in electrical power measurements.) Whenever
the potential of a metal is measured, the reference potential used in the
measurement must also be specified.
Some common reference electrodes are the Saturated Calomel Electrode
(SCE), the Standard Hydrogen Electrode (SHE), and the Silver/Silver-Sulfide
electrode (SSSE). It is agreed that the SHE electrode is the universal
reference electrode standard and the potential of this electrode is arbitrarily
assigned a voltage of 0.000 V. The SHE electrode is not widely used because it
is more difficult to use and maintain than alternative reference electrodes.
The potential of the SCE reference electrode is about 241 millivolts above the
SHE potential. The SSS reference electrode has a reference potential about 890
millivolts below the SHE potential in typical cooking liquor, but the exact
potential of this electrode depends on the sulfide concentration.
Fina.1 Report -9-
0
1
4

a
o
j
n

v
p
a
d
o
T
8
A
U
p

(
9
2
B
T
O
A

.
1
0
)

W
P
i
T
~
U
a
i
o
d

a
t
p

g
o

3
u
m
i
s
a
0 c
o
r
Z
4
1
I
I
V
M

U
3
J
.
S
3
9
I
a
U
,
,
,
S
3
A
1
V
A

1
1
V
S
X
O
G

9
N
m
H
.
L
I
s
.
.
l
.
0
*
4
;

a

I
n
S

J
a
'
-
3

a

o
u
i
3
3
1
-
H
o
n
o
u
I
1
3
3
:
1
a
H
i
v
l
i
n
S
-
N
l
Project 3589
The voltage of a metal measured with any reference electrode can be
referred back to the universal standard (the SHE potential) by subtracting
(algebraic subtraction) the reference potential of the reference electrode from
the voltage measured at the voltmeter. For example, a steel immersed in cooking
liquor may have a measured potential of -1000 mVSCE, which means that the poten-
tial of the steel/liquor interface is one thousand millivolts below the standard
potential established at a Calomel reference electrode. But the Calomel
reference potential is 241 mV above the SHE potential. The same steel would
have a potential of -759 mVSHE [i.e., -1000 - (-241) = -759 mV] if the Standard
Hydrogen Electrode was used as the reference electrode.
The potential that spontaneously occurs at wetted metal surfaces is
called the rest potential (or open circuit potential) and is often given the
symbol, Ecorr. As with all corrosion potentials, Ecorr must be referred to a
specific reference electrode potential. Ecorr is established by a complex
interaction between the wetted metal and the contacting electrolyte. The Ecorr
of a material exposed to an electrolyte may vary with time, depending on
changing conditions at the wetted surface.
The potential at a metal surface may be artificially shifted from the
rest potential in two ways - by connecting the metal to a DC power supply or by
adding oxidizing or reducing species to the electrolyte. The shift in potential
from Ecorr is called a polarization. Ecorr can be increased relative to the
usual rest potential by doping the electrolyte with oxidants - species that
will readily consume electrons at the metal surface. The potential can also be
increased above Ecorr by connecting the metal to a battery whose polarity and
voltage make it a sink for electrons. Potentials below Ecorr can be achieved by
-11-
Final Report
Project 3589
adding reducing species to the environment or by connecting the metal to a bat-
tery acting as a source of electrons. In most instances, it is more convenient
to change potentials with a DC power supply.
When a power supply is used to control potentials, it will act as
either a sink or a source of electrons for the metal electrode; an electrical
current will flow from (or to) the power supply as the potential is shifted from
,the rest potential. This current must flow through the wetted metal surface,
through the electrolyte, and back to the power supply through another metal
immersed in the electrolyte (Fig. 5). This second metal electrode, often
called the counter electrode, is connected to the second terminal of the power
supply when potentials are being shifted. The counter electrode is only needed
when the potential of a metal is artificially shifted from the rest potential by
connection to a battery or other DC power source. The current required to shift
the potential can be measured by an ammeter connected between the power supply
and the counter electrode.
A special DC power supply used to polarize metals in corrosion tests is
called a potentiostat. When controlled potential tests are performed, the test
specimen and the other required electrodes are connected to a potentiostat as
shown in Fig. 5. The desired potential is set by adjusting the set-point
control on the potentiostat. The potentiostat passes current of appropriate
polarity and magnitude to hold the desired potential throughout the test.
The rest potential, Ecorr, measured at the surface of a steel surface
exposed to caustic cooking liquors is an important quantity since Ecorr can lie
within the range required for cracking. The potentials measured during polari-
zation of a steel electrode, and the current required to achieve those potentials
Final Report
-12-
Project 3589
Figure 5.
i
Schematic diagram of the electrical current flow associated
with a polarization of a metal to potentials above Ecorrs
i.e., an anodic polarization.
CATHOD
Final 'Re~port -13-
. , ^
a A - p Z ? a 3 u a ~ I d x 8 T T T I A s u o ~ f j n T S P T O U Z T a 0 T 4 q 3 O a P X 4 U T P 9 9 a ~ u m T T ? 9 3 s ' S M T T ;
9 A T ~ 8 o : 0 o d J o u o . T w u u m o j a q p : f n o q 4 1 T m L 1 8 A - f l Z 9 O P ' O i i 0 3 ) , I t p S T 8 3 8 U 3 0 T V : ) T d A ~
S T . 1 O T A I 4 a q s T U J , * P a S V 9 I 3 U T S T 1 T V 1 u a ~ o d a q s 9 j u a a a n o P 9 T ~ d d v B u T s u a a o u J u p
A q p ~ ~ 9 z a q S T T V 3 9 m O A I O U V * 9 * 2 J T U T U A O 4 S S U ' p 8 3 V 1 8 U 9 S 9 a V S O A a n o Z
. U o T l u z T a R ~ o d o ~ p o u p u a q m . p a . z j u n o z u e 8 a P J O - p q a q J o s a d A - a a a u j ~ , 8 I A O J 2 : ) q T 9
U v U T P 8 I S I I I I T T 9 8 U J o i o T q a I 4 q T U 0 T 8 4 q 0 O l a T g l O l u J 3 4
2
T S U T 9 T q v I a p T s u o z
9 P T A o j I d ' S & A l n O u O T I B Z T I a o d 3 j P O u v A 1 I I j n o T j I v d ' s a A a f l o u 0 T ~ I ~ z T . T O d
* A o f a q p a q ~ j o s a p s I s u o T i d a o x e e a v
a ' a W 3 q ~ n o q l v I ' s e o I v z u o ~ j n T O S S i P 8 S V 9 i 3 P S U O - P I z T 4 U T o d z p o q i v 9 T i A ' u O 0 p
- f l T O S S T P T V I 3 9 U J O 9 1 I v 1 a I 4 l a S P I D I U J A T T I O U 9 2 S l 1 O T I l 2 z . 1 1 o d o 4 p o u V a f l T I A a o
s n o e u v i u o d s 9 q l m O T 9 4 S T V I P u a o d 9 A 9 T q 3 V S U O T ~ I ~ z T I I 1 d 3 T p o W 3 v o B T T 4 A ' T T l
- u a l o d 3 s a 1 a 9 4 2 u v 4 l a 9 1
2
T M S T V I p u a l o d o l S u O T I U Z T I B ~ o d S M & 0 1 4 O A a f l 3 u O T I W z T . I T o d
: ) T p o U u l e * u V . m ) b s / s d m v J o 9 2 T u n U T ' A ~ T s u 9 P : U I u a a f l n 3 I U T L 1 : I o o ~ j P O J 2 9 T a
a i j ~ J o v a a v 9 3 v ; i n s p a s o d x a o i q a A q I u a a a n o a q 2 U T P T A T P A q P O Z T T I m I O u A [ T ~ u n a S f T
q u a i a f l o P O T ~ d d e a U J - - -
2
T A U T u m o 4 q s S I 9 P O 1 : 1 3 [ 8 z I a ~ u n l o - 9 4 ~ l P u L . A i d d n s e a m o d
a 4 j u e a a m l q p a j o a u u o D i a l a m m v u p A q p a I f n s v a U 1 S T P U B ' T T i u B : o d U B A T A J 3
o i A
1
d d n s I a m o d D a I a q A q P 9 P T A O a d q u a i a f l o a q S T j u a i i n l o t U J ' B A o q v p a q T J O ~ P
s e p e a n l s v e m l p u e P a T T O a l U 0 ~ S T T U T l u a l o d a 8 U j * P 9 P 2 O 3 9 J S T ( j u e I i n o P a T ~ d d I )
2 0 1 T . a q I 8 T T q m S T V T 1 u a j o d u 1 3 5 o 3 R I ~ S 0 T l u a l o d v 2 u p u u m v z 2 o i d A q P u p 8 U u a i i p
a i v s a A a r L 3 u 0 T ~ I Z T . 1 T O d ' U O T I ~ a I T P a A T ~ I U 2 U 2 0 a A T I T s o d v U T 3 1 a M I f O J
P 8 U u v z S A T T U D T 1 I U M 9 S A S S T 1 T v ~ u a l o d a q u a 8 4 m p a u ~ v q q o v l p ( 3 u a I a z n 3 P a T T d d I )
S O T s n s a a A 1 T v 3 u a j d J O 3 O 1 J d 3 T m W 1 T a v
2
O T T m 8 S V S T a A a n l 3 u o T l v z T a e I o d V
S M A N f 1 N o U Y Z I X V ' I 0 d
s e l a o n b T T B U T N O O D a 4 1 U T T ~ l 9 q J o a o T A v 4 8 q T V T ~ 0 ~ 3 T a ' 4 l a q T A O S P
2 a o d a e x T U U T A ~ V - 6 8 S E 2 p 3 8 0 1 d
Project 3589
Ecorr
I-
z
QU
0
on
Ecorr
Ecorr
log CURRENT
Figure 6. Three types of electrochemical behavior as represented by anodic
polarization curves, (a) active, (b) active-passive, (c) passive.
ACTIVE
ACTIVE-PASSIVE
PASSIVE
Final Report
-15-
Project 3589
corrosion. Passive metals are protected by surface films and show little
increase in dissolution current with increased potential. Stainless steels
are examples of materials made passive by the presence of a chromium-rich sur-
face film that protects the underlying metal. Active/passive behavior occurs
when a metal initially exhibits active corrosion behavior but reverts to
passive behavior once the potential is raised above a certain threshold level
called the active/passive transition potential. The onset of passivity in
active-passive metals is associated with formation of a film on the surface of a
metal which protects the underlying metal from corrosion. The film also reduces
the applied current required to achieve higher potentials, which is a charac-
teristic of passive behavior.
''~' : ' Active/passive behavior is important from a caustic cracking viewpoint
because cracking is frequently observed at potentials where there is a tran-
sition from active to passive behavior (Fig. 7). The film that forms at the
transition from active to passive behavior forms slowly and does not provide the
level of protection to the underlying metal provided by passive films formed at
somewhat higher potentials. As discussed earlier, films with poor integrity
promote cracking by the slip/dissolution mechanism. Therefore, metals which
undergo active/passive transitions are candidates for stress corrosion cracking
when their rest potentials are in the active/passive range.
FAST-SLOW POLARIZATION AS AN ACCELERATED SCC TEST
Since polarization curves depict the film-forming characteristics of a
corroding metal, polarization curves are often used to predict SCC behavior. As
discussed above, an active-passive transition is a good indicator of SCC suscep-
tibility, but it is often difficult to assign limits to the potential range
associated with cracking in this active-passive zone.
Final Report -16-
Final Report
a,
0,
C,
(5
log APPLIED CURRENT
Typical cracking range for metals with active-passive corrosion
behavior, as with carbon steels exposed to kraft liquors.
U
0
.4
-J
z
0
Ecorr
Figure 7.
Project 3589 -17-
A
Project 3589
The fast-slow scan technique is an attempt to bracket the potential
range for cracking with greater precision than can be achieved with a single
polarization curve.
4
In this approach, the metal is subjected to two separate
polarization scans, one fast and one much slower, and the polarization curves
from the two scans are compared as shown in Fig. 8. Although rapid scanning
through anodic potentials usually requires a slightly higher current at the same
potential, the differences between the fast and slow scan curves are generally
small. However, large differences in the current density during fast and slow
scans sometimes indicate the range of potentials where cracking occurs. This
difference between fast and slow scan curves can be used to show the potential
range where cracking is likely to occur.
Thus, fast and slow scan polarization curves generated on pressure
vessel steels immersed in cooking liquors might be exploited as an accelerated
test to identify the potential range for cracking. This approach was examined
in the PPRIC program.
SLOW STRAIN RATE TESTING
One of the accelerated tests used in both the PPRIC and Battelle studies
was the slow strain rate test (SSRT). This test has been found to be very useful
for characterizing SCC susceptibility in a number of circumstances. With the
SSRT method, specimens of the metal of interest are slowly strained while
immersed in an environment that might induce stress corrosion cracking. Crack-
ing susceptibility is determined by comparing the results of the slow tensile
tests in air or other inert environment with those derived from a slow tensile
test in a potentially aggressive environment. In some cases, a small cyclic
load is superimposed on the slowly increasing tensile load to promote cracking.
Final-Report
-18-
Project 3589
._
C .
0
Ecorr
Slow Scan
'I
I
I
I
I
I
t
sec
$
Fast Scan
log Applied Current
Figure 8. Schematic diagram showing the potential zone associatedwith
cracking as revealed by the fast/slow scan technique.
-19- FihnaL Rep~ort
Project 3589
In all cases, the specimens are strained to failure and cracking susceptibility
is assessed by comparing the stress-strain curves and examining the physical
appearance of failed test specimens.
If the environment causes stress corrosion cracking of the metal, the
SCC susceptibility will be revealed by differences in behavior of specimens
tested in inert and crack-promoting environments. The effects of stress corro-
sion cracking on mechanical properties are shown schematically in the load-
elongation curves for tensile tests with and without cracking, in Fig. 9. When
stress corrosion cracking occurs during the SSRT test, the total elongation of
the test specimen at the point of final rupture is much reduced, since cracks
initiate and grow early in the test, severing the specimen before normal elonga-
tions can develop.
Stress corrosion cracking susceptibility may also be revealed in a SSRT
test by a reduction in the extent of necking of the gage section of the tensile
specimen. In tensile tests without SCC, extensive necking usually occurs and
there is a large reduction in the cross-sectional area of the specimen at the
point of failure. If stress corrosion cracking occurs, failure occurs by crack
growth before extensive necking can develop. One measure of the degree of
necking is the percent reduction in cross-sectional area (RA) of the gage sec-
tion of the tensile specimen at the point of fracture
RA = 100 x (Initial Area - Final Area)/Initial Area
Tensile tests without SCC are marked by extensive necking and a large RA,
whereas tests conducted under stress corrosion cracking conditions will have a
smaller RA.'
Final Repor't -20-
Project 3589
'0
0
__
Elongation
A schematic example of the load-elongation curves for SSRT
tensile tests in crack-promoting and inert environments.
Figure 9.
-21- Fina.I Reportt
Project 3589
Another indicator of cracking susceptibility is a loss of tensile
strength - the maximum load applied during the tensile test divided by the ori-
ginal cross-sectional area of the gage section. SCC can be detected in SSRT
testing by a diminished tensile strength compared to the normal tensile strength
developed during a SSRT test in air or other inert environment.
A final indicator of stress corrosion cracking susceptibility is the
formation of many small cracks on the surface of the gage section close to the
point of rupture. These secondary cracks are normally absent from tests in
inert environments that do not cause stress corrosion cracking, but appear when
the environment induces cracking.
The most sensitive measure of SCC susceptibility in an SSRT test is the
appearance of secondary cracks on the gage section in the vicinity of the point
of failure. Loss of elongation and reduction in area are also fairly sensitive
indicators of SCC susceptibility in the SSRT test. Loss of strength is relati-
vely insensitive as an indicator of SCC susceptibility, as extensive secondary
cracking must occur soon after the start of the SSRT test for significant
effects on tensile strength to become apparent.
SSRT tests are particularly useful to determine the range of poten-
tials where cracking is possible. An artificial potential can be imposed on the
test specimen during the slow strain rate tests to assess SCC susceptibility at
that potential. SSRT tests conducted at several different potentials will
establish the range of potentials where SCC is expected. Rest potential measure-
ments will then indicate whether stress corrosion cracking will occur during
exposure without polarization. The SSRT test data also indicate the minimum
potential that must be achieved to prevent cracking by anodic protection. This
approach was used extensively by PPRIC in their studies of digester cracking
conditions.
Final Repb6rt -22-
Final Report
The SSRT test is effective as an accelerated SCC test method because
the slow deformation imposed on the test specimen is similar to the slow defor-
mation process thought to occur at the crack tip during caustic cracking. The
slow stretching of the test specimen ruptures the protective film on the metal
surface and promotes the localized dissolution processes that are responsible
for crack growth. Many cracks initiate and grow at the ruptured film sites,
leading to a multitude of secondary cracks and corresponding losses of strength
and ductility.
The SSRT test is recognized as a fairly severe test of the SCC
resistance. SSRT testing is useful to screen for conditions that COULD cause
cracking, but such cracking may not occur, or may not be as severe, under actual
service conditions.
CRACK GROWTH RATE TESTING
The rate of crack growth under severe static loading conditions is
another good indicator of the stress corrosion cracking susceptibility of a
metal exposed to an aggressive environment. Crack growth rates are usually
measured using precracked specimens that are preloaded in a controlled fashion
and then exposed to the environment for a period of time. The average rate of
crack growth is determined by measuring the difference in crack length before
and after exposure and dividing by the exposure time to obtain the average crack
velocity.
The driving force for crack growth obviously depends on the stress
applied to the specimen during the exposure, but also depends on the length of
the crack and the geometry of the test specimen. All of these factors which
Project 3589 -23-
Project 3589
influence crack growth rates are included in a quantity called the stress inten-
sity, K, which is a measure of the driving force for crack propagation.
Laboratory tests to determine crack growth velocities are usually conducted on
specimens whose.stress intensities can be calculated according to published for-
mulas..
The double cantilever beam (DCB) specimen shown in Fig. 10 is commonly
used for crack growth studies; it was used extensively by PPRIC in their DCRC
research projects. The DCB specimen is mechanically precracked before exposure
to the test solution to provide a sharp crack tip where further cracking will be
easily induced. The specimen is loaded by forcing the crack faces apart in
controlled fashion. Usually, a wedge is forced into the notch of the DCB speci-
men or a bolt is torqued to spread the notch faces apart. The distance
separating the crack faces along the line of loading is called the crack opening
displacement, V. The stress intensity for a DCB specimen is given in Fig. 10
as a function of the crack opening displacement, crack length, and geometry of
the specimen. As cracks grow in the bolt-loaded DCB specimen, the crack length
increases but the load is relaxed even more quickly; consequently, the stress
intensity rapidly decreases as the crack length increases during a typical crack
growth test (see inset, Fig. 10). .'
The crack velocity depends on the stress intensity in a complicated
way, as shown schematically in Fig. 11. Note that the crack velocity is plotted
on a logarithmic scale to accommodate the large changes in crack velocity that
are usually observed.. When the .applied stress intensity is low -because the
applied loads are small or the crack is short - the crack velocity is strongly
dependent on the stress intensity. In this region, small increases in stress
intensity result in large increases in crack velocity. At somewhat higher
Final'Report -24-
'
d0t
L
.
C
D
0
D
O
U
B
L
E

C
A
N
T
I
L
E
V
E
R

B
E
A
M

S
P
E
C
I
M
E
N
k
,
=
K
a
F
i
g
u
r
e

1
0
.

T
h
e

D
o
u
b
l
e

C
a
n
t
i
l
e
v
e
r

B
e
a
m

(
D
C
B
)

s
p
e
c
i
m
e
n

w
i
t
h

a

f
o
r
m
u
l
a
f
o
r

s
t
r
e
s
s

i
n
t
e
n
s
i
t
y

a
s

a

f
u
n
c
t
i
o
n

o
f

c
r
a
c
k

o
p
e
n
i
n
g
,
c
r
a
c
k

l
e
n
g
t
h
,

a
n
d

s
p
e
c
i
m
e
n

g
e
o
m
e
t
r
y
.
V
E
h

(
3
h

(
a

+

0
.
6
h
)
2
+

h
3
)
4
(
(
a

+

0
.
6
h
)
3
+

h
2
a
)
E

=

Y
o
u
n
g
'
s

M
o
d
u
l
u
s
h

=

H
a
l
f

H
e
i
g
h
t
a

=

C
r
a
c
k

L
e
n
g
t
h
V

=

C
r
a
c
k

O
p
e
n
i
n
g
D
i
s
p
l
a
c
e
m
e
n
t
II
I
-
a
.
:
3
0I
-
A
P
o
1
0
r
t
Project 3589
0
-I
LU
W
0
.. 0
.O
STRESS INTENSITY, K
Figure 11. Schematic diagram showing the typical dependence of crack
velocity on stress intensity.
Final Report -26-
4i
I
KISCC
Project 3589
stress intensities, crack velocity becomes independent of stress intensity. In
this zone, commonly referred to as the plateau region, the velocity of cracking
is fairly constant over a wide range of stress intensities and crack growth
appears to be controlled by the rate of diffusion of aggressive species down the
length of the crack. At still higher stress intensities, crack velocities
increase once again because of mechanical overload effects.
Below a critical stress intensity level, no crack growth is observed.
The stress intensity marking the transition from the no-crack to crack-growth
regimes is called the threshold stress intensity, KIsc. If the stress inten-
sity lies below KIscc because the crack length is short or because the applied
stress is low, cracks will not propagate. Threshold stress intensities are
usually determined by allowing crack growth to occur in DCB specimens (at
steadily decreasing stress intensities) until no further growth is detected; the
stress intensity where crack growth ceases is KI
sec
Crack growth rate testing using DCB specimens can also be done under
controlled potential conditions to map the effect of potential on crack velocity.
In this case, the DCB specimen is connected to a potentiostat to control the
potential during the crack growth exposure.
The crack velocities measured in the laboratory can be used to estimate
the crack velocity expected in actual service equipment, provided appropriate
caution is exercised. Stress intensities can be assigned to cracked components
in operating equipment such as digesters, provided the potential, geometry of
the crack, and the magnitude of the stress is known (including all residual
stresses). The velocity of crack growth expected in service situations can
then be estimated by consulting the laboratory data relating crack velocity to
Fina.1 Report -27-
Project 3589
applied stress-intensity. However, care must be taken to ensure that the con-
ditions are the same in both the laboratory and field circumstances, since the
master curves for crack velocity will depend on variables such as potential,
solution composition, and metallurgical structure, to name a few.
CIRCULAR PATCH TEST SPECIMEN
The circular patch test specimen, sometimes called the ring weld
coupon, is designed to retain the high tensile residual stresses that are pre-
sent in a plate when it is welded under constraint. The specimen was originally
designed for studies of cold cracking of welds but has since found use as a
specimen for studies of stress corrosion cracking in weldments.
13- 14
The cir-
cular patch test specimen was used by PAPRICAN in its investigation of the
effectiveness of surface treatments to prevent caustic cracking of welds.
1 2
The test specimen is prepared by placing deep multipass welds of cir-
cular shape on both sides of a thick plate which is constrained against movement
during the welding process. Typically, constraint is provided by temporarily
welding the plate to the welding table while the ring welds are made. The
geometry of the specimen is shown schematically in cross-section in Fig. 12.
Tensile residual stresses as high as 70 ksi were measured in the weld metal of
circular patch test specimens. The heat-affected zone in the base metal near
the weld exhibited tensile residual stresses on the order of 50 ksi. These
residual stresses are similar to those expected (and measured) in continuous
digester welds.
1 2
The circular patch test specimen is a self-stressed test specimen
suitable for studies of caustic stress corrosion cracking. In the PAPRICAN
studies, the circular patch specimens were exposed to simulated cooking liquors
I
Final Report -28-
Final Report
Section A
Circular Patch Weld Coupon
Figure 12. Schematic diagram of the circular patch test specimen.
Project 3589 -29-
* 2 u p p e n R J 2 U f l U 9 a A f d U T B A T I D S U 8 9 8 1 9 A s i i o t u p d s j a q a o o 3 p a T i d d e s p o t p a m
a ~ e d a a a m o S u o f l T p u o D p a p ~ a n - s w a q p U T p e s o d x a s u a m T p a d s l 0 f l U a D U O p B A J a s q o
s I A 2 U T 2 o i o ~ f u o J B o f l O D s s e i s p ~ d u n - 2 u J l p i D a 9 f l p u 3 o l u a s o q o s i w n l u a l o d : 1 i
: 1 J o d a H d i T E u T d - o f - 6 8 S f ' : 0 a o a d
Final Report
LITERATURE CITED
1. Yeske, R. Stress corrosion cracking of continuous digesters for kraft
pulping. The Institute of Paper Chemistry, Final Report, Project 3544,
Jan., 1983.
2. Moskal, M. Measurement of residual stress in a continuous digester. Proc.
4th International Conference on Corrosion in the Pulp and Paper Industry,
Stockholm, NACE, 1983. p. 32.
3. Yeske, R. Liquor conditions in Kamyr continuous digesters. The Institute
of Paper Chemistry, Final Report, Project 3552, Aug., 1984.
4. Parkins, R. Fundamental aspects of stress corrosion cracking. NACE,
Houston, 1969. p. 601.
5. Singbeil, D.; Garner, A. First semiannual progress report of the research
program to investigate cracking of continuous digesters. Pulp Paper Res.
Inst. Can., July, 1983.
6. Singbeil, D.; Garner, A. Second semiannual progress report of the research
program to investigate cracking of continuous digesters. Pulp Paper Res.
Inst. Can., Jan., 1984.
7. Singbeil, D.; Garner, A. Final semiannual progress report of the research
program to investigate cracking of continuous digesters. Pulp Paper Res.
Inst. Can., Sept., 1984.
8. Pednekar, S.; Payer, J.; Beavers, J. First semiannual progress report on
evaluation of repair methods for SCC in continuous digesters. Battelle
Columbus Laboratories, Oct., 1983.
9. Pednekar, S.; Payer, J. Second semiannual progress report on evaluation
of repair methods for SCC in continuous digesters. Battelle Columbus
Laboratories, June, 1984.
10. Pednekar, S.; Payer, J. Third semiannual progress report on evaluation of
repair methods for SCC in continuous digesters. Battelle Columbus
Laboratories, Aug., 1984.
11. Pednekar, S.; Payer, J. Final report on evaluation of repair methods for
SCC in continuous digesters. Battelle Columbus Laboratories, Aug., 1984.
12. Singbeil, D.; Garner, A. Stress Corrosion Testing of Full Thickness
Digester Weldments. A Report to the TAPPI/IPC Digester Cracking Research
Committee, Pulp and Paper Res. Inst., Canada, Sept. 8, 1986.
13. Vagi, J.; Meister, R.; Randall, M. Weldment evaluation methods. OMIC
Report, Defense Metals Information Center, Battelle Columbus Laboratories,
Columbus, OH, 1968.
Project 3589 -31-
Project 3589
14. Standard Practice for Preparation of Stress Corrosion Test Specimens for
Weldments, ASTM G58-85, American Society for Testing Materials.
THE INSTITUTE OF PAPER CHEMISTRY
Ronald A. Yeske
Senior Research Associate
Engineering Materials/Corrosion Section
Engineering Division
Final Report
-32-
Project 3589
GLOSSARY
This is a glossary of terms used in connection with stress corrosion
cracking.
Active - Term applied to a metal that is experiencing active corrosion, i.e.,
dissolution unimpeded by the presence of a protective, passivating surface
film.
Anode - An electrode in an electrochemical cell where positive current enters
the electrolyte and oxidation (e.g., dissolution) occurs releasing
electrons.
Anodic Polarization - A shift in the electrochemical potential to potentials
higher than the spontaneous rest potential.
Anodic Protection - An electrochemical technique to prevent corrosion or
corrosion-assisted cracking of a metal by artificially raising its
potential. Potential shifts are achieved using a DC power supply to pass
electrical current through the wetted surface of the anodically protected
metal.
Carbon Steel - A steel containing less than 2% carbon and only residual
quantities of other elements, except those added for deoxidation (e.g.,
Si and Mn).
Cathodic Polarization - A shift in the electrochemical potential to potentials
lower than the spontaneous rest potential.
Cathode - An electrode in an electrochemical cell where reduction occurs
(consuming electrons) and positive current passes from the electrolyte
into the electrode. In anodic protection, a cathode is installed to
allow passage of current between the anodically protected structure and
the DC power supply.
Caustic Cracking - A form of stress corrosion cracking that occurs in carbon
steel exposed to hot caustic environments while simultaneously subjected
to high tensile stress; also called caustic embrittlement.
Current Density - An electric current passing through an electrode in an
electrochemical cell, divided by the surface area of the electrode;
typical units are mA/cm
2
.
DCB Specimen - A Double Cantilever Beam specimen is a test specimen that is
used for measurement of crack growth rates. The specimen is precracked
and then prestressed to promote crack growth during exposure to an
appropriate environment.
Ecorr - The rest potential that is spontaneously achieved when a metal is
immersed in an electrolyte.
-33- Final-Re'poft
A D V I D P u v ' t p B U 9 V J L - z ' s s a ~ i s p a T I d d v B u T A T O A u T s m i a l J j o n p o a d v
s e e e s s e a d i c a ' 4 l m O I 2 ' ) 8 I O V O J O J 9 0 0 U T A T I P T v D J u v t q D i 9 t U t A 3 T S u 9 3 U I s 8 a 8 i S
9 2 A O 1 3 0 9 T 9 a q v p d o a d d l e u v 0 3 T I R : 1 M O ~ q - p i d a z n s
v J O a i n s o d x a 2 u T a f l p 8 S s a r S 9 T S u 9 2 D j I R 3 S v a o u a s o a d a q A q p a o n p u T
S T R 3 8 m J 0 B u p p ~ v I O P 8 3 S T S S V - u O T S O J J Z O D J O U U O J V - U T p p 8 I 1 u O T 8 O . J I O 0 s s g a l S
* P I V P u v ~ s T R T I U 9 1 t o d
8 u 9 2 J 9 J 9 T v I 9 A T u f l a q s 9 ~ 3 n j T p s u o o 4 D ~ 4 p u v S 3 T O A 0 0 0 . 0 a q o 3 p o u f l o p
S T T 8 T l U u O o d 9 S O 4 A 9 P O I ~ l z T q ' - q U q 1 q J q V - ( 3 H s ) B P O - 1 3 0 9 U 9 2 0 O I P A H P . I v P u v : S
' T P T ~ u a o d 2 H S a q a A o q v A m ~ p T S T T v T l u a l o d
9 u a ~ a ; 8 2 O S O ' q m 8 P J 3 T 9 0 u 9 I J 8 I ' V - ( R D S ) 9 P O J 3 3 T a T 8 W O T B D p a I v t I ~ v S
a I O D ' 8 3 A T O I 2 : 1 9 T U V U T P a s l ~ m m i F
S T [ 8 V I M V ' - ' q M ~ P O A 9 T 4 0 9 A T s f l o a u v u o d s S T 3 8 4 1 T v T l u a l o d 8 U 7 T v 8 P u 9 O d 3 S O - d
3 u a m e a n s p a m a q 3 z o ; N j v m q u a t u q 8 2 v 2 T O A V s v p a s n a q u p o l v w i I T v T u a 1 o d
a T q T o n p o j d a a ' O T q v s v s 3 T q l T q x a 4
3
T q m O P J 0 T U V - 8 P O ' ~ z 9 T a 9 D u ~ l 9 J 9 '
g t p 0 3 3 u 9 3 v P v 9 3 4 O a 3 3 8 T O 9 q 3 P U B L 1 A 0 1 3 Z ) 8 T 8 U I U T P 9 s l I m m T T V 1 9 m V
u 8 a a m a a z u 8 3 9 J T P I 8 T I U 8 3 o d a q ! a 2 I V 1 t O A 1 0 9 D . 0 J O A T 1 j O m 0 ~ j ; 8 1 U V - T V T I U ~ O I
* q u e j a z n o ; o . 9 2 u s s u d A q p a s n v z T B T ~ u a l o d O P O J 3 T a U T 9 2 u 8 4 Z I O U L U O T ~ v z T I R T ~ c {
* s 9 l v z u O T S 6 1 1 O
3
S m O T S
A ~ ~ T T ~ u e q s q n s 3 v I n l l T J e o g j a f s B A T ~ Z o a ~ o d v J O u O T ~ v m 1 O j 8 1 U . - U O T 3 V A T S S ` S a
e a ~ v a l s q n s T V 3 0 m a 4 3 J O U O T S O I I O D J O 9 3 V B R S m " O T S A T T U T l u e I s q n s
2 B 4 2 m T T J a o u ; i n s v A q P a 1 O A 0 0 a z e I n s T V 8 9 J O 9 2 3 9 u j - D A T S S I a
* T P T ~ u a o d 3 S 9 I a 8 - T v T 3 U 9 3 O d j T n l
3
T D u a d o
- a j n ' z n a I T R Z ~ T T V I O M I I U T S 9 0 j a a : U T
a a i q l o a o s a j i o p u n o q U T I I
2
u ~ a 3 9 ' s a n o o 3 v q 3
2
u T N O 9 1 0
2
u T N p ~ i D . 1 I T f l u v 2 l j u I
- S u T p l e m B u T I n p p a T i d d v l e a q a q A q p a a a 3 v E u a a q S v q T V 3 9 m e s v q a q 3
J O 9 a n l a n Z 1 S O a z T U 1 B q a l W ' 1 P T a m V 0 3 3 u 9 3 v f P g 9 9 a v 9 q l - 9 U O Z P 9 3 ) 9 J V - 3 V H
* ~ 1 n l 3 n a l ~ s
T v T ~ 1 f l T T V 3 9 m 1 I U T S u ~ v a R 3 u a u f I p v u a ~ a m ~ q 9 D V J 1 8 I U T a t - X a v z p u n o g U T V I I
I l T V T J 3 9 9 u T T I I : 3 9 A I D A I o d v d n i 8 ) l S l 8 ( I f T T P - S U T V I I
9 B I T 0 1 3 0 9 T O u o m m o o 0 i A q P 9 3 i : 9 M S T V I ; ' m I R T T U I T s s T p u a a A 3 9 q
l ~ v 3 U O z T v 3 T I l 8 T 9 ; 0 I ~ n s a 1 v s u S u J f l a n o o u o I s o i I o D - u o T s o a a o o 0 T U V A T V 0
U S " T 2 u I j u d T a T j I u d p u I l a l I o a l o o l u v U T P a s 1 a I I 1 T T h u l V - 9 P O a 3 3 9 T
~ ~ i o d a ~ ~ ~ ~ ~ ~ I I U T & ~ ~ ~ ~ ~ 6 8 9 3 0 9 P ' . A3 a o d a ' a T v u T a
Project 3589 -35- Final Report
Threshold Stress Intensity - The stress intensity below which stress corrosion
cracks will not propagate; K sc
"scc
[PST HASELTON LIBRARY
li lii I~ 1 ~1111ll~lllll~l 1 1 1 1 1 IIIIIIDlhIIl
5 0602 01064723 0