23. - 25. 5.

2012, Brno, Czech Republic, EU

SENSITIVITY TO INTERGRANULAR ATTACK KINETICS OF HIGH-ALLOYED AUSTENITIC

STAINLESS STEELS WITH COPPER

Václav ŠEFL, Jaroslav BYSTRIANSKÝ

Vysoká škola chemicko-technologická, Technická 5, 166 28 Praha 6 vaclav.sefl@vscht.cz,

jaroslav.bystriansky@vscht.cz

Abstract
The sensitivity to intergranular attack of stainless austenitic steels is caused by gradient of chromium
concentration in the structure. This is caused by precipitation of carbides on the grain boundaries and thus
drop of chromium concentration in the adjacent region.
The degree of sensitization was studied using double-loop electropotentiokinetic reactivation method in 0.5M
H2SO4 with addition of NH4SCN. The samples of austenitic stainless steel AISI Super304H were aged at
different temperatures (620-720°C) and times (1-1000h). The structure of the samples was studied using
optical and scanning electron microscopy. The results showed that even the steel stabilized with Nb is prone
to sensitization; the experimental rate of sensitization fits the Larson-Miller equation.
Key words:
steel, intergranular, corrosion, precipitation

1. INTRODUCTION
Intergranular corrosion is a localized attack caused by structural inhomogeneity. The inhomogeneity can be
either local depletion of certain alloying element, precipitation of secondary phase or segregation. All of
these result in concentration gradient of certain element in the structure of the material. If the local
concentration of chromium in the case of stainless steels, drops below 14% wt., the material is unable to
form passive layer in this area the surface becomes active and it corrodes with high corrosion rates. Material
is therefore “sensitized”.

1.1. Sensitization
During the thermal or thermo-mechanical processes, precipitates can form in the steel structure. In the
austenitic steel, the precipitated phase causing sensitization is the chromium carbide (Cr23C6). The
precipitation of the carbide causes transport of chromium from surrounding area to the nucleation site.
The combination of precipitates varies with steel compositions, exposure length and temperature. There can
be large variety of carbides (M23C6, M6C), nitrides (Cr2N, TiN), carbonitrides (MX), Laves-phases (Fe2Mo,
Fe2Nb), δ-phase (FeNi)x(CrMo)y, χ-phase Fe36Cr12Mo10, G-phase (Ti4C2S2) and in the case of copper alloyed
steels even the ε-phase [1-9].
The composition of the steel does not only affect the composition of the secondary phases, but mainly their
solubility in the solid solution. This is more affected by lattice structure than the element composition. This
explains the difference in the sensitization temperature range between ferritic and austenitic steels. The
austenitic steels sensitization temperature range is 450-850°C [10, 11], and the range for the ferritic steels is
430-930°C [12]. The differences between the two are different diffusivities of carbon and other elements in
the given lattice. The higher carbon diffusivity in the ferritic lattice allows nucleation of the precipitate at lower
temperature while the low diffusivity of carbon in austenitic lattice requires higher temperatures to start the
nucleation.
 23. - 25. 5. 2012, Brno, Czech Republic, EU

1.2. Causes and prevention

The most common cause of sensitization is a wrong thermal treatment regime - i.e. exposure of the material
which is prone to sensitization in the temperature zone where carbides with high chromium content can form.
The “bulk” sensitization, e.g. exposure in the power plant, causes growth of the carbides on the grain
boundaries in all of the material. It should be noted, that chromium diffusion from the grains during long-term
high temperature can equalize the differences between the grain boundary and bulk grain chromium
concentration, thus causing “self-healing” [3, 4, 13]

1.3. Parameters affecting sensitization/desensitization

The degree of sensitization is proportional to the carbon content (L-grade steels) in the structure and can be
prevented by alloying with an element which forms carbides more easily than chromium, such as titanium or
niobium. These carbides begin to form at around 800-900°C and deplete the available carbon, which is thus
not available to form chromium carbides. The minimizing of carbon content is technically more demanding,
requires further raffination of steel and the material can still be sensitized during prolonged exposures [10,
14, 15].
Other parameter known to affect the sensitization/desensitization is cold work prior to the sensitization.
According to some authors [13, 16, 17], the cold work promotes both sensitization and desensitization of the
material (AISI 304). This is probably due to the increase of both short and long-distance chromium diffusion.
This was explained by increase of dislocation density and higher number of vacancies which promote the
long-distance transport, and grain refinement which enhances short-distance transport due to shorter
diffusive lengths.

1.4. Nucleation of carbides

The intergranular corrosion has severe effect on mechanical properties. The net of chromium carbides
alongside the grain boundaries causes drop-off of the grains and provides a net for a potential crack growth.
The nucleation of carbides on the grain boundaries is free-energy driven. The grain boundaries are basically
lattice imperfections with higher surface energies and thus are the preferential sites for nucleation. The twin-
boundary interfaces are generally more resistant compared to the grain boundaries [10].

The aim of the work was to determine the parameters of sensitization of steel AISI Super 304H using the
double-loop electro potentiokinetic reactivation. The goal is to determine the maximum sensitization for each
temperature and length of exposure.

2. EXPERIMENT

2.1. Samples
The material used for all the experiments was creep resistant austenitic stainless steel AISI Super 304H. The
composition is summarized in Tab. 1.

Tab. 1: Composition of the sample

C Si Mn Cu Cr Ni Nb N
Element
0.08 0.2 0.8 3 18 9 0.1 0.1
 23. - 25. 5. 2012, Brno, Czech Republic, EU

2.2. Thermal aging

The samples were thermally worked in an induction furnace as follows:
1. solution treatment - heating to 1250°C/30 min, water quenched
2. thermal aging: 620-720°C for 1-1000h.
Thermal aging simulates the exposure in a real LMP  T  C  log t   
power plant. The real
exposure does not have stable exposure temperature, fluctuations caused by start-up and failures in the
circuit can cause overheating of the material.

2.3. Sensitization quantification

The double-loop electro potentiokinetic reactivation method (DL-EPR) was used to quantify the degree of
sensitization. The method consists of anodic polarization of the sample to the passive region and then back
to the corrosion potential. The potential range and exact composition of the electrolyte was modified to
match the higher corrosion resistance of the material. We used cyclic polarization range -700 to +500mV/Ecor
and electrolyte composition of 0.5M H2SO4 + 0.01/0.001M NH4SCN. The active dissolution and reactivation
peak from the curve was integrated (Fig. 1) and the degree of sensitization (DOS) was calculated from
electrical charges using:

Qr (1)
DOS 
Qa

2.4. Structure study

The sensitization was assessed from the structure of the material. The samples were ground using grinding
paper up to P2500 roughness and then polished with TOPOL solution. As prepared surface was
potentiostatically etched in 10% wt. (NH4)2S2O8 at +6V/Ecor. The structure was studied using optical
3 microscope Zeis Axio Observer.
10
The metallography samples were then studied using
10 2 electron scanning microscope TESCAN VEGA with
backscatter electron detector.
Current density [A.m-2]

101
3. RESULTS AND DISCUSSION
100
The sensitization of the samples with different thermal
history is visible from Fig.-Fig. . It is visible, that the
10-1
sensitivity results are affected by the composition of
testing electrolyte. In the beginning we even used the
10-2
0.1M NH4SCN, but the forming oxygen made the
sensitization curves unreliable in most cases. The
10-3
curve support show the previously mentioned
phenomenon of sensitization and subsequent “self-
10-4
-0.8 -0.4 0 0.4 0.8 healing”. It is also visible, that the sensitization time
and thus the time to begin nucleation are higher for
Potential [V/ACLE]
lower temperatures. The chromium diffusion, which is
the limiting process of M23C6 forming, is temperature
Fig. 1: Integration of active and repassivation
driven and can be fitted by Larsson-Miller parameter:
peak of the potentiodynamic curve
(2)
 23. - 25. 5. 2012, Brno, Czech Republic, EU

60
50
0.5M H2SO4 0.01M NH4SCN
0.5M H2SO4 0.001M NH4SCN
40

Sensitization Qr/Qp [%]

Sensitization Qr/Qp [%]

40
30

20
20

10
0.5M H2SO4 0.01M NH4SCN
0.5M H2SO4 0.001M NH4SCN
0
0
0 200 400 600 800 1000
0 200 400 600 800 1000
Aging time [h]
Aging time [h]

Fig. 2: Sensitization as function of aging time; Fig. 3: Sensitization as function of aging time;
temperature 620°C temperature 670°C
100

80 The maximum measured degree of sensitization

of the sample aged at 670°C for 100h was
Sensitization Qr/Qp [%]

38/50% depending on the composition of

60 electrolyte. The same degree of sensitization for
the samples aged at 620/720°C was calculated
to be 3.5 hours, resp. 3500h. While the first
40
value (720°C) is in good agreement with the
experimental results (Fig. 3-4), the other sample,
20 aged at 620°C (Fig. 2), requires further thermal
0.5M H2SO4 0.01M NH4SCN aging.
0.5M H2SO4 0.001M NH4SCN
The samples with highest degree of
0 sensitization were then prepared to observe the
0 10 20 30 40 50 structure. For the sake of keeping this short, we
Aging time [h] only show the pictures from the scanning
electron microscope. The Fig. 5 shows the
Fig. 4: Sensitization as function of aging time; structure of sample aged at 620°C for 100h.
temperature 720°C There is clearly visible ditch structure with
carbides (dark lines) spread around the grain
boundaries and within the grain itself (grain carbide decoration). There are also two twins with carbides on
the incoherent interfaces and almost no precipitates at the coherent interfaces (longer sides of the twins).
Same phenomenon was observed on twins of the other samples (Fig. 6, 7); however there was almost no
grain boundary decoration. All three structures contain bright spots of NbC carbide which is distributed
alongside the former deformation texture. The annealing temperature was not high enough to dissolve these.
 23. - 25. 5. 2012, Brno, Czech Republic, EU

Fig. 5: Structure of the sample aged at 620°C for Fig. 6: Structure of the sample aged at 670°C for
1000h - detail of the grain boundary 100h

4. CONCLUSION
We have shown the effect of thermal aging on
austenitic stainless steel AISI Super304H.
Our electrochemical experiments using double-loop
potentiokinetic reactivation method showed that the
material is prone to sensitization - the time to reach
the maximum sensitization can be fitted by Larsson -
Miller parameter. The maximum degree of
sensitization differs depending on the electrolyte
composition - however it seems that electrolyte with
the higher activator content (0.01M NH4SCN) is too
aggressive for the material and thus yields unlikely
results (95% DOS). The maximum degree of
sensitization in the other electrolyte was
37.5/50/61% for samples aged at 620/670/720°C.
This suggests that higher temperatures cause higher
DOS probably due to the higher short-range
chromium diffusion rates. The structure all the
samples with highest DOS show a ditch structure; in
case of the 620°C/1000h aged sample the grain
Fig. 7: Structure of the sample aged at 670°C for boundary is decorated with carbides.
4h Our data showed that even the material designed for
exposure in aggressive high temperature
environments is prone to sensitization and that the sensitization and desensitization is temperature and
 23. - 25. 5. 2012, Brno, Czech Republic, EU

exposure time driven. This is especially important for practical application; the start-up or failures cause
overheating of the material. In the worst case scenario, the material can be sensitized during the start-up and
so-sensitized material can be then operated at much lower temperature - the desensitization (self-healing) in
this case can take up to several thousand hours.
The future work, currently in progress, will compare data from this and more conventional methods to
determine degree of sensitization (Streich, Strauss, oxalic acid test etc.) and will focus on further
mathematical modeling of such processes.

5. ACKNOWLEDGEMENT
This paper was created in the projects of MPO FR-TI1/086 and with financial support
from MSMT No 21/2012

6. LITERATURE
[1] MINAMI. Y.. KIMURA. H.. IHARA. Y. Microstructural changes in austenitic stainless steels during long-term aging.
Materials Science and Technology. 1986. vol. 2. p. 795-806.
[2] SOURMAIL. T. Precipitation in creep resistant austenitic stainless steels. Materials Science and Technology.
2001. vol. 17. p. 1-14.
[3] GONZALEZ-RODRIGUEZ. J.G.. CASALES. M.. ESPINOZA MEDINA. M.A.. ANGELES-CHAVEZ. C..
IZQUIERDO. G.. GUARDIAN. R. Microstructural evolution of Alloy 690 during sensitization at 700 °C. Materials
Characterization. 2003. vol. 51. no. 5. p. 309-314.
[4] SANDERSON. S.J. Sensitization times for intergranular corrosion in Alloy 800. Metal Science. 1977. vol. 11. no.
1. p. 23-30.
[5] BECKITT. F.R.. CLARK. B.R. The shape and mechanism of formation of M23C6 carbide in austenite. Acta
Metallurgica. 1967. vol. 15. no. 1. p. 113-129.
[6] ADAMSON. J.P.. MARTIN. J.W. The nucleation of M23C6 carbide particles in the grain boundaries of an
austenitic stainless steel. Acta Metallurgica. 1971. vol. 19. no. 10. p. 1015-1018.
[7] TRILLO. E.A.. MURR. L.E. A TEM investigation of M23C6 carbide precipitation behaviour on varying grain
boundary misorientations in 304 stainless steels. Journal of Materials Science. 1998. vol. 33. no. 5. p. 1263-
1271.
[8] GOLPAYEGANI. A.. LIU. F.. SVENSSON. H.. ANDERSSON. M.. ANDRÉN. H.-O. Microstructure of a Creep-
Resistant 10 Pct Chromium Steel Containing 250 ppm Boron. Metallurgical and Materials Transactions A. 2011.
vol. 42. no. 4. p. 940-951.
[9] ESCRIBA. D.M.. MATERNA-MORRIS. E.. PLAUT. R.L.. PADILHA. A.F. Chi-phase precipitation in a duplex
stainless steel. Materials Characterization. 2009. vol. 60. no. 11. p. 1214-1219.
[10] ENGELBERG. D.L.. Intergranular Corrosion. in: TONY. J.A.R. (Ed.) Shreir's Corrosion. Elsevier. Oxford. 2010.
pp. 810-827.
[11] MARSHALL. P.: Austenitic stainless steels: microstructure and mechanical properties. Elsevier Applied Science
Publishers. (1984)
[12] STREICHER. M.A.. Theory and application of evaluation tests for detecting susceptibility to intergranular attack in
stainless steels and related alloys - problems and opportunities. in. ASTM. Philadelphia. 1978. pp. 3-84.
[13] SINGH. R.. CHATTORAJ. I.. KUMAR. A.. RAVIKUMAR. B.. DEY. P. The effects of cold working on sensitization
and intergranular corrosion behavior of AISI 304 stainless steel. Metallurgical and Materials Transactions A.
2003. vol. 34. no. 11. p. 2441-2447.
[14] LIMA. A.S.. NASCIMENTO. A.M.. ABREU. H.F.G.. DE LIMA-NETO. P. Sensitization evaluation of the austenitic
stainless steel AISI 304L. 316L. 321 and 347. Journal of Materials Science. 2005. vol. 40. no. 1. p. 139-144.
[15] MATULA. M.. HYSPECKA. L.. SVOBODA. M.. VODAREK. V.. DAGBERT. C.. GALLAND. J.. STONAWSKA. Z..
TUMA. L. Intergranular corrosion of AISI 316L steel. Materials Characterization. 2001. vol. 46. no. 2–3. p. 203-
210.
[16] BRIANT. C.L.. MULFORD. R.A.. HALL. E.L. SENSITIZATION OF AUSTENITIC STAINLESS STEELS. I.
CONTROLLED PURITY ALLOYS. Corrosion. 1982. vol. 38. no. 9. p. 468-477.
[17] BRIANT. C.. RITTER. A. The effect of martensite on the sensitization of low carbon 304 stainless steel.
Metallurgical and Materials Transactions A. 1981. vol. 12. no. 5. p. 910-913.