15th North American Waste to Energy Conference May May 21-23, 2007, Miami, Florida USA

NAWTEC15-3220

High Temperature Corrosion Resistance of Different Commercial Alloys Under Various Corrosive Environments

Shang-Hsiu Lee and Marco J. Castaldi Department of Earth & Environmental Engineering (HKSM)

Columbia University, New York, NY 10027

ABSTRACT

High temperature corrosion is a major operating problem because it results in unscheduled shutdowns in Waste-to-Energy (WTE) plants and accounts for a significant fraction of the total operating cost of WTE plants. Due to the heterogeneous nature of municipal solid waste (MSW) fuel and the presence of aggressive elements such as sulfur and chlorine, WTE plants have higher corrosion rates than coal-fired power plants which operate at higher temperature. To reduce corrosion rates while maximizing the heat recovery efficiency has long been a critical task for WTE operators.

Past researchers focused on high temperature corrosion mechanisms and have identified important factors which affect the corrosion rate [1-4]. Also, there have been many laboratory tests seeking to classify the effects of these corrosion factors. However, many tests were performed under isothermal conditions where temperatures of flue gas and metal surface were the same and did not incorporate the synergistic effect of the thermal gradient between environment (flue gas) and metal surface. This paper presents a corrosion resistance test using an apparatus that can maintain a well controlled thermal gradient between the environment and the surface of the metals tested for corrosion resistance. Two commercial substrates (steels SA213-Til and NSSER-4) were tested under different corrosive environments. The post-test investigation consisted of mass loss measurement of tested coupons, observation of cross-sectional morphology by scanning electron microscopy (SEM), and elemental analysis of corrosion products by energy dispersive spectrometry (EDS).

The stainless steel NSSER-4 showed good corrosion resistance within the metal temperature range of 500 DC to 630 DC. The alloy steel SA213-Til had an acceptable corrosion resistance at metal temperatures up to 540 DC, and the performance decreased dramatically at higher temperatures.

INTRODUCTION

High temperature corrosion mechanisms on waterwall and superheater in Waste-to-Energy (WTE) plants have been investigated extensively in laboratory and field tests. Many factors such as flue gas composltlon, operating gas and metal temperature, fluctuations of flue gas temperature, thermal gradient between flue gas and metal surface, and characteristics of molten salts deposits are all considered to be crucial factor for high temperature corrosion.

Work done previously has been done mostly to compare and evaluate the corrosion tendencies of materials by controlling one of the corrosion factors mentioned above [5-7]. However, these tests may not be adequate for forecasting long term and synergistic effects of these corrosion factors on the tube life. Especially, some dynamic factors,

199

such as fluctuations in the flue gas temperature and thermal gradient between the gas and metal surface, which are difficult to be reproduced and controlled in laboratory tests have been found to be accountable for the breakdown of the protective oxide scales on metal surface and attendant increased corrosion [8]. Also, the thermal gradient between the gas and metal surface has been observed to strongly influence the deposition behaviors and rates of molten salts that contain corrosive compounds from WTE combustion [1].

In order to elucidate the synergistic effects of these corrosion factors, the authors developed an apparatus which can maintain a thermal gradient between a representative WTE combustion gas and tested samples that are maintained at representative waterwall and superheater temperatures. In this

Copyright © 2007 by ASME

paper, two commercial steels (SA213-Til and NSSER-4) were tested under controlled thennal gradients under experimentally controlled corrosive gases. Future tests will incorporate factors of temperature fluctuations of flue gas and molten salt deposits.

EXPERIMENTAL PROCEDURE

Two commercial steels, SA213-Til and NSSER-4 have been carried out in the test. It is well known that SA213-Til has been widely used as the base tube of superheater. NSSER-4 is developed by a Japanese steel company which claims its good resistance to chlorine corrosion. The chemical compositions of these two steels are summarized in Table 1. Both steels were cut to dimensions of 1 x l xO.078 in by a water-cooled machine. The preparation of samples followed the standard procedures of ASTM G 1-03 which included degreasing in an organic solvent, grinding with 120-grit SiC paper, ultrasonic cleaning after each of the above steps, and drying at 100 °C for one hour. After that the clean, dry samples were measured and weighted.

Figure 1 shows the configuration of the apparatus. Corrosion tests were carried out in a

furnace at a gas temperature of 7S0oC and 100 hours duration. The furnace had a constant temperature zone of 13 inches for exposing the test samples. A sample carrier made of a square stainless steel tube was placed in the furnace. Cooling air was passed through the tube and exhausted outside of the furnace into a vent. A schematic cross-section of the furnace and sample carrier is illustrated as Fig. 2. The sample carrier had six indents to place samples and each indent was coupled to a thermocouple to monitor the sample temperature. The desired lower sample temperature was attained by controlling the flow rate of the cooling air through the tubular passage, thus inducing the desired thermal gradient between outer gas flow and metal surface.

The temperature profiles of six samples and their average temperatures during the test are shown in Fig. 3. During the 100-hour test, each sample was maintained at stable temperature. By controlling the flow rate of the cooling air, surface temperatures of six samples were controlled from SOO-630oC, and the thermal gradient (at furnace temperature of 7S0°C) was controlled from 120-2S0°C. This setup enabled a single test to yield five plots of temperature versus time, with one duplicate set for comparison.

Table 1. Chemical Composition (wt. %) of tested samples

Steel SA213-Ti1 NSSER-4

C 0.08 0.04

Si 0.28 2.S

Mn 0.43 0.8

p 0.014

S Ni 0.002

13.1

Fig.1 Configuration of the apparatus

200

Cr LOS 17.3

Mo 0.S2 2.S

Fe bal. bal.

Copyright © 2007 by ASME

Inlet

front rear

- - ----- .. ---------- .. -------'\ Fluid tlow:::: CSTR Synthetic combustion gas Tgas:::: 750 °c / - - ---- - - - - - - - - - .. - - - - - - - - - - - -+< '\ .I Outlet .... -- - - - - - - - ........ -- - - - - - - - - - - ... � ... -- - - - -- - -

Thermocouples Sample ;!' T sample:::: 630°C

-------- --------- ----------- -- -- � Cooling air .... ------ ---- -- - -- -------- -

Fig.2 The schematic cross-section of the furnace and sample carrier

1000 .,---------------------,

800

IT 600 o�

ci E (!!. 400

200

coupon 1 coupon 2 coupon 3 coupon 4 coupon 5 coupon 6

O�-,---._--_.--�---._---r� o 20 40 60 80 100

Time, [hr)

Fig.3 Temperature profiles of six tested samples

The composition of synthetic corrosive gas used in the test consisted of 8% O2, 12% CO2, 500ppmv HCI, 100ppmv S02, 15% water vapor, and balanced with N2 at a total gas flow rate of 500 mllmin. The gas was preheated up to 200 °c before injecting into the furnace.

After the tests, samples were cooled down in the furnace and were then prepared for corrosion product analysis, metallographic corrosion rate analysis, and mass loss corrosion rate measurement. Analysis of the surface morphology and elementary composition of the corrosion product was conducted by using a scanning electron

201

microscope (SEM) equipped with an energy dispersive spectroscopy (EDS) unit.

The measurement of mass loss corrosion rate measurement followed the standard procedure of ASTM GI-03. A cleaning cycle that combined the procedures of light brushing, ultrasonic cleaning, chemical cleaning, and mass loss measurement was repeated several times until the corrosion products were removed completely. Final sample weights were measured to the nearest 0.001 mg and adjusted for blank weight losses in the cleaning process. Corrosion rates for the bare samples were calculated using mass loss, exposed area, and test

Copyright © 2007 by ASME

duration.

Finally, the metal samples were sectioned, mounted, ground, and polished for metallographic corrosion rate analysis. The analysis of cross-sectional morphology was done by using the SEM.

RESULTS and DISCUSSIONS

1. NSSER-4

Figure 4(a) is a photograph of a metal sample after a 100-hr test at metal temperature of 630°C; Figure 4(b) is the SEM image of the external topography of the scale formed on the sample surface. It can be seen that the corroded areas

(scales) spread on the surface unevenly and exhibit a certain level of adhesion which can not be peeled off easily.

The elementary analysis of the surface of each sample is shown as Fig. 5(a), (b), (c), and (d). The Fe and Cr concentrations of less corroded areas are higher than that of corroded areas while the ° and S concentrations exhibit the opposite situation. In addition, there is not a clear correlation between compositions of scales and metal temperatures. According to thermodynamic calculations, these scales are expected to be formed mostly Fe203, FeS, and Cr203. The other elements such as Si and Ni do not show similar trends, and CI is either not identified or exists at a very low concentration on samples' surfaces.

Fig.4 Surface morphologies of the sample for 100 hrs.

(a) Fe content 100

(b) 0 content 12

80

I II 60 "" P l 40 'i ..

.. I I

20 ...

.. corroded area .. less corroded area

10 I : corroded .,e. I I less corroded area

I

� * I 6

I 'i .. I J ""

l>

temp. ('C) temp. ('C)

202 Copyright © 2007 by ASME

(c) 5 content (d) Cr content 40 35

A 35

I 30

A I] • corroded area

I 4 less corroded area 30 25

25 20

!j � � ..

20 i

15 � �

15 II 10

10 I t� �

I A comxled area A

i 5 A less corroded area

·5 480 500 520 540 560 580 600 620 640 660 480 500 520 540 560 580 600 620 640 660

temp. (oC) temp. (OC)

Fig.5 Elementary analyses of the samples' surfaces: (a) Fe, (b) 0, (c) S, and (d) Cr

Figure 6 shows the cross sectional SEM image of the remaining sample substrate after chemical cleaning. The accompanying table summarizes the corresponding wt. % composition appearing in Figure 6. High CI concentration is found on the surface of the remaining substrate, and it decreases with increasing distance below the sample surface while Cr and Fe concentrations reveal opposite tendencies. According to the thermodynamic calculation, it is assumed that reaction products on the scale/metal interface consist of iron chlorides. Combining with the results of elementary analysis of samples' surfaces mentioned above, this helps to describe the corrosion mechanism as follows:

203

chloride (or chlorine) penetrates the oxide scale through cracks and pores of the oxide scale due to the chemical potential/temperature gradient and accumulates at the metal/oxide scale [9].

The cross sectional SEM images of the remaining sample substrate are shown in Fig. 7(a) and (b), for metal temperatures of 630 °C and 502 °C, respectively. It can be seen that the sample substrate is oxidized, and the crack and cavity occur after the metal chloride diffuses outward through the scale.

Copyright © 2007 by AS ME

Composition wt% Area 0 Si Cl Cr Fe Mn Ni P S

a 8.85 0.01 39.53 2.28 46.99 0.00 1.48 0.35 0.52 b 0.00 0.00 39.38 2.94 50.09 0.00 2.19 0.49 0.67 c 0.24 0.27 41.75 3.47 52.24 0.00 1.11 0.29 0.64 d 0.00 2.38 3.16 20.48 70.79 0.00 1.36 0.36 1.47 e 0.00 2.14 1.70 18.51 74.52 0.00 1.08 0.46 1.59 f 0.00 2.32 1.06 18.97 74.36 0.11 1.21 0.54 1.44

g 0.00 2.39 1.84 19.12 71.43 0.00 0.07 0.34 1.55

Fig.6 Cross sectional SEM image of sample and the elementary analysis of each spot

Fig.7 Cross sectional SEM images of remaining sample substrates

2. SA213-TlI

Figure 8 is a photograph of the sample after a 100-hour test at the metal temperature of 600°C. The whole surface of the sample is corroded and forms a crusty external scale which can be peeled off easily after cools down.

The elementary analysis of each sample's surface is shown in Fig. 9(a), (b), (c), and (d). Both the Fe and 0 concentrations of the outer scales are higher than that of inner scales. Higher concentrations of S and Cr are found in the inner scale than in the outer scale. The Fe concentration of the inner scale decreases as the metal temperature increases while the 0 concentration shows an opposite trend. Therefore, it can be assumed that when the metal temperature increases

more Fe from inner scales is oxidized.

3. Comparison of mass loss corrosion rate

Figure 10 shows the mass loss corrosion rate of these two metals. The stainless steel NSSER-4 shows a good corrosion resistance of less than 10 mils per year within the metal temperature range of from 500°C to 630°C. When the metal temperature is below 540°C, the corrosion rate of alloy steel SA213-T II is about 30 mils per year and increases dramatically when the metal temperature is higher than 560°C. Therefore, this alloy steel is not designed to apply at this high temperature, a result that is in agreement with the practical situation reported from WTE plants [10].

Fig.8 Surface morphologies of the sample for 100 hrs.

204 Copyright © 2007 by ASME

(a) Fe content (b) 0 content 100 1.8

I 1.6

I I •

I 95 1.4 • • • 1.2

11 90 1.0 •

l 85 1 !I l 0.8 .. .. J J 0.6

1 II 80 2 0.4

75 1 1 I 0.2

• oUle, scale I l · oUle"caleJ o inner scale 0.0 • o inner scale

70 -0.2

460 480 500 520 540 560 580 600 620 640 660 460 480 500 520 540 580 580 600 620 640 660

temp. (0C) temp. (oC)

(c) 5 content (d) Cr content 14 "

I I � OUIe,scolel

I • l:! 12 inner scat. 12 OUIer scale I 0 inner scale

y1 10 10

1 £

l � I 12 6 'i 'i

•

� • •

8 • • 0 • • • • •

-2 460 480 500 520 540 580 580 600 620 640 660 460 480 500 520 540 560 580 600 620 640 660

temp. (oC) temp. ('C)

Fig.9 Elementary analyses of the samples' surfaces; (a) Fe, (b) 0, and (c) S

140

120

I • NSSER-4 I •

� • SA213 T11 ., QI 100 >. ... QI a. .!!! 80

I ••

.! 60 .,

a::: c 40 0 'iii • • 0 t 20 0 U • ••

• • 0

460 480 500 520 540 560 580 600 620 640 660

Metal Temperature (0C)

Fig. lO Comparison of mass loss corrosion rate

205 Copyright © 2007 by ASME

CONCLUSIONS

Two commercial steels were tested under the corrosive environment for 100 hrs. The stainless steel NSSER-4 showed a good corrosion resistance within the metal temperature range of from 500°C to 630°C. The alloy steel SA213-TIl had an acceptable corrosion resistance below the metal temperature of 540°C, and the performance degraded dramatically when the metal temperature was increased further.

It was not the objective of these tests to compare the corrosion resistance of these two steels since the alloy steel SA213-TIl is usually used as tubing with other alloys coated on it, and also the costs of these alloys are not comparable. The effect of thermal gradient will be clearer in following tests that will compare the result of applying molten salts on metal surface with that of without molten salts. In addition, the metal temperature in future tests needs to be controlled at a lower range, from 400°C to 450°C, which is typical of conditions in WTE plants, by improving the design of the apparatus.

REFERENCES

1. Kawahara, Y., High temperature corrosion

mechanisms and effect of alloying elements for

materials used in waste incineration

environment. Corrosion Science, 2002. 44(2):

p. 223.

2. Covino, B.S., Jr. , et aI., The Role of Ash

Deposits in the High Temperature Corrosion of

Boiler Tubes, in Corrosion 2003. 2003, NACE

International, 1440 South Creek, Houston, TX

77084-4906 USA: San Diego, CA

3. Radernarkers, P., W. Hesseling, and lv.d.

Wetering, Review on corrosion in waste

incinerators, and possible effect of bromine.

2002, TNO Industrial Technology.

4. F. H. Stott, c.Y.S., The influence of HCI on the

oxidation of iron at elevated temperatures.

Materials and Corrosion, 2000. 51(5): p.

277-286.

206

5. Mohanty, B.P. and D.A. Shores, Role of

chlorides in hot corrosion of a cast Fe-Cr-Ni

alloy. Part I: Experimental studies. Corrosion

Science, 2004. 46(12): p. 2893.

6. N. Otsuka, Y.F.Y.K.T.H., Laboratory corrosion

tests for Simulating fireside wastage of

superheater materials in waste incinerators.

7.

8.

Materials and Corrosion, 2000. 51(4): p.

236-241.

Spiegel, M., Influence of gas phase

composition on the Hot Corrosion of steels and

nickel-based alloys beneath a

(Ca-Na-K)-sulfate mixture containing PbS04

and ZnS04. Materials and Corrosion, 2000.

51(5): p. 303-312.

Kawahara, Y., Evaluation of high-temperature

corrosion life using temperature gradient

corrosion test with thermal cycle component in

waste combustion environments. Materials and

Corrosion, 2006. 57(1): p. 60-72.

9. Kofstad, P., High Temperature Corrosion.

1988: ELSEVIRE APPLIED SCIENCE

PUBLISHERS LTD.

10. Albina, D.O., Theory and experience on

corrosion of waterwall and superheater tubes

of Waste-To-Energy facilities, in Department of

Earth and Environmental Engineering. 2005,

Columbia University.

Copyright © 2007 by ASME