Abstract

The sudden collapse of the lining refractory in a waste incineration plant during a regular
maintenance has been investigated. Preliminarily visual inspection showed corrosion of the
support hangers that have been installed to enhance the stability and the strength of the lining
refractory. One of the ends of the 310 SS support hanger was exposed to the incinerator hot
gas steam with a temperature of 1100 oC. Visual, optical and SEM inspection have been
conducted on the failed support hangers to figure out the cause of the failure. Upon
examination, the support hangers showed the formation of thick oxide scale and a phase
transformation in some parts. The second phase particles are believed to be sigma-phase
particles with some traces of δ-ferrite that gave the steel its magnetic properties. Sigma phase
particles are very brittle, but cause no problem if the temperature is kept at the range in which
sigma-phase forms. However, impact loading can lead to catastrophic failure for the
sigmatized stainless steels.

Introduction:
During routine shut down of a water incinerator plant to repair the worn areas of the
refractory lining, approximately 12m2 collapsed without warning from one of the vertical
walls of the plant. An initial examination of the failure revealed that the refractory support
hangers made from 310 SS had corroded leading to the collapse of the refractory lining on the
incinerator wall. When in operation, the incinerator had been used to burn a wide variety of
hazardous chemicals at temperature above 1100 oC. The refractory lining is made up of three
layers; an inner insulating board, a second layer of light insulating refractory and a third outer
layer of heavier and more erosion resistant hot face refractory. Running through the refractory
lining are stainless steel support hangers, which are screwed into the incinerator wall, figure1.
The hangers give the refractory lining the strength it requires to support itself [1].

Before the failure, the lining and the supports were in service for five years and some of the
lining has worn out already so the tips of the hangers are exposed to the hot gas steam
temperature of about 1100 oC. The failure occurred while the operators were jack-hammering
the lining to remove the heat-affected refractory.

Analysis:
The Hot Steam:
The exhaust gas generated from the incineration of solid wastes contains corrosive
substances, such as dust, acid gases (SOx, HCl, NOx, CO2, and other gases), and air which is
readily available for the combustion process. Such oxidizing environments are highly
corrosive to metals and especially stainless steels with the most corrosive substances are HCl
and SO2.

310 Stainless Steel:

The support hangers are made of 310SS which is an austenitic stainless steel with high creep
strengths and high resistance to high-temperature corrosion [2]. The composition of the 310SS
is shown in table1.

Table 1: Composition ranges for 310 grade stainless steel [3]

Grade C Mn Si P S Cr Mo Ni N
310 min. - - - - - 24.0 - 19.0 -

max. 0.25 2.00 1.50 0.045 0.030 26.0 22.0

1
This type of stainless steel is generally used at temperatures starting from about 800 or 900
°C [3] and designed to resist oxidation in continuous service at temperatures up to 1150°C [4] in
both oxidizing and carburizing atmospheres. The typical heat treatment for such stainless
steel is annealing for maximum corrosion resistance. Typical applications for 310SS include:
furnace parts, heat exchangers, and oil burner parts.

Visual Inspection of the Support Hangers:

The support hangers, and as mentioned above, are made of 310SS with a nominal cross
section of 8mm and shaped in a V-shape as shown in figure 1. Different specimens have been
taken from the support hanger to be tested under the optical and scanning electron
microscopes. The locations of the different specimens are shown in figure 2. After the failure
of the support hangers, the thickness measurements were noted down and tabulated in table 2;
table 2 shows the measurements for only one of the support hangers. The measurements of
thickness were taken each 30mm from the end of the support hanger near the incinerator wall.
Most notably, most of the support hangers have failed on the same location where the hot
face refractory contacts the thermal insulating refractory as shown in figures 3 and 4. The
failed side is the lower side in figure 2. At the location of failure, we notice thick oxide film
formed and this can be noted from table 2 at a distance of 150mm from the exposed end of
the support hanger. Unfortunately, the fracture surface have oxidized already and no further
visual inspection can be carried out. When the support hangers were tested with a magnet,
they showed a high magnetism with the magnetic properties increase from the pivoted end
toward the failed end near the hot gas steam. This indicates that there was a phase change in
the stainless steel since stainless steels are non-magnetic. Building on that and on the
thickness of the oxide scale formed, we can conclude that the corrosion product is not a
protective Cr2O3 layer.

Table 2: Measurement of thickness of support hanger 2 after failure (distances taken from the exposed end)
Support Hanger 2
Left- side (Upper side) Right-side (Lower-side)
Intervals (mm) T1 (mm) T2 (mm) T (mm) Intervals (mm) T1 (mm) T2 (mm) T (mm)
30 30
60 6 5.9 5.95 60
90 6.9 7 6.95 90
120 8 8.1 8.05 120
150 8.2 8.1 8.15 150 11.6 12.6 12.1
180 8.1 8.1 8.1 180 8.2 8.2 8.2
210 8.1 8.1 8.1 210 8.2 8.1 8.15
240 8.4 8.2 8.3 240 8.2 8.3 8.25
Average 7.67 7.64 7.66 Average 9.05 9.30 9.18

Optical Microscope:
For testing under the microscopes, different specimens have been cut out from one of the
failed hangers; the location of each specimen is shown in figure 2. The specimens were
painted to distinguish them; white, yellow, blue, and red starting from the exposed end
toward the pivoted one as shown in figure 2. The observations for each specimen are as
follows:
White: this specimen is from the nearest location to the exposed end. Visual inspection of
the specimen shows that it is with the smallest diameter among all specimens. This
indicates high corrosion product has formed in this area; most of the oxide scale has fallen
off so we can only see very thin intermittent oxide layer under the microscope. Some
internal oxidation has been noted; this internal precipitates will most probably be Cr2O3. We

2
 can notice the formation of second phase particles all the way to the centre of the specimen
with the shapes vary between needle, big needle, and some globular shapes (figures 5 and
6).

Yellow: if we move further toward the pivoted end of the support hanger, we will notice
similar observations with very thin intermittent oxide layer with small amount of internal
precipitates of Cr2O3. However, a distinct layer of second phase particles have formed
beneath the oxide layer with fewer particles toward the centre.

Blue: if we take another step toward the pivoted end of the support hanger; we can notice
almost similar observations to those found in the yellow specimens. The layer of the formed
second phase particles is almost the same width of the oxide layer.

Red: this is the last specimen that has been taken from the support hanger near the pivoted
end. This specimen is with the largest diameter where only small amount of oxide layer has
formed on the outer surface of the specimen. Similarly, only small dispersed amount of
internal precipitates has formed. Most notably, there is no second phase particle formation
in this specimen. This microstructure is typical to that of ordinary stainless steels.

We tested one more specimen under the optical microscope which is the cross section of the
support hanger from the nearest location to the incinerator wall. From this specimen we
noticed that the crack is brittle; no elongation lips, and had propagated along the grain
boundaries rather than across them. (Unfortunately, no figures provided since I could not find
the specimens to take more micrographs!)

From the optical micrographs we have concluded that the metal of the hanger was not
showing typical stainless steel behavior at the time of the failure due to the very thick oxide
layer formed on the metal. In addition, we noticed the formation of the second phase particles
in the bulk of the metal, and the presence of the second phase particles increases with the
increase of temperature along the refractory from no formation at the pivoted end to 'the
highest density with equal distribution' at the exposed end. Not all of the hangers failed at the
same time; some of them failed prior the main failure since there were already oxidized. The
other ones failed because they were not able to withstand the excessive load applied on them.
No clue has been found on the effect of the refractory type on the corrosion morphology;
although the corrosion increases as the temperature increases. The crack propagation was
intergranular rather than transgranular.

Scanning Electron Microscope:

We examined the white and yellow specimens under the SEM to get better micrographs and
to conduct an elemental analysis of the different phases we obtained using the optical
microscope. Figure 7 shows the cross section of the white specimen when tested under the
microscope. Different spectra have been used to figure out the composition of the different
parts and phases present. From figure7-spectrum1 we notice that the dense oxide layer
formed is mainly composed of Cr, Fe, and O. this indicates that this layer is a spinel layer
with some chromia. This spinel is what caused the fast growth of the thick oxide layer with a
thickness of about 3.4mm (figure 12). Internal precipitates present in the alloy are chromia as
shown in figure7-spectrum3. When the second phase particles were examined, they showed
high concentration of Cr and the area adjacent to these particles are Cr-depleted zones which
indicates that there was Cr migration in the alloy as shown in figure7-spectra4 and 5. To find
what exactly the compositions of the different second phase particles are, we conducted an

3
elemental analysis for these particles. Although they have different shapes; globular, big
needle, and small needle shapes, they have exactly the same composition as shown figure8-
spectra1, 2, and 4. Again, the adjacent zones are Cr-depleted zones (spectrum 3).

Also, we have conducted similar analysis for the yellow specimen. The observations for this
specimen are similar to those found for the white specimen. The thick oxide layer is a spinel
with some traces of chromia (figure9-spectrum1). Internal precipitates are chromia (figure9-
spectra 2, 3, and 4). The dark internal precipitates at location 5 are not a new phase; they are
Si-rich particles in which Si was introduced during the polishing process of the specimen as
shown in figure9-specturm 5. Like the case for the white specimen, the different-in-shape
second phase particles are with the same compositions which are Cr-rich particles (figure10-
sepctra 1, and 3) and the adjacent zones are Cr-depleted zones (figure10-spectrum 2).

Finally, we took a part from the support hanger that was embedded in the refractory left with
a thick oxide layer not fallen off yet and we tested that under the SEM. Figure11 shows the
location from which the sample has been taken. From figure 12, we notice that the thickness
of the oxide scale is 3.4mm which is very thick and unusual for stainless steel as discussed
earlier. When we tested several locations across the oxide scale we got exactly the same
composition which is spinel with high amounts of Cr, Fe, and O; therefore, only one
spectrum is shown in figure 12. In a closer look (figure13), we notice that the oxide scale
starts to spall; this explains the small amount of oxide scale observed for the white and
yellow specimens under the optical microscope as discussed above. Spectrum 1 shows high C
present in this area because of the epoxy resin entered the interface between the oxide and the
metal when the oxide scale started to detach. Spectra 2, 3, and 4 show the spinel composition
of the oxide layer, Cr-depleted zone, and Cr-rich particles respectively.

Second Phase Particles:

As discussed above, the stainless steel has a single phase, austenite. However, the specimens
tested showed the formation of second phase particles and their presence depends on the
temperature distribution; as the temperature of the support hanger increases, the presence of
the second phase particles increases. Phase changes in the stainless steel can influence the
mechanical as well as the chemical properties of the stainless steel [5]. The possible phase
change at high temperature for stainless steel is sigma-phase.

The sigma-phase is a brittle, Cr-Mo [4] or Cr-Fe rich [6] intermetallic phase. At temperatures
between approximately 600 and 900 °C and for prolonged exposure time, sigma phase
precipitates in high Cr, Mo, or Si-containing stainless steels; Grades 310, 314, 316, and 317 [2]
are among the most susceptible austenitic stainless steels. Hau et al state that: "[…] it would
appear that there should be little consequence as long as the affected components
continuously operate at the elevated temperature. However, cracking could occur if the
components were impact loaded or excessively stressed during maintenance work" [7]. Sigma
phase depletes the chromium from the matrix and the main characteristics of these sigma-
phase particles are: non-magnetic, very low impact strength, and very low intergranular,
pitting and crevice corrosion resistance [4], [8]. Table 3 shows the second phase particles that
are most likely to form in stainless steels. From this table and for sigma-phase, we notice that
the formation of sigma phase is faster when formed from δ-ferrite. This is very important
since δ-ferrite is magnetic and the steel showed high magnetization as mentioned earlier.
However, the mechanism of nucleation is still a matter of controversy, particularly on the role
of δ-ferrite and M23C6 in the nucleation process. Studies have reported its formation
associated with or without the dissolution of carbides [6]. This is strongly supports our

4
 conclusion about the type of the second phase particles since we did not observe any carbides
in the steel. Therefore, we can conclude that the second phase particles are mainly sigma-
particles with some traces of δ-ferrite phase in the alloy that gave the steel its magnetic
property.

Table 3: Most common second phase particles in stainless steels [4]

Recommendations:
The main causes of the failure are the formation of second phase particles and the impact
loading caused by jack-hammering of the lining refractory during maintenance. Therefore:
• Normal austenitic stainless steels should not be used anymore. A good alternative of
310SS is 310LN with nitrogen and 347L with the Nb that will prevent the formation
of carbides and second phase particles. Cold working of the stainless steels was found
to initiate the formation of second phase particles whereas annealing is found to
suppress their formation.
• Maintenance interval should be reduced for better inspection and anticipation of
proper failures in the refractory.
• The use of hammering should be stopped as the excessive impact loading caused the
sudden failure of the lining refractory.
• The failure occurred after five years of service at which the ends of the support
hangers were exposed to the hot gas steam with very high temperature almost near
their limits (1150 oC); therefore, the replacement schedule should account for the
minimum thickness of the lining refractory that will not cause the exposure of the
support hangers to the hot gas steam.

Conclusion
From the analysis of the failed support hangers of the waste incinerator plant, we found that
the cause of the failure is the corroded support hangers which are made of 310SS. However,
the oxide layer thickness is unusual for austenitic stainless steels. Closer investigation
revealed that there are second phase particles in the metal. No consensus about the type of
these phase particles has been approached, but depending on many literatures and published
papers, they are believed to be sigma-phase particles with some δ-ferrite that gave the steel

5
 the magnetism. The use of the 310SS or any austenitic stainless steel should be stopped in the
incinerator plants. A good alternative is the 310LN with the nitrogen addition or the 347L
with Nb alloying that both will suppress the formation of second phase particles and carbides
in the austenitic stainless steels. Also, the maintenance schedule and practice should be
improved to minimize the intervals between maintenances and the hammering technique for
removal of the worn lining refractory should be prohibited.

References:
1. Paul Jordan, "Case Study: High Temperature Components"; University of Manchester,
2007
2. "Stainless Steel - High Temperature Resistance";
[http://www.azom.com/details.asp?ArticleID=1175]
3. "Stainless Steel - Grade 310";
[http://www.azom.com/details.asp?ArticleID=966#_Composition]
4. ASM Handbook, Volume 13: Corrosion
5. "Phase Changes in High-Temperature Service ";
[http://httd.njuct.edu.cn/MatWeb/gas/ka_ht/ht_phcha.htm]
6. T.Sourmail; "Precipitates in creep resistant austenitic stainless steels";
Cambridge University;
[http://www.msm.cam.ac.uk/phase-trans/2003/sourmail.review/index.html]
7. Jorge Hau et al; "Sigma Phase Embrittlement of Stainless Steel in FCC Service";
[www.cap-eng.com/news.asp?show=art&artid=235]
8. "Examinations to the sigma-phase"
[http://www.metallograf.de/start-eng.htm?untersuchungen-
eng/sigmaphase/sigmaphase.htm]

6
Appendix
Figures addressed in the report:

Figure 1: Schematic of a refractory support hanger and its relationship relative to the layers of refractory.

Figure 2: Schematic diagram showing the locations of the samples cut from the failed refractory support
hangers.

7
Figure 3: The failure location for support hanger 2 Figure 4: The failure location for support hanger 3

Figure 5: White specimen cross section under optical microscope (10x)

Figure 6: White specimen cross section under optical microscope (20x)

8
4

5 3

Figure 7: SEM micrograph of the white specimen with EDX analysis

9
Figure 8: SEM micrograph of the second phase particles in the white specimen with EDX analysis

10
 1

Figure 9: SEM micrograph of he yellow specimen with EDX analysis

11
 1

Figure10: SEM micrograph of the second phase particles in the yellow specimen with EDX analysis

Figure 11: Refractory from which the specimen with thick oxide layer has been taken

12
 2 3 4
1

Figure 12: The specimen took from the refractory (figure11)

1 4

2 3

Figure 13: Oxide/Metal interface for

specimen in figure 12

13