Mitigation of High Temperature Corrosion in

Waste-to-Energy Power Plants

Timothy T. Sharobem

Submitted in partial fulfillment of the

requirements for the degree of
Doctor of Philosophy
in the Graduate School of Arts and Sciences

COLUMBIA UNIVERSITY

2017

i
 © 2016
Timothy T. Sharobem
All rights reserved

i
 ABSTRACT

Mitigation of High Temperature Corrosion in Waste-to-Energy Power Plants

Timothy T. Sharobem

Waste-to-energy (WTE) is the environmentally preferred method of managing post-

recycling wastes. In this process, municipal solid waste is combusted under controlled conditions

to generate steam and electricity. Waste is by nature heterogeneous and has a substantially high

composition of chlorine (0.47-0.72 wt%) as compared to other solid fuels used for power

production. During combustion, chlorine is converted to hydrogen chloride and metal chlorides,

which can accelerate the high temperature corrosion of boiler surfaces, especially superheater

tubes. This corrosion can significantly affect plant efficiency and profitability by causing

unplanned shutdowns or preemptively forcing operators to limit steam temperatures.

The following work focuses on the role of chlorine compounds on boiler tube corrosion

and investigates approaches for minimizing its effects. The corrosion behavior was studied by

conducting laboratory furnace tests on alloys of current and future interest to the WTE industry.

Test specimens were coupons machined from boiler tubes to a nominal area of 3.2 cm2 (0.5 in2).

An chemical environment was introduced in an electrical furnace that replicates the fireside of

superheater tube. This included a mixed gas stream with O2, CO2, H2O, HCl, SO2, and N2, and

temperatures ranging between 400-550°C (752-1022°F). For some experiments, a salt layer was

applied to the coupons with a loading of 4.0 ±10% mg/ cm2 to understand the behavior of the

effects of metal chlorides. Following each experiment, the corrosion rate was determined by

taking the mass loss as specified in an American Standard Testing Method (ASTM) protocol,

G1-09. Additional insights were obtained by characterizing the coupons via scanning electron

ii
microscopy (SEM) and elemental dispersive spectroscopy (EDS). Additionally, the corrosion

scale and salt layer were characterized via powder X-ray diffraction (XRD).

The addition of 800 ppm of hydrogen chloride (HCl) gas to a mixed gas oxidizing

environment accelerated the corrosion rate of SA178A (Fe-0.1C) at 500°C (932°F) as

determined by the change in the parabolic rate constant over a period of 72 hours, from 0.18 to

1.7 μm2/h (3.0 E-03 to 2.5 E-02 mil2/h). The findings from the EDS and XRD scale analyses

were compared to other literature and thermodynamic calculations that showed that effect that

HCl accelerates corrosion via an active oxidation mechanism.

A parametric study was performed on the effect of hydrogen chloride on three alloys,

SA178A, SA 213-T22 (2.5 Cr-1 Mo-Fe) and NSSER-4 (Fe-17Cr-13Ni). Varying the

concentration from 400 ppm to 800 ppm at 500°C increased the mean mass loss by 17.5%, as

compared to the 60% increase from 0 to 400 ppm. For each alloy, the mass loss increased

sharply with temperature between 450, 500, and 500°C, with corresponding apparent activation

energies of Ea NSSER- 4 53 kJ/mol, Ea SA213 T22 110 kJ/mol, and Ea SA178A 111 kJ/mol. The lower

apparent activation energy for NSSER-4 demonstrates that effect of hydrogen chloride is

mitigated with austenitic alloys versus carbon steel or low alloyed steel. In a comparative study

between isothermal and temperature gradient tests, it was also shown that the corrosion of

SA178A was not impacted by a temperature gradient up to 250 °C.

Another important chlorine compound in WTE boilers are metal chlorides, which are

readily contained in fly ash and boiler deposits. Using sodium chloride as a surrogate

compound, the corrosion behavior under chloride salts was investigated by applying a salt layer

(4.0 mg/cm2) on coupon surfaces. Corrosion under the chloride layer was much more severe

than below the HCl-containing atmospheres alone. The mass loss for the commercial steels was

iii
increased by more than an order of magnitude. Based on SEM and XRD coupon and corrosion

product characterization, this behavior was the result of a second active oxidation mechanism in

which sodium chloride reacts with and depletes protective oxides such as chromium (II) oxide.

The WTE furnace tests with the sodium chloride layer were executed for six different

Ni-Cr coatings, including Inconel 625 (Ni-Cr-Mo), SW1600, SW1641 (Ni-Cr-Mo-B-Si) and

Colmonoy 88 and SP 99 (Ni-Cr-B-W). The primary corrosion attack observed was pitting

located under the original salt layer. Colmonoy 88, showed superior corrosion resistance with

mass losses between 0.3-3.1 mg/cm2 between 450-550°C as compared to the Ni-Cr-Mo, and Ni-

Cr-B-So coatings which has mass losses between 10-30 mg/cm2. The enhanced corrosion

performance of Colmonoy 88 and SP 99 was attributed to the alloying addition of tungsten,

which had been previously shown in literature to also improve the pitting resistance for Ni-Cr in

aqueous environments.

The corrosion behavior under metal chlorides was compared with metal sulfates, which

are also prominent in WTE fly ash and boiler deposits. The application of sulfate salts on

coupon surfaces was shown to semi-protective on WTE boiler tube surfaces up to temperatures

of 550°C. The mass loss for carbon steel and Fe-17Cr-13Ni (NSSER-4) below sodium sulfate

was an order of magnitude lower than under sodium chloride. These results motivated

experiments aimed at sulfating chloride boiler deposits by increases the sulfur/chlorine gas ratio

(SO2/HCl) in WTE fuel gas. The SO2/HCl ratio was modified between 0.3 to 0.6 and 1.0

respectively. By increasing the SO2/HCl ratio, the sodium chloride layer applied on the coupon

surface was converted from a chloride rich salt to a sulfate rich and was shown to dramatically

reduce the corrosion of tube alloys up to 500°C. The impact of sulfating the alloy was most

prominent with alloys with high mass loss under the sodium chloride layer. Tests showed a

iv
reduction in the corrosion rates of SA213 T22 (37%), Inconel 625 (23%), and NSSER-4 (27%).

At 550 °C, there was no trend with respect to increases of the ratio, which suggests that other

corrosion reactions are faster than the rate of sulfation.

Finally, the annualized cost factor was defined and proposed as a method for replacing

current superheater alloys with alternative materials, such as those tested in this thesis. From this

discussion it was calculated that the installation of a Colmonoy 88 protected superheater can cost

approximately 1.4 times the cost of an Inconel 625 cladded replacement, or as much as 4.3 times

the cost of a T22 superheater tube and remain a cost effective option.

v
Table of Contents

List of Figures………………………………………….…………………………………….……v

List of Tables……………………………………………………………………………………...x

Acknowledgment……...…………………………………………………………………………xii

1. Introduction…………………………………………………………………………………..1

1.1 Background………………………………………………………………………………..1

1.2 Waste-to-energy technology………………………………………………………………4

1.3 Corrosion problem in WTE boilers……………………………..…………………………5

1.4 Thesis Structure……………………………………………………………...……………8

2. High Temperature Corrosion Concepts…………………………………………………..11

2.1 Oxidation of metals………………………………………………………………………11

2.2 Oxide formation and growth……………………………………………..........................12

2.3 High temperature corrosion kinetics …..………………………………………………...14

2.4 Superheater environment…………………………………………………...……………17

2.5 Corrosion affected by hydrogen chloride………………………………………………..19

2.6 Corrosion affected by metal chlorides ……………………………………………..……22

2.7 Corrosion affected by metal sulfates………………………………………………….….24

2.8 Conclusions…………………………...………………………………………………….25

3. Solutions for Mitigating High Temperature Corrosion………………………………….27

i
 3.1 Introduction…………………………………………………………….…………….…..27

3.2 Material selection………………………………………………………………...………28

3.3 Boiler modifications………………………………………………….………………......30

3.4 Chemical modification of fly ash……………………………………………….……..…32

3.5 Calcium hydroxide injection……………………………………………………………..36

3.6 Cleaning technologies……………………………………………………………………38

4. Experimental methods……………………………………………………………………...39

4.1 Summary……...………………………………………………………………………….39

4.2 Alloys……….……………………………………………………………………………39

4.3 Coupon preparation………………………………………………………………………42

4.4 Corrosion test rig………………………………………………………………………....42

4.5 Salt layer application……………………………………………………………………..45

4.6 Quantifying extent of corrosion………………………………………………………….46

4.7 Characterization techniques……………………………………………………………...49

4.8 Research Objectives…………………………………………………………………….53

5. Parametric Studies on Hydrogen Chloride Corrosion……………………………...……54

5.1 Summary…………………………………………………………………………………54

5.2 Introduction…………………………………………………………………..…..………55

5.3 Effect of hydrogen chloride……………………………………………………………...57

5.4 Studies with SA178A, SA213 T22, and NSSER-4………………………………….......66

5.5 Comparison of isothermal versus temperature gradient………………………………....70

ii
 5.6 Conclusions………………………………………………………………………………72

6. High Temperature Corrosion of Alloys below Chloride Salt Deposits…………….……73

6.1 Summary………………………………………………………………………………....73

6.2 Introduction……………………………………………………………………………...74

6.3 Studies with SA178A, SA213 T22, and NSSER-4…………………………...…………75

6.4 Studies with nickel based coatings…………………………………………………..…..85

6.5 Conclusions…………………………………………………………………………...….95

7. The Effect of Sulfur on High Temperature Corrosion…………………………………...96

7.1 Summary…………………………………………………………………………...…….96

7.2 Introduction…………………………………………………………………………..…..97

7.3 Industrial experience……………………………………………………………………..99

7.4 Effect of sulfate layer…………………………………………………………………...100

7.5 Tests varying the SO2/HCl ratio………………………………………………………..103

7.6 Effect of SO2/HCl ratio on various alloys below chloride salt layer………………...…110

7.7 Conclusions……………………………………………………………………………..117

8. Economics Impacts of Corrosion mitigation……..…………………………………..….121

8.1 Introduction…………………………………………………………………………….121

8.2 Annualized cost factor………………………………………………………………….121

8.3 Saving from extended tube lifetimes…………………………………...………………124

8.4 Cost of corrosion additives………………………………………..……………………125

iii
 8.5 Conclusions…………………………………………………………………………..…127

9. Conclusions and Suggestions for Future Work…………….………...…………………128

9.1 Conclusions……………………………………………………………………………..128

9.2 Suggestions for the WTE Industry……………………………...………………………130

9.3 Future work………………………………………..……………………………………131

References………………………………………………...……………………………………133

Appendix: Images of Test Coupons………………………………………………….……...…143

iv
List of Figures

Figure 1-1: Rates of MSW generated, recycling and landfilling based on four waste-to-energy

growth scenarios………………………………………………………………..…………2

Figure 1-2: Avoided methane emission under four WTE growth scenarios…………………..….3

Figure 1-3: MSW electricity generating capacity in the United States…………………………...4

Figure 1-4: Typical Waste-to-Energy boiler configuration……………………………………….6

Figure 2-1: Ellingham-Richardson diagram………………………………………………..……13

Figure 2-2: Schematic of oxide growth in a pure oxygen environment showing ionic

diffusion across the scale and defects in the scale that accelerate the diffusion rate

or allow for molecular transport ………………………………………………...………14

Figure 2-3: Oxidation kinetics for linear and parabolic oxidation……………………………….17

Figure 2-4: Schematic of active oxidation mechanism caused by hydrogen chloride gas……....21

Figure 2-5: Predominance diagram of Fe-O-Cl system at 500°C ……………………………….22

Figure 2-6: Schematic overview of various corrosion attack in WTE boiler tubes……………..27

Figure 3-1: Schematic of GKS boiler in Schweinfurt, Germany…………………………...…..33

Figure 3-2: Accumulation of deposits on chloride trap……………………………….…...…….33

Figure 3-3: Vattenfall Chlorout concept……………………………….………………………..35

Figure 3-4: Boiler tubes without and with Chlorout injection……………………..…………….35

Figure 3-5: The Gotaverken Miljo sulfur recirculation concept…………………………………36

Figure 3-6: Schematic of corrosion test system for calcium hydroxide spray…………….…….37

Figure 4-1: Example of WTE superheater boiler tubes …………………………………………40

v
Figure 4-2: Schematic of corrosion test rig………………………………………………........…43

Figure 4-3: Photograph of the corrosion test rig…………………………………………………43

Figure 4-4: Temperature profile of corrosion test rig at setpoint of 500°C with

500 mL/min flow of nitrogen……………………………………………………….……46

Figure 4-5: Hitachi TM 3000 scanning electron microscope……………………………………51

Figure 4-6: X’Pert powder X-ray diffractometer…………………………………………...……52

Figure 5-1: Mean mass loss of SA178A versus time (0-72 h) at 500°C under two synthetic flue

gas environments (0 ppmv HCl and 800 ppmv HCl)……………………….………...…58

Figure 5-2: Square of mass loss of SA178A versus time under two synthetic flue gas

environments (0 ppmv HCl and 800 ppmv HCl…………………………………………59

Figure 5-3: Images of SA178A coupons following (a) 50 hours under 800 ppmv HCl and

(b) 50 hours under 0 ppmv HCl gas mix…………………………………………....……60

Figure 5-4: XRD pattern of scale recovered from SA178A coupon at 500°C under flue gas

containing 0 ppm HCl ………………………………………………………………...…63

Figure 5-5: XRD pattern of scale recovered from SA178A coupon at 500°C under flue gas with

800 ppmv HCl …………………………………………………………………...………64

Figure 5-6: Mean mass loss of SA178A after 24 hours at 500°C under flue gas mixtures with

0,400, and 800 ppmv HCl…………………………………………………………..……65

Figure 5-7: Corrosion rates of NSSER-4, SA178A, and SA213 T22 at 450 -550°C, under WTE

flue gas containing 800 ppm HCl………………………………..………………………68

Figure 5-8: Equilibrium vapor pressures of solid metal chlorides as a function

of temperature....................................................................................................................69

Figure 6-1: Average mass loss of SA178A, SA213 T22, and NSSER-4 below NaCl layer and

vi
 WTE flue gas at 450°, 500°, and 550 °C…………………………………………..…….76

Figure 6-2: Coupons of SA213 T22 and NSSER-4 below NaCl layer following 24 hours

at 450, 500, and 550°C…………………………………………………………..……….76

Figure 6-3(a-b): Surface SEM-EDS of SA213 T22 coupon below sodium chloride layer at

450°C and 550°C ………...……………………………………………………….……..78

Figure 6-4 (a-b): Surface SEM-EDS of NSSER-4 coupon below sodium chloride layer at

450°C and 550°C…………………………………………………………………..…….79

Figure 6-5: XRD pattern of SA 213 T22 scale following at 500°C. ……………...……………80

Figure 6-6: XRD pattern of SA 213 T22 scale following at 550°C. ………………...…………81

Figure 6-7: XRD pattern of powder recovered from NSSER-4 coupon (500°)…………..……81

Figure 6-7: XRD pattern of powder recovered from NSSER-4 coupon (550°)…………..……82

Figure 6-9: Illustration of an active oxidation mechanism induced by alkali chlorides………....85

Figure 6-10: Average mass loss of Ni-coatings, Inconel 625, Colmonoy 88, SP 99, SW1600,

and SW1641 applied to SA213 T22 following 24 hour exposure below NaCl layer and

WTE flue gas at 450°, 500°, and 550°C…………………………………………………88

Figure 6-11. XRD pattern of scale recovered from Inconel 625 coating at 500°C....…………...90

Figure 6-12. Surface SEM-EDS of Inconel 625 scale at 500°C…………………………………90

Figure 6-13. SEM-EDS of surface of SW1600 and SW1641 coupons at 550°C…………….…94

Figure 7-1: Mass Loss of SA178A and NSSER-4 in the presence of WTE flue gas and below

NaCl and Na2SO4 at 500°C after 24 hours……………………………………….……102

Figure 7-2: NSSER-4 coupons after 24 hours at 500°C below sodium chloride

and sodium sulfate deposits ………………………………………………………...….102

Figure 7-3: Mass loss of NSSER-4 versus temperature (450-700°C) below

WTE flue gas and Na2SO4 deposits……………………….............................................103

vii
Figure 7-4: Corrosion of SA178A versus SO2/HCl ratios 0.1-2.0; 500°C and 50 hours…….…104

Figure 7-5: Fe-O-Cl predominance diagram, 500°C……………………………………….…..105

Figure 7-6: Fe-O-S predominance diagram, 500°C…………………………………….………106

Figure 7-7: XRD pattern of scale of T22 recovered after 50 hours………………………….…107

Figure 7-8: Comparison of corrosion rate of NSSER-4 vs SO2/HCl at 500 °C below chloride

and sulfate layers……………………………………………………………………….108

Figure 7-9: The corrosion rate of NSSER-4 below NaCl layer vs. the SO2/HCl ratio at 500°C;

the effect of water vapor….………………………………….…………………………109

Figure 7-10: Corrosion rate of SA178A and NSSER-4 vs. SO2/HCl ratio at 500 °C…………110

Figure 7-11: The effect of SO2/HCl on the mass loss of SA 213 T22 at 450°, 500°,

and 550°C………………………………………………………………………………111

Figure 7-12: The effect of SO2/HCl on the mass loss of Inconel 625 at 450°, 500°,

and 550°C……………………………………………………...…………………….…112

Figure 7-13: The effect of SO2/HCl on the mass loss of Colmonoy 88 HVOF at 450°, 500°,

and 550°C………………………………………………………………………………112

Figure 7-14: Salt coverage and EDS values for Inconel 625 after 24 hour tests

at SO2/HCl = 0.3………………………………..………………………………………114

Figure 7-15: Salt coverage and EDS values for NSSER-4 after 24 hour tests

at SO2/HCl = 0.3…………………………………………………………….…………114

Figure 7-16 (a-c): SEM- EDS values for SW1600 at SO2/HCl = 0.3, 450°C, 500°C,

and 550° C…………………………………………………………………...…………115

Figure 7-17(a-c). SEM and EDS cross section mapping for SW1600 at 450°C , 500°C,

and 550°, at SO2/HCl=0.3…………………………..……………………..……………118

viii
Figure 8-1: Estimated avoided costs by replacing SA 213 T22 with nickel coatings

(three year calculation)…………………………………..…………………………..…126

ix
List of Tables

Table 1-1: Chloride concentrations in MSW………………………………...……………………8

Table 2-1: Ultimate Analysis of MSW Fuel Based on US MSW Composition……......………..18

Table 2-2: Typical range for flue gas composition in WTE boilers.……………………..……...19

Table 2-3: Crystalline compounds in WTE fly ash identified by XRD……………….…....……19

Table 2-4: Thermodynamic data of reverse Deacon reaction (T = 200-1000°C)…………….….20

Table 2-5: Melting temperature of pure salts and common WTE eutectics …………….………26

Table 3-1: Common superheater tube alloys in WTE plants……………………;………………29

Table 4-1: Elemental composition of investigated alloys…………………………..……………31

Table 4-2: Elemental composition of investigated protective coatings.........................................41

Table 4-3: Densities of alloys……………………………………………………..…………….44

Table 5-1: Chlorine content of various solid fuels……………………………………………….55

Table 5-2: Rate constants for SA178A, 500°C (932°F) under corrosive gas mixtures……….…59

Table 5-3: Mass change comparison for three test conditions for SA178A at 500°C………..….61

Table 5-4: EDS Analysis SA178A corrosion layer and below layer……...…………….……….64

Table 5-5: Comparison of the types of mass losses for alloys, following 24 hour exposure…...70

Table 5-6: Apparent activation energy for corrosion of alloy under WTE flue gas……….……70

Table 5-7: Comparison of corrosion rate in isothermal and temperature gradient

environment (ΔT = 240°C)……………………………………………………………...72

Table 6-1: Cross section image of coatings below NaCl layer at 500°C and compositions of

products adjacent to metal surface…………………………………………….......……..92

x
Table 6-2: Comparison of average corrosion depth measured by coupon cross section

and mass loss obtained from ASTM method…………………………………….………93

Table 7-1: Sulfur content of various solid fuels ………………………………………………....97

Table 7-2: Thermodynamic data of sulfur dioxide oxidation………………………………..….98

Table 7-3: Corrosion rate of alloys measured in Battelle plant test…………………………….100

Table 8-1: Examples of cost factor for SH tube alloys…………………………………………122

Table 8-2: Life factor calculated for various Ni coatings based on mass loss data……………123

Table 8-3: Cost of maintenance reported by 7 WTE plants based on 2004 WTERT Survey on

Boiler Corrosion…………………………………….………………………………..…127

xi
Acknowledgements

First, I would like to thank my thesis advisor, Professor Marco J. Castaldi for his

guidance, patience, and support during the pursuit of my doctoral degree. I have been inspired by

his diligence and unwavering passion toward research and chemical engineering. I am forever

grateful for his confidence in me. I would also like to especially thank Professor Nickolas J.

Themelis for his invaluable and generous support on this research. His expertise and knowledge

in the field of waste management was an invaluable asset to my studies.

Many thanks also are due to my esteemed committee members: Professor Paul F. Duby,

Professor Robert J. Farrauto, and Professor Siu-Wai Chan. Especially, Professor Duby has been

a longtime mentor during my tenure at Columbia University, which began as an undergraduate

transfer student in chemical engineering. He taught me innumerable, practical lessons on how to

approach experimental research. I have also been fortunate to have been lectured by Professors

Farrauto and Chan.

I gratefully acknowledge the financial support of the Waste‐to‐Energy Research and

Technology Council (WTERT) of the Earth Engineering Center at Columbia University and its

sponsors. Without them, this academic pursuit would not have been possible. I would like to

express gratitude especially to Covanta Energy and the leadership of Steve Goff. I would also

like to acknowledge Shang-Hsiu Lee, Eric Hanson, and Andrey Tovstukha for their valuable

technical input and their generosity in providing materials for my experimental work.

I would also like to thank both former and current members of the Combustion and

Catalysis Laboratory at the City College of New York and at Columbia University. I would like

xii
to thank Anuta Belova; Federico Barrai; McKenzie Kohn; Amanda Simson; Naomi Klinghoffer;

Garrett Fitzgerald; Melis Duyar; Alex Frank; Christiane Janke; Simona Ciuta; Marcella Lusardi;

Jeffrey Leblanc; Stephen Crowley; Swanand Tupsakhare; Robyn Smith; Michaela Wagar;

Demetra Tsiamsis; and Deepak Kumar Sharma. Every member in this group has served a unique

role throughout my doctoral research. In particular, Deepak Sharma spent many hours at the end

of my doctoral work running furnace experiments and operating analytical equipment.

Other colleagues from Columbia University who provided a great deal of friendship and

moral support include Thomas Ferguson; Kyle Fricker; Xiaozhou Zhou; Rob Van Haaren;

Ranjith Kharvel; Jennifer McAdoo; Constantine Spanos; and Angela Zheng. Thank you for your

comradery and friendship.

Finally, I cannot thank my family and friends enough for the years of love and

encouragement. I am grateful for my beautiful mother, Mary, for her unconditional love and

unceasing prayers. Secondly, I am fortunate to have a brother, Steven, whose friendship and

example has been a positive influence in my life. Lastly, I would like to thank my “adopted

brothers” Bill Slattery, Brendan O’Dea, and Christian Chiavetta for their friendship.

Thank God for all of you.

xiii
Chapter 1: Introduction

1.1 Background

For millennia humans have had to manage waste. Initially, this concerned human

excrement and sewage but in time also included municipal solid wastes (MSW), which are

streams of matter generated from residential activities [1]. In the United States an average

person generates 3.3 kg (7.2 lbs) of MSW per day which can include a variety of items such as

paper, glass, metals, plastics, food wastes, among others. In total, the US manages 389 million

tons annually [2]. Based on current domestic habits, these values will only increase over time.

After reuse and recycling there are only two methods that can manage waste at the rates

at which it is generated, either disposal in a landfill or thermal treatment for material and energy

recovery [3]. The latter is more commonly known as to waste-to-energy (WTE), in which MSW

is combusted under controlled conditions, reducing its volume by 90%. The heat from the

combustion gases is transferred through metal surfaces to steam which powers turbines for

production of electricity [4].

Under the current state of waste management, landfilling and waste-to-energy will co-

exist for the foreseeable future. However, in the face of global climate change, thermal treatment

will become a more important approach for handling waste. Many studies have shown the

environmental benefits of waste-to-energy [5-7]. One benefit, in particular, is the avoidance of

methane which is released from landfills due to biological activity with MSW [8]. In the United

States alone, landfill gas is the second largest source of anthropogenic emissions of methane and

only 18% of it is captured and utilized as an energy source [9].

1
 In 2007 a study between the Earth Engineering Center at Columbia University and

NASA’s Goddard Institute for Space Studies demonstrated how increasing the capacity of waste-

to-energy can reduce methane emissions [10]. The major results are shown in figure 1-1 and

figure 1-2. The study developed four scenarios of WTE growth, ranging from very conservative,

in which the year 2000-2007 growth in capacity was assumed to remain at a constant rate of

2.5% through to the year 2030, to assumed increases in the rate of growth of WTE capacity of

5%, 7.5% and 10% per year, respectively, between 2010-2030. The only way to reduce landfill

greenhouse gas emissions to the year 2030 is by achieving a 7.5% annual growth in WTE

capacity globally.

Figure 1-1. Rates of MSW generated, recycling and landfilling based on four WTE growth
scenarios. The reference case is an annual capacity growth rate of 2.5% [10].

2
Figure 1-2. Impact of WTE growth on net methane missions from MSW. Impact of WTE
growth scenarios on MSW landfilled is shown in figure 1-1 [10].

Another cited benefit of WTE is its classification as a renewable source of electricity by the

Energy Policy Act of 2005, the US Department of Energy, and 24 state governments [11]. It was

proven through carbon-14 measurements of (ASTM D6866 protocol) carbon dioxide in WTE

stack gas that 64-66% of the total carbon in MSW is biogenic [12].

WTE has a significant advantage over other renewable electricity sources in that it is a

mature technology which already provides baseload electricity and is located in populated areas

where the power is most needed. Figure 1-3 is from a 2011 US Energy Information

Administration (EIA) report on WTE plants in the United States. This shows the current

electricity generating capacity by state. Most of the facilities operate in the northeastern US and

in Florida. In total, 87 facilities convert approximately 30 million tons of MSW into 15 billion

kWh [13].

3
 Figure 1-3. MSW electricity generating capacity in the United States [13]

1.2 Waste-to-energy technology

Analyses on the combustible portion of MSW have shown that an elemental composition

can be expressed by the formula C6H10O4, which has a calculated heating value of 23.5 MJ/kg

[14]. In reality, MSW does not consist solely of combustible materials. It has significant

moisture content (about ~20% in the U.S.) and a significant inorganic fraction comprised of

metals, and glass, also about 20% by mass. In practice, the actual heating value of MSW is on

the order of about 11.6 MJ/kg (~5000 Btu/lb) [15] or half the heating value of the combustible

materials. Given the heterogeneous nature of waste, and variability in precipitation, the value of

5000 Btu/lb, is highly variable.

WTE plants are classified in two broad categories based on the characteristics of the fuel,

mass burn and refuse-derived fuel (RDF) combustion systems. Mass burn systems are more

common, 80% of the WTE plants worldwide, and are simpler in design, since they process waste

“as received”, directly without presorting. RDF systems aim to produce a more uniform material

(and a higher heating value) by shredding, removing ferrous metal, and some additional

4
screening to remove grit or non-ferrous metal recovery [16]. The RDF technology was

surpassed by mass burn facilities in the 1980s, as they were able to process larger tonnage and

had the advantages of lower capital costs. Since the fuel is more heterogeneous, one tradeoff for

mass burn plants is that it has a lower thermal efficiency [17].

A typical mass burn system is shown in figure 1-4. Waste is introduced via a moving

grate which slowly moves the waste through the furnace and tumbles, exposing as many particles

to the combustion air as possible, as they undergo drying, combustion, and burnout stages before

the residue is discharged. The most common grate is the reciprocating grate and combustion air

is injected into a furnace as underfire air, below the grate, and as overfire air, into the flame

above the grate. Underfire air is injected into the hoppers below the grate to keep the grate cool

and promote combustion. Generally, most mass burn furnaces operate at 50-100% excess air,

with 50-70% of the total air fed as underfire air and the balance as overfire air. [18]. Afterward

the flue gas passes through an integral waterwall boiler, superheater, and economizer, producing

steam. The cooler flue gas is then cleaned by air pollution control (APC) equipment to remove

fine particulates (fly ash), acid gases, and any remaining traces of organic compounds [19]. The

burned out residue is discharged from the grate and is known more commonly as bottom ash.

1.3 Corrosion problem in WTE boilers

The combustion of MSW is different than that of any other fuels because of its

heterogeneous nature. Besides its high moisture content, MSW contains elements such as

chlorine, sulfur, alkali metals like sodium and potassium, and heavy metals like zinc and lead.

During combustion these species can be liberated in the flue gas and produce an environment

5
which can cause fouling of heat transfer surfaces and even worse depletion of the heat transfer

surfaces and tube failure.

Figure 1-4. Typical Waste-to-Energy boiler configuration [20]

High temperature corrosion in WTE boilers typically occurs in three sensitive areas: the

top of the refractory lining in the first pass, the transition of the first to the second pass, and tube

bundles of the screen, evaporator and superheater [21]. The most aggressive corrosion, i.e. the

fastest corrosion rate, occurs in the superheater as this is the metal surface which has the highest

temperature. The corrosion behavior is generally referred to as high temperature corrosion- as it

involves the transformation of the metal in an environment absent of a liquid electrolyte (which

would be considered aqueous corrosion). The corrosion behavior is also referred to as fireside

6
corrosion, which indicates that the important gas-liquid-solid interaction occur between the outer

surface of the boiler tubes with the flue gas environment [22]

The corrosion problems in WTE plants are realized in three ways. First, it is one of the

common causes of shutdowns. It has been estimated that corrosion accounts for 70% of plant

shutdowns [23]. Secondly, corrosion impacts plant profitability and the health of the waste-to-

energy industry. It has been estimated that corrosion-related maintenance costs account for a

third of the annual maintenance budget and can be as high as the 10% of the annual turnover

[23]. Finally, the threat of unmanageable high temperature corrosion limits the thermal

efficiency by forcing plants to operate at lower steam, boiler tube temperatures. It is not

uncommon for example to have a WTE operate at 18-25% efficiency versus a coal fired plants

which are on the order of 30% [24]. This is largely because the aggressive environment

produced by the combustion of waste, most notably the presence of chlorine compounds

Previous literature and plant experience have identified chlorine, as the most problematic

component of MSW, given its relatively high concentration in MSW and the various forms in

which it is liberated from the furnace which can interact boiler alloys. MSW on average contains

between 0.47-0.72 wt% chlorine [25]. Table 1-1 below is data adapted from a study estimating

the chlorine content in MSW and shows a chlorine concentration within the specified range at

0.58 wt%. Plastics, largely due to polyvinyl chloride, PVC, are a significant source of chlorine,

since they contain the highest chlorine per mass at 2.5 wt%. Paper, organics, and biomass have

considerable chlorine content which can be either a part of organic (e.g. polyvinyl chloride) or

inorganic compounds (e.g. sodium chloride) [25].

7
 Table 1-1. Chlorine concentrations in MSW [25]

Waste % of component Cl concentration Cl in MSW

component in MSW (average) (g/kg) (g/kg)
Paper 29.4 2 0.59
Plastics 10.4 25 2.6
Organics 34.7 3 1.04
Textiles 4.4 12.5 0.55
Wood 2.7 12.5 0.34
Misc. combustible 5 12.5 0.63
Glass and metals 4 0.6 0.02
Metals 4.7 0.6 0.03
Misc. non-
combustible 4.7 0.6 0.03
Total 100 5.82 (0.58%)

Recognition that chlorine in the refuse is the primary source of corrosion problems came

slowly to the WTE community. Unlike other combustion based power plants, WTE power plants

were not commonplace until the 1960s. As the pioneer WTE plants ran into severe corrosion

problems, these were initially attributed to similar conditions occurring in coal plants (which are

more related to various sulfur compounds). The largest body of work on WTE corrosion was

initiated by engineers at Battelle Memorial Institute, the Electric Power Research Institute

(EPRI) and jointly by the American Society of Mechanical Engineers (ASME) and the US

Department of Energy (DOE) [26]. The largest collection of the development of that work can be

found by ASME’s Center for Research and Technology (CRTD) Report Volume 38 that was co-

published with the U.S. Department of Energy.

1.4 Thesis Structure

The following thesis is aimed to identify the major causes of superheater tube corrosion

in waste fired plants; understand the mechanisms of the underlying chemical reactions; and

8
recommend solutions that will allow superheated steam temperatures in WTE plants of 500°C

and higher. The organizational structure of this work is as follows:

Chapter 2 provides background information elaborating on the chlorine corrosion

problem in WTE plants, with particular interest to the reactions which occur with the superheater

boiler tubes. Chapter 3 provides background on what has been developed in terms of preventive

measures, including material selection, high temperature coatings, sacrificial surfaces, and

additive injections. This literature review helps to provide a comprehensive understanding of the

current areas of corrosion research and helps promote the experimental work discussed in later

chapters. In particular, the role of alternative steels, high temperature coatings, and corrosion

additives are explored experimentally and discussed in chapter 5-7.

Chapter 4 provides details on the execution of the laboratory corrosion testing, with the

intent that the reader can reproduce the same experiments and analysis. Details are provided on

the alloys tested, the synthesized WTE superheater fireside corrosion environment, the

determination of the extent of corrosion and the characterization techniques.

Chapter 5 discusses the corrosion behavior of hydrogen chloride for three commercials

steels: SA 178A, SA 213 T22, and NSSER-4. The effects of temperature, temperature gradient,

and the concentration of hydrogen chloride are provided. This is followed by chapter 6 which

focuses uniquely on the behavior below chloride salts, with sodium chloride used as a surrogate

compound. As will be discussed, these salts are the some of the major components of WTE fly

ash which have been shown to accelerate metal corrosion. In particular, this chapter will

investigate conditions at which the chloride salt is at temperatures below the first melting

9
temperature (FMT). The commercials steels from chapter 5 are utilized for this discussion as

well as 10 nickel based coatings- which may have promising

Chapter 7 investigates the impact of sulfur compounds- in particular sulfate salts and

SO2/SO3 on the chlorine-metal interactions discussed in chapters 5 and 6. These studies help

understand how variation in the waste feed (the sulfur, chlorine composition) can impact boiler

corrosion at temperatures between 450-550°C. Furthermore, the results discussed in this chapter

have relevance to the chemical additive technologies discussed in chapter 3, namely the sulfur

recirculation technology currently tested by Gotaverken Miljo and the Chlorout additive,

developed by Vatenfall.

Lastly, chapter 8 provides some additional context to some of the experimental results

acquired in this thesis by discussing the economic impact of utilizing different alloys in the

superheater and incorporating corrosion additives as part of boiler maintenance. Chapter 9

presents the conclusion of this study as well as recommendations for future work.

10
Chapter 2: High Temperature Corrosion Concepts

2.1 Oxidation of metals

With the exception of some areas in its first pass, WTE boilers contain an oxidizing gas

environment, since MSW is fired with excess air. In the case of a bare (virgin) metal surface in

contact with oxygen, metal oxide formation proceeds according to the following reaction:

M(s) + O2 (g)  MO2 (s) (2-1)

or more generally

xM +1/2 (yO2) → MxOy (2-2)

The spontaneous behavior is predicted by a negative change in the Gibbs free energy for

the reaction at a given temperature, which is expressed as:

ΔG0 = ΔH0 - TΔS0 (2-3)

where ΔH0 is the change in enthalpy, ΔS0 is the entropy change in the system, and T is the

absolute temperature [27]. The Gibbs free energy can also be expressed in terms of the

equilibrium product, which is expressed as the ratio of the activities of the products to the

activity of the reactants. Specifically for reaction 2-1, this is a ratio of the metal oxide activity

(the products) to the oxygen potential (equilibrium pressure), and the metal activity (the

reactants)

ΔG0 = -RT ln (aMO2 /aM Po2) (2-4)

11
In this expression, aMO and aM are the activities of the oxide and the metal respectively and PO2 is

the partial pressure of oxygen. This dictates that an oxide will form on the surface of a metal

when the oxygen potential in the environment is greater than the oxygen partial pressure in

equilibrium with the oxide. In general, activity is the product of the mole fraction and the activity

coefficient. Typically, ideal behavior is assumed for solids and activity of the metal and the

metal oxide is assigned the value of unit. Applying this assumption, the Gibbs free energy can be

expressed simply as a function of temperature and the partial pressure of oxygen.

ΔG0 = RT ln (Po2) (2-5)

The two expressions for the Gibbs free energy, such as 2-3 and 2-5, are plotted and can

be obtained from Ellingham diagrams (the inclusion of the equilibrium dissociation pressure

“Po2” is known as an Ellingham-Richardson diagram). An example of this diagram is shown in

figure 2-1 [28]. As discussed, the Gibbs free energy has linear relationship with temperature with

a positive slope, which implies that oxidation reactions take place with a decrease in entropy, i.e.,

the product has a higher order than the reactants. Changes in slope in the diagram, such as that

for MgO indicate a phase change, melting, boiling or a change in structure. On the right of the

plot additional nomographic scales are shown on the right of the plot to express the equilibrium

pressure of oxidants, such as O2, H2O, and CO2 [29].

2.2 Oxide formation and growth

is typically spontaneous at room temperature, the reaction kinetics are so slow that the

consequences often are negligible. However, if the temperature is raised the rate of reaction may

increase many orders of magnitude. Oxidation of a metal surface is considered to begin with the

12
Figure 2-1. Ellingham-Richardson diagram. Additional nomographic scales are shown on the
right to express the equilibrium pressure of oxidants, such as O2, H2O, and CO2 [28].

13
adsorption of oxygen molecules from the atmosphere, followed by nucleation and thereafter the

formation of a thin oxide layer as explained by Kofstad [30]. Once a thin film exists, oxidation

can proceed only by the diffusion of reactants through the film. There are many factors

influencing the oxide growth and a major one is the properties of the oxide layer. A defect-rich

or porous oxide scale will be a less efficient diffusion barrier compared to a lattice with low

defect concentrations [31]. The basic principles of oxide growth are illustrated in Figure 2-2.

Figure 2-2. Schematic of oxide growth in a pure oxygen environment showing ionic diffusion
across the scale (left) and defects in the scale that accelerate the diffusion rate or allow for
molecular transport (right) [31].

2.3 High temperature corrosion kinetics

One of the limitations of the Ellingham diagram (figure 2-1) is that it does not take into

account the kinetics of reaction. They can only be used to determine whether under certain

conditions oxide formation is possible. There are three types of kinetic rate laws associated with

the growth of an oxide layer. These kinetic expressions are also generally applied to all high

temperature corrosion behavior, even in mixed gas environments in which more than one

oxidizing agent is present.

14
 Logarithmic Law usually represents oxidation in a thin layer regime. In the case of most

metals heated at low temperatures, the kinetics follows logarithmic behavior. The rate of reaction

rises very fast in the beginning and then slows down, either following a direct or inverse

logarithmic law [32]. An example of the expression of the direct law is:

x = K* log (t) + A (2-6)

where x is the change in weight as a result of oxidation or thickness of the oxide formed or the

amount of oxygen consumed per unit area, and t is time.

Linear Equation is used to express a kinetic regime in which the rate of oxidation

remains constant with time. In that kinetic regime, the oxidation is independent of the amount of

the gas or metal previously consumed in the reaction. In this case, the rate of reaction is directly

proportional to time and is expressed as follows:

dx/dt = K*t (2-7)

x = K*t + D (2-8)

The linear kinetic regime can sometime occur when an oxide layer undergoes spallation, in

which the oxide layer growing on the surface is cracked, exposing unreacted metal underneath.

This is also referred to as breakaway [33].

Parabolic Oxidation is the most common form of oxidation kinetics. This expresses that

the oxide growth rate is decreasing with time. The rate of reaction is therefore inversely

proportional to the scale thickness or the weight of the oxide formed. This is represented as:

dx/dt = Kp /x (2-6)

15
or after integration:

x2 = 2 Kp t + C (2-7)

where x is either: the change in weight as result of oxidation, thickness of the oxide formed, the

amount of oxygen consumed per unit surface area of the metal, or the amount of metal converted

to the oxide, t denotes the exposure time, Kp is the parabolic rate constant, and C is the

integration constant.

Most metal and alloys follow parabolic kinetics at elevated temperatures. The oxide

growth process is governed the rate limiting step, namely the diffusion of ions or electrons

through the initially formed scale [34], as discussed in section 2.2. The rate constant, Kp, follows

an Arrhenius relationship and can be therefore be expressed as a function of temperature [34].

Kp = A* exp (-EA / RT) (2-8)

where R is a constant 8.314 J/mol-K and T is absolute temperature (Kelvin) and EA is the

apparent activation energy. Figure 2-3 demonstrates what the oxide measurement would appear

as over time in the different kinetic regimes.

16
 Figure 2-3. Oxidation kinetics for linear and parabolic oxidation [34]

2.4 Superheater environment

In addition to the oxidizing environment, the combustion of MSW contains several

species which if liberated into the flue gas can participate in corrosion reactions- with the alloy

surface and with the oxide layers. Table 2-1 is a summary of the ultimate analysis performed on

various waste streams found in U.S. MSW, reported on a dry basis [35]. Other than the

significant concentration of chlorine, there is a large concentration of ash, which can contain

alkali metals such as sodium and potassium or heavy metals, such as zinc and lead. There is also

an appreciable amount of sulfur in waste streams like textiles, but they are not as common in

U.S. MSW. There is roughly 5-10 times more chlorine in MSW than sulfur [36].

The gas environment in a WTE boiler will depend on the components of MSW that were

fed into the furnace, the moisture content of waste, and the percentage of excess air fired. Table

2-2 provides approximate ranges for the major gas species in WTE flue gas, taken from studies

17
from various WTE plants [36]. In WTE superheaters, the metal temperatures typically range

400-450°C while the gas entering that pass is around 700-750°C. In addition to the species listed

in table 2-2, there are others compounds present in the flue gas which can deposit or condense

downstream onto boiler tubes. Some of these compounds have been identified by analyzing the

boiler ash during shutdowns using X-ray diffraction (XRD), which is capable of identifying

crystalline compounds. A list of the common species and the studies they were taken from is

provided in table 2-3.

Table 2-1. Ultimate Analysis of MSW Fuel Based on US MSW Composition [35]

Table 2-2. Typical range for flue gas composition in WTE boilers [36]

O2 (%) CO2 (%) H2O (%) SO2 (ppmv) HCl (ppmv)

5-11 8-14 10-20 0-150 250-1300

18
 Table 2-3. Crystalline compounds in WTE fly ash identified by XRD [37-39]

Ref Fly Ash Compounds

[37] NaCl, KCl, CaSO4
[38] NaCl, KCl, CaSO4, SiO2
[39] KCl, NaCl, Na2SO4, Na2Ca2(SO4)3

2.5 Corrosion affected by hydrogen chloride

Most oxidation follows parabolic kinetics because the oxide layer becomes protective,

that it mitigates additional oxidation underneath the formed scale. Chlorides can accelerate the

oxidation of alloys by participating in reactions which yield the oxide layer un-protective,

containing cracks or pores. The corrosion effect caused by HCl in an oxidizing environment is

shown in figure 2-4. It is sometimes referred to as the chlorine cycle, as chlorine gas is

regenerated after a series of reactions between oxygen, the oxide layer, and the metal [40].

HCl in the gas phase reacts with oxygen to form chlorine. In the reaction below

HCl (g) + 0.5 O2 (g) = H2O (g) + Cl2 (g) (2-9)

This reaction is sometimes referred as the reverse-Deacon reaction [40]. Chlorine (Cl2) is

thermodynamically favored over HCl at lower temperatures in an oxygen environment, as shown

by the thermodynamic data (provided by HSC 5.1) in table 2-4.

Chlorine gas penetrates through the scale to the scale/metal interface. This step will

occur if the oxide layer non-protective or semi non-protective. This would depend on the alloy or

other gases under which the alloy is exposed. At the scale-metal interface, chlorine reacts to form

19
metal chlorides. The presence of metal chlorides is generally not found in the scale, except at the

metal-scale interface. This is because the oxide species is thermodynamically preferred over the

chloride species in an environment in which sufficient oxygen is present. This is observed in a

metal/alloy-O-Cl predominance diagram. An example of a predominance diagram for iron, Fe-

O-Cl , at 500°C is shown in figure 2-5. The dotted lines connect the partial pressures of oxygen

and chlorine that are present in the bulk of WTE flue gas.

Table 2-4. Thermodynamic data of reverse Deacon reaction (T = 200-1000°C)

Temperature (°C) Gibbs Free Energy, ΔG (kJ/mol)

200 -26.5
300 -19.8
400 -13.1
500 -6.4
600 0.4
700 7.1
800 13.8
900 20.5
1000 27.2

20
 Figure 2-4. Schematic of active oxidation mechanism caused by hydrogen chloride gas [40]

If the temperature is high enough, the metal chloride formed at the metal-scale interface

has a high enough vapor pressure to diffuse outward through the scale. Examples of the

equilibrium vapor pressure of chlorides are shown in table 2-4. The diffusing metal chloride

vapor diffuses outward spontaneously reacts with the available oxygen to form metal oxide and

release chlorine gas. An example of this is shown in the below between Iron (II) chloride and

oxygen.

2FeCl2 (g) + 0.5O2 (g)  2Fe2O3 (s) + 2Cl2 (g) (2-10)

Chlorine is re-released which enables the gas either to participate in this network of reactions

again, or to be swept by the bulk gas away from the metal surface. The active oxidation is

21
generally accelerated in oxidizing-chlorinating environments in which the vapor pressure of

metal chlorides is significant, typically 10-4 atm or greater [40].

2.6 Corrosion affected by metal chlorides

Alkali chlorides, such as KCl and NaCl, are some of the most common superheater tube

deposits as shown in table 2-3. These compounds can participate in a few corrosion reactions.

The first is similar to the active oxidation mechanism discussed for HCl. A schematic for the

behavior is shown later in this thesis in figure 6-9. The key feature is the formation of mixed

oxides which provide an even, non-protective scale, and re-release Cl2 (g).

Alkali chlorides can also release Cl2 (g) when they are reacted converted to alkali sulfates in

environments containing SO2, O2,, and water via a sulfating reaction. Some authors have

discussed that this reaction can reduce corrosion in WTE boilers despite the release of chlorine

gas, speculating that the conversion of chlorides to sulfates reduces molten salt corrosion [42].

Chloride salts in contact with boiler tube alloy and scale can potentially form low melting

point eutectics (i.e., salt solutions characterized by the lowest possible melting point) which may

dissolve (“flux”) the oxide layer that is protecting the metal surface [43]. Corrosion by molten

salts is often termed ‘hot corrosion’ and is divided into type I and type II [44]. The type I hot

corrosion refers to temperatures above the melting point of the condensed salt, while in type II a

molten salt is formed below the melting point of the pure salt. In the latter case, the melting point

is lowered due to dissolution of formed corrosion products. Table 2-5 gives a list of several

chloride and sulfate species that may exist on the surface of boiler tubes during operation, the

melting point of the pure species, and the concentration and temperatures of some common

mixtures [45].

22
 Figure 2-5. Predominance diagram of Fe-O-Cl system at 500°C

Hot corrosion studies have suggested a fluxing mechanism where an otherwise protective

oxide dissolves in the fused salt and precipitates as non-protective particles within the salt or at

the salt/gas interface. Due to fast transport in the liquid phase, corrosion by molten salts can have

rapid kinetics [46].

Previous authors have shown that iron, nickel and chromium oxides are soluble in molten

alkali chlorides and also that the solubility is dependent on the specific gas composition [47].

WTE field exposures have identified fused zinc and lead chlorides as a major concern, since

these are known to have low melting points and form low melting eutectic mixtures with other

salts [48]. In WTE operation, it is expected that even more complex mixtures, not found in

binary or ternary phase diagrams, may be present.

23
2.7 Corrosion affected by sulfates

Sulfates can induce corrosion on WTE boiler tubes if a molten mixture is present.

Individually sulfates are not an active species on WTE boiler tubes, as the metal temperatures are

not hot enough. At elevated temperatures above 550° as has been experienced in coal-fired

boilers, sulfates can from alkali metal trisulfates in the presence of SO2 and O2 [49]. The

corrosion mechanism of those species is the sulfur analog of the chloride active oxidation that is

shown and discussed in further detail in figure 6-9. A schematic showing the different tube

temperature regimes and the corrosion phenomena that dominates those sections is represented in

figure 2-6.

2.8 Conclusion

There are three major chloride high temperature corrosion mechanisms: active oxidation

from HCl, active oxidation from alkali chlorides, and molten salt attack. Each mechanism

accelerates tube corrosion by removing or weakening the oxide layer which is formed on the

tube to naturally protect itself in the surrounding environment. In order to prevent this corrosion,

strategies need to be adopted to eliminate both HCl and metal chlorides, especially alkali

chlorides. Chapter 3 discusses lessons that can be transferred from biomass boilers to WTE

boilers on alkali chloride mitigation. In addition, other prevention methods and metallurgical

studies are discussed.

24
Figure 2-6. Schematic overview of possible corrosion attack in WTE boiler tubes

25
 Table 2-5. Melting temperature of pure salts and common WTE eutectics [50]

First Melting
Melting Salt Mixtures
Compounds Temperature
Temperature (°C) (wt% mix)
(FMT), °C

ZnCl2 283 KCl-ZnCl2 (48:52) 230

PbCl2 489 NaCl-ZnCl2 262

FeCl2 673 KCl-FeCl2 (45.8:54.2) 355

ZnSO4 730 NaCl-FeCl2 (50:50) 370

KCl 775 K2SO4-ZnSO4-Na2SO4 388

NaCl 801 KCl-PbCl2 (21:79) 412

CrCl2 821 NaCl-PbCl2 (17:83) 415
Na2SO4 884 NaCl-CrCl2 (53.7-46.3) 437
NiCl2 1001 KCl-CrCl2 (30-70) 462
K2SO4 1076 KCl-NiCl2 508

CaSO4 1400 NaCl-NiCl2 560

PbSO4 1170 KCl-NaCl 657

Note: for salt mixture, the wt% is 50:50 unless otherwise specified

26
Chapter 3: Solutions for mitigating chloride corrosion

3.1 Introduction

The steam temperature in a power plant is the maximum value for which the superheater

design can produce electricity economically. This involves the assessment of two interrelated

costs--the initial investment and the maintenance costs necessary to minimize operating

problems, outages, and replacements [51]. Determining the optimum steam temperature is

usually based on a number of factors including: (a) the alloy choice, (b) the surface area required

to provide the superheater steam temperature, (c) the gas temperature zone in which the

superheater tubes are located; (d) the permissible steam pressure drop in the tubes—which

controls the steam flow; and (g) their arrangement in the boiler [52].

In order to promote higher steam temperatures and, therefore, higher thermal efficiency

in WTE plants, modifications in the above factors are necessary in order to minimize the high

temperature corrosion phenomena. In general, these measures can be classified as primary or

secondary. Primary measures are modifications to the process conditions whereas secondary are

metallurgical solutions [53]. Some of these solutions are suited for new plant designs, whereas

others can be affected during the plant shutdown period. In this chapter, the modifications

suggested focus specifically on mitigating high temperature corrosion mechanisms of chlorides.

Some of the methods described in this chapter are based on the experience of biomass-fired

power plants, which like WTE plants have significant chloride corrosion problems.

27
3.2 Material Selection

Most of the materials used in boilers are steels, the majority of which are low alloy or

carbon steels. Carbon steels are alloys of iron with about 0.05 to 1 wt% carbon. Low alloys

steels include other elements at compositions of about 2 wt% or less for improvements of

mechanical properties. Two examples of these materials, SA 213 T22 and SA 213 T11, are listed

in Table 3-1, which lists common alloys found in WTE superheater tubes- based on plant survey

and correspondence with the WTE industry. Low alloy steels are attractive for their relative

expense, while possessing a wide range of strength and hardness [54]. Because of these benefits

these alloys serve as the base metal for superheater in highly aggressive environments with

another alloy—a protective layer applied to it.

While typically 2.5 times more expensive [55] than low alloy steels, stainless steels are

an attractive alternative because of their increased oxidation resistance. By definition stainless

steels contain at least 10.5% chromium. Oxidation resistance increases with increasing

chromium content. They are classified as ferritic, austenitic, duplex, martensitic, and

precipitation hardening. Each classification has its own special characteristics, which are well

known in the literature. Austenitic stainless steels are widely used in industry for its corrosion

resistant properties. The addition of nickel to iron-chromium alloys stabilize the face-centered

cubic austenite phase and works with chromium synergistically to improve corrosion resistance-

by forming a thin, chromium rich oxide layer on the gas-metal surface [56, 57]. In WTE power

plants, the most common austenitic steel for boiler applications is TP 310H, the composition of

which can be found in Table 3-1.

In general, a practical approach of extending the lifetime of boiler tubes is the application

of a highly alloyed material with a thickness less than or equal to 3 mm [58]. This is an enticing

28
option because of difficulties associated with mechanical properties, workability, and high

material price. The role of coatings is to provide a metal surface composition which will react

with the environment to produce the most protective scale possible. Nickel has intrinsically low

corrosion rates in oxidizing-chlorinating environments compared with the other major base

metals in high temperature alloys. Additions of chromium, molybdenum, and tungsten to nickel

based alloys have all shown to improve oxidation resistance [59]. Two of the more common

deployed nickel based alloys are provided in Table 3-1. To bring perspective on cost, the

purchase cost of an Inconel 625 cladded superheater is approximately 4.5 times for the same area

[60]. Based on the corrosion performance and lifetime, this cost can be considered prohibitive or

attractive.

Table 3-1. Common superheater tube alloys in WTE power plants [61]

Coatings can be produced on boiler tubes by several methods: co-extrusion, weld

overlaying, diffusion treatment, thermal spraying, and laser cladding [62]. The application

method can sometimes diminish the corrosion performance of a material. For example, weld

overlay Ni-Cr-Mo coatings (such as Inconel 625) have not performed well at high temperature

WTE superheaters because of iron migration to coating from the base metal due to the high heat

load during welding [63]. Such migration does not occur for high velocity oxy-fuel coatings

29
(HVOF). Laser cladding, similarly to weld overlay, is prone to iron migration, as a result of high

dilution with the base material [64,65].

Table 3-2 provides several examples of different types of coatings that have been

investigated at WTE superheater conditions along with the specific parameters.

3.3 Boiler design modifications

3.3.1 Extending the length of the passage between furnace and superheater

In Europe, several WTE power plants, have deployed methods for removing alkali

chlorides. Two Swedish plants in Södertälje and in Norrköpinga , extended the length of passage

between the furnace and the superheater via addition of a separate radiant pass. This allows for

the undesirable ash components to fall out of the flue gas before reaching the superheater.

Additionally, sulfur granules are added to the furnace to alter the S/Cl ratio (see section 3.4)

which increases the melting temperatures of the ash, because sulfates have higher melting

temperatures than chlorides. The additional pass allows for a longer residence time for the

completion of the sulfation reaction. Water cannons are installed on the sides of the empty pass

to clean the walls and spring hammers remove agglomerated deposits. This arrangement allows

for a tolerable superheated steam temperature to be as high as 540°C [66].

30
 Table 3-2. Studies of coatings investigated in WTE environments

Alloy(s) Temperatures (°C) Environment Ref

Molten deposits of tin, lead,
Inconel 625 650 zinc chloride 67
eutectic mixture of PbCl2-KCl
Fe-12Cr-Mo-V 500-600 in air 68
HVOF Coatings: 75Cr3C2- air oxidation with fly ash from
25NiCr; biomass-
Cr3C2; NiCrSi; Ni20Cr15Mo;
WC-1 450 fired boiler (high Ca, K, Cl) 69

HVOF Coating: 75% Cr3C2-25%

Ni20Cr 450 biomass fired boiler fly ash 70

HVOF wired sprayed coating:

Al/80% Ni-20%Cr 400 field test, WTE plant 71

HCl, SO2 containing

Laser cladding Alloy 625 300-500 synthesized flue gas 72

Sealed flame spray, HVOF, laser HCl, SO2 containing

cladding of Alloy 625 400-500 synthesized flue gas 73

HVOF sprayed NiCr, NiCrBSi, HCl, SO2 containing

Cr3C2-NiCr 500 synthesized flue gas 74

31
 3.3.2 Chloride Trap- collecting fly ash on screen tubes prior to the superheater

In Schweinfurt, Germany, the GKS WTE plant has an extended boiler pass that was

retrofitted with a section of screen tubes. As the flue gas flow turns into the extended pass, the

flow encounters a section of screen tubes, which are maintained in the temperature range of 300-

400°C (Figure 3-1). Low melting ash components condense on the walls of the empty pass. This

modification has been referred to as “chloride trap”, as the deposits are rich in chlorides. An

image of the deposits is shown in Figure 3-2. Deposits are removed daily via the horizontal wash

system “HWS” which is shown in Figure 3-1 [75].

Figure 3-1. Schematic of GKS Boiler in Schweinfurt, Germany [76]

3.4 Chemically modifying fly ash

3.4.1 Co-firing MSW with high sulfur coal or elemental sulfur

Firing high ash coal along with MSW (or biomass) can modify the WTE fly ash chemical

composition. The combustion of high sulfur coal generates a significant amount of acidic ash

particles (SiO , Al O ) [77]. These amorphous, high surface area particles provide sites for the
2 2 3

alkali chlorides to trap and collect, instead of passing through with the rest of the flue gas as

32
 Figure 3-2. Accumulation of deposits on chloride trap [76]

vapor. In the presence of higher SO2 /SO3, the chlorides are converted to sulfates which are much

less likely to form melts or undergo hot corrosion with boiler tube alloys [78]. While this

measure has been shown to be effective, co-firing of coal in a WTE boiler as a means for

corrosion reduction, is not considered to be an economically viable option [79].

3.4.2. Ammonium sulfate additive for corrosion and NOx reduction

The purpose of ammonium sulfate injection is to obtain the same benefits as co-firing

with high sulfur coal mentioned above, but it is economically more viable. This is achieved by

the dual benefit of increasing the SO3/SO2 levels in the boiler and, also, providing an additional

function in selective non-catalytic reduction of NOx molecules (SNCR). This method is

commercialized in Europe by Vattenfall and is called the Chlorout corrosion additive [80]. The

injected aqueous solution of ammonium sulfate dissociates into NH3 and SO3 in the following

reaction:

(NH4)2SO4 (aq) → 2NH3(g) + SO3(g) + H2O(g) (3-1)

33
SO3 reacts with alkali chlorides and converts them into alkali sulfates and HCl.

SO3(g) + H2O(g) + 2KCl(g) → 2HCl(g) + K2SO4(s,g) (3-2)

An example of the change in the deposition in the superheater is shown in the photographs of

Figure 3-4 which show the same tube after 1000 hour exposure b) with continuous injection of

the Chlorout additive.

As noted above, the released ammonia contributes to the SNCR reaction in reaction 3

shown below.

4NH3(g) + 4NO (g) + O2(g) →4N2(g) + 6H2O (g) (3-3)

The Chlorout technology also includes a novel alkali chloride measuring technique,

called IACM (In-situ Alkali Chloride Monitor), which measures the concentration of gaseous

alkali chlorides (KCl and NaCl) on-line. The measurement principle is based on molecular

absorption over a certain wavelength interval. IACM enables monitoring and control the

incoming fuel mix as well as control of the required amount of the ChlorOut additive.

3.4.3 Sulfur recirculation

Gotaverken Miljo, a Babcock Wilcox affiliate in Sweden, also increased SO2

concentrations in WTE boilers without combusting additional sulfur or adding sulfur

compounds, by removing SO2 in the second stage of the wet flue gas scrubbing system and

recycling the captured sulfur compounds into the furnace via a patented “sulfur recirculation

process.” [81].

34
 Figure 3-3. Vattenfall Chlorout concept [80].

With ChlorOut Without ChlorOut

Figure 3-4. Boiler tubes without and with Chlorout injection [80].

The second wet scrubbing stage (depicted as WS II in Figure 3-5) uses NaOH solution as

the desulfurization agent. During preliminary testing, H2O2 was used to react with SO2 (g) to

form sulfuric acid

H2O2 (l) + SO2 (g) - H2SO4 (l) (3-4)

35
 The sulfuric acid produced in the scrubber was injected into the furnace through a tuyere

with atomization air. In the furnace conditions, the sulfuric acid readily decomposes to SO2:

H2SO4 (l)  H2O (g) + ½ O2 (g) + SO2 (g) (3-5)

In this sulfur recirculation, the mean concentrations of sulfur in the flue gas were raised from 270

mg/Nm3 to ~800 mg/Nm3. In the sulfated ash environment, substantial reductions in tube

wastage (between 60-80%) were observed after 1000 hours of field exposure for Inconel 625

(give composition same as Sanicro), Sanicro 28 (Fe-27 Cr-31 Ni), and 15 Mo (low alloy steel) at

superheater tube metal temperature of 525 °C.

Figure 3-5. The Gotaverken Miljo sulfur recirculation concept [81].

3.5 Calcium hydroxide injection to removed HCl and SO2

Injection of chemicals such as calcium hydroxide or magnesium hydroxide Ca(OH)2 or

Mg(OH)2 into combustion furnaces has been proposed by the Earth Engineering Center and

others as a means of sequestering HCl and decreasing its concentration in the flue gas. This

36
method is still under investigation, but preliminary results have shown it to be effective [82]. An

example of a testing system which demonstrated the beneficial effect of this method is shown in

Figure 3-6, adapted from the PhD thesis of Shang-Hsiu Lee.

The chemical reaction involved in the proposed process for chlorine sequestration is the

same as that used in the semi‐dry (or “dry”) scrubber section of the Air Pollution Control in

WTE power plants. In order to meet the MACT regulations, most WTE

Figure 3-6. Schematic of corrosion test system for calcium hydroxide injection [82].

facilities, worldwide, have adopted the semi‐dry scrubber system for acid gases and particulate

emissions. In a semi‐dry scrubber system, a finely atomized lime slurry of Ca(OH)2 is injected

in the scrubber through spray lance/nozzle assemblies to react with the acid gases and generate

calcium sulfite/sulfate/chloride reaction products and has a removal efficiency of greater than

95% [82].

WTE power plants consume between 9.1 to 13.6 kg of lime (about 95% CaO) per ton of

MSW processed [83]. The CaO is converted to Ca(OH)2, yielding a concentration of about 10

wt.% Ca(OH)2. This slurry is then atomized into the scrubber system with dilution water. The

37
flowrate of the dilution water is controlled to maintain the desired outlet temperature of the

scrubber system, which is typically about 150° C. Adding the dilution water just prior to the

atomizers reduces the Ca(OH)2 concentration to just a few percent. The proposed injecting

location for the sequestration process is either at a level right above the combustion grate or in

front of the superheater section [83].

3.6 Improvements to a cleaning system

Cleaning technologies are implemented to maintain the effectiveness of the surfaces and

prevent pluggage of gas passes. Steam and air are the most commonly used media for this

application. One disadvantage from sootblowing is that localized erosion and corrosion can

occur in areas that are extraordinarily cleaned by the blowing medium. On remedy to that

problem has been the installation of tube shields.

A mechanical rapping system can be used to complement sootblowing, to limit the

number of cleaning cycles. In this system, a number of anvils strike pins to impart acceleration

through the superheater tube assembly. The purpose is to remove the bulk of the ash while

leaving a light layer so as to provide some protection to the tube [84].

38
Chapter 4: Experimental methods

4.1 Summary

In this chapter the details of a custom designed laboratory-scale corrosion apparatus and

experiments are provided. The alloys investigated in this thesis were selected based on relevance

to WTE boilers and through literature review and discussions with numerous companies in the

WTE industry. Coupon test samples with dimensions of 2.5 cm x 1. 3 cm x 0.3 cm (0.5”x 1.0” x

0.3”) were fabricated from select alloys provided by WTE vendors and operating companies An

array of coupons were subjected to an environment replicating the flue-gas conditions typically

present in WTE superheater ducts for approximately 24 hours. The corrosion that developed

during the testing was quantified by determining the total mass loss from the coupon following

the completion of the test according to an ASTM procedure for corrosion product removal. This

procedure enabled the calculation of a corrosion rate that permitted extrapolation to anticipated

wastage per year. In addition to the corrosion rate, several techniques were used to assess the

physical and chemical changes in the samples, including: scanning electron microscopy (SEM);

elemental dispersive spectroscopy (EDS); and X-ray diffraction (XRD)

4.2 Alloys

Three commercial steels and ten high temperature coatings were investigated in this

thesis. The steels which were studied at length in chapters 5-7 are shown in table 4-1. Two of the

alloys (SA 178A and SA 213 T22) have been widely utilized in WTE boilers and another

(NSSER-4, Fe-17Cr-13Ni) has been considered for future use.

39
 With the exception of NSSER-4, the test alloys were provided by plant operators in the

form of tube section, like those shown in figure 4.1. The tubes were machined into test coupons.

In addition to their industrial relevance, these particular alloys were chosen because of their

contrasting metallurgical composition for a thorough understanding of the corrosion mechanisms

at WTE superheater conditions.

Figure 4-1 Example of WTE superheater boiler tubes

SA213 T22 is used commonly in superheaters either individually or with a coating [88].

SA 178A is a carbon steel that is more commonly utilized as a waterwall alloy [85] but can

sometimes be implemented if superheaters if cladded. Both alloys are considered a benchmark

because it demonstrated poor corrosion resistance. (For the most of the analysis moving forward

in this thesis, SA 213 T22 will be used as the benchmark). NSSER-4 is a stainless steel that has

not been currently implemented into WTE power plants to date but was identified in previous

experimental work as having favorable chloride corrosion resistance [86].

Additionally, six nickel coatings were investigated for this thesis and are discussed in

chapter 6. They were comprised of 5 unique materials: Inconel 625, Colmonoy 88, SP 99,

SW1600, and SW1641. The other coating was Colmonoy 88 applied by a laser weld, so that the

performance of the application technology can also be compared. SP 99 has a very similar

40
 Table 4-1. Elemental composition of investigated base metals [87-89]

Alloy C Cr Ni Mn Fe Mo W

SA178 A* 0.07 0.02 0.05 0.47 bal 0.01

SA213 T22 0.1 2.25 bal 1

NSSER-4 0.04 17.3 13.1 0.8 bal

SA178A also contains: Si 0.06, P 0.001, Nb 0.002, Cu 0.119, Al 0.026, Ca 0.001, N 0.08, Ti 0.003, and V 0.002

Table 4-2. Elemental composition of investigated protective coatings

Composition
Alloy Process Ni Cr C B Si Mo W Fe
Inconel 625 laser Balance 21.5 0.1 10 5.6
Colmonoy 88 HVOF Balance 15 0.8 17.3 10.9
Colmonoy 88 laser Balance 15 0.8 17.3 10.9
SP 99 laser Balance 15 3 4 1.0 17.3 3.5
SW 1600 TS and IH Balance 15 3.1 4.3 2.5
SW 1641 TS and IH Balance 37.1 3.6 3.4 3
Composition
Substrate Ni Cr C B Si Mo W Fe
SA213 T22 2.25 0.1 1 Balance

Process: HVOF = high velocity oxygen fuel coating; TS: thermal spray; IH: induction heating

composition to colmonoy, with the addition of 1 wt% Molybdenum. SW1600 and SW1641

belong to a group of NiCrMoBSi alloys. The thickness of each coating was between 1500-2000

microns (1.5-2.0 mm) and was applied to the same substrate material SA 213 T22.

41
4.3 Coupon preparation

The coupons were fabricated in the Combustion and Catalysis Lab (CCL) either using a

water cooled saw or a cut-off (chop) saw equipped with an alumina-silicon carbide blade. The

typical coupon dimensions are 2.5 cm x 1. 3 cm x 0.3 cm (0.5”x 1.0” x 0.3”) allowing placement

of multiple samples in the corrosion test rig that is described in more detail in section 4.4. The

sample coupons were prepared for testing following the ASTM G1‐03 procedure for corrosion

testing. This includes imprinting the designation code into the sample surface for clear

identification post-test; grinding the sample edges and faces using a 120‐grit surface; degreasing

with an organic solvent (reagent grade acetone) in ultrasonic baths; and drying in air at 100°C for

one hour. After preparation, the clean, dry sample was weighed in a Mettler Toledo Analytical

balance (XS64), with readability to ± 0.1 mg.

4.4 Corrosion Test Rig

A schematic of the experimental setup is shown in Figure 4-2 and a photo of the actual

corrosion test rig is shown in Figure 4-3. Testing was performed in a sealed, single zone tube

furnace (Lindberg 55035) in a 25 mm (~1”) ID quartz tube. The test coupons were placed on a

316L stainless steel plate (8” x 0.75”) to prevent corrosion of the base metal. This technique

ensured that the corrosion observed occurred only on the exposed surface containing the metal or

alloy of interest. The furnace was operated by PID temperature controllers (Omega Engineering,

Inc CN 7800) connected to a K-type thermocouple (Omega Engineering, Inc., KQXL‐116U‐18)

which provided temperature resolution of ± 0.1C.

42
Figure 4-2 Schematic of corrosion test rig.

Figure 4-3 Photograph of corrosion test rig

43
 The gas delivery section shown in Figure 4-2 is represented via cylinder outlines

connected to mass flow controllers (MFC). This permitted the use of pure gases that were

independently controlled to yield a desired mixture. The gases were procured from Tech Air Inc.

and were certified ultra-high-purity grade. The gaseous environment introduced into the furnace

was intended to replicate a WTE flue gas mixture. Incorporating suggestions from WTE

engineers, the base composition of the WTE flue gas mixture consisted of 8% O2, 12% CO2, 800

ppmv HCl, 100 ppmv SO2, 15% water vapor, with a balance of N2. For each experiment, the

total gas flow rate was constant at 500 mL/min. Upon exiting the furnace, the gas mixture was

vented into a fume hood. Given the corrosive nature of the gases used in this experiment and the

health risks associated with short-term exposure to SO2 and HCl [91,92], proper personal

protective equipment and protocols were followed.

One of the advantages of using pure gases and concentrated mixtures is to enable an

understanding of the effects of varying one gas species against another. In particular, the effect

of varying SO2 and HCl are of great interest to WTE operators because they have been identified

as major contributors to hot corrosion events. As such concentrated gas cylinders containing

5000 ppmv HCl in nitrogen and 3000 ppmv of SO2 in nitrogen were used to vary the HCl and

SO2 concentration within controllable flowrates allowing precise ratios to be established.

Additionally, a total flowrate of 500 mL/min was maintained in each experiment by modifying

the nitrogen balance. Flowrates were controlled via mass flow controllers (Aalborg Instruments

and Controls, Inc., SKUW‐183356). The synthetic flue gas mixture was preheated prior to

entering the furnace to 350° C ± 20°C using electrical resistive heating elements (Omega

Engineering, Inc., FGS101‐060) wrapped securely around the inlet transfer line as shown in

Figure 4.2 and controlled by variable transformers (Stacco 1050 B).

44
 Water vapor was introduced into the test rig by vaporizing deionized water (conductivity

≤ 1.5 μS/cm) in a separate furnace upstream of the heated inlet transfer tubing. The water flow

rate was controlled via a micro syringe pump (New Era-1000). The steam furnace was controlled

via a variable transformer with the temperature monitored by a k-type thermocouple.

Temperature data was recorded by wiring the thermocouple and temperature controllers to a data

acquisition board (Omega Engineering, Inc. OMB-DAQ-55) providing continuous readings.

Figure 4-4 shows a temperature profile generated for the test furnace at a desired set point

of 500°C. The temperature measured is within 2°C of the set point over a large portion of the

heated zone of the furnace, exceeding six inches in the length. Since the test coupons are 1.0

inch in length, it was determined that six coupons can be tested per experiment experiencing a

uniform temperature field. The temperatures used in the isothermal experiments correspond with

the metal temperatures typically measured in WTE superheater tubes. A discussion comparing

the effects of a thermal gradient corrosion test with an isothermal test is provided in chapter 5.

4.5 Salt Layer Application

In addition to the gas environment, the effects of chloride and sulfate salts were also

examined. As discussed in Chapters 2 and 3, some of the most common deposits found on

superheater tubes during shutdown consist of alkali chlorides and alkali sulfates. Discussions and

literature survey resulted in the use of sodium chloride and sodium sulfate being selected for this

study as the surrogate species to probe the effect of chloride and sulfate deposits.

45
 Figure 4-4. Temperature profile of electric furnace at set point of 500°C with
500 mL/min flow of nitrogen

Salt layers were applied to coupons via successive wetting and drying in accordance with

previous designed protocols for hot corrosion testing [90]. An aqueous solution of 100 grams per

liter NaCl or 50 grams per liter Na2SO4 was applied to one side of the ASTM prepared surface of

the test coupon via a micropipette and dried in an air environment at 120-130°C. The application

of the salt layer was adjusted so that a loading of 4.0 ± 0.1 mg salt/cm2 coupon area was used in

each experiment and is further detailed in Chapter 5.

4.6 Quantifying extent of corrosion

After each corrosion experiment, the samples were left to cool in the furnace under

nitrogen flow- to limit spallation of the corrosion product as much as possible. The coupons are

then characterized using the techniques discussed in section 4.7 which then enables the extent of

corrosion to be quantified.

46
 Uniform corrosion rates are represented as a loss of metal thickness as a function of time.

This value can be directly measured from experimental data, or as is often the case, it can be

calculated from mass loss data. The mass loss is defined as the difference between the original

mass of the specimen (i.e. the sample weight after the ASTM preparation procedure) and the

mass of the specimen after exposure and after following the removal of the corrosion product.

The corrosion product removal is performed following the same ASTM G1-03 method

discussed before for preparing the corrosion samples. The corrosion product removal uses

mechanical and chemical processes. Mechanical cleaning includes scrapping, scrubbing with a

non-metallic bristle brush, and possibly placing the coupons in an ultrasonic cleaner depending

on the type of corrosion product and the adhesion to the specimen. Throughout the testing it was

observed that mechanical cleaning was not sufficient to remove all of the developed corrosion

product post-testing. This necessitated the use of an additional chemical cleaning step.

In the ASTM procedure, there are multiple cleaning procedures for each type of alloy.

The choice of which procedures to follow is partly a matter of trial and error. The chemical

cleaning is typically accompanied with mechanical cleaning (brushing or placing in an

ultrasonicator bath). Cleaning may require multiple cycles until a constant weight is obtained.

The chemical procedure involves immersion of the corrosion test specimen in a specific solution

that is designed to remove the corrosion products with minimal dissolution of underlying metal

that was not corroded during the test. The following steps are the generally accepted chemical

procedures that were adopted for the alloys tested in this thesis [91].

It is critical to address however, that this procedure should remove only the corrosion

products and not result in the removal of a base metal. To determine the mass loss of the base

47
metal, the cleaning procedure should be replicated on an control specimen. If there is a mass loss

associated with the method chosen, and it is reliable, that value should be corrected when

assessing the mass loss. Typically when applying the chemical procedure, at most 0.5 mg mass

was corrected from the virgin metal loss- which were the case for low alloy steels and the zine

and sodium hydroxide cleaning at 80° C.

It also should be noted that if the primary type of corrosion is pitting, a mass loss

measurement is generally not a good indication of the extent of pitting- unless there is also

uniform corrosion and the ratio of pitting corrosion to uniform corrosion is small. To evaluate

pitting corrosion, another ASTM procedure [92] is recommended, and the pitting can be further

quantified by microscopy techniques. In either case however, reporting the mass change per

specimen area (e.g. mg/cm2) is acceptable.

Once the final (and corrected) mass loss is obtained, the mean corrosion rate may then be

obtained by applying the following equation:

Corrosion rate = [K * w] / [A* t * ρ] (4-1)

where t = exposure time (in hours), A is the area of the coupon (cm2), ρ is the alloy density (in

g/cm3), and w is the mass loss (in milligrams or grams). The densities of the investigated alloy,

assembled from various sources are shown in table 4-3. The rate constant, K, is a constant which

can be changed depending on what values are of interest, typically either penetration

(mm/mil/μm) per time unit. The mass loss per unit area per unit time can also be acceptable.

48
 Table 4-3. Densities of alloys [93-94]

Alloy Density (g/cm3)

Base metals
SA 178A 7.86
SA 213 T22 7.85
NSSER-4 7.98
Protective Coatings
Inconel 625 8.44
Colmonoy 88
9.89
(HVOF)
Colmonoy 88
(laser) 9.89
SP 99 9.89
SW1600 7.73
SW1641 6.31

4.7 Characterization techniques

In addition to determining the mass loss and the corrosion rate, there are a variety of

analytical techniques which have been used to identify the chemical and physical changes within

a sample. This section provides a brief description of each of the various analytical techniques

used in the discussion of this thesis.

4.7.1 Scanning electron microscopy (SEM)

Scanning electron microscopy (SEM) is utilized for surface and cross section imaging. It

is a widely used technique to identify physical changes occurring at the microscale level. SEM

operates by scanning a sample surface with a beam of electrons, which are generated by an

electron gun and focused with magnetic lenses down to a diameter of about 0.01 nm. The

electrons interact with atoms at the surface of the sample, leading to emission of new electrons.

The emitted electrons are then collected and counted with a detector. Some SEM units have

49
included two different modes: either detecting secondary electron emission or backscattered

electron emission [95]. Both modes provide distinct advantages to the user. The secondary

electrons are generated either by the primary beam or any scattered electron that passes near the

sample surface. They are emitted after several collisions with most of their energy emerging

from the sample surface [96]. As a result, secondary electrons generate high resolution images.

Backscattered electrons are reflected back with minimum amount of collisions with surface

atoms. However, in a single collision with atoms of different atomic weights, the loss of energy

will be different and backscattered electrons will have different energies. So this mode provides

images that give a contrast between different elements, in particular for heavier elements [97].

The SEM used mostly in this thesis work is the Hitachi TM-3000, a tabletop unit, shown

in Figure 4-5. The Hitachi TM 3000 has an operating voltage of either 5 or 15 kV with

magnification up between 15X- 30,000 X and is equipped with back scattered detector.

4.7.2 Energy dispersive spectroscopy (EDS)

SEM can be combined with a number of different techniques for chemical analysis, the

most common being energy dispersive spectroscopy (EDS). This allows for an elemental

footprint for most of the elements present on the surface of a sample or the surface of a sample’s

cross section. When the electron beam from the SEM interacts with the surface, X-ray photons

50
 Figure 4-5. Hitachi TM 3000 scanning electron microscope

are generated and radiated. The energy of the radiating photons corresponds to a transition

energy which is characteristic for each element (K , K , L , L lines). These X-rays are then

passed to a semiconductor detector which converts the X-rays of the various elements into

electrical signals which are then passed to a multichannel analyzer that separates the various

elements into different channels depending on energy.

There are limitations to EDS, most notably it only has a detection limit of elements that

are 0.1 wt% or greater [104]. Therefore, the technique is often considered semi-quantitative.

Nevertheless it can be useful to provide trends and ratios among major species. The EDS

accompanied with the Hitachi TM 3000 is the Bruker Quantax 50.

4.7.3 X-Ray Diffraction

Crystalline salts (e.g. NaCl and Na2SO4) and other corrosion products are analyzed by

powder X-ray diffraction. Identifying these compounds is very useful in ascertaining reaction

mechanisms.

51
 The X-ray source consists of a tungsten filament cathode which generates electrons when

heated. When the electron beam impacts the anode, it gives rise to x-rays which are

monochromatic before reaching the specimen. The atoms in a crystal are a periodic array

of coherent scattering centers and thus can diffract light. The interaction of X-rays with a

powdered sample creates secondary “diffracted” beams of X-rays related to interplanar spacings

(d) in the crystalline powder according to Bragg’s law, nλ = 2d sinθ, where n is an integer, λ is

the wavelength of the X-rays, d is the interplanar spacing generating the diffraction, and θ is the

diffraction angle [98].

In this work a powder XRD (X’Pert MRD) was used with CuKα-radiation (λ=1.54178 Å)

source. The X’Pert software, HighScore Plus, contains a variety of databases for diffraction

patterns, most notably from NIST (National Institute of Standards and Technology). This

database was used to identify the compounds obtained in the diffraction patterns. A picture of the

X’Pert MRD powder diffraction unit is shown in Figure 4-6.

Figure 4-6. X’Pert X-ray Diffractometer

52
4.8 Research Objectives

4.8.1 Problem statement

The ultimate objective of this research is to propose solutions which can allow waste-to-

energy power plants to operate boilers for longer periods of time without corrosion related

shutdowns and operate at superheater steam temperatures of 500°C and higher. This is

accomplished by providing a scientific understanding of surface reactions related to chloride

corrosion and the impact of variables such as temperature; alloying composition, salt deposit,

and SO2/HCl flue gas ratio.

4.8.2 Scientific and engineering questions

The scientific and engineering questions to be addressed by this research are:

(1) What are the likely mechanisms of high temperature corrosion of boiler alloys in the

presence of WTE flue gas? Under chloride salt deposits?

(2) What is the relationship between temperatures on corrosion under chloride deposits?

(3) What types of alloys are most suitable for corrosion resistance under aggressive WTE

superheater conditions?

(4) What is the relationship of SO2/HCl to the high temperature corrosion of alloys?

(5) What are the recommendations that can be given to the WTE Industry in order to mitigate

corrosion?

53
Chapter 5: Parametric studies on hydrogen chloride
corrosion

5.1 Summary

The following is an experimental investigation into the effects of hydrogen chloride gas

on the high temperature corrosion of steels at superheater boiler tube temperatures (450-550°).

The experiments focused on two common boiler tube alloys: carbon steel SA178A (Fe-0.1C),

low alloy steel SA213 T22 (Fe-2.25 Cr- 1.0 Mo) and a promising material for WTE boiler tubes,

NSSER-4 (Fe-17Cr-13 Ni).

In an oxidizing environment at high temperature, the presence of hydrogen chloride gas,

accelerates the corrosion steels. The parabolic rate constant of SA178A increases from 0.17 to

1.8 μm2/h with the addition of 800 ppmv of hydrogen chloride to the synthetic flue gas. The

acceleration was inferred to be caused by the active oxidation as shown through physical

observations, corrosion product layer characterization (XRD, EDS) and thermodynamic

calculations. Activate oxidation can be mitigated with alloying additions of chromium and

nickel, as demonstrated with experiment using NSSER-4 (Fe-17Cr-13Ni). The austenitic steel

had nearly half (53 kJ/mol) the apparent activation energy of the carbon and low alloyed steel

respectively (110-111 kJ/mol).

In the remainder of the chapter some additional parameters were investigated on the HCl

corrosion behavior. It was determined that concentration of hydrogen chloride in the flue gas had

a significant difference in the thickness of scale formed on SA178A a between 0 and 400 ppmv

at 500 °C and a moderate effect between 400-800 ppmv. Studies investigating the impact of a

54
thermal gradient on SA178A also showed that the corrosion rate had a minimal change in mass

change in a gas-metal temperature gradient of 260°C as compared to isothermal environment.

5.2 Introduction

The average mass percentage of chlorine in MSW is compared to other solid fuel sources

in table 5-1. Several biomass and coal types are listed to demonstrate the variability. MSW has a

narrower chlorine range between and with the exception of a few sources (e.g. Illinois River

Basin coal, switchgrass) have significantly higher chlorine content. In chapter 1, table 1-1, it was

shown that the chlorine content in MSW is contributed from various waste streams including

plastics (0.26 wt%), biomass (0.34 wt%), textiles (0.55 wt%) and paper and paperboard (0.06

wt%). Thus, the presence of chloride compounds cannot be avoided in WTE power plants

without drastically modifying the fuel composition and plant operation.

Table 5-1. Chlorine content of various solid fuels [99-100]

Percent Chlorine
Fuel Subcategory
(Weight %)

bark, coniferous trees 0.02

straw 0.18
Biomass
switchgrass 0.79
yard wastes 0.11
Appalachian coal 0.1
Coal Illinois Basin coal 0.32
Pacific River Basin coal 0.01
MSW 0.47-0.72

55
 Chlorinated compounds in MSW are converted to hydrogen chloride gas (HCl) in the

combustion environment. Hydrogen chloride may form chlorine gas, Cl2, in an oxidizing

environment. Below 600°C, chloride gas Cl2 is the thermodynamically favored species [101]. In

addition to HCl and Cl2, metal chlorides are also present in the form of vapor or condensed

species. It has been suggested in the literature that these species form as a result hydrogen

chloride reacting with the inorganic fraction of MSW [110]. Two examples of common metal

chlorides, sodium and zinc, and their formation reactions are shown in equations 5-1 and 5-2

below.

Na2O (s,g) + 2HCl(g)  2NaCl(s,g) + H2O(g) (5-1)

ZnO (s,g) + 2HCl(g)  ZnCl2(s,g) + H2O(g) (5-2)

The complexity of corrosion environment in WTE boilers is increased by the participation of the

different types of chloride compounds.

The experimental investigation in chapters 5, 6 and 7, is intended to provide an

understanding on the corrosion phenomena by identifying the effects of individual parameters on

high temperature corrosion. Chapter 5 mostly focuses on the behavior of hydrogen chloride

while chapter 6 focuses on the effect of metal chlorides- using sodium chloride as a surrogate

compound. Chapter 7 in part investigates the behavior of both chlorine compounds under the

influence of changes in the flue gas environment-specifically due to sulfur.

56
5.3 The effect of hydrogen chloride

The effect of hydrogen chloride on high temperature corrosion was observed by exposing

coupons of SA178A carbon steel in an electric furnace over various exposure times at 500°C. In

one set of experiments the coupons were subject to a representative WTE flue gas mixture,

consisting of 15% H2O, 12% CO2, 8% O2, 100 ppmv SO2, and 800 ppmv HCl and N2 balance for

a total flow rate of 500 milliliters per minute. In the second set of experiments, hydrogen

chloride was removed from the gas mixture and nitrogen was added to keep the total flowrate in

both test conditions constant. The coupons were recovered from the furnace after 12, 24, 50, and

72 hours respectively.

Figure 5-1 shows the average mass loss obtained from the coupons of SA178A under the

two gas environments. The mass loss is obtained by removing the corrosion product from the

coupon surface, following the method specified in ASTM G1-03 discussed in chapter 4. The

values in the figure are normalized over the specimen area (mg/cm2) to account for small

changes in the coupon area (typically 3.45 cm2) and represent the average of six coupons.

The profile of the mass loss versus time shows largely different behavior between the two gas

environments. The profile of the gas environment containing 800 ppm HCl has steeper increase

at 12 hours and appears to follow a parabolic trend up to 72 hours. The profile of the gas

environment containing 0 ppm HCl has more moderate increase and appears to plateau after 24

hours. The values from figure 5-1 can be used to determine the kinetic behavior of the high

temperature corrosion of SA178A under the two different corrosive environments. As discussed

in chapter 2, high temperature corrosion typically follows a parabolic rate law. This can be

determined by plotting square of the weight change over time, and determining whether the best

57
fit is linear or parabolic. Both the 800 ppm HCl and 0 ppm HCl exhibited a linear profile with the

square of the weight over time. Based on the data it can be concluded that the corrosion under

these two environments follow a parabolic rate law. The slope of the plot of the weight squared

per time is the parabolic rate constant and is reported in Table 5-2. Additionally, the units are

expressed in thickness squared over time. Based on these results, the corrosion rate of SA178A is

11 times greater at 500°C under the 800 ppm HCl environment than the 0 ppm HCl environment.

Figure 5-1. Mean mass loss of SA178A versus time (0-72 h) at 500°C under two synthetic flue
gas environments (0 ppmv HCl and 800 ppmv HCl)

58
 While the corrosion rate is evaluated by determining the total mass removed (intended to

be the total corrosion product), it is worth noting that the mass losses are evaluated in two stages

which can provide additional insight on the corrosion behavior. First, the corrosion product is

lightly scraped and/or brushed to remove scales or powder from the coupon surface. Secondly,

the coupons are subjected to chemical cleaning, whereby the corrosion products are dissolved

with minimal loss of the base metal.

Figure 5-2. Square of mass loss of SA178A versus time under two synthetic flue gas
environments (0 ppmv HCl and 800 ppmv HCl

Table 5-2. Rate constants for SA178A, 500°C (932°F) under corrosive gas mixtures

Test Environment Parabolic rate constant, 500°C

mg2/cm2-h μm2/h mil2/h

0 ppmv HCl 0.11 0.18 3.0E-04

800 ppmv HCl 1.18 1.7 2.5E-03

59
 An example of this mass loss breakdown is shown in table 5-3. The mass loss of SA178A

under a WTE mix with 0 ppm HCl after 72 hours has nearly the same overall mass loss (i.e.

corrosion layer thickness) as SA178A coupon at 12 hours under 800 ppm HCl. Yet in obtaining

the same overall mass loss, the mass loss contributions are significantly different. The coupon

under the HCl environment has lost a significant mass due to mechanical cleaning, whereas the

coupon under 72 h in a HCl absent environment, required chemical dissolution of the corrosion

product.

It can be inferred that a higher a chemical removal suggests a more adherent corrosion

product. Thus, the corrosion layer in the absence of HCl forms a more adherent corrosion layer

to the carbon steel coupons. Figure 5-3 demonstrates this behavior by comparing SA178A

coupons after 50 hour experiments. The coupon under the HCl environment appears to have

cracks and has already lost mass during the process of cooling down from test temperature to

room temperature.

Figure 5-3. Images of SA178A coupons following (a) 50 hours under 800 ppmv HCl and
(b) 50 hours under 0 ppmv HCl gas mix

In general, coupons that have a thicker corrosion layer are more susceptible to having a

layer peel or scrap off the coupon during temperature fluctuation or when a test specimen is

cooled down to room temperature. This behavior is known as spallation. There are various

60
 Table 5-3. Mass change comparison for three test conditions for SA178A at 500°C

Spallation &
Total mass loss mechanical removal Chemical removal
Test condition (mg/cm2) (mg/cm2) (mg/cm2)

0 ppm HCl, 72 h 2.1 1.7 0.4

800 ppm HCl, 12 h 2.0 0.2 1.8

800 ppm HCl, 50 h 8.4 5.7 2.7

factors which can cause spallation including epitaxial stress, the mode of the diffusing species,

and the difference in thermal expansion coefficients between the corrosion layer and the alloy or

the types of oxide formed in the corrosion products [102].

The tendency for certain oxides to promote spallation can be predicted using the Pilling

Bedworth ratio (PBR)- which is defined as the ratio of the molar volume of the metal oxide to

the molar volume of the metal substrate. Equation 5-3 shows the expression of PBR where M

stands for the atomic or molar mass, ρ is the species density and n is the atoms of metal per mole

of oxide. If there is a large difference in the volume of the oxide versus the metal the oxide is

more likely to spall. Therefore, ratios much less than or greater than 1 would be candidates for

spallation.

PBR = Voxide / Vmetal = Moxide * ρmetal / n* Mmetal * ρoxide (5-3)

Thermodynamic calculations indicate that Fe2O3 and Fe3O4 are the predominant corrosion

products for high temperature corrosion of SA178A (99 wt% Fe), except at the metal-oxide

interface where the partial pressure of oxygen may be <10-10 bar, at which iron (II) chloride

would be able to form. This is shown in a Fe-O-Cl predominance diagram at the test temperature

61
of 500°C, which considers the thermodynamic equilibrium between the potential iron oxides and

chlorides as well as iron at a fixed temperature. The Fe-O-Cl diagram at 500°C was provided in

figure 2-3 and is displayed again below.

Applying this understanding, the PBR calculated for compounds Fe2O3 and Fe3O4 on Fe

was determined to be 1.02 and 1.2 respectively. As these values are approximately close to 1,

they are typically considered an adherent oxide. Therefore in the case of SA178A and comparing

the test conditions, PBR is not a good marker for explaining spallation of SA178A coupons.

Thus, other phenomena should be considered.

Figure 2-3. Predominance Diagram of Fe-O-Cl system at 500°C

Figures 5-4 and figures 5-5 show the XRD pattern for the scales recovered from SA178A

coupons 500° C under the flue gas environments with and without hydrogen chloride. The

patterns for the scale are shown on top in red, followed by the peaks of standard compounds

62
(ICDD software, X’Pert instrument software) in blue below. Both scales were identified to

include two types of iron oxides, maghemite and magnetite. This is in agreement with the

thermodynamic calculation previously discussed. Additionally, the patterns did not reveal other

compounds which would be able to explain why the corrosion product under the HCl containing

flue gas was nearly 2.5 times larger in mass.

Figure 5-4. XRD pattern of scale recovered from SA178A coupon at 500°C under flue gas
containing 0 ppm HCl (with patterns of highest matched compounds)

Additionally, elemental dispersive x-ray spectroscopy (EDS) spot analysis was performed on two

SA178A coupons in the HCl containing environment was performed above and below the

spalled corrosion product. An average of 3-spots was taken for both coupons and the results

which are shown in Table 5-2. While there is a small detection of sulfur below the corrosion

layer, chlorine was not detected. Chlorides were therefore not detected in the corrosion layer

surface neither in the scale analysis, nor below. It should be noted however, that the

63
Figure 5-5. XRD pattern of scale recovered from SA178A coupon at 500°C under flue gas with
800 ppmv HCl (with patterns of highest matched compounds)

analysis provide in Table 5-4 is qualitative in that the corrosion layer removal depends on the

layer adherence. This does not conclusively eliminate the presence of chlorine or other elements.

The oxide layer, as demonstrated in Table 5-3 is not fully recoverable without the use of a

chemical solvent or a polishing method (the latter case would likely remove virgin material).

Table 5-4. EDS Analysis of SA178A corrosion layer and below layer
(Values are normalized wt%)

Below Corrosion Corrosion Layer

Element Layer (wt%) (wt%)

Iron 86 62

Oxygen 13.9 38

Sulfur 0.1 0

Chlorine 0 0

64
 This analysis would be further supported by performing a cross-section SEM-EDS

analysis. Other studies have performed this type of study in the past and were able to identify

the presence of chlorine below the scale at the interface of the virgin metal [103]. Nevertheless,

the results above are in agreement with the literature that the addition of hydrogen chloride

engages in the active oxidation mechanism.

5.3.1 Effect of hydrogen chloride concentration on corrosion rate

The effect of the hydrogen chloride concentration was investigated at 0, 400 and 800

ppmv and is shown in figure 5-6. Like the experiments discussed in figure 5-1, the HCl

concentration was modified and replaced with nitrogen to maintain the same the total gas

flowrate. The mass losses were obtained following 24 hour exposure at 500°C. The error bars

represent the standard deviation from the mean of the mass loss obtained from the sample

coupons.

Figure 5-6. Mean mass loss of SA178A after 24 hours at 500°C under flue gas mixtures with
0,400, and 800 ppmv HCl

65
 There is a significant increase between 0 and 400 ppmv HCl, as the value of the average

mass loss is doubled from 1.6 to 3.3 mg/cm2. The average value obtained at 800 ppmv had a

marginally higher loss at 4 mg/cm2. These values, the average of 6 coupons per experiment, had

a variation of 0.4-0.6 mg/cm2, which is demonstrated with the error bars.

The effect of corrosion due to hydrogen chloride is consistent with findings from other

authors which investigated the hydrogen chloride effect at 500 ppm and 1000 ppm on alloys

between 500-600 °C respectively [104] under a similar WTE flue gas environment. In that

study, the corrosion rate of SA178A was elevated by 70% between 0 ppm HCl and 500 ppm HCl

and elevation by only 7% between 500 ppmv and 1000 ppmv [114]. This is compared to the

data in figure 5-6 which shows a 106% elevation between 0 ppm and 400 ppm HCl and a 20%

increase between 400 ppm and 800 ppm. This behavior suggests that there is a possible diffusion

limitation above 500 ppm HCl.

5.4 Comparison of alloy behavior

Figure 5-7 shows the mass loss of SA178A, SA213 T22 (Fe-2.25Cr-1Mo) and NSSER-4

(Fe-17Cr-13Ni) under a constant gas environment containing 800 ppm HCl at 450, 500, and

500°C respectively after 24 hours 500°C respectively after 24 hours. The mass loss of the

carbon and low-alloyed steel was approximately an order of magnitude greater than the stainless

steel, 0.25 mg/cm2 at 500°C compared with 3.6 mg/cm2 and 3.9 mg/cm2 for T22 and SA178A.

An image of NSSER-4 coupon vs. the SA213 T22 coupons (figure 5-8) demonstrates the stark

contrast between the two corrosion products.

66
 The NSSER-4 corrosion product is a thin film, and only required chemical cleaning in

order to determine the total mass loss, whereas the SA 213 T22 exhibited mass loss distribution

to SA178A between spallation and mechanical removal to chemical removal. The corrosion

resistance of NSSER-4 (Fe-17Cr-13Ni) is consistent with the active oxidation mechanism. The

formation of metal chlorides is considered a paramount part of the corrosion process [105]

Considering the equilibrium FeCl2(s) = FeCl2(g), which would the single type of metal chloride

formed in the case SA178A and SA 213 T22, it evident from the calculation provided by figure

5-9 that the equilibrium vapor pressure of FeCl2 is 2 orders of magnitude higher at 500 °C than

nickel and chromium chlorides.

A difference in the corrosion behavior of the various steels at the different temperatures

can be quantified in terms of an apparent activation energy. As discussed in chapter 2, the kinetic

temperature dependence which follows an Arrhenius-type equation-the theory of which is

discussed in chapter 2.

The activation energy is usually obtained by the slope of the log of the rate constant

versus the inverse of absolute temperature. Using the data from figure 5-7, the parabolic

constant cannot be determined, as all of the data points at the three temperatures were obtained at

the same exposure time of 24 hours. Instead of the parabolic rate constant, the penetrations of

these three materials after the 24‐hour test were used to plot their temperature dependence. The

densities of the steels- which can be found in chapter 4- were used to convert the mass loss

(mg/cm2) from figure 5-7 into a penetration (units of microns). The mass loss is already

assumed to be uniform based on normalizing the mass loss per area of the coupon.

67
 Figure 5-7. Corrosion rates of NSSER-4, SA178A, and SA213 T22 at 450 -550°C, under
WTE flue gas containing 800 ppm HCl

The apparent activation energy calculated for SA178A, SA213 T22, and NSSER-4 is

reported in Table 5-6. The low alloys steels are on the order of 110-111 kJ/mol, more than

double the value for NSSER-4, 53 kJ/mol. Therefore the lower alloyed steeled are more

sensitive to temperature than the austenitic stainless steels. The values for the apparent activation

energy are comparable with the values obtained by Lee (2009) after 100-hour experiments

determined that the apparent activated energies of SA178A and NSSER-4 were 149 kJ/mol and

the 59 kJ/mol respectively between 486-586°C. The higher value for the apparent activation

energy for SA178A can be explained by the results at 586°C which had a 7-fold increase of the

68
parabolic rate constant. Removing the value at 586°C, lowers the apparent activation energy to

117 kJ/mol.

Figure 5-8. Equilibrium vapor pressures of solid metal chlorides as a function of temperature
[106]

This agreement interesting to note as that the apparent activation energy for SA178A and

NSSER-4 between those results and the result in this section were acquired over different time

intervals, for 24 and 100 hours respectively. Therefore, it can be concluded, that the corrosion for

SA 178A and NSSER-4 are in the same kinetic regime at 24 hours and 100 hours.

69
Table 5-5. Comparison of the types of mass losses for alloys, following 24 hour exposure

Total mass loss Mechanical removal Chemical removal

Alloy (mg/cm2) (mg/cm2) (mg/cm2)

SA178A 3.9 3.3 0.6

SA213 T22 3.6 2.3 1.1

NSSER-4 0.25 0 0.25

Table 5-6. Apparent activation energy for corrosion of alloy under WTE flue gas

Alloy Apparent activation energy

(kJ/mol)

SA178A 111
SA213 T22 110
NSSER-4 53

5.5 Comparison of isothermal vs. temperature gradient testing

Another interesting variable can be investigated from these results, is the impact of a

temperature gradient on furnace experiments. One of the variables which exist in real boiler

environments which are difficult to reproduce in laboratory experiments is the temperature

gradient between the gas and metal surface. Typically for a WTE superheater, the flue gas and

the tube temperature van vary between 200-300°C. In reality that temperature difference also

changes rapidly, based on the behavior in the boiler.

There have been experiments which have looked at high temperature corrosion in WTE

environments under a gas-metal temperature gradient as well with a temperature fluctuation.

70
Lee (2009) looked at the corrosion of SA178A under a gas-metal temperature of ~250°C, by

flowing air into a hollowed sample carrier for the alloy coupons within a furnace.

Comparatively, Kawahara (2006) conducted tests with gas-metal temperature gradients

between100-300 and also examined the gradient ramp rate between 100-300 °C.

Kawahara (2006) examined the effect of thermal gradient on the corrosion of SA178A

under a WTE flue gas environment and identified that thermal gradient as high as 200°C had a

minimal effect on accelerating the corrosion of SA178A [107]. The rate was affected at

fluctuations exceeding 200°C/min over 50 hours, but only marginally increasing the thickness

loss by 15% from 0.052 mm to 0.045 mm at a metal temperature of 550°C.

. The corrosion rate – as defined as the mass loss per area per time- is compared in Table

5-7 for the isothermal experiments reported in this chapter with the constant thermal gradient.

The results are in virtually agreement with one another. This suggests that the spallation

observed from the corrosion tests under the HCl environment likely occurs upon cooling and not

during the experiment temperature. Alternatively, the broken layer would allow a significantly

higher corrosion rate. Therefore, it is the metal temperature which most strongly governs the

high temperature corrosion behavior.

71
Table 5-7. Comparison of corrosion rate in isothermal and temperature gradient environment
(ΔT = 240°C)

Metal Gas-Metal Corrosion rate

Alloy Temperature (°C) Temperature Gradient (°C) (mg/cm2/h)

SA178A (c.s) 510 240°C (Gas ~750°C) 6.60E-02

SA178A (c.s) 500 0°C (isothermal) 6.37E-02

5.6 Conclusions

The high temperature corrosion of hydrogen chloride was examined for steels between

450-550°C. The primary corrosion mechanism was inferred to be active oxidation, based on

various observations, the product characterization, and thermodynamic calculations discussed in

this chapter. The oxidation induced by hydrogen chloride can be mitigated by replacing low

alloyed steel with alloys containing chromium, as demonstrated with NSSER-4 (Fe-17Cr-13Ni).

The addition of chromium and nickel has been noted in literature to form a more protective,

passive oxide layer. The higher alloyed material also had a lower sensitivity to temperature as

quantified by lower activation energy between 450-550°C. NSSER-4 had nearly half (53 kJ/mol)

the value of the carbon and low alloyed steel respectively (110-111 kJ/mol).

72
Chapter 6: High Temperature Corrosion of Alloys
below Chloride Salts Deposits

6.1 Summary

In this chapter, the corrosion behavior below chloride salts was investigated with sodium

chloride as a surrogate compound. Furnace tests were performed on carbon steel SA17A, low-

alloyed steel SA213 T22, austenitic steel NSSER-4 (Fe-17Cr-13Ni) and nickel based coatings

(Inconel 625, SP 99, SW1600, SW1641, and colmonoy 88) on SA 213 T22.. Tests were

conducted under a simulated WTE flue gas environment, with a NaCl loading of 4.0 mg/cm2±

10% at 450°, 500°, and 550°C for 24 hours.

The addition of the sodium chloride layer accelerated the corrosion of SA178A, SA 213

T22, and NSSER-4 by nearly an order of magnitude below the previous gas exposure containing

800 ppmv HCl. It was determined through a combination of SEM-EDS and XRD that the sodium

chloride participates in two reactions. The first is the sulfation reaction in which sodium chloride

is converted to sodium sulfate. The second reaction is the formation of sodium chromate, which

has been suggested in other literature as a type of active oxidation which can accelerate the

corrosion of alloys.

The sodium chloride corrosion tests were the then executed for several nickel based

coatings, Inconel 625, Colmonoy 88, SP 99, SW1600 and SW1641 applied on SA213 T22. The

tests confirmed similar behavior to that which was observed for the steels, namely the onset of

the sulfation reaction and the formation of sodium chromate. Inconel 625, SW1600, and SW1641

were shown to have a significant corrosion penetration below the salt layer, where sodium and

sulfur were detected with chromium and oxygen, suggesting the attack of chromium rich oxide

73
layers. The colmonoy 88 applied coatings, conversely, demonstrated excellent corrosion

resistance. It is suggested that the alloying addition of tungsten helps to mitigate the effective

pitting corrosion caused by localized depletion of the passive chromium rich oxide layer by

sodium chloride.

6. 2. Introduction

Corrosion of alloys below chloride-containing ash deposits is considered to be even

more severe than in a chlorine-containing atmosphere without deposits. Gaseous HCl is less

corrosive than alkali chloride deposits. These compounds can accelerate corrosion, even at

temperatures below first melting temperature (FMT) [118-120]. The atmosphere in contact with

the deposits and the corrosion layer affects reactions which can alter the deposit and the partial

pressure of oxidizing agents near the metal [108].

In the following experiments, the behavior of the various alloys was investigated under

the sodium chloride layer at the three temperatures discussed in chapter 5 (450°, 500°, and 550°C

respectively). The samples were covered with NaCl by droplet addition of an aqueous solution

and dried at 110°C to evaporate the moisture. The coupons were weighed afterward in order to

report the mass addition from the salt layer and were controlled to 4.0 mg/cm2. The coupons

intentionally had parts of the surface uncovered to allow for surface characterization of the

coupon surface with and without a chloride layer.

74
6.3. Studies with SA178A, SA213 T22, and NSSER-4

6.3.1 Mass change and observations

Steel coupons underwent significant corrosion under the sodium chloride salt layer,

following 24 hour experiments under the WTE flue gas exposure. The average mass loss in

mg/cm2, of the coupons of SA178A, SA213 T22, and NSSER-4 at 450°, 500° and 550°C is

shown figure 6-1. The values of are an average of 6 coupons. The values of SA178A and SA

213 T22 are similar to each another varying between 33-65 mg/cm2 for the temperature range

investigated. NSSER-4 had a marginal mass loss of 0.6 mg/cm2, at 450 °C but increased

significantly with temperature. At 450°C, NSSER-4 had a factor of 55 smaller mass loss than

SA178A, at 500 °C it was 23 and by 550°C it was a factor of 11 smaller. For industrial purposes,

the austenitic steel, Ni-17Cr-13 Ni, showed once again an excellent corrosion resistance under an

aggressive oxidizing-chlorinating environment.

The mass loss of three steels under the sodium chloride layer can be compared with the

results in chapter 5, in figure 5-7, which displayed the mass loss under the HCl-containing flue

gas alone. The values varied from 2 mg/cm2 to 6 mg/cm2 between 450°C and 550°C. Therefore,

as expected, the corrosion below the sodium chloride layer was even more severe, by an order of

magnitude or more.

Examples of coupons from SA213 T22 and NSSER-4 are shown in figure 6-2. SA 213

T22 and SA178A were virtually identical in appearance, so SA178A is not shown. At 450°C,

some oxidation is visible around and under the white deposit- which was originally part of the

sodium chloride layer that was applied to the coupons. After the 24 hours tests at 500°C, the

appearance of a white deposit was still visible on the NSSER-4 coupon but nearly undetectable

for the carbon and low alloy steel. The corrosion layer on SA178A and SA213 T22 was

75
 Figure 6-1. Average mass loss of SA178A, SA213 T22, and NSSER-4 under NaCl layer
and WTE flue gas at 450°, 500°, and 550 °C

Alloy 450°C 500°C 550°C

SA213 T22

NSSER-4

Figure 6-2. Coupons of SA213 T22 and NSSER-4 under NaCl layer following 24 experiments
at 450, 500, and 550°C

sufficiently large to exhibit spallation when cooled to room temperature. The spalled corrosion

layer was harvested and ground into powder and analyzed via X-ray diffraction- the results of

76
which are discussed in section 6.3.3. At 550°C the NSSER-4 coupons also exhibited a powdery

residue but still did not produce a recoverable corrosion layer. The residue that could be

recovered was also studied.

6.3.2 Surface characterization

The surface of coupons SA 213 T22 and NSSER-4 coupons was also investigated by

SEM-EDS. Figures 6-3 shows the micrograph and the elemental composition of the SA 213 T22

at 450 and 550°C. The surface image and elemental analysis are consistent with the physical

observations in Figure 6-2. The EDS analysis also suggests that the sodium chloride that was

present on the coupon surface from sodium chloride is now a composed of chloride and sulfate.

This suggests that some of the sodium chloride salt layer has reacted with the sulfur in the

synthetic WTE flue gas mixture to form sodium sulfate. This was compared to the surface SEM-

EDS of NSSER-4, figure 6-4 which only showed sulfur on the surface. Additionally, it was

observed that no chromium was found on the surface of the coupon around the salt at 550 C,

while some was present at the lower temperature.

6.3.3 XRD of corrosion products

As mentioned in section 6.3.1, the spallation of SA213 T22 coupons was retrieved

following the experiment. The scales were ground with an agate mortar and pestle and analyzed

via powder x-ray diffraction. The pattern from the scales recovered at 450°C and 550°C are

shown in figures 6-5 and 6-6. Both samples detected a mixture of iron oxides and sodium

chromate. The scale at 550° also detected chromium iron oxide. Neither sample detected sodium

chloride or sodium sulfate.

77
Figure 6-3 (a). Surface SEM-EDS of SA213 T22 coupon following 24 hour experiment under
sodium chloride layer at 450°C. Surface shows a mixed chlorine-sulfur surface, suggesting the
conversion of NaCl to Na2SO4. Iron, oxygen, and alloying element M also detected.

Figure 6-3 (b). Surface SEM-EDS of SA213 T22 coupon following 24 hour experiment under
sodium chloride layer at 550°C. Surface shows surface that is rich with iron oxides with a small
mass percent of sulfur and sodium. Chlorine is barely detected.

78
Figure 6-4 (a). Surface SEM-EDS of NSSER-4 coupon following 24 hour experiment under
sodium chloride layer at 450°C. Surface largely shows the salt layer on the coupons surface
which detects no chlorine, but 16.9 wt% sulfur.

Figure 6-4 (b). Surface SEM-EDS of NSSER-4 coupon following 24 hour experiment under
sodium chloride layer at 550°C. Surface largely shows the sodium and sulfur from the salt layer
surface.

The mass recovered from the scales was relatively large, about 80-90 mg such that the

entire mass was not analyzed in the run. This may explain why surface analysis detected the

presence of sulfur compounds and the XRD pattern did not identify a sulfur compound. It should

79
also be stated that one of the limitations of EDS is the penetration depth of ~100 nm for

backscattering detector. Therefore while sulfur may have been present, the remnant of the salt

may not have been present. Meanwhile, as the corrosion products of NSSER-4 coupon were too

small to recover at 450°C, the oxide layer was thin enough such that the only powder recovered

was sodium sulfate.

Fe3O4

Fe2O3

Na2Cr2O4

Figure 6-5. XRD pattern of scale recovered from SA 213 T22 following 24 hour experiment at
450°C. Peaks identified include iron oxides: Fe2O3 and Fe3O4 and a sodium chromium oxide

80
 scale

CrFeO3

Na2Cr2O4

Fe2O3

FeO

Figure 6-6. XRD pattern of scale recovered from SA 213 T22 following 24 hour experiment at
550°C. Peaks identified include iron oxides: Fe2O3, chromium iron oxide,

scale

Na2SO4

Figure 6-7. XRD pattern of powder recovered from NSSER-4 coupon following 24 hours
experiment at 450°C.

81
 scale

Fe2O3

Figure 6-8. XRD pattern of powder recovered from NSSER-4 coupon following 24 hour
` experiment at 550°C.

6.3.4 Discussion

Based on the high temperature corrosion literature review that was conveyed in chapter 2,

there are three possible options for sodium chloride to accelerate corrosion at 450-550°C. The

first is the formation of a molten chloride mixture, which in turn fluxes, or dissolves the

surrounding metal oxides. Table 2-5 in chapter 2 identified some common molten chloride

mixtures. Based on the temperature investigated, the only mixture that could exist under the

conditions investigated here is the FeCl2-NaCl which has a first melting temperature at 370°C.

As stated in previous chapters, iron chlorides are only stable in environments with very low

partial pressure of oxygen. According to the Fe-O-Cl predominance diagram at 500°C, the partial

pressure of oxygen at the oxide-metal interface would need to be on the order of 10-14 bar for

formation of iron chloride species to be thermodynamically possible. So while the formation of

82
the low melting mixture cannot be ruled out it would occur where virtually no oxygen is present.

An additional considereation for a molten phase is NaCl-Na2SO4 since the product

characterization in figure 6-3 identified the conversion of sodium chloride to sodium sulfate.

However, the FMT temperature of a NaCl-Na2SO4 is at 657°C. Therefore, it is highly unlikely

that the elevated corrosion rates for SA178A, SA213 T22, and NSSER-4 shown in figure 6-1

would attributed to a molten salt corrosion. A second possible reaction mechanism is the

sulfation reaction. Reaction 6-1 below demonstrates the reaction stoichiometry. 1 mole of

chlorine gas is released for every 2 moles of sodium chloride converted to sulfate. This release

of additional Cl2 alone may not necessarily accelerate the high temperature corrosion of alloy,

since it has been already

2 NaCl + SO2 + O2 = Na2SO4 + Cl2 (6-1)

2 NaCl + Fe2O3 + 0.5 O2 = Na2Fe2O4 + Cl2 (6-2)

2 NaCl + Cr2O3 + 2.5 O2 = 2Na2CrO4 + 2Cl2 (6-3)

At 500°C: ΔG (rxn 6-1) = -100 kJ/mol;

ΔG (rxn 6-2) = -54 kJ/mol;
ΔG (rxn 6-3) = -287 kJ/mol

shown in Chapter 5 that effectively doubling the concentration of HCl from 400 ppmv to 800

ppmv at 500°C, only increases the mass loss by about 20%. However, some in the literature

83
have suggested that the sulfation reaction can accelerate corrosion if the chlorine gas diffuses

from the deposits into the oxide layer [109]. However, as it will be shown in chapter 7, sulfation

reaction has demonstrated in some cases to mitigate the corrosion of alloys.

The third reaction that is considered was previously discussed in chapter 2, in which

alkali chlorides (e.g. NaCl) react with iron or chromium oxides. The reactions being discussed

here—and shown in the schematic, figure 6-8. As in the case of the sulfation reaction (reaction 6-

1) chlorine gas is released, this can be similar to the active oxidation mechanism discussed for

HCl. Unlike the behavior of HCl, this reaction consumed an oxide that was present on the alloy

surface, thus depleting a protective barrier to the virgin metal surface. Based on the XRD pattern

in figures 6-5 and 6-6 identifying the formation of sodium chromate, it is most likely that the

acceleration under the sodium chloride layer is caused by these types of reactions. As will be

shown in the following section, the formation of sodium chromate was observed again for

Inconel 625 coupons which further support this claim.

Although the XRD pattern recovered from the small powder from the NSSER-4 coupons

was inconclusive, in terms of identifying a probable mechanism, it should be noted that the

powder analysis and the surface EDS are marked by a significantly high concentration of iron

and a low concentration of chromium. This may suggest that the chromium layer which is grown

on the surface of the alloy was depleted by sodium chloride allowing for the outward growth of

iron oxides. Further characterization would be needed to confirm this hypothesis- specifically, a

cross sectional image of a NSSER-4 coupon with an elemental mapping to identify where iron,

chromium, and nickel and perhaps chloride resides.

84
 Figure 6-9. Illustration of an active oxidation mechanism induced by alkali chlorides

6.4. Studies with nickel coatings

The experiments with SA178A, SA 213 T22, NSSER-4 not only provided insight on the

corrosion behavior under sodium chloride layer, it served as the validation of 24 hour furnace

test as an approach for investigating the corrosion behavior of multiples alloys in a relatively

short period of time. This test was executed to rank the corrosion resistance of various nickel

coatings applied to SA 213 T22. The composition of the alloys and the coating process was

previously shown in Table 4-2 and is repeated below as matter of convenience to the reader.

The results of six coating are discussed in the following section. They were comprised of

5 unique materials: Inconel 625, Colmonoy 88, SP 99, SW1600, and SW1641. The other coating

was Colmonoy 88 applied by a laser weld, so that the performance of the application technology

can also be compared. SP 99 has a very similar composition to colmonoy, with the addition of 1

wt% Molybdenum. SW1600 and SW1641 belong to a group of NiCrMoBSi alloys. For these

85
experiments, the coupons were fabricated with base metal, SA 213 T22 included. The sodium

chloride layer was applied only to the top surface of the coupons, with the coating material. As

SA213 T22 has been extensively studied at these temperatures and gas conditions, the mass loss

of the SA213 T22 was subtracted from the overall mass loss to report the true mass loss of the

coated material. The values for the mass loss are those from the gas environment (figure 5-7), as

sodium chloride was not applied to the base metal. Careful consideration was also taken to polish

the base metal and flatten the coupon and measure is area with a caliper to have an accurate area

for the mass correction. This method for reporting the mass loss was validated by the cross

section measurements of the corrosion images in Table 6-2. The rank and relative proportion of

the largest measured corrosion depth is, outside of one outlier, consistent with the trend in mass

loss measurements.

Table 4-2. Elemental composition of investigated protective coatings

Composition
Alloy Process Ni Cr C B Si Mo W Fe
Inconel 625 laser Balance 21.5 0.1 10 5.6
Colmonoy 88 HVOF Balance 15 0.8 17.3 10.9
Colmonoy 88 laser Balance 15 0.8 17.3 10.9
SP 99 laser Balance 15 3 4 1.0 17.3 3.5
SW 1600 TS and IH Balance 15 3.1 4.3 2.5
SW 1641 TS and IH Balance 37.1 3.6 3.4 3
Composition
Substrate Ni Cr C B Si Mo W Fe
SA213 T22 2.25 0.1 1 Balance

86
 6.4.1. Alloy ranking

The results of coating tests are provided in figure 6-7 for all three test temperatures. The

mass loss of the base metal, SA213 T22, was also included for comparison. As expected, the

coating mass losses were lower than SA 213 T22, with the exception of SW1600. (It should be

noted that since SW1600 is a more dense material than SA213 T22, the calculated corrosion

penetration, corrosion rate would be smaller, even if the mass change were identical). The mass

loss for the coatings typically varied between 10-40 mg/cm2 between the tested temperature

range. The hierarchy of the coatings was the same at each temperature, i.e. the best performing

alloy had the lowest mass loss at each temperature and the worst performing alloy had the

greatest mass loss at each temperature.

High velocity oxygen fuel applied (HVOF) colmonoy 88 was the superior material with a

mass loss ranging between 0.3-3.1 mg/cm2. This equates to a corrosion rate of 5.0 E-04 mil/h at a

the temperature of 450°C. The mass loss acquired for the HVOF colmonoy 88 was a factor of ten

lower than the next resistant coating which was the laser cladded colmonoy 88.

6.4.2 Mass loss and observations

The images of coupon surfaces following the furnace experiments can be found in

Appendix A. The sodium salt layer behaved in some ways similarly to the experiments on low

alloyed steels and NSSER-4. The sodium salt layer remained largely visible at 450° and 500°C

and was then indistinguishable at 550° C. One difference with the salt on the nickel coatings

was the appearance of a color change from white to yellow. This was observed most

prominently for SW1600.

87
 Each coating itself underwent color changes after the furnace test, with the exception of

the HVOF colmonoy 88 which was initially dark. The most intense color changes on the coating

surface were observed at or around the sodium salt layer. For some alloys the color change was

a black film (laser colmonoy 88), and red for other coatings (e.g. SP 99). Inconel 625 and SW

1600 had what can be described as a macroscale pit forming around the sodium salt deposit at

450° and 500°C. At 550 °C, the surface appeared to have a uniform corrosion.

Figure 6-10. Average mass loss of Ni-coatings, Inconel 625, Colmonoy 88, SP 99, SW1600,
and SW1641 applied to SA213 T22 following 24 hour exposure under NaCl layer and WTE flue
gas at 450°, 500°, and 550°C

88
 6.4.3 Product analysis

With the exception of Inconel 625, the corrosion products could not be removed from the

coupon without compromising the coating surface. The corrosion products from Inconel 625

were grounded and studied with XRD. The pattern of the scale is provided in figure 6-7. The

other coatings were embedded in POLYCAST-128 metallographic resin and a cross section

SEM-EDS analysis was performed. The image for the nickel coatings at 500° C are shown in

Table 6-1. The elemental composition of the corrosion products adjacent to the coating surface is

also provided, as the average of four EDS spot analysis.

The peaks in the Inconel 625 scale include nickel oxide, sodium sulfate, sodium chromate

and iron oxide (magnetite). The detection of sodium chromate is an indication of a reaction

between sodium chloride and chromium oxide as previously discussed in section 6.3.4.

Additionally, iron is observed on the coupon surface via SEM-EDS analysis in figure 6-8. The

combination of a depleted chromium oxide and the onset of iron oxide forming on the coupon

surface can help explain the high mass loss measured for the Inconel coupons. The depletion of

the protective chromium oxide layer allows for outward growth of the iron oxide layer to the

surface. The formation of sodium chromate may also suggest why a yellow residue was observed

on the salt layer on the surface of coupons like SW1600. Sodium chromate is a yellow solid at

room temperature, whereas sodium sulfate and sodium chloride are white [110].

Three salient observations can be conveyed from the cross section elemental analysis in

Table 6-1. The first is that similar to Inconel 625, SW1600 and SW1641 show an unexpected

89
 scale

NiO

Na2SO4

Na2Cr2O4

Fe3O4

Figure 6-11. XRD pattern of scale recovered from Inconel 625 coating at 500°C.

Figure 6-12. Surface SEM-EDS of Inconel 625 scale at 500°C. The iron content on the surface
averaged 3-4 wt% in spots like the one shown above.

90
value for iron, 3 and 5 wt% respectively. Unlike Inconel 625, which contains iron, the source of

in SW1600 and SW1641 is likely from an impurity which was introduced during the coating

process. The cross section image confirms that the coating was still intact below the corrosion

penetration. It should be noted also that iron is even measured on the surface of the SW1600 and

SW1641 coupons. It is most notably observed at 550 °C one detects iron at 1 and 3.6 wt% on the

surface at 550°C (figure 6-9). The second observation is that SW1600 and SW1641 also had a

comparatively lower value for chromium at 1 and 4 wt% while the sodium and sulfur peaks were

discernibly higher. In the elemental mapping for these images, shown in chapter 7, the fissures

found below the salt layer (most notably seen in SW1641) feature chromium, oxygen, sulfur, and

sodium. Based on the behavior observed for Inconel 625, it can be inferred that the sodium

chloride attacks the chromium rich oxide layer.

The second important observation is that the three coatings which have the lowest mass

loss have one common feature: the alloying of tungsten and substantial concentration of

tungsten found adjacent to the metal surface. While not commonly cited in high temperature

corrosion literature, tungsten is cited elsewhere for improving resistance against pitting corrosion

[111-112]. Simply defined, pitting corrosion is a localized form of attack that results in a rapid

penetration at small discrete areas. Pitting corrosion results from a failure of a passive film. It

can be argued that the high temperature corrosion of alkali chlorides is a form of pitting

corrosion. Tungsten alloying benefits are often compared with molybdenum, as they are both

strong carbide former sand solid solution strengtheners [113]. In terms of corrosion resistance,

molybdenum and tungsten have been shown to increase the passivity of austenitic steels and

nickel alloys but stabilizing microstructure. Despite their similarities, it is apparent, especially

91
Table 6-1. Cross section image of coatings Colmonoy 88, SP 99, SW1600 and SW161 under
NaCl layer at 500°C and compositions of corrosion products adjacent to metal surface

Corrosion products
Coating Cross section image (500°C) adjacent to metal
(wt%)

Colmonoy88
(HVOF) 32 O, 5Na, 10W 4 S,
15Cr, 5 Fe, 26 Ni

Colmonoy88 31 O, 11Na, 7W, 7S,

(laser) 14 Cr, 6Fe. 28 Ni

32 O, 4Na, 3 Si, 10
SP 99 W, 4 S, 1Cl, 12Cr, 5
Fe, 27 Ni

40O, 24 Ni, 22 S, 10
Na, 1 Cr, 5 Fe
SW1600

92
 37O, 31 Ni, 3Fe,
SW1641 6Na,18S, 1 Cl, 4Cr

Table 6-2. Comparison of corrosion depth measured by coupon cross section (average of 3
penetration depths) and mass loss obtained from ASTM method. The depths are consistent with
the mass loss reported, with the exception of the laser cladded colmonoy 88

Comparison of average corrosion depth

and mass loss at 500°C

Coating Mass loss (mg/cm2) Depth, μm (mil)

Colmonoy 88 (HVOF) 1.8 2.5 (0.1)

SP 99 10.4 21.4 (0.83)
Colmonoy 88 (laser) 11.6 40.4 (1.6)
SW1641 21 58.0 (2.3)
SW1600 53 101 (4)

when comparing the results of Inconel 625 (which contains ~10% Mo) the addition of tungsten is

more effective in increasing the corrosion resistance under alkali chlorides. It would be of

interest to quantify whether there is a relationship that can be expressed between the corrosion

performance under chloride salts and the percentage of tungsten in nickel-chromium alloys. A

relationship for pitting corrosion in iron (III) chloride solution has been developed for austenitic

steels and nickel alloys. For the latter, the resistance to pitting corrosion (PREN) is expressed as

PREN= % Cr + 3* (%Mo + %W) [126]. While this relationship does not hold for the data

provided in this chapter, it can serve as the impetus for a future study on various nickel coating

with varying additions of molybdenum and tungsten.

93
Figure 6-13. SEM-EDS of surface of SW1600 (top) and SW1641 (bottom) coupons at 550°C.
EDS. Surface EDS shows the presence of iron on both surfaces, which is unexpected.

Lastly, the results in figure 6-7 revealed that the HVOF application of colmonoy 88 had ten-fold

improvement in corrosion resistance over laser welded coating. The elemental composition

adjacent to the alloy surface does not provide enough information to distinguish why HVOF

performed significantly better than the laser cladding. The cross section measurement for the

laser colmonoy coating did identify one location under the chloride layer which has a more

94
significant penetration than under the HVOF coating. This was realized by the skewed average

penetration reported in table 6-2. Thermally sprayed coatings are mainly attacked through oxides

and voids at splat boundaries. The extent of the oxides at splat boundaries is controlled by

reducing the spraying temperature [114]. It is recommended based on the performance of the

HVOF coating in this chapter, to continue with the current spray conditions.

6.5. Conclusions

One of the common chlorine compounds present in WTE boilers are alkali chlorides.

These compounds are one of the major components in fly ash. As shown in this chapter, the

deposition of alkali chlorides on boiler tube surfaces readily engage in high temperature

corrosion mechanisms The sodium chloride induced corrosion occurs below the salt layer and

has a proclivity for depleting alloy surfaces of protective and semi-protective oxides. The extent

of corrosion is more aggressive that hydrogen chloride and is quantifiable for many alloys even

after 24 hours of simulated exposure. One of the ways to mitigate the corrosion induced by

sodium chloride, is to add elements such as tungsten and molybdenum which have been shown

in other corrosion literature to have reduce pitting corrosion. In the tests discussed in chapter 6,

tungsten was especially effective, as part of Colmonoy 88.

95
Chapter 7: The Effect of Sulfur on High Temperature
Corrosion

7.1. Summary

In this chapter, the effect of sulfur on WTE boiler corrosion is discussed. Sulfur is a

minor component of MSW, an order of magnitude less than chlorides. Industrial experience has

shown that the introduction of sulfur sources (e.g. coal, sulfur additives) into WTE boilers

replaces boiler tube deposits rich in chlorides with sulfates and subsequently corrodes at much

lower rates. These trends have prompted several research endeavors into understanding the

mechanisms and developing commercial approaches for its use.

Experiments have demonstrated that sulfates are semi-protective on WTE boiler tube

surfaces up to temperatures of 550°C. The corrosion rates for carbon steel and Fe-17Cr-13Ni

(NSSER-4) under sodium sulfate had an order of magnitude less corrosion than under sodium

chloride. The remainder of the chapter investigates the how the change of sulfur in the gas

composition, expressed using the SO2/HCl ratio, impacts the high temperature corrosion of

alloys. Typical levels of the SO2/HCl flue gas ratio in WTE boilers range between 0.1-0.3 [115].

As the ratio is increased above 0.3, the experiments have shown that the corrosion of several

candidate superheater tube alloys is reduced at temperatures as high as 500°C. The limitations on

this behavior and differences with industrial experience are discussed.

96
7.2 Introduction

MSW and biomass also have significantly lower quantities of sulfur when compared to

other fossil fuels such as utility coals. Table 7-1 presents typical ranges of sulfur content for

utility coals, various biomass feedstocks, and MSW. The sulfur content of MSW is generally

about an order of magnitude less than other sources. Table 2-1 in chapter 2 showed that the vast

majority of sulfur in MSW is found in textiles and rubbers.

Table 7-1. Sulfur content of various solid fuels [116,117].

Fuels Subcategory Sulfur (Percent)

Biomass poplar wood 0.01-0.03
switchgrass 0.07 - 0.11
yard wastes 0.1-0.3
Fossil Fuels coal (low rank; lignite/sub-
bituminous) 1.0 - 3.0
coal (high rank
bituminous/anthracite) 0.5 - 1.5
oil (typical distillate) 0.2 - 1.2
Municipal Solid Waste
0.05-0.1

In general, sulfur bounded in solid fuels is present in one of three forms, pyritic, organic,

and or sulfate [1]. Specifically regarding MSW combusted with excess air, the likely sulfur

compounds liberated into the boiler consist of oxides and sulfates. SO2 , sulfur dioxide, is

typically the largest form of sulfur present, but can oxidize to SO3- the thermodynamically

favored form below 700°C in the presence of abundant oxygen, as described by equation 7-1.

SO2 + 0.5 O2  SO3 (7-1)

97
Table 7-2 indicates the thermodynamic predominance of SO2 and SO3. This shows the Gibbs

free energy of the conversion of sulfur dioxide to sulfur trioxide. This reaction is

thermodynamically favorable at low temperatures up till about 800 °C. It is also well known that

the homogeneous oxidation of SO2 to SO3 is kinetically limited. It has been suggested that metal

oxides present in the boiler as well as in the WTE fly ash from MSW combustion can catalyze

this reaction [118,119]. Similar to chlorides, sulfates are typically found in boiler deposits, both

in the waterwall and in the superheater. The most common sulfates are either heavy metals such

as lead or zinc or alkali/alkali-earth metals such as sodium, potassium, and calcium.

The corrosion at high temperatures attributed to sulfur compounds has been well

investigated [120,121]. In addition to reacting with heat exchange surfaces, sulfur compounds

can react with chlorine compounds present in the process gas and in the boiler deposits,

influencing different corrosion reactions to occur. The following section is a description of the

industrial experience between sulfur and chlorine compound which provided the experimental

investigation in the rest of this chapter.

Table 7-2. Thermodynamic data of sulfur dioxide oxidation

SO2 (g) + 0.5 O2 = SO3 (g)

Temperature (°C) ΔG (kJ/mol)
300 -90
400 -71
500 -52
600 -33
700 -15
800 3
900 22
1000 40

98
7.3 Industrial Experience

In the 1970’s as part of work commissioned by the Solid Waste Management Office of

the US Environmental Protection Agency (US EPA), the Battelle Memorial Institute performed

tests to discern the effects of chlorine and sulfur compounds on WTE corrosion. The finding

was made that the addition of elemental sulfur with MSW trended with a substantial reduction in

boiler corrosion- as determined by corrosion probes [122]. Those results gave rise to a series of

tests to investigate the relationship how controlled additions of elemental sulfur, and later co-

firing MSW with high-sulfur coals, affected boiler corrosion.

Table 7-3 demonstrates the results obtained from the corrosion study of two metal tube

alloys (carbon steel A10 and stainless steel 316) following 1000 hours inside the boiler. The

alloys were a part of corrosion probes- consisting of alloys rings fixed to a hollowed tube to

allow for air flow. The air flow was designed to control the metal rings to an average temperature

within the ranges of operating superheater tubes. The corrosion rate was determined by the

thickness loss measured in the alloy rings after 1000 hours, and normalized to a reference

condition- which was the combustion of raw refuse. The rates were plotted as a function of the

SO2/HCl, which demonstrate the addition of sulfur into the furnace. These studies concluded that

at investigated temperature 370 °C and 593 °C--the addition of sulfur yielded a lower corrosion

rate. These results have served as the motivation for various approaches to mitigate chloride

corrosion that are discussed in chapter 3.

99
 Table 7-3. Corrosion rate of alloys measured in Battelle plant test (mils/hour) [123]

Carbon Steel SS310 (25 Cr 20 Ni)

SO2/HCl 370°C 482°C 593°C SO2/HCl 370°C 482°C 593°C
0.3 0.28 0.33 0.55 0.3 0.04 0.14 0.23
1.5 0.22 0.25 0.32 1.5 0.02 0.03 0.03

To understand the effect on the tube thickness, the deposition on the alloys rings was

analyzed via XRD and identified that smaller corrosion layers were observed under sulfate

deposits. The hypothesis is that the presence of sufficient SO2 in the furnace atmosphere, the

sulfation of metal chlorides would occur. A general expression of the sulfation reaction is shown

below in equation 7-2.

2MCl (s,g) + SO2 (g) + 0.5 O2 (g) + H2O(g)  M2SO4 (s) + 2HCl (7-2)

In the remainder of this chapter, laboratory experiments demonstrate the phenomena

between sulfur and chlorine compounds in WTE superheaters. The objective is to elucidate the

behavior of sulfur injection as a method for mitigating high temperature chloride corrosion.

Several factors are examined, including the effects of deposit, temperature, alloy scale with

respect to the SO2/HCl ratio.

7.4 The effect of the sulfate layer

Furnace tests were performed to distinguish the different corrosion behavior of alloys

under synthetic boiler tube deposits at superheater conditions. Pure compounds were used to

investigate the behavior related to compounds, groups of compounds commonly found in boiler

100
deposits. Sodium sulfate was applied on coupons of SA 178A (carbon steel) and NSSER-4 (17

Cr-13 Ni-Fe) at the same loading (4.0 mg/cm2 ±10%) which was tested in chapter 5 and 6.

Sodium sulfate was utilized in these experiments since sodium chloride was chosen as the

surrogate metal chloride compound, as sodium sulfate would be the expected compound if

sodium chloride were to undergo a sulfation reaction. The coupons were subjected to a typical

WTE gas mix with SO2/HCl =0.3 at a temperature of 500°C for 24 hours.

Figure 7-1 displays the mean mass loss of six coupons of SA178A and NSSER-4 under

following the furnace tests. The mean mass loss of NSSER-4 is shown magnified by a factor of

10 to be visible on the same axes as the carbon steel specimens. The mass loss of the sulfate

covered coupons had a mass loss similar or slightly higher than the coupons exposed to the gas

environment and was about an order of magnitude smaller than the corrosion rates obtained

under a 100% chloride deposit. Examples of NSSER-4 from these experiments are shown in

figure 7-2. A white residue is present on the coupons which were coated with sodium sulfate,

which are the unreacted sulfate. In contrast, the deposited sodium chloride is no longer present

on the coupon and a thicker mass change was observed. Given the changes in the corrosion rate,

and the observation of the surface, the sulfate layer appears to prevent further corrosion and

remain nearly intact on the surface.

The stability of the sodium sulfate layer was obtained by repeating the previous 24 hour

tests under the WTE gas over a wide for a range of temperatures. Figure 7-3 shows the mean

mass loss of NSSER-4 coated with sodium sulfate following 24 hour experiments at 450°C,

500°C, 550°C, and 700°C respectively. 700°C was utilized as an extreme temperature. The

results of NSSER-4 under the sulfate layer are plotted alongside the mass loss of NSSER-4 in the

absence of a salt deposit. Between 450 and 500°C the mass loss is nearly identical, and slightly

101
diverges at 550°C. It is at this temperature the sulfate layer becomes highly corrosive. At 700°C,

the sulfate is very aggressive and accelerates the mass loss by more than a factor of 4.

Figure 7-1. Mass Loss of SA178A and NSSER-4 in the presence of WTE flue gas and under
NaCl and Na2SO4 at 500°C after 24 hours

Figure 7-2. NSSER-4 coupons following 24 hour experiments at 500°C under sodium chloride
and under sodium sulfate deposits

102
Figure 7-3. Mass loss of NSSER-4 versus temperature (450-700°C) after 24 hour exposure to
WTE flue gas and flue gas and Na2SO4 deposits

At 550° C and above, the sulfate later no longer provides protection for the coupon surfaces. It

has been discussed in chapter 2 that sulfate induced high temperature corrosion is a factor at

these temperature due to the formation of pyrosulfates- which can flux oxide layers. [124].

7.5 Tests varying the SO2/HCl ratio

The relationship between the SO2/HCl ratio and corrosion of alloys was investigated for

several variables: the type of deposit (sulfate, chloride) compared to virgin material, water

concentration, temperature, and alloy type. These tests were designed to understand the potential

mitigation mechanism(s) that may be occur. Additionally, these tests can be added to the library

of how different boiler materials and operating conditions respond to changes in the

sulfur/chlorine ratio in the combustion feed.

103
 For all of the experiments discussed in the proceeding sections, the SO2/HCl ratio was

modified by adjusting the SO2 concentration relative to a constant HCl concentration of 800

ppmv. In doing so modification of the SO2/HCl ratio would replicate the field test discussed at

the beginning of this chapter. Unless otherwise specified (experiments in section 7.5.1), all

experiments discussed are 24 hour furnace tests- similar to those discussed in chapter 6.

7.5.1. Effect of different salt layers

In Figure 7-4 the SO2/HCl concentration is varied between 0.1-2.0 at 500° C for coupons

of SA178A. Ratios between 0.1-0.3 can be considered in the range of normal plant operation,

whereas higher than 0.3 can be considered sulfur addition. In these initial tests, no deposit (i.e.

no salt layer) was applied to the coupons. Similarly the tests were repeated for select ratios at the

shorter exposure time of 12.5 hours-to determine whether any variations were visible at smaller

scales. The data can be considered a flat line between the various flue gas ratios, which is to say

the average corrosion (mass loss) of the alloys was not impacted by varying the flue gas ratio.

Figure 7-4. Corrosion of SA178A versus SO2/HCl ratios 0.1-2.0; 500°C and 50 hours

104
 The data in figure 7-4 are consistent with thermodynamic predominance diagrams

provided in figure 7-5 and 7-6 respectively. These diagrams consider the thermodynamic

equilibrium for a metal in a mixed gas environment at a specified temperature. The diagrams

were constructed using HSC Chemistry software. Figure 7-6 corresponds to Fe-O-Cl at 500°C

and figure 7-7 corresponds to Fe-O-S at 500°C. As SA178A is a low carbon steel, iron is >99

wt% of its composition. Both diagrams indicate that the predominant compound found in the

corrosion layer is Fe2O3. Fe3O4 may be present at lower partial pressures of oxygen, between the

corrosion layer and the virgin alloy interface. It is also possible at very low partial pressures,

interfacing between Fe3O4 and the virgin alloy interface iron sulfides and/or iron (II) chloride.

Figure 7-5. Fe-O-Cl predominance diagram, 500°C

105
 Figure 7-6. Fe-O-S predominance diagram, 500°C

Based on these diagrams, there is no difference in product composition at higher

SO2/HCl. Even as SO2 is increased and the driving force for SO2 to diffuse to the oxide-metal

interface is also increased, the predominant products will remain the same as long as sufficient

oxygen is present. This analysis is confirmed by considering the XRD pattern of the scales

generated from SA178A at different SO2/HCl ratios and comparing?? Them?. This comparison

is shown in figure 7-7.

106
 Fe3O4

Figure7-7. XRD pattern of scale of T22 recovered after 50 hours

Figure 7-8 shows the average corrosion rate for NSSER-4 at various SO2/HCl ratios

under sodium sulfate and sodium chloride after 24 hours at 500°C. The corrosion rates of the

coupons under the sodium sulfate layer are about an order of magnitude smaller than the values

under the sodium chloride. The values are invariant with the gas composition. Comparatively the

corrosion under the sodium chloride shows a modest, yet discernible decrease at every ratio. The

oxide layer is 27% smaller at SO2/HCl = 2 than the layer measured at SO2/HCl = 0.1.

This result agrees with the findings provided in earlier plant studies that the SO2/HCl is

able to suppress the corrosion of alloys under chloride deposits. Furthermore the sulfate layer at

500° C, which has been shown to be relatively inert, maintains this behavior at 500°C as the SO2

concentration in the synthetic flue gas is varied between 100 ppmv to 1600 ppmv.

107
Figure 7-8. The corrosion rate of NSSER-4 vs SO2/HCl at 500 °C under chloride and sulfate
layers

Some literature had suggested sulfates in the presence of SO2 and O2 at temperatures of 500°C or

greater were susceptible to alkali tri-sulfates [125]. Under these conditions, the molten

compound can dissolve the protective oxide layer, further accelerating corrosion.

7.5.2 Effect of water

Figure 7-10 shows the corrosion rate of NSSER-4 under a sodium chloride layer in two

different gas environments. One data set is the typical gas environment, replotted from figure 7-9

(15% H2O). The other set is the same experiments performed in an anhydrous (0 vol% H2O)

environment. The purpose of the study was to determine whether the postulated sulfation

reaction of alkali chlorides proceeds as equation 7-3 or as equation 7-4 or both- and whether this

has an impact on alloy corrosion rate. Under the anhydrous experiments, the corrosion rate is

similar to the reference condition and remains stagnant the same value. Under the moist

environment, as mentioned previously the rate is suppressed with respect to an increase in the

108
SO2/HCl ratio. The findings from figure 7-9, coupled with literature suggest that conversion of

alkali chlorides follows equation 7-3, that is sodium forms a sulfate and releases HCl gas.

Figure 7-9. The corrosion rate of NSSER-4 under a NaCl layer vs. the SO2/HCl ratio at 500°C;
the effect of water vapor

2NaCl (s,g) + SO2 (g) + 0.5 O2 (g) + H2O(g)  Na2SO4 (s) + 2HCl (g) (7-3)

2NaCl (s,g) + SO2 (g) + O2 (g) +  Na2SO4 (s) + Cl2(g) (7-4)

Previous literature has investigated the sulfation reaction under anhydrous conditions at

temperatures between 100-650°C. Different products were formed on the NaCl surface at

different ranges of temperature and Cl2 was detected in the gas phase. Experiments in the

presence of water vapor yielded the expected products of Na2SO4 and HCl [126].

109
7.6 Effect of SO2/HCl ratio on various alloys under chloride salt layer

In chapter 6 an alloy comparison was made among multiple candidate superheater

materials and tube coatings. In this chapter the comparisons between those materials are

continued, focusing on the influence of the SO2/HCl ratio on the high temperature corrosion

behavior.

Figure 7-10 are initial results with the materials that were investigated in the previous

studies. These results demonstrate the relationship between the corrosion change and the

SO2/HCl are not the same for each alloy. A much steeper effect (an 80% decrease) was observed

for the carbon steel as compared to NSSER-4 (27%) over the entire range. Whereas the corrosion

rate continues to drop up to SO2/HCl = 2 for the carbon steel, the profile appears to plateau for

NSSER-4 after SO2/HCl = 1.4.

Figure 7-10. Corrosion rate of SA178A and NSSER-4 vs. SO2/HCl ratio at 500 °C

For the remainder of the comparative alloy study, three SO2/HCl ratios and three

temperatures were chosen. SO2/HCl ratios of 0.3, 0.6, and 1.0 were chosen as the median range

of those investigated in the preliminary tests. It was determined that within this range, a trend

110
would be discernable because of the industrial experience increasing the sulfur/chlorine ratio as

similar levels. The temperatures of 450°, 500°, and 550°C were chosen as a way of

complementing the analysis discussed in chapter six.

A plot of the mass loss per area for SA213 T22 (Fe-2.25Cr-1Mo) under a sodium

chloride layer versus the SO2/HCl ratio is shown in figure 7-11. At 450°C and at 500 °C (the

bottom 2 lines) positive trends are observed. The largest change occurs at 450°C between

SO2/HCl = 0.3 to 1.0, a 38% reduction. At 500°C, the rate drops to 28% within those same

conditions. At 550°C, no relationship was defined- as the change was within experimental error.

Figure 7-11. The effect of SO2/HCl on the mass loss of SA 213 T22 at 450°, 500°, 550° C.

Inconel 625 laser weld on T22 is shown in figure 7-12. As shown in chapter 6, Inconel

reduces the mass loss obtained from post-test analysis by nearly 50%. Similarly to T22, Inconel

exhibits a trend at 450°C between SO2/HCl = 0.3 to 1.0, a 38% reduction. At higher temperatures

111
however the trend is indistinguishable. The most different behavior of the materials investigated

belongs to HVOF applied Colmonoy 88 which did not demonstrate any trend. The rates for

Colmonoy88 are comparatively much lower than any of the materials investigated in this study-

but were still measurable within the sensitivity of the laboratory equipment.

Figure 7-12. The effect of SO2/HCl on the mass loss of Inconel 625 at 450°, 500°, 550° C.

Figure 7-13. The effect of SO2/HCl on the mass loss of Colmonoy 88 HVOF at 450°, 500°,
550° C.

112
 There are several insights that can be discussed from the tests of the different alloys at the

different temperatures. The first is that increasing SO2/HCl appears to that have greater impact

on alloys with higher corrosion rates. SA178A and T22 corrosion were the most sensitive to

changes to the ratio. NSSER-4 and Inconel 625 with moderately high corrosion had measurable,

yet smaller sensitivity to the ratio, whereas Colmonoy88 had none.

One possible explanation for the sensitivity could be attributed to the alloys surface and

whether sulfation was promoted by different metals. This was refuted by comparing the salt

surface of the Inconel 625 (figure 7-14) and stainless surface (figure 7-15) respectively as shown

with accompanying EDS analysis. The images reveal the salt layer present on the surface of the

coupon following the 24 hour experiment. For both alloys the EDS measurement indicates no

presence of chloride on the surface and a significant concentration of sulfur- suggesting that the

layer on the surface has completely reacted with the gas environment. Considering the alloying

content the trend cannot be attributed solely to one metal on the surface.

Another interesting observation is that the sodium chloride was not expected to

completely react to sodium sulfate under these tests after 24 hours. Previous literature on the

sulfation of solid sodium chloride at 500 C for an SO2 partial pressure of 0.0003 atm predicted a

conversion of 8% at 450 °C, 16% at 500 °C, and 19% at 550 °C [127]. Those experiments were

conducted in a flow through reactor with NaCl placed on a porous quartz glass support. In the

case of these experiments, either iron, chromium or nickel oxide supported the sodium chloride.

Another important observation is that although there were different corrosion trends with the

SO2/HCl at different temperatures, this was not impacted by the presence or lack of the sodium

sulfate on the surface. Figure 7-16 show the images of the salt layer on SW 1600, another Ni

coating, at each temperature. In all the cases, the surface chloride concentration was negligible

113
after 24 hours. That does not suggest however that the chlorides not participate in the alloy

corrosion.

Figure 7-14. Salt coverage and EDS values for Inconel 625 after 24 hour tests at SO2/HCl = 0.3

Figure 7-15. Salt coverage and EDS values for NSSER-4 after 24 hour tests at SO2/HCl = 0.3

114
Figure 7-16 (a). SEM- EDS values for SW1600 at SO2/HCl = 0.3, 450°C

Figure 7-16(b). SEM- EDS values for SW1600 at SO2/HCl = 0.3, 500°C

115
 Figure 7-16(c). SEM- EDS values for SW1600 at SO2/HCl = 0.3, 550°C

The images in figures 7-17 (a-c) are the cross section taken for a nickel overlay on T22,

after 24 hours furnace tests at 450 °C and 550°C respectively for SO2 /HCl = 0.3 The images are

couples with EDS mapping to show the presence of individual elements displayed in the image.

The substrate T22 is not shown in the images- only the coating, corrosion layer, salt layer and

resin-used to preserve the corrosion layer for cross section.

At each temperature the salt layer is visible, indicated by the presence of sodium. This is

most noticeable at 450 C, where the least corrosion has occurred. In the area marked by sodium,

sulfur and oxygen is co-present, reflecting the EDS data shown for the surface in figures 16 and

18 respectively. Below the deposit, the cross section demonstrates large fissures in the coating

which looks characteristically like the diffusion observed with chlorine. These fissures occur in

areas on the coating where chromium is present. At 450° and 500°C, chlorine is not present in

the fissures, but sulfur and oxygen. At 550°C however, chlorine is present where the largest

fissures are present, along with sulfur, oxygen, and chromium. The chlorine is detected below

116
sodium and below sulfur. The detection of chlorine at higher temperatures may suggest why the

SO2/HCl did not have an impact at 550 °C.

The images at 550° C, may also explain the discrepancy between the laboratory data and

field data. In section 7-3, it was reported that increases in the SO2/HCl were effective in reducing

corrosion in WTE power plants as high for boiler tubes as high as 593 °C. In the plant, sulfur is

co-fired with MSW allowing for the sulfation of chlorides in the gas phase, prior to condensing

and reacting with boiler tubes. In the case of the laboratory, the chlorides are initially present on

the alloy surface and can equally participate in corrosion or sulfation reactions. The data suggests

that the chloride-metal oxide reactions may be faster than the sulfation reaction at 550° C.

7.7 Conclusions

Experiments have shown that sulfates are semi-protective on WTE boiler tube surfaces

up to temperatures of 550°C. The mass loss for carbon steel and Fe-17Cr-13Ni (NSSER-4) under

sodium sulfate had an order of magnitude less corrosion than under sodium chloride. Using this

behavior, the SO2/HCl ratio in the synthetic flue gas was increased from 0.3 to 0.6 and 1.0

respectively to determine the impact on sodium chloride corrosion. The surface of coupons was

readily converted from chlorine to sulfur demonstrating the effect of SO2. As the ratio is

increased above 0.3, the experiments have shown that the corrosion of several candidate

superheater tube alloys is reduced at temperatures as high as 500°C. The impact of sulfating the

alloy was observed most prominently with alloys that were aggressively corroded by the chloride

layer. At 550 °C, there was no trend with respect to increases of the ratio, which suggest that

other corrosion reactions, faster than the sulfation reaction, govern the corrosion rate.

117
 Figure 7-17(a). SEM and EDS Mapping for SW1600 at 450°C , SO2/HCl = 0.3

Notes: The elemental fingerprint of sodium and sulfur are in the same locations in this cross
section maps. The presence of chlorine was not detected. Either it was not present at all (total
conversion of chloride to sulfate) or chlorine is present at such a low amount it was in the noise
of the EDS data.

118
 Figure 7-17(b). SEM and EDS Mapping for SW1600 at 500°C , SO2/HCl = 0.3

Notes: Similar to the 450°C map, sodium and sulfur are the present on surface with no Cl
present. The oxidize layer under sulfate appears rich in chromium.

119
 Figure 7-17(c). SEM and EDS Mapping for SW1600 at 550°C , SO2/HCl = 0.3

Notes: There is considerable sulfur present, but chlorine is also present below the sulfur and
below the sodium and it is residing with oxygen and chromium

120
Chapter 8: Economic Impacts of Corrosion Mitigation

8.1 Introduction

This chapter discusses some of the economic impact that corrosion mitigation can have

on a WTE plant. The annualized cost factor is evaluated for the alloys investigated in chapter 6

against a benchmark superheater tube material, SA 213 T22. The analysis is used to identify

relative installation costs at which certain materials are feasible. Additional impacts discussed in

this chapter include savings from reduced downtime and failure and the price of corrosion

additives.

8.2 Annualized cost factor

Generally defined, the annual cost is defined as the payment of owning and operating an

asset over its entire lifetime. The equivalent annual cost (EAC) is calculated by dividing the net

present value of the asset (i.e. equipment) over the annuity factor. This type of calculation

requires information on installation costs, tax rate, acceptable rate of return, depreciation, and the

estimate lifetime of the equipment [128].

Given the competitive nature of the power and WTE industry the information necessary

to calculate the equivalent annual cost, specifically the installation costs, are not readily

available. Instead, an less accurate but convenient method can be applied to distinguish the

economic feasibility of using one type of material in a superheater over another by calculating

the annualized cost factor- which is presented in equation 8-1 below, which is the ratio of a cost

121
factor and life factors for a given a material. An annualized cost factor less than 1 would make

the material an attractive option.

Annualized cost factor = cost factor/life factor (8-1)

The cost factor considers the purchase price of a material and labor cost for installation.

Some examples of typical cost factors are shown below. The cost factors are relative to a

benchmark material, which in the case of superheater boiler tubes is SA 213 T22--which was

been heavily investigated in this work. For example, a cost factor of 2.5 for TP 310H means that

the cost and installation of a superheater with that stainless steel will be approximately 2.5 times

more expensive. Examples of cost factors are shown in Table 8-1 for common utility boiler

alloys TP 310H and Inconel 625 [142].

Table 8-1. Examples of cost factor for SH tube alloys [129]

Cost Factors (Relative to SA 213 T22)

TP 310H 2.5
Inconel 625 2.4

There are additional factor that can be weighed into the cost factor: For example: the cost of a

heavy wall superheater is approximately 10% more costly than a normal boiler. In the case of

coatings, the cost of laser cladding is approximately 1.6 times the cost of thermal spray [130].

The life factor is defined as the ratio of the wastage rate of a material with the wastage

rate of a baseline or benchmark material. Once again, the benchmark material for this evaluation

122
is SA 213 T22, which also happens to have the highest wastage rate of any material tested in this

thesis (equivalent to SA178A).

Using the mass loss data from chapter 6, the life factor is estimated for the coatings

investigated in chapter 6 at the lowest and highest temperatures investigated 450°C (842°F) and

550°C (1022°F) and is displayed in Table 8-2. The highest factors belong to HVOF colmonoy

88, which had superior corrosion resistance. Although the cost factors are not known for all of

these materials, the information is available for Inconel 625, in table 8-2, and has a CF= 2.4.

Using the most conservative life factor (i.e. the life factor at 550°C), the annualized cost factor of

Inconel 625 is 1.2. As that value is greater than 1, Inconel 625 is not a desirable material in the

environment investigated in chapter 6.

Conservatively, using the value of the laser cladding instead of the HVOF coating, it can

also be shown that the installation of a colmonoy 88 protected superheater can be as much as 4.3

times the cost of a T22 superheater tube and be feasible, or approximately 1.4 times the cost of

an Inconel 625 superheater replacement.

Table 8-2. Life factor calculated for Ni coatings based on mass loss data

Life Factor (Relative to SA 213 T22)

Coating 450°C (842°F) 550°C (1022°F)
Inconel 625 1.7 2
Colmonoy 88 (HVOF) 110 21
Colmonoy 88 (laser) 7.7 4.3
SP 99 6.1 3.5
SW1600 3.1 1.2
SW1641 1.0 1.1

123
8.3 Savings from extended tube lifetimes

An estimation of the costs associated with a given forced outage was published in the

ASME CRTD Vol 38 report for Waste Incineration plants. The calculation can be using some

simplifying assumptions, which are listed below [131].

The cost of a forced outage

The cost of forced outage has several factors. The minimum time required to repair a failed tube,

provided that the tube location is accessible and the replacement piece is a stock item, is 12 to 16

hours. This time assumes 4 to 6 hours for the boiler to cool down, 4 hours to make the repair, and

4 to 6 hours for hydro-testing the boiler and heat up.

Cost of repairing the failure

Two boilermakers are typically used at an hourly rate each (with overtime) x tb hours on site per

outage + cost of materials. The boilermakers are required to be on-site to make preparations, to

carry out the repair, and to be available through the hydrotest. It is assumed that the normal

operating staff provide the labor necessary for providing access to the failure and performing the

boiler operating functions, so that their time is not charged to the outage

Buying power, replacement power

The cost of replacement power is calculated as the cost of natural gas required to augment the

firing of the remaining boilers to maintain output.

124
 This is expressed as: steam flow to be replaced (lb/hr) x enthalpy for operating steam

temperature (MBtu/lb) x cost of natural gas (SC/MBtu) x duration of outage (hr)]/[utilization

efficiency (e)].

Based on the model used by Joyner and Lai [132], it can be estimated that a WTE plant

spends about $170,000 per failure. Using the life factor in table 8-2, the potential savings by

replacing SA213 T22 with other materials is shown in figure 8-1 for a three-year period. For

example, since the corrosion resistance of Inconel 625 is a factor of 2 better than the resistance of

SA 213 T22, it assumed that the frequency between failures will also double. Based on these

assumptions it can be shown that installation of a colmonoy 88 coated boiler can save about

$130,000 year in avoided costs

8.4 Corrosion Additives

Another consideration is the use of corrosion additives to mitigate corrosion- such as the

addition of a sulfur containing chemicals, to increase the SO2/HCl ratio. There are a few

technologies that are operating at commercial scale, such as the ChlorOut TM ammonium sulfate

based additive by Vattenall or the sulfur recirculation technology by Gotaverken Mijo The

commercial scale for Chlorout is predominantly for biomass fired plants, 50 MWe or smaller,

which have a higher alkali content in their feed and have significant alkali chloride corrosion

issues.

125
Figure 8-1. Estimated avoided costs by replacing SA 213 T22 with nickel based boiler coatings
(three year calculation)

For this exercise, the optimistic assumption is made that a new additive cannot exceed the

annual plant cost toward on corrosion maintenance. Table 8-3 shows the corrosion maintenance

cost of several WTE power plants from a WTERT survey in 2004, corrected for inflation in 2016

dollars. Normalizing the cost with the tonnage of waste burned annually (tonnage of waste/day x

days of annual operation), the value is estimated (in 2016 dollars) at $1.68 per ton MSW

combusted. It should also be considered that the cost for the extra Ca(OH)2 needed to remove

the excess SO2 would be $0.21/ton combusted. This assumes that the sulfur dosage would

increase the SO2 concentration in the boiler by a factor of 3 and that the cost of lime is

$30/metric ton. Correcting for the cost of the additional sulfur removal and the acceptable cost of

the corrosion additive would need to not exceed $1.47 ton of MSW combusted. Depreciation

cost of any additional infrastructure and the variable cost of the additive should have to compete

with this value.

126
Table 8-3. Cost of maintenance reported by 7 WTE power plants 2004 WTERT Survey on
Boiler Corrosion [145]

Surveyed WTE Plants (2004 WTERT Survey)

A B C D E F G
corrosion
maintenance
($/year) 50000 100000 18000 145000 253000 320000 99120
Capacity (ton
MSW day) 255 250 242 460 608 250 264
Days of operation
(per year) 330 346 327 332 306 328 344
Cost/ton
combusted ($/ton) 0.59 1.16 0.23 0.95 1.36 3.90 1.09
Average $ 1.33

8.5 Conclusions

This chapter discussed some of the economic impacts from the corrosion testing

performed in previous chapters. The annualized cost factor was defined and discussed as a way

of justifying replacement of expensive, yet more corrosion resistant alloys. The corrosion

resistance of coating colmonoy 88 and SP 99 can be attractive options if costs are not

prohibitive. Based on its relative corrosion performance colmonoy 88 can afford an installation

cost as much as 4.3 times the cost of a T22 superheater tube or approximately 1.4 times the cost

of an Inconel 625 superheater replacement, and be economically feasible.

127
Chapter 9: Conclusions and Suggestions for Future
Work

9.1 Conclusions

Waste-to-energy is the environmentally preferred approach for managing post-recycled

wastes. One of the pressing issues related to waste-to-energy technology is the problem of high

temperature corrosion in boilers. The combustion of municipal solid wastes produces a highly

corrosive environment which can react with, and deteriorate, heat transfer surfaces. Chlorine,

which is readily present in MSW (0.47-0.72 wt%), has been identified as the important corrosive

species.

The experimental work discussed in this thesis investigated the corrosion behavior of

chlorine in waste-to-energy boilers, with specific interest on superheater tubes. The goal of this

research was to provide solutions which can enable WTE operators to reduce frequency of

corrosion related shutdowns and operate at higher steam temperatures, thus increasing electricity

production and overall efficiency. This was achieved through furnace experiments designed to

replicate the fireside corrosion environment of WTE boilers, with a synthesized flue gas and

chloride deposits at temperatures between 450-550°C (842-1022°F). The behavior was

investigated for three commercial steels: carbon steel (Fe-0.1 C), low alloyed steel (SA 213

T22), an austenitic steel (Fe-17.1Cr-13Ni), and 6 different nickel coatings applied on SA 213

T22.

The introduction of 800 ppmv of hydrogen chloride in flue gas increased the parabolic

rate constant (0-72 h) of SA178A from 0.18 to 1.7 μm2/h. This was caused by the active

oxidation mechanism- in which the metal forms non-protective, less adherent oxide. This

128
mechanism was inferred through a combination of physical observation, product characterization

via XRD and EDS, and agreement with thermodynamic calculations. The active oxidation

affected by hydrogen chloride can be mitigated by replacing iron with chromium and nickel, as

was demonstrated with NSSER-4 (Fe-17Cr-13Ni), which had an apparent activation energy (53

kJ/mol), that was half the apparent activation energy of the carbon and low alloyed steels (110-

111 kJ/mol).

Another important chlorine compound in WTE boilers are chloride salts, which are

readily contained in fly ash. Using sodium chloride as a surrogate compound, the corrosion

behavior under chloride salts was investigated by applying a salt layer (4.0 mg/cm2) on coupon

surfaces. Corrosion under the chloride layer was much more severe than under the HCl-

containing atmospheres alone. The mass loss for the commercial steels was accelerated by more

than an order of magnitude. Based on SEM and XRD coupon and corrosion product

characterization, the acceleration of the mass loss was inferred to be caused by another active

oxidation mechanism in which the sodium chloride consumes protective or semi-protective

oxides. The accelerated corrosion rates were realizable even after 24 hour furnace tests. This

observation led to the development of an furnace test which could be used to rank new alloys.

The furnace tests were executed for six different nickel coatings, including Inconel 625

(Ni-Cr-Mo), SW1600, SW1641 (Ni-Cr-Mo-B-Si) and Colmonoy 88 and SP 99 (Ni-Cr-B-W).

Colmonoy 88, showed superior corrosion resistance to the chloride environment with mass

losses between 0.3-3.1 mg/cm2 between 450-550°C. The mass loss of the Ni-Cr-Mo, and Ni-Cr-

B-So were shown to be significantly higher between 10-30 mg/cm2 The improved corrosion

resistance was attributed to the alloying addition of tungsten, and to a lesser extent molybdenum,

which has been shown in other literature to improve the corrosion resistance of pitting corrosion.

129
 Experiments have shown that sulfates are semi-protective on WTE boiler tube surfaces

up to temperatures of 550°C. The mass loss for carbon steel and Fe-17Cr-13Ni (NSSER-4) under

sodium sulfate was an order of magnitude lower than corrosion under sodium chloride. On the

basis of these results, the SO2/HCl ratio in the synthetic flue gas was increased from 0.3 to 0.6

and 1.0 respectively to determine the impact on sodium chloride corrosion. The surface of

coupons was readily converted from chlorine to sulfur demonstrating the effect of SO2. As the

ratio is increased above 0.3, the experiments showed that the corrosion of several candidate

superheater tube alloys is reduced at temperatures as high as 500°C. The impact of sulfating the

alloy was observed most prominently with alloys that were aggressively corroded by the chloride

layer. At 550 °C, there was no trend with respect to increases of the ratio, which suggests that

other corrosion reactions were faster than the rate of sulfation.

9.2 Recommendations for the WTE industry

Based on the results of this research, several recommendations can be made for the WTE

industry. The first is that colmonoy 88 and SP 99 should be investigated for longer term testing

and/or field tests based on their corrosion resistance. It would especially be of interest to

determine whether the HVOF application of Colmonoy 88 performs can withstand the erosion

and thermal fluctuations that were not simulated in this work. The second recommendation is for

WTE plants to explore the installation of sacrificial surfaces – similar to the concept of the

chloride trap discussed in chapter 3. The chloride trap is a section of screen tubes, which are

maintained in the temperature range of 300-400°C. Low melting ash-typically composed of

chlorides-are designed to condense on the walls of the empty pass. The removal of metal

chloride deposits would not eliminate the high temperature corrosion attack of HCl but, as shown

130
in chapter 6, the removal of metal chlorides can reduce the wastage rate of superheater alloys by

nearly an order of magnitude. Finally, the third recommendation would be to investigate sulfur

additive technologies that would have dual purpose- similar to the “Chlorout” additive discussed

in chapter 3, which is used for corrosion mitigation and for NOx reduction.

9.3 Suggestions for future work

Better understanding of the chlorine cycle (Chapter 5)

Although it is obvious that HCl accelerates the high temperature corrosion of metals, the

mechanism is not entirely clear- in particular the diffusion of HCl and metal chlorides. HCl-

induced corrosion could be analyzed in more detail with some advanced techniques such as high

temperature mass spectroscopy to analyze the different volatile species and quantify their impact

on the corrosion attack [132]. More detailed post-exposure analyses such as GDOES, SEM and

TEM may be also helpful in understanding the interaction between chlorine and the various

alloying elements and how this affects the scale growth.

Behavior of other chloride deposits (Chapter 6)

Laboratory studies investigating alloy performance in the presence of molten and solid

chlorides are of great interest. In particular, a comparison can be made on the high temperature

corrosion behavior of individual chloride salts such as potassium and calcium chloride with the

results of sodium chloride. Furthermore, a test can be developed in which alloys are exposed to

a synthetic fly ash mixture that is designed to form a melt at test temperatures. This would allow

for short (~24 hour) duration experiments to rank alloys that have been assumed to have

excellent corrosion resistance. A mixture of 40 wt% K2SO4, 40 wt% Na2SO4, 10 wt% KCl, and

131
10 wt% NaCl would be a possibility as well as mixtures containing heavy metal chlorides such

as zinc and lead chlorides, to simulate corrosion for lower temperature superheater tube alloys.

Comparison of metal coating methods (Chapter 6)

This thesis compared the performance of thermal spray and laser cladding of colmonoy

88 on SA 213 T22, and a discernible difference was observed. It would be valuable to compare

to the difference between thermally spray and laser cladding coating methods for multiple alloys.

This would help optimize both the alloy selection and parameters of the application process. In

general HVOF is a cheaper than laser cladding and has a lower heat input. But laser cladding

layers have a thicker build up and are considered more robust [133].

The effect of alloying additions of Cr +W + Mo on pitting corrosion resistance

It would be of interested to quantify whether there is a relationship that can be expressed

between the corrosion performance under chloride salts and the alloying composition pf nickel-

chromium alloys. An empirical value called PREN, Pitting resistance equivalent number, has

been developed for pitting corrosion for austenitic steels and nickel alloys in aqueous chloride

environment. PREN did fit well with data acquired in chapter 6, but its relevance to other

applications is an impetus for a further study. The design criteria for those experiments would be

the change in mechanical properties a result of changing the alloy composition.

132
 References

1. Environmental Protection Agency: Advancing Sustainable Management, Assessing Trends in

Material Generation, Recycling and Disposal in the United States (June 2015)

2. D. Shin (2014) Generation and Disposition of Municipal Solid Waste (MSW) in the United States
– National Survey (MS Thesis) Retrieved from Columbia University libraries

3. N. B. Klinghoffer, and Castaldi, M.J. (2013) Waste to energy (WTE): an introduction. Waste to
Energy Conversion Technology (pp. 3-9) Philadelphia, PA: Woodhead Publishing Limited.

4. N. J. Themelis, Y.H. Kim, and M.H. Brady, Energy recovery from New York City municipal
solid wastes. Waste Management & Research, 2002. 20(3): p. 223‐233.

5. C. S. Psomopoulos, A. Bourka, and N.J. Themelis, “Waste-to-energy: A review of the status and
benefits in USA”, Waste Management 29 (2009) 1718–1724.

6. N.J. Themelis, N.J., Y.H. Kim, and M.H. Brady, Energy recovery from New York City
municipal solid wastes. Waste Management & Research, 2002. 20(3): p. 223‐233.

7. Integrated Waste Services Association web site, http://www.wte.org/

8. P. Simmons, N. Goldstein, S. Kaufman, N.J. Themelis, J. Thompson, The State of Garbage,

BioCycle, 1 (2006) 26-43.

9. N.J. Themelis, P.A. Ulloa, Methane generation in landfills, Renewable Energy, 32 (2007) 1243-
1257.

10. E. Matthews and N.J. Themelis, Potential for reducing global methane emissions from landfills,
2000-2030 in 11th International Waste Management and Landfill Symposium, Sardinia, 2007.

11. Bogner et al., Mitigation of global greenhouse gas emissions from waste: conclusions and
strategies from the Intergovernmental Panel on Climate Change, 4th assessment, Waste
Management and Research, 2008, 26: 11-32, wmr.sagepub.com/cgi/reprint/26/1/11.

12. American Standard of Testing Methods (ASTM) D866 testing method

13. Energy Information Administration: Energy Explained: Waste-to-energy (MSW)

http://www.eia.gov/Energyexplained/?page=biomass_waste_to_energy#tab2

133
14. C. Ducharme, Analysis of a Thermal Plasma Assisted Waste-to-Energy Processes, 18th North
American Waste To Energy Conference (NAWTEC 18) 2010. Florida, USA.

15. S. Goff. (2011). Energy from Waste Retrieved from http://www.seas.columbia.edu/

earth/wtert/sofos/homepage

16. L. M. Grillo (2013) Municipal Solid Waste Combustion Plants Waste to Energy Conversion
Technology (pp. 3-9) Philadelphia, PA: Woodhead Publishing Limited.

17. J. Gittinger and W. Arvan (1998), Considerations for the design of RDF-fired refuse boilers,
presented to Power-gen Europe 1998, Milan.

18. C. Velzy and L. Grillo, (2007a), Fuels and Furnace, in Avallone, E, Baumeister, T, and Sadegh,
A, Marks’ Standard Handbook for Mechanical Engineers, 11th edition, New York, McGraw
Hill, pp. 7-48 – 7-53.

19. C. Velzy and L. Grillo, (2007b), Waste to energy combustion, in Kreith F and Goswami, D,
Handbook of Energy Efficiency and Renewable Energy, Boca Raton, CRC Press, pp 24-1 – 24-
42.

20. F. Rinaldi and B. Najafi, “Temperature Measurement in WTE Boilers Using Suction
Pyrometers” Sensors 2013, 13, 15633-15655.

21. Albina, D.O., K. Millrath, and N.J. Themelis. Effects of Feed Compositions on Boiler Corrosion
in Waste‐to‐Energy Plants. in 12th North American Waste To Energy Conference (NAWTEC
12). 2003. Florida, USA.

22. P. R. Roberge, P.R., Handbook of Corrosion Engineering. 2000: Mc Graw Hill.

23. D.N. French, Metallurgical Failures in Fossil Fired Boilers, 2nd ed., John Wiley & Sons Inc.

24. S.H. Lee., N. J. Themelis, and M. Castaldi, High‐Temperature Corrosion in Waste‐ to‐Energy
Boilers. Journal of Thermal Spray Technology, 2007. 16(1): p. 104‐110.

25. N.J., Themelis, Chlorine Balance in a Waste‐To‐Energy Facility. 2005, Earth Engineering
Center.

26. A.S. Khanna (2002), High Temperature Oxidation and Corrosion, ASM International

134
27. Jones, D.A., Principles and Prevention of Corrosion, 2nd Edition, Upper Saddle River, NJ
Prentice Hall 1996

28. P. Koftsad (1988), High Temperature Corrosion, Elsevier Applied Science, London

29. P. Viklund, (2011), PhD Thesis, KTH

30. C. Wagner, Journal of the Electrochemical Society 103, p.627-633 (1956)

31. R. Roberge, P.R., Handbook of Corrosion Engineering. 2000: McGraw Hill.

32. A.S. Khanna (2002), High Temperature Oxidation and Corrosion, ASM International

33. C. Wagner, Corrosion Science 9, p.91-109 (1969).

34. P. Viklund, “High Temperature Corrosion during Mass Incineration, “ (2011), PhD Thesis,
KTH

35. P. Rademakers, W. Hesseling and J van de Wetering, “Review on Corrosion in Waste

Incinerators, and possible effects of bromine.” TNO Report, 2002

36. S.H. Lee., N. J. Themelis, and M. Castaldi, High‐Temperature Corrosion in Waste‐ to‐Energy
Boilers. Journal of Thermal Spray Technology, 2007. 16(1): p. 104‐110.

37. J. Pettersson, C. Pettersson, N. Folkeson, L.-G. Johansson, E. Skog, J.-E. Svensson, Materials
Science Forum 522-523, p.563-570 (2006)

38. Y. Kawahara, Corrosion Science 44, p.223-245 (2002).

39. H. J. Grabke, E. Reese and M. Spiegel, Corrosion Science 37, p.1023-1043 (1995).

40. A. Zahs., Spiegel, M. and Grabke J.H, “ Chloridation and oxidation of iron, chromium, nickel
and their alloys in chloridizing and oxidizing atmospheres at 400-700°C, Corrosion Science
(42), 1093-1122, 2000

41. D.O Albina (2005), MS Thesis, Columbia University.

42. C. Pettersson, J. Pettersson, H. Asteman, J.-E. Svensson, L.-G. Johansson, Corrosion Science
48, p.1368-1378 (2006).

135
43. H. J. Grabke, Materials and Corrosion, 54, p.736-746 (2003).

44. P. Henderson, P. Szakalos, R. Pettersson, C. Andersson and J. Högberg, Materials and

Corrosion 57, p.128-134 (2006)

45. E. M. Levin, C. R. Robbing, H. F. McMurdie, Phase diagrams for ceramists, The American
Ceramic Society Inc. (1969)

46. S. C. Cha, M. Spiegel, Materials and Corrosion 57, p.159-164 (2006)

47. A. Zahs., Spiegel, M. and Grabke J.H, “ Chloridation and oxidation of iron, chromium, nickel
and their alloys in chloridizing and oxidizing atmospheres at 400-700°C, Corrosion Science
(42), 1093-1122, 2000

48. G.Y. Lai “Corrosion Mechanism and Alloy Performance in Waste-to-Energy Boiler
Combustion Environment”, 12th North American Waste-to-Energy Conference, Georgia USA,
2004

49. Y. Kawahara,., High temperature corrosion mechanisms and effect of alloying elements for
materials used in waste incineration environment. Corrosion Science, 2002. 44(2): p. 223.

50. J. Pettersson, C. Pettersson, N. Folkeson, L.-G. Johansson, E. Skog, J.-E. Svensson, Materials
Science Forum 522-523, p.563-570 (2006)

51. Babcock & Wilcox, Steam: its generation and use 40th edition

52. Adams, B., et al. Seghers Boiler Prism: a Proven Primary Measure against High Temperature
Boiler Corrosion. in 12th North American Waste to Energy Conference. 2004. Savannah,
U.S.A.

53. D.A. Jones, Principles and Prevention of Corrosion, 2nd Edition, Upper Saddle River, NJ
Prentice Hall 1996

54. G. Wright, Metals Handbook, Vol 13. Corrosion, 9th edition, ASM International, Metals Park,
OH, p. 97, 1987.

55. E. Hanson and M. Turner, “Second year Comparison of Superheater Metal Wastage Rates
Utilizing Various Boiler Tube Alloys in a Waste-to-Energy Facility”, Proceedings of the 11
Annual North American Waste to Energy Conference (NAWTEC 11), 2003.

136
56. M.A. Uusitalo, P.M.J. Vuoristo, Chlorine corrosion of thermally sprayed coatings at elevated
temperatures, in: E. Lugsheider (Ed.), International Thermal Spray Conference, Essen,
Germany, 4–6 March 2002, DVS, pp. 429–434.

57. G.Y Lai, High Temperature Corrosion of Engineering Alloys, ASM International, Materials
Park, OH, 1990.

58. Bu-Qian Wang, Erosion-corrosion of coatings by biomass-fried boiler fly ash, Wear, 1995, 188
p.40-48.

59. R. Smith, Chemical Process: Design and Integration, Wiley and Sons.

60. Y. Kawahara, M. Kira, Corrosion prevention of waterwall tube by field metal spraying in
municipal waste incineration plants.

61. M.A. Uusitalo, Corrosion Science 46 (2004).

62. M. Spiegel, R. Warnecke, Performance of thermal spraycoatings under waste incineration

conditions, in: Corrosion 2001, Houston, TX, Nace, paper 182.

63. L.N. Moskowitz, Journal of Thermal Spray Technology 2 (1993) 21.

64. K. Dopler, H. Kreye, R. Schwetzke, Journal of Thermal Spray Technology 9 (2000) 407.

65. S. Sharp, Sharp Consulting, “Energy from Biomass- Lessons Learned from European Boilers”

2011 PEERS Conference

66. F.J. Perez, J. Nieto, J.A. Trilleros, Materials Science Forum 522-523 (2006) 531-538.

67. E. Otero, A. Pardo, F.J. Perez, Oxidation of Metals, 49 (5/6) (1998) 467-484.

68. Bu-Qian Wang, Wear, 1995, 188, p 40-48.

69. Y. Kawahara, M. Kira, Corrosion, 1997, 53(3), p 241-251.

70. V. Fantini, Laser Cladding: a new technology for corrosion and erosion of boiler tubes, global
solutions, Proceedings of the International Thermal Spray Conference (2007), 1120-1124.

71. J.M. Guilemany. M. Torrell, Journal of Thermal Spray Technology, 2008, 17(2), p 254-262

72. S.C. Cha, H.W. Gudenau, GT Bayer, Materials and Corrosion, 53 (2002) 195-205.

73. Bu-Qian Wang, Wear 1999 (1996) 268-274.

137
74. Ragnar Warnecke, Bernd Benker, Christian Deuerling, Ferdinand Haider, Siegried Horn,
Jürgen Maguhn, Volker Müller, Hermann Nordsieck, Barbara Waldmann and Ralf
Zimmermann, “The mechanisms of corrosion and how to avoid them”, presented at 5th ISWA
Beacon conference on Waste-to-Energy. October 25-26, 2007, Malmö, Sweden.

75. Sharp, Sharp Consulting, “Energy from Biomass- Lessons Learned from European Boilers.”

76. E. Bjorkman, Energy & Fuels 1997, 11, 1026-1032

77. A. Jones, Principles and Prevention of Corrosion, 2nd Edition, Upper Saddle River, NJ Prentice
Hall 1996

78. D.O Albina (2005), MS Thesis, Columbia University

79. Vattefall, “ChlorOut Concept”, http://chlorout.vattenfall.com/en/our-concept.htm.

80. S. Andersson, J. Froitzheim, E. Larsson and J.Pettersson, “Sulphur recirculation for increased
electricity production in MSWI” Proceedings: International Symposium on Energy from
Biomass and Waste Venice, Italy; 8-11 November 2010.

81. Wark, K., C.F. Warner, and W.T. Davis, Air Pollution: Its Origin and Control 3ed.
1997:Prentice Hall.

82. S. Goff, Dry scrubbing operation in WTEs Corrosion Research, J.T. Nickolas, Editor. 2008:
New Jersey.

83. S.H. Lee, PhD Thesis (2009), Columbia University

84. Babcock & Wilcox, Steam: its generation and use 40th edition

85. S.H. Lee., N. J. Themelis, and M. Castaldi, High‐Temperature Corrosion in Waste‐ to‐Energy
Boilers. Journal of Thermal Spray Technology, 2007. 16(1): p. 104‐110.

86. MatWeb Website: Colmonoy

87. MatWeb Website: SP 99

88. P. J. Goodhew, J. Humphrey, R. Beanland, Electron Microscopy and Analysis 3rd Ed., Taylor
and Francis, London (2001

89. A.S. Khanna (2002), High Temperature Oxidation and Corrosion, ASM International.

90. Hitachi: http://www.microscopy.ou.edu/hitachi-3000.html

91. http://www.charfac.umn.edu/instruments/eds_on_sem_primer.pdf.

138
92. D. Brune, R. Hellborg, H. J. Whitlow, O. Hunderi, Surface Characterization: A User’s Guide.

93. G. Sorrel, Materials at High Temperature, 14. No. 3. 1997. 137-150.

94. AEBOM: Wood fuels handbook

95. Y. Yudovich , M.P. Ketris International Journal of Coal Geology 67 (2006) 127–144.

96. A. Zahs, M. Spiegel, H. J. Grabke, Corrosion Science 42, p.1093-1122 (2000)

97. D. Young, (2008) High Temperature Oxidation and Corrosion of Metals, Elsiever.

98. Viklund et al. Oxidation of Metals August 2011, Volume 76, Issue 1, pp 111-112.

99. M. Spiegel, Materials and Corrosion 50, p.373-393 (1999).

100. N. Otsuka, Corrosion Science 50, p.1627-1636 (2008).

101. 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.

102. K. Salmenoja, K. Mäkelä, “Prevention of Superheater Corrosion in the Combustion of

Biofuels”, Paper No. 238. Corrosion 2000 Orlando, USA (2000).

103. N. Folkeson, J. Pettersson, C. Pettersson, L.-G. Johansson, E. Skog, B.-Å. Andersson, S.

Enestam, J. Tuiremo, A. Jonasson, B. Heikne, J.-E. Svensson, Materials Science Forum, 595-
598, p.289-297 (2008).

104. Zahs, M. Spiegel, H.J. Grabke, Corrosion Science 42 (2000) 1093.

105. R. Pettersson, L. Liu, J. Sund, Corrosion Engineering, Science and Technology 40 (3), p. 211-
216 (2005).

106. H. J. Grabke, E. Reese and M. Spiegel, Corrosion Science 37,(1995) p.1023-1043.

107. CRC Handbook of Chemistry.

108. J. S. Kim, P. J. Xiang, and K. Y. Kim, Corrosion, February 2005, Vol. 61, No. 2, pp. 174-
183.

109. H.J. Park and H.W. Lee, Int. J. Electrochem. Sci., 9 (2014) 6687 – 6698.

110. R.A. Antunes and M. C.L. de Oliveira , Corrosion Science 76 (2013) 6–26.

139
111. M.A. Uusitalo et al. Corrosion Science 46 (2004) 1311–1331.

112. https://web.anl.gov/PCS/acsfuel/preprint%20archive/Files/20_2_PHILADELPHIA_04-
75_0094.pdf.

113. S. Agblevor, Energy & Fuels 1995, 9, 635-640.

114. Classification and comparison of municipal solid waste based on thermochemical

characteristics Journal of the Air & Waste Management Association Volume 64, Issue 5, 2014.

115. David J. Bayless a , Ashikur R. Khan a , Srinivas Tanneer a & Rajkumar Birru, “An
Alternative to Additional SO3 Injection for Fly Ash Conditioning”, Journal of the Air & Waste
Management Association, 50:2, 169-174.

116. Catalytic oxidation of SO2 to SO3 in the presence of fly ash US Patent 5011516S.

117. Bryers, R. Flagging, Slagging, Fouling, and high Temperature Corrosion of Heat Transfer
Surfaces due to Impurities in Steam Raising Fuels, Progress in Energy and Combustion Science
Vol 22, pp 29-120, 1996.

118. P. Koftsad, High Temperature Corrosion, 2nd Ed, Elsevier, New York, 1988.

119. Krause, H.H. Effects of Flue Gas Temperature and Composition on Corrosion from Reuse
Firing, Nace Conference 1991, Paper 242.

120. Rademarkers, P., W. Hesseling, and J.v.d. Wetering, Review on corrosion in waste
incinerators, and possible effect of bromine. 2002, TNO Industrial Technology.

121. Karlsson A, Moller P and Johansen V, “Iron and Steel Corrosion in a System of O2, SO2 and
Alkali Chloride. The Formation of Low Melting Point Salt Mixture”, Corrosion Science (30),
pp. 153-158, Pergamon Press, 1990.

122. Harb JN and Smith EE, “Fireside Corrosion in PC-Fired Boilers”, Prog. Energy Combustion
Science, 1990, Vol.16, pp 169-190.

123. Fielder. W.; Stearns, C.; Kohl, F. Reaction of NaCl with Gaseous SO3, SO2, and O2.
J.Electrochem. Soc. 1984 131, 2414-2417.

124. Boonsngsup, L.; Iisa, K.; Federick, W. Kinetics of the Sulfation of NaCl at Combustion
Conditions, Ind. Eng. Chem. Res., 1997, 36 (10), pp 4212–4216.

125. D.A. Jones, Principles and Prevention of Corrosion, 2nd Edition, Upper Saddle River, NJ
Prentice Hall 1996.

140
126. E. Hanson and M. Turner, “Second year Comparison of Superheater Metal Wastage Rates
Utilizing Various Boiler Tube Alloys in a Waste-to-Energy Facility”, Proceedings of the 11
Annual North American Waste to Energy Conference (NAWTEC 11), 2003.

127. Hayden Laser Services:

https://de356l4tocdyu.cloudfront.net/pdf/LAM2012%20Presentation%2011.pdf.

128. ASME Center for Research and Technology Development (CRTD) Volume 38 (1996):
Assessment of Factor Affecting Boiler Tube Lifetimes in Waste-Fired Steam Generators: New
Opportunities for Research and Development.

129. D. Joyner and G. Y. Lai, “Economic Impacts and solutions for waste-to-energy corroson
management,” Proceedsings of the 7th Annual North American Waste to Energy Conference
(NAWTEC 7), 1999.

130. P. Viklund, (2011), PhD Thesis, KTH.

131. M.A. Uusitalo et al. Corrosion Science 46 (2004) 1311–1331.

132. R.A. Antunes and M. C.L. de Oliveira , Corrosion Science 76 (2013) 6–26.

141
Appendix: Images of test coupons

Images of coatings following 24 h experiments under NaCl layer and WTE synthetic flue gas
mixture at 450, 500, and 550°C

Coating 450°C 500°C 550°C

Inconel 625

Colmonoy 88
HVOF

Colmonoy 88
laser

SP 99

SW1600

SW1641

142